United States Office of Water EPA821-B-01-004 Environmental Protection (4303) September 2001 Agency Final Report: Interlaboratory Variability Study of EPA Short-term Chronic and Acute Whole Effluent Toxicity Test Methods, Vol.1 ------- Disclaimer This final report has been reviewed by the Analytical Methods Staff in the Engineering and Analysis Division within the USEPA Office of Water and EPA's WET Interlaboratory Variability Workgroup. Mention of company names, trade names, or commercial products in this report does not constitute endorsement or recommendation for use. Questions or comments regarding this report should be addressed to: William A. Telliard, Director Analytical Methods Staff Engineering and Analysis Division (4303) Office of Science and Technology U.S. EPA Office of Water Ariel Rios Building 1200 Pennsylvania Avenue, NW Washington, DC 20460 ------- TABLE OF CONTENTS Acknowledgement xi Executive Summary xii List of Acronyms and Abbreviations xv Glossary xvi 1.0 INTRODUCTION AND BACKGROUND 1 1.1 Regulatory Background 1 1.2 The WET Variability Study 2 1.3 Other EPA Documents 2 2.0 STUDY DESIGN AND OBJECTIVES 4 2.1 Objectives 4 2.2 General Study Design 4 2.2.1 Study Management 4 2.2.2 Methods Evaluated 5 2.2.3 Laboratories 7 2.2.4 Samples 8 2.2.5 Schedule 8 3.0 LABORATORY PROCUREMENT 11 3.1 Identification and Solicitation of Potential Referee Laboratories 11 3.2 Identification and Solicitation of Potential Participant Laboratories 11 3.3 Prequalification Requirements for Participant Laboratories 12 3.3.1 Prequalification Documentation 12 3.3.2 Prequalification Determination 14 3.4 Prequalification and Selection of Referee Laboratories 15 3.5 Selection of Participant Laboratories 15 3.6 Participant Laboratory Meeting 16 4.0 PRELIMINARY TESTING 19 4.1 Part 1 - Background Testing 19 4.2 Part 2 - Range-finding 21 4.3 Part 3 - Holding Time Testing 22 4.4 Part 4 - Final Preliminary Testing 23 in ------- 5.0 SAMPLE PREPARATION 24 5.1 Freshwater Methods 24 5.2 Marine Methods 28 5.2.1 Mysidopsis Chronic and Sheepshead Acute and Chronic Test Methods 28 5.2.2 Silverside Acute and Chronic Test Methods 30 5.3 Problems Encountered in Sample Preparation 31 6.0 PACKAGING AND DISTRIBUTION OF TEST SAMPLES 33 6.1 Sample Distribution Scheme 33 6.2 Packaging and Shipment of Samples 33 6.3 Sample Tracking 35 6.4 Problems Encountered in Sample Distribution 39 7.0 INTERLABORATORY TESTING 40 7.1 General Testing Requirements 40 7.2 Method-Specific Requirements 43 8.0 DATA REPORTING AND EVALUATION 56 8.1 Report Submission 56 8.2 Data Review 59 8.2.1 Data Package Receipt and Gross Completeness Check 59 8.2.2 Data Accuracy and Quality Check 59 8.2.3 Effect Concentration Recalculation and Verification 62 9.0 RESULTS 65 9.1 Analysis of Results 65 9.1.1 Valid Tests 65 9.1.2 Successful Test Completion Rate 65 9.1.3 False Positive Rate 66 9.1.4 Precision 66 9.2 Ceriodaphnia Acute Test Method Results 70 9.2.1 Successful Test Completion Rate 70 9.2.2 False Positive Rate 70 9.2.3 Precision 70 9.3 Ceriodaphnia Chronic Test Method Results 76 9.3.1 Successful Test Completion Rate 76 9.3.2 False Positive Rate 77 9.3.3 Precision 77 9.4 Fathead Acute Test Method Results 87 9.4.1 Successful Test Completion Rate 87 9.4.2 False Positive Rate 87 iv ------- 9.4.3 Precision 87 9.5 Fathead Chronic Test Method Results 93 9.5.1 Successful Test Completion Rate 93 9.5.2 False Positive Rate 93 9.5.3 Precision 94 9.6 Selenastrum Chronic Test Method Results 104 9.6.1 Successful Test Completion Rate 104 9.6.2 False Positive Rate 104 9.6.3 Precision 105 9.7 Mysidopsis Chronic Test Method Results 117 9.7.1 Successful Test Completion Rate 117 9.7.2 False Positive Rate 117 9.7.3 Precision 118 9.8 Sheepshead Acute Test Method Results 124 9.8.1 Successful Test Completion Rate 124 9.8.2 False Positive Rate 124 9.8.3 Precision 124 9.9 Sheepshead Chronic Test Method Results 127 9.9.1 Successful Test Completion Rate 127 9.9.2 False Positive Rate 127 9.9.3 Precision 128 9.10 Silverside Acute Test Method Results 134 9.10.1 Successful Test Completion Rate 134 9.10.2 False Positive Rate 134 9.10.3 Precision 134 9.11 Silverside Chronic Test Method Results 139 9.11.1 Successful Test Completion Rate 139 9.11.2 False Positive Rate 139 9.11.3 Precision 139 9.12 Champia Chronic Preliminary Testing Results 146 9.13 Holmesimysis Acute Preliminary Testing Results 146 9.14 Results Summary 146 9.14.1 Successful Test Completion Rate 147 9.14.2 False Positive Rate 148 9.14.3 Precision 149 10.0 REFERENCES 152 VOLUME 2: APPENDIX A. WET Variability Study Plan B. Participant Laboratory Standard Operating Procedures ------- C. List of Referee and Participant Laboratories D. Preliminary Testing Results E. Analysis of Percent Minimum Significant Differences F. Method Performance Including Referee Laboratory Data VI ------- LIST OF TABLES Table 2.1. Four phases and specific objectives of the WET Variability Study 5 Table 2.2. WET test methods included in the WET Variability Study 6 Table 2.3. Endpoints and effect concentrations evaluated for each test method in the WET Variability Study 7 Table 2.4. Final participant laboratory schedule for interlaboratory testing during the WET Variability Study 9 Table 3.1. Summary of referee laboratory solicitation 11 Table 3.2. Referee laboratories selected for the WET Variability Study 16 Table 3.3. Results of participant laboratory prequalification and selection 18 Table 4.1. Summary of preliminary testing requirements 20 Table 5.1. Description of samples used for freshwater methods in the WET Variability Study 25 Table 5.2. Spiking concentrations in reference toxicant ampule samples 26 Table 5.3. Effluent sample volumes collected for interlaboratory testing 27 Table 5.4. Receiving water sample volumes collected for interlaboratory testing 27 Table 5.5. Description of samples used for marine methods in the WET Variability Study 29 Table 6.1. Sample distribution scheme for the WET Variability Study 34 Table 6.2. Number and volume of samples required for freshwater methods in the WET Variability Study 36 Table 6.3. Number and volume of samples required for marine methods in the WET Variability Study. 37 Table 6.4. Episode numbers and sample numbers used in the WET Variability Study 38 Table 7.1. Summary of test conditions and test acceptability criteria for the Ceriodaphnia acute test method 44 Table 7.2. Summary of test conditions and test acceptability criteria for the Ceriodaphnia chronic test method 45 Table 7.3. Summary of test conditions and test acceptability criteria for the fathead acute test method. 46 Table 7.4. Summary of test conditions and test acceptability criteria for the fathead chronic test method. 47 Table 7.5. Summary of test conditions and test acceptability criteria for the Selenastrum chronic test method 48 Table 7.6. Summary of test conditions and test acceptability criteria for the Mysidopsis chronic test method 49 Table 7.7. Summary of test conditions and test acceptability criteria for the sheepshead acute test method. 50 Table 7.8. Summary of test conditions and test acceptability criteria for the sheepshead chronic test method 51 Table 7.9. Summary of test conditions and test acceptability criteria for the silverside acute test method. 52 vn ------- Table 7.10. Summary of test conditions and test acceptability criteria for the silverside chronic test method 53 Table 7.11. Summary of test conditions and test acceptability criteria for the Champia chronic test method 54 Table 7.12. Summary of test conditions and test acceptability criteria for the Holmesimysis acute test method 55 Table 8.1. Report due dates 57 Table 8.2. Data reporting elements 57 Table 8.3. Test data qualifier flags 60 Table 8.4. Sample results affected by EPA guidance on concentration-response relationships (USEPA, 2000a) 63 Table 9.1. Test results identified as potential outliers by ASTM h statistics 69 Table 9.2. Results for Ceriodaphnia acute test method performed on blank samples 72 Table 9.3. Results for Ceriodaphnia acute test method performed on reference toxicant samples 73 Table 9.4. Results for Ceriodaphnia acute test method performed on effluent samples 74 Table 9.5. Results for Ceriodaphnia acute test method performed on receiving water samples 75 Table 9.6. Precision of point estimates from the Ceriodaphnia acute test method 75 Table 9.7. Results for Ceriodaphnia chronic test method performed on blank samples 79 Table 9.8. Results for Ceriodaphnia chronic test method performed on reference toxicant samples. . . 81 Table 9.9. Results for Ceriodaphnia chronic test method performed on effluent samples 83 Table 9.10. Results for Ceriodaphnia chronic test method performed on receiving water samples. ... 85 Table 9.11. Precision of point estimates from the Ceriodaphnia chronic test method 86 Table 9.12. Precision of NOEC values from the Ceriodaphnia chronic test method 86 Table 9.13. Results for fathead acute test method performed on blank samples 88 Table 9.14. Results for fathead acute test method performed on reference toxicant samples 89 Table 9.15. Results for fathead acute test method performed on effluent samples 91 Table 9.16. Results for fathead acute test method performed on receiving water samples 92 Table 9.17. Precision of point estimates from the fathead acute test method 92 Table 9.18. Results for fathead chronic test method performed on blank samples 96 Table 9.19. Results for fathead chronic test method performed on reference toxicant samples 98 Table 9.20. Results for fathead chronic test method performed on effluent samples 100 Table 9.21. Results for fathead chronic test method performed on receiving water samples 102 Table 9.22. Precision of point estimates from the fathead chronic test method 103 Table 9.23. Precision of NOEC values from the fathead chronic test method 103 Table 9.24. Results for Selenastrum chronic test method performed on blank samples with EDTA. . 107 Table 9.25. Results for Selenastrum chronic test method performed on blank samples without EDTA. 108 Table 9.26. Results for Selenastrum chronic test method performed on reference toxicant samples with EDTA 109 Table 9.27. Results for Selenastrum chronic test method performed on reference toxicant samples without EDTA 110 viii ------- Table 9.28. Results for Selenastrum chronic test method performed on effluent samples with EDTA. Ill Table 9.29. Results for Selenastrum chronic test method performed on effluent samples without EDTA. 112 Table 9.30. Results for Selenastrum chronic test method performed on receiving water samples with EDTA 113 Table 9.31. Results for Selenastrum chronic test method performed on receiving water samples without EDTA 114 Table 9.32. Precision of point estimates from the Selenastrum chronic test method conducted with EDTA 115 Table 9.33. Precision of point estimates from the Selenastrum chronic test method conducted without EDTA 115 Table 9.34. Precision of NOEC values from the Selenastrum chronic test method 116 Table 9.35. Results forMysidopsis chronic test method performed on blank samples 119 Table 9.36. Results for Mysidopsis chronic test method performed on reference toxicant samples. . . 120 Table 9.37. Results for Mysidopsis chronic test method performed on effluent samples 121 Table 9.38. Results for Mysidopsis chronic test method performed on receiving water samples 122 Table 9.39. Precision of point estimates from the Mysidopsis chronic test method 123 Table 9.40. Precision of NOEC values from the Mysidopsis chronic test method 123 Table 9.41. Results for sheepshead acute test method performed on blank samples 125 Table 9.42. Results for sheepshead acute test method performed on reference toxicant samples 125 Table 9.43. Results for sheepshead acute test method performed on effluent samples 126 Table 9.44. Results for sheepshead acute test method performed on receiving water samples 126 Table 9.45. Precision of point estimates from the sheepshead acute test method 127 Table 9.46. Results for sheepshead chronic test method performed on blank samples 129 Table 9.47. Results for sheepshead chronic test method performed on reference toxicant samples. . . 130 Table 9.48. Results for sheepshead chronic test method performed on effluent samples 131 Table 9.49. Results for sheepshead chronic test method performed on receiving water samples 132 Table 9.50. Precision of point estimates from the sheepshead chronic test method 133 Table 9.51. Precision of NOEC values from the sheepshead chronic test method 133 Table 9.52. Results for silverside acute test method performed on blank samples 135 Table 9.53. Results for silverside acute test method performed on reference toxicant samples 136 Table 9.54. Results for silverside acute test method performed on effluent samples 137 Table 9.55. Results for silverside acute test method performed on receiving water samples 138 Table 9.56. Precision of point estimates from the silverside acute test method 138 Table 9.57. Results for silverside chronic test method performed on blank samples 141 Table 9.58. Results for silverside chronic test method performed on reference toxicant samples. . . . 142 Table 9.59. Results for silverside chronic test method performed on effluent samples 143 Table 9.60. Results for silverside chronic test method performed on receiving water samples 144 Table 9.61. Precision of point estimates from the silverside chronic test method 145 Table 9.62. Precision of NOEC values from the silverside chronic test method 145 IX ------- Table 9.63. Successful test completion rates for test methods evaluated in the WET Variability Study. 147 Table 9.64. False positive rates for test methods evaluated in the WET Variability Study 148 Table 9.65. Within-laboratory, between-laboratory, and total variability observed for test methods evaluated in the WET Variability Study 150 Table 9.66. Summarized precision estimates (CVs) for test methods evaluated in the WET Variability Study 151 LIST OF FIGURES Figure 3.1 Participant laboratory selection process 17 ------- Acknowledgement This report was prepared by DynCorp Information and Enterprise Technology, Inc., under the direction of William A. Telliard, Director of the Analytical Methods Staff of the Engineering and Analysis Division within EPA's Office of Water. EPA would like to acknowledge the technical input provided by members of EPA's WET Interlaboratory Variability Workgroup throughout the design and implementation of the WET Variability Study and throughout the preparation of this final report. EPA thanks the contribution of contractor and referee laboratory support for their involvement in the implementation of the WET Variability Study and preparation of this report. EPA also thanks participant laboratories listed in Appendix C for their involvement in the WET Variability Study. In addition, EPA thanks the Association of Metropolitan Sewerage Agencies (AMSA) for coordinating the participation of non-EPA sponsored laboratories involved in the study. EPA WET Interlaboratory Variability Workgroup William Telliard, USEPA Office of Water, Washington, DC Marion Kelly, USEPA Office of Water, Washington, DC Teresa Norberg-King, USEPA Office of Research and Development, Duluth, MN Florence Fulk, USEPA Office of Research and Development, Cincinnati, OH John Fox, USEPA Office of Water, Washington, DC Stephen Sweeney, USEPA Office of General Counsel, Washington, DC Kathryn Greenwald, USEPA Office of Enforcement and Compliance Assurance, Washington, DC Contractor Support Robert N. Brent, Ph.D., DynCorp Information and Enterprise Technology, Inc., Alexandria, VA Julie Reemelin, DynCorp Information and Enterprise Technology, Inc., Alexandria, VA Brian Rusignuolo, DynCorp Information and Enterprise Technology, Inc., Alexandria, VA William Peltier, Watkinsville, GA Referee Laboratories EA Engineering, Science and Technology, Inc., Sparks, MD Ogden Environmental and Energy Services, Inc., San Diego, CA MEC Analytical, Inc., Tiburon, CA EnviroSystems, Inc., Hampton, NH XI ------- Executive Summary In 1995, the U.S. Environmental Protection Agency (EPA) promulgated 17 whole effluent toxicity (WET) test methods for use in monitoring toxicity under the National Pollutant Discharge Elimination System (NPDES) [60 FR 53529; October 16, 1995]. As part of a settlement agreement (Edison Electric Institute et al. v. USEPA. Settlement Agreement, July 24, 1998) to resolve a judicial challenge to this rulemaking, EPA conducted an interlaboratory variability study of 12 EPA short-term chronic and acute whole effluent toxicity test methods (the WET Variability Study). This report presents the results of the WET Variability Study. The purpose of the WET Variability Study was to characterize (1) interlaboratory variability, (2) the rate of successful test completion, and (3) the rate of false positive incidence for the following 12 approved WET test methods: • Cladoceran, Ceriodaphnia dubia, acute test (Ceriodaphnia acute) • Cladoceran, Ceriodaphnia dubia, survival and reproduction test (Ceriodaphnia chronic) • Fathead minnow, Pimephales promelas, acute test (fathead acute) • Fathead minnow, Pimephales promelas, larval survival and growth test (fathead chronic) • Green alga, Selenastrum capricornutum, growth test (Selenastrum chronic) • Mysid, Mysidopsis bahia, survival, growth, and fecundity test (Mysidopsis chronic) • Sheepshead minnow, Cyprinodon variegatus, acute test (sheepshead acute) • Sheepshead minnow, Cyprinodon variegatus, larval survival and growth test (sheepshead chronic) • Inland silverside, Menidia beryllina, acute test (silverside acute) • Inland silverside, Menidia beryllina, larval survival and growth test (silverside chronic) • Red macroalga, Champia parvula, reproduction test (Champia chronic) • Mysid, Holmesimysis costata, acute test (Holmesimysis acute) For two of these methods (the Champia chronic and Holmesimysis acute test methods), EPA was unable to obtain interlaboratory data due to laboratory unavailability (see Section 2.1). Intralaboratory data were obtained for the Champia chronic method, but no valid intralaboratory or interlaboratory data were obtained for the Holmesimysis acute method. For each of the remaining 10 methods, EPA selected a minimum of 7 and a maximum of 20 participant laboratories to constitute a "base" study design. Additional volunteer or externally sponsored laboratories (above 20) participated on a more limited basis as part of an "extended" study design (see Section 3). In total, 55 participant laboratories were involved in the study, with 7 to 35 participant laboratories testing each method. Each participating laboratory was required to prequalify for the study by documenting that their capacity and capabilities, experience and proficiency, and quality assurance and quality control systems met the needs of the study (see Section 3). Four referee laboratories also were involved in the WET Variability Study (see Section 3). For each test method, a referee laboratory was responsible for conducting preliminary testing (see Section 4), collecting and preparing homogeneous bulk test samples (see Section 5), and distributing "blind" test sample xn ------- aliquots to participant laboratories (see Section 6). Referee laboratories prepared the following four test sample types for each test method: blank sample; reference toxicant sample; municipal or industrial effluent sample; and receiving water sample. Referee laboratories distributed some combination of these test sample types to participant laboratories for testing. Laboratories participating in the base study design received 4 blind test samples as whole volume (volume necessary to conduct the test) or ampule (to mix and dilute to a required volume) samples; laboratories participating in the extended study design received 3 blind test samples as ampule samples. Participant laboratories were required to analyze each blind test sample according to the promulgated WET test method manuals and specific instructions in participant laboratory standard operating procedures (SOPs) developed for the study (see Section 7 and Appendix B). Interlaboratory testing was conducted from September 1999 through April 2000. In total, the WET Variability Study generated interlaboratory precision data from testing more than 700 blind samples among 55 participant laboratories. Following testing, participant laboratories submitted all test data on analyzed samples to the Sample Control Center (SCC) operated by DynCorp Information & Enterprise Technology, Inc., for independent review and calculation (see Section 8). SCC biologists reviewed test data to verify that all pertinent information was provided, tests were conducted in accordance with the WET method manuals and the WET Variability Study SOP, and test results were accurately calculated. Following test review, results were compiled and method performance characteristics (interlaboratory variability, successful test completion rate, and false positive rate) were calculated for each WET test method (see Section 9). Table 1 displays summarized results from the WET Variability Study. Successful test completion rates were greater than 90% for all WET test methods except the Ceriodaphnia chronic (82%) and Selenastrum chronic (63.6% with Ethylenediaminetetraacetic acid [EDTA] and 65.9% without EDTA) test methods. False positive rates were less than 5% for all WET test methods except the Selenastrum chronic test method conducted without EDTA (33.3%). Interlaboratory variability was described by the coefficient of variation (CV) calculated for point estimates. Interlaboratory CVs of LC50s (median lethal effect concentrations) ranged from 20.0% to 38.5% for acute test methods. Interlaboratory CVs of IC25s (25% inhibition concentrations) ranged from 10.5% to 58.5% for chronic test methods. Xlll ------- Table 1. Summarized test results from EPA's WET Variability Study. Test method Ceriodaphnia acute Ceriodaphnia chronic Fathead acute Fathead chronic Selenastrum chronic (with EDTA)° Selenastrum chronic (without EDTA)° Mysidopsis chronic Sheepshead acute Sheepshead chronic Silverside acute Silverside chronic Champia chronic6 Holmesimysis acute6 Successful test completion rate (%) 95.2 82.0 100 98.0 63.6 65.9 97.7d 100 100 94.4 100 ND ND False positive rate3 (%) 0.00 3.70 0.00 4.35 0.00 33.3 0.00 0.00 0.00 0.00 0.00 ND ND Intel-laboratory Precision (%CV)b 29.0 35.0 20.0 20.9 34.3 58.5 41.3 26.0 10.5 38.5 43.8 NDf ND a False positive rates reported for each method represent the higher of false positive rates observed for hypothesis testing results or point estimates. b Coefficients of variation (CVs) reported for each method represent the CV of LC50 values for acute test methods and IC25 values for chronic test methods. CVs reported are based on total variance and averaged across sample types. c The Selenastrum chronic test method was conducted with and without ethylenediaminetetraacetic acid (EDTA) as a component of the nutrients added to test and control treatments. d Successful test completion for the optional fecundity endpoint was 50%. e ND = not determined. Due to insufficient laboratory support, interlaboratory data were not obtained for the Champia chronic and Holmesimysis acute test methods. f While interlaboratory test data were not obtained for the Champia chronic method, intralaboratory data was obtained from the referee laboratory. Intralaboratory CVs were 27.6%, 49.7%, and 50.0% for reference toxicant, receiving water, and effluent sample types, respectively. xiv ------- List of Acronyms and Abbreviations ASTM CFR CuSO4 CV CWA DMRQA DO BAD EC50 EDTA EPA FR HDPP IC25 IC50 ICp KC1 LC50 LOEC MHSF MSD NOEC NPDES ORD PMSD QA QC SAS sec SETAC SOP SOW TAG WET YCT American Society for Testing and Materials Code of Federal Regulations Copper sulfate Coefficient of variation Clean Water Act Discharge Monitoring Report Quality Assurance Dissolved oxygen Engineering and Analysis Division 50% effect concentration Ethylenediaminetetraacetic acid Environmental Protection Agency Federal Register High-density polypropylene 25% inhibition concentration 50% inhibition concentration Percentage inhibition concentration Potassium chloride Median lethal effect concentration Lowest observed effect concentration Moderately-hard synthetic freshwater Minimum significant difference No observed effect concentration National Pollutant Discharge Elimination System Office of Research and Development Percent minimum significant difference Quality assurance Quality control Statistical Analysis Software Sample Control Center Society of Environmental Toxicology and Chemistry Standard operating procedure Statement of work Test acceptability criteria Whole effluent toxicity Yeast-cerophyl-trout chow mixture xv ------- Glossary Accuracy is used in this document only to describe the quality of being free from mistakes and errors. In other test method performance applications, accuracy of a test method is the closeness of agreement between measured values and an accepted reference (or known) value; however, accuracy in this sense cannot be determined for whole effluent toxicity test methods. Acute Toxicity Test is a test to determine the concentration of effluent or ambient waters that causes an adverse effect (usually death) on a group of test organisms during a short-term exposure (e.g., 24, 48, or 96 hours). Algal Suspension is a homogenized mixture of alga and liquid media. Aliquot is a subsample of a larger homogenized sample. Ambient Laboratory Illumination is the general lighting condition occurring daily in the laboratory. Ampule Sample is a small volume (generally 100 mL) liquid sample that is reconstituted at participant laboratories to provide the necessary test sample volume. Base Study Design used in the WET Variability Study consisted of participant laboratories receiving four blind test samples, which included some combination of the four test sample types (blank, reference toxicant, effluent, and receiving water). The base study design included a minimum of 7 and maximum of 20 participant laboratories, with up to 9 EPA-sponsored laboratories and up to 11 non-EPA-sponsored laboratories. Between-Laboratory Variability is the variability of test results from different laboratories using the same test method and analyzing the same test material. Between-laboratory variability, as used in this document, does not include the within-laboratory component of variance. Between-laboratory coefficients of variation (CVs) are CVs calculated based on solely the between-laboratory component of variance. Bioassay Grade is a rating on chemicals or chemical mixtures that have been tested and do not contain interferences to various bioassay tests. Blank is a non-toxic sample prepared as the typical synthetic control dilution water for the method being tested. Blind Sample is a sample that is of unknown composition to the testing laboratory. xvi ------- Brood is a group of offspring produced from a female in a single reproductive event. Bulk Sample is a large homogenized volume of a collected or prepared test sample. In this study, referee laboratories prepared bulk samples that were divided and distributed to participant laboratories for testing. Chronic Toxicity Test is a short-term test in which sublethal effects (e.g., reduced growth or reproduction) are usually measured in addition to lethality. Coefficient of Variation (CV) is a standard statistical measure of the relative variation of a distribution or set of data, defined as the standard deviation divided by the mean. It is also called the relative standard deviation (RSD). The CV can be used as a measure of precision within (within-laboratory) and between (between-laboratory) laboratories, or among replicates for each treatment concentration. In this study, within-laboratory, between-laboratory, and total CVs were calculated. The within-laboratory CV is used to express just the variability between duplicate samples tested in the same laboratory. The between- laboratory CV is used to express just the variability between laboratories testing duplicate samples. The total CV is used to express the total interlaboratory variability of the results, including both within- laboratory and between-laboratory components of variability. Control Chart is a chronological graphical representation of the results from reference toxicant tests performed with the same reference toxicant, test species, test conditions, and endpoints repeated by the same laboratory. Culture is an ongoing, reproducing population of organisms maintained in a laboratory to provide the laboratory with a supply of test organisms. Data Qualifier Flag is an identifier for tests with deviations in test conditions, sample holding times, sample temperatures, test acceptability criteria, or test water quality. Effect Concentration (EC) is a point estimate of the toxicant concentration that would cause an observable adverse effect (e.g., death, immobilization, or serious incapacitation) in a given percent of the test organisms, calculated from a continuous model (e.g., Probit Model). EC25 is a point estimate of the toxicant concentration that would cause an observable adverse effect in 25% of the test organisms. Effluent is wastewater discharged from a facility. Endpoint is the final measurement of a biological response (e.g., reproduction, growth, or survival). Episode is a designation to group samples prepared for a specific test method. XVII ------- Extended Study Design used in the WET Variability Study consisted of participant laboratories receiving three blind ampule samples, which included some combination of blank and reference toxicant samples. The extended study design included all non-EPA-sponsored laboratories not included in the base study design (those above the 20 laboratory maximum for the base study design). False Positive is a test result that indicates toxicity in a non-toxic or blank sample. False Positive Rate is the percentage of valid test results that indicate toxicity in blank samples. Forty Fathoms® Artificial Sea Salts is a commercially available chemical mixture of dry reagents that is used to create synthetic seawater. h Statistic is a calculated parameter defined by ASTM that represents the consistency of test results from laboratory to laboratory. Holding Time is the elapsed time from the end of sample collection to the first use of the sample. Because bulk samples were prepared in the laboratory for this study, the holding time began when sample aliquots were divided for distribution to participant laboratories. Hypothesis Testing is a statistical technique (e.g., Dunnett's test) for determining whether a tested concentration is statistically different from the control. Effect concentrations determined from hypothesis testing are NOEC and LOEC values. The two hypotheses commonly tested in WET testing are: Null hypothesis (H0): The effluent is not toxic. Alternative hypothesis (Ha): The effluent is toxic. Inhibition Concentration (1C) is a point estimate of the toxicant concentration that would cause a given percent reduction in a non-lethal biological measurement (e.g., reproduction or growth), calculated from a continuous model, (i.e., Interpolation Method). IC25 is a point estimate of the toxicant concentration that would cause a 25-percent reduction in a non-lethal biological measurement. Interlaboratory Variability is the variability of test results from different laboratories using the same test method and analyzing the same test material. Interlaboratory variability, as used in this document, includes both within-laboratory and between-laboratory components of variance. Interlaboratory coefficients of variation (CVs) are CVs calculated based on the total variance of results for a given method, endpoint, and sample type. k Statistic is a calculated parameter defined by ASTM that represents the consistency of within- laboratory precision from laboratory to laboratory. xvm ------- Lowest Observed Effect Concentration (LOEC) is the lowest concentration of an effluent or toxicant that results in adverse effects on the test organisms (i.e., where the values for the observed endpoints are statistically different from the control). Minimum Significant Difference (MSD) is the magnitude of difference from control where the null hypothesis is rejected in a statistical test comparing a treatment with a control (i.e., the smallest difference between control and treatment responses that can be determined as statistically significant). MSD is based on the number of replicates, control performance, and power of the test. Moderately Hard Synthetic Freshwater (MHSF) is water prepared from deionized water and reagent grade chemicals as described in the WET method manuals to produce a hardness of 80 to 100 mg CaCO3/L. No Observed Effect Concentration (NOEC) is the highest tested concentration of an effluent or toxicant that causes no observable adverse effect on the test organisms (i.e., the highest concentration of toxicant at which the values for the observed responses are not statistically different from the controls). National Pollutant Discharge Elimination System (NPDES) program regulates discharges to the nation's waters. Discharge permits issued under the NPDES program are required by EPA regulation to contain, where necessary, effluent limits based on water quality criteria for the protection of aquatic life and human health. Outlier is an extreme observation that is divergent from other observations of the same parameter. In this study, ASTM h and k statistics were used to identify potential outliers. Participant Laboratory is a laboratory selected to conduct a specific test method and report data for use in the study. Percent Minimum Significant Difference (PMSD) is the magnitude of difference from control, expressed as a percentage of the control response, where the null hypothesis is rejected in a statistical test comparing a treatment with a control (i.e., the smallest difference between control and treatment responses, expressed as a percentage of the control response, that can be determined as statistically significant). Photoperiod is the diurnal cycle of light and darkness to which test organisms are exposed. Point Estimate is a statistical estimate of some amount of adverse effect derived from a mathematical model that assumes a continuous concentration-response relationship. xix ------- Precision is a measure of reproducibility within a data set. Precision can be measured both within a laboratory (within-laboratory) and between laboratories (between-laboratory) using the same test method and toxicant. Prequalification is a the process of determining whether laboratories meet specific requirements for participation in the study. Quality Assurance (QA) is a practice in toxicity testing that addresses all activities affecting the quality of the final effluent toxicity data. QA includes practices such as effluent sampling and handling, source and condition of test organisms, equipment condition, test conditions, instrument calibration, replication, use of reference toxicants, recordkeeping, and data evaluation. Quality Control (QC) is the set of more focused, routine, day-to-day activities carried out as part of the overall QA program. Range-finding is preliminary testing to determine the range of toxicant concentrations that produce a targeted range of effects. Receiving Water is water into which wastewater flows. Reconstitute is to mix and dilute an ampule sample to the required volume for use in a toxicity test. Referee Laboratory is a laboratory selected to provide support for the testing of a specific test method in the study. The referee laboratory was responsible for conducting preliminary testing and for collecting, preparing, and distributing test samples to participant laboratory. Reference Toxicant is a known toxic chemical that is routinely tested to evaluate the consistency and precision of toxicity tests. Reference toxicant testing is a component of the quality assurance/quality control program in WET testing. Reference Toxicant Test is a check of the sensitivity of the test organisms and the suitability of the test methodology. Reference toxicant data are part of a routine QA/QC program to evaluate the performance of laboratory personnel and the robustness and sensitivity of the test organisms. Sample Code is a unique identifying number for each sample. Spiking is the addition of a reference toxicant to a sample matrix in order to elicit a toxicological effect. Static Non-renewal Test is a toxicity test that exposes test organisms to the same test solution (without renewal) for the duration of the test. xx ------- Static Renewal Test is a toxicity test that exposes test organisms to a fresh solution of the sample at regular prescribed intervals (typically every 24 or 48 hours). Successful Test Completion Rate is the percentage of initiated and properly terminated tests that met the test acceptability criteria as specified in the WET method manuals. Synthetic Seawater is artificial saltwater that is prepared by adding commercial sea salts or reagent grade chemicals to deionized water. Test Acceptability Criteria (TAG) are specific criteria for determining whether toxicity test results are acceptable. The effluent and reference toxicant must meet specific criteria as defined in the test method (e.g., for the Ceriodaphnia dubia survival and reproduction test, the criteria are as follows: the test must achieve at least 80 percent survival and an average of 15 young per surviving female in the controls). Traffic Report Form is a form to document the chain-of-custody for each test sample shipped. The traffic report form identifies the episode number, sample number, name and address of the referee laboratory, name and address of the participant laboratory, date shipped, airbill number, tests requested, and pre-shipment sample information (sample preparation date and initial water chemistry). Valid Test is a test that met the required test acceptability criteria for the method as stated in the WET method manuals. Variance is a measure of the dispersion in a set of values, defined as the sum of the squared deviations divided by their total number. Whole Effluent Toxicity (WET) is the total toxic effect of an effluent measured directly with a toxicity test. Whole Volume Sample is a sample that is provided in the appropriate volume for direct use in a toxicity test. Within-laboratory Variability is the variability of test results from the same laboratory using the same test method and analyzing the same test material. Within-laboratory coefficients of variation (CVs) are CVs calculated based on solely the within-laboratory component of variance. xxi ------- 1.0 INTRODUCTION AND BACKGROUND Whole effluent toxicity (WET) is defined as "the aggregate toxic effect of an effluent measured directly by an aquatic toxicity test" [54 FR 23686; June 2, 1989]. WET tests expose aquatic organisms to a range of effluent concentrations under controlled laboratory conditions. Exposure durations generally range from 24 hours (acute tests) to 7 days or more (short-term chronic or chronic tests). At the end of the exposure period, biological endpoints such as survival, growth, reproduction, or fertilization are measured in each effluent concentration and a control treatment. Toxicity of the effluent is determined by statistically comparing (either by hypothesis testing or point estimation) measured responses between the control and various effluent concentrations. Test results are expressed as the concentration of effluent estimated to produce a given effect (i.e., effect concentration). Effect concentrations such as the NOEC (No Observed Effect Concentration), LOEC (Lowest Observed Effect Concentration), LC50 (median lethal effect concentration), EC50 (median effect concentration), or IC25 (25% inhibition concentration) are commonly used to report the results of WET tests. 1.1 Regulatory Background The Clean Water Act (CWA) was enacted in 1972 with the objective of "restoring the chemical, physical, and biological integrity of the Nation's waters." Along with other goals, CWA section 101(a)(3) states that "it is the national policy that the discharge of toxic pollutants in toxic amounts be prohibited." The U.S. Environmental Protection Agency (EPA) has pursued this goal through implementation of the water quality standards program and the National Pollutant Discharge Elimination System (NPDES) permitting program. These programs have adopted an integrated strategy of water quality-based toxics control that includes three approaches: chemical-specific control, whole effluent toxicity (WET) control, and biological criteria/bioassessment (USEPA, 1991). To implement this strategy, States are encouraged to define numeric or narrative water quality standards that include chemical-specific criteria, criteria for WET, and biological criteria. Some states have included numeric criteria for WET, while others have relied on narrative criteria. These water quality standards and criteria are maintained by controlling the discharge of pollutants through the NPDES permitting program. When a discharge causes or has a reasonable potential to cause or contribute to the excursion above of numeric or narrative water quality standards, a NPDES permit limit will be issued to control the discharge. Permit limits for WET are established if the discharge has a reasonable potential to cause or contribute to the excursion above of water quality standards for WET. In 1995, EPA approved 17 WET test methods for use in NPDES permit monitoring [60 FR 53529; October 16, 1995]. The EPA-approved WET test methods resulted from many years of development and testing by EPA, States, municipalities, academia, and the regulated community. These WET test methods measure the acute and short-term chronic toxicity of effluents and receiving waters to aquatic plants, invertebrates, and vertebrates from freshwater and marine environments. WET test methods approved for use in NPDES monitoring are listed in 40 CFR §136.3, Table IA, and standardized test procedures for 1 ------- conducting the approved WET tests are published in the following three test method manuals (the WET method manuals). • USEPA, Methods for Measuring the Acute Toxicity of Effluents and Receiving Waters to Freshwater and Marine Organisms, Fourth Edition, EPA-600-4-90-027F, August 1993 • USEPA, Short-Term Methods for Estimating the Chronic Toxicity of Effluents and Receiving Water to Freshwater Organisms, Third Edition, EPA-600-4-91-002, July 1994 • USEPA, Short-Term Methods for Estimating the Chronic Toxicity of Effluents and Receiving Water to Marine and Estuarine Organisms, Second Edition, EPA-600-4-91-003, July 1994 1.2 The WET Variability Study Following the promulgation of WET test methods in 1995 [60 FR 53529; October 16, 1995], various parties filed suit to challenge the rulemaking. To resolve that litigation, EPA entered into settlement agreements with the various parties. In a 1998 settlement agreement (Edison Electric Institute et al. v. USEPA. Settlement Agreement, July 24, 1998), EPA agreed to conduct an interlaboratory variability study of 12 EPA short-term chronic and acute whole effluent toxicity test methods (the WET Variability Study). EPA conducted the WET Variability Study from September 1999 to April 2000 and published preliminary results from the study in October 2000 (USEPA, 2000b; USEPA, 2000c). In 2001, EPA submitted the preliminary results of the study for expert peer review. This document incorporates peer review comments (where appropriate) and presents the final results of the WET Variability Study. EPA intends to comment on the significance of these results in subsequent rulemaking and propose to ratify or withdraw each of the WET test methods evaluated in the WET Variability Study. 1.3 Other EPA Documents The WET method manuals (USEPA, 1993; USEPA, 1994a; USEPA, 1994b) were published in 1993 and 1994 and incorporated by reference in the 1995 WET rule [60 FR 53529; October 16, 1995]. Following this rulemaking, EPA issued clarifications to the WET test methods on April 10, 1996, via a memorandum from Tudor Davies, Director of EPA's Office of Science and Technology. This memorandum, titled "Clarifications Regarding Flexibility in 40 CFR Part 136 Whole Effluent Toxicity (WET) Test Methods" (USEPA, 1996), provided clarification on the following WET test issues: pH and ammonia control, temperature, hardness, test dilution concentrations, and acceptance criteria for Champia parvula. In February 1999, EPA published an errata to the WET method manuals (USEPA, 1999). This errata amended the approved WET test methods to correct typographical errors and omissions, provide technical clarification, and establish consistency among the 1995 WET rule language and the WET method manuals. ------- In June 2000, EPA published a guidance document titled, "Understanding and Accounting for Method Variability in Whole Effluent Toxicity Applications Under the National Pollutant Discharge Elimination System Program" (USEPA, 2000d). This guidance document quantified WET test variability, evaluated statistical methods for deriving WET permit conditions, and provided guidance on accounting for and minimizing WET test variability and its effects on the regulatory process. In July 2000, EPA published a guidance document titled, "Method Guidance and Recommendations for Whole Effluent Toxicity (WET) Testing (40 CFR Part 136)" (USEPA, 2000a). This document included guidance and recommendations on nominal error rate adjustments, confidence intervals, concentration- response relationships, dilution series selection, and selection of an acceptable dilution water for WET testing. The documents mentioned above were used in the conduct of WET tests and review of WET test data reported in this study. The WET method manuals (USEPA, 1993; USEPA, 1994a; USEPA, 1994b), clarification memo (USEPA, 1996), and errata document (USEPA, 1999) were used by referee and participant laboratories in the conduct of WET tests during the WET Variability Study. The two guidance documents published in 2000 were not available to laboratories at the time of testing, but were used by EPA in the review and analysis of WET test results from the study. ------- 2.0 STUDY DESIGN AND OBJECTIVES 2.1 Objectives In conducting the WET Variability Study, EPA's primary objectives were to: • characterize the interlaboratory variability of 12 WET test methods through the determination of CVs for LCSOs and IC25s and ranges for NOEC values, • characterize the rate at which participating laboratories successfully completed WET tests initiated (successful test completion rate), and • characterize the rate at which WET tests indicate "toxicity" is present when measuring non-toxic samples (false positive rate). EPA developed a preliminary study plan to meet these objectives, and made this study plan available for public and peer review comment on October 9, 1998. EPA then revised the study plan in response to comments received and published a final study plan on June 11, 1999 (see Appendix A). The WET Variability Study was conducted according to the final study plan with the following exceptions. • The minimum number of participant laboratories was reduced from nine to seven for the sheepshead acute and chronic test methods. Only seven laboratories qualified to participate in these methods. A minimum of seven participant laboratories still satisfied the data quality objective requiring a minimum of six complete and useable data sets for each WET test method. • No interlaboratory testing was conducted for the Champia chronic and Holmesimysis acute test methods. The interlaboratory testing phase was canceled for these test methods due to insufficient participant laboratory support. Only one participant laboratory could be procured for the Champia chronic test method, and only two participant laboratories could be procured for the Holmesimysis acute test method. This report presents only referee laboratory preliminary testing results for these two methods. 2.2 General Study Design The WET Variability Study was conducted in five phases designed to accomplish the overall study objectives. These phases are described in Table 2.1 and discussed in more detail in Sections 3-9 of this report. 2.2.1 Study Management The WET Variability Study was directed by EPA with contractor support provided by DynCorp Information & Enterprise Technology, Inc., under the Sample Control Center (SCC) contract. EPA Office of Water's Engineering and Analysis Division (EAD) and EPA's Office of Research and Development (ORD) provided overall management and technical oversight of the study. Laboratory procurement, day-to-day management, coordination of study activities, data review, and preparation of the preliminary and final study reports were performed by SCC under EAD and ORD guidance. SCC ------- contracted four referee laboratories to conduct preliminary testing and prepare and distribute blind test samples to 55 participant laboratories (see Appendix C for a list of referee and participant laboratories). Participant laboratories conducted WET tests and submitted data reports to SCC. Table 2.1. Four phases and specific objectives of the WET Variability Study. Phase Objectives 1 - Laboratory Procurement (See Section 3) Identify potential referee and participant laboratories to support the study Prequalify and select referee laboratories for Phases 2, 3, and 4 Prequalify and select participant laboratories for Phase 4 of the study 2 - Preliminary Testing (See Section 4) Determine the suitability of selected effluent and receiving water sample matrices for use in the study through characterization of physical, chemical, and toxicological properties of the test sample Determine the appropriate spiking concentrations for reference toxicant samples to achieve the desired range of toxicity Determine the persistence of toxicity in effluent and receiving water test samples Assess whether the desired range of sample toxicity will be maintained in test samples following shipping and handling 3 - Sample Preparation and Distribution (See Sections 5 and 6) Collect effluent and receiving water samples Prepare blank, reference toxicant, effluent, and receiving water samples for use by referee and participant laboratories in Phase 4 Minimize variability between samples prepared for and distributed to each participant laboratory in Phase 4 Distribute blind test samples to all qualified laboratories for initial use within 36 hours of individual sample shipment from the referee laboratories 4 - Intel-laboratory Testing (See Section 7) Obtain interlaboratory test data for each WET method using four test sample types 5 - Data Review and Analysis (See Sections 8 and 9) Review test data and calculate test results Evaluate precision of the test methods, the rate at which laboratories successfully completed tests initiated, and the rate at which the tests indicate "toxicity" is present when measuring non-toxic samples 2.2.2 Methods Evaluated EPA evaluated 12 of the 17 promulgated WET test methods in the WET Variability Study. These included two acute freshwater methods, three short-term chronic freshwater methods, three acute marine methods, and four short-term chronic marine methods. The test methods evaluated in the WET Variability Study are displayed in Table 2.2, and the endpoints and effect concentrations evaluated for each test method are displayed in Table 2.3. Each test method was conducted in accordance with the WET method manuals (USEPA, 1993; USEPA, 1994a; USEPA, 1994b), and as appropriate, specific guidance provided in the WET Variability Study participant laboratory standard operating procedure ------- (SOP) documents (Appendix B), the WET method manuals errata (USEPA, 1999), and clarifications provided in an April 10, 1996, memorandum from Tudor Davies, Director of EPA's Office of Science and Technology (USEPA, 1996). Table 2.2. WET test methods included in the WET Variability Study. Test method Cladoceran, Ceriodaphnia dubia, acute test0 Cladoceran, Ceriodaphnia dubia, survival and reproduction testd Fathead minnow, Pimephales promelas, acute test0 Fathead minnow, Pimephales promelas, larval survival and growth test4 Green alga, Selenastrum capricornutum, growth testd Mysid, Mysidopsis bahia, survival, growth, and fecundity test6 Sheepshead minnow, Cyprinodon variegatus, acute test0 Sheepshead minnow, Cyprinodon variegatus, larval survival and growth test6 Inland silverside, Menidia beryllina, acute test0 Inland silverside, Menidia beryllina, larval survival and growth test6 Red macroalga, Champia parvula, reproduction test6 Mysid, Holmesimysis costata, acute test°'f Common test method name3 Ceriodaphnia acute Ceriodaphnia chronic Fathead acute Fathead chronic Selenastrum chronic Mysidopsis chronic Sheepshead acute Sheepshead chronic Silverside acute Silverside chronic Champia chronic Holmesimysis acute Test method number* - 1002.0 - 1000.0 1003.0 1007.0 - 1004.0 - 1006.0 1009.0 - Test type freshwater acute freshwater chronic freshwater acute freshwater chronic freshwater chronic marine chronic marine acute marine chronic marine acute marine chronic marine chronic marine acute a Common test method names were used in this report to refer to the test methods in the WET Variability Study. b Test method numbers were not designated for acute test methods in USEPA, 1993. c USEPA, Methods for Measuring the Acute Toxicity of Effluents and Receiving Waters to Freshwater and Marine Organisms, Fourth Edition, EPA-600-4-90-027F, August 1993 d USEPA, Short-Term Methods for Estimating the Chronic Toxicity of Effluents and Receiving Water to Freshwater Organisms, Third Edition, EPA-600-4-91-002, July 1994 e USEPA, Short-Term Methods for Estimating the Chronic Toxicity of Effluents and Receiving Water to Marine and Estuarine Organisms, Second Edition, EPA-600-4-91-003, July 1994 f The EPA-appro ved acute test with Holmesimysis costata was performed using the acute test procedures for Mysidopsis bahia and test conditions optimized for//, costata. ------- Table 2.3. Endpoints and effect concentrations evaluated for each test method in the WET Variability Study. Test method Ceriodaphnia acute Ceriodaphnia chronic Fathead acute Fathead chronic Selenastrum chronic Mysidopsis chronic Sheepshead acute Sheepshead chronic Silverside acute Silverside chronic Champia chronic Holmesimysis acute Acute tests Survival LC50 X - X - - - X - X - - X Test duration (hours) 48 - 96 - - - 96 - 96 - - 96 Short-Term Chronic Tests Survival LC50 NOEC - X - X - X - X - X - - Growth IC25 NOEC - - - X X X - X - X - - Reproduction IC25 NOEC - X - - - X - - - - X - Test duration (days) - 8a - 7 4 7 - 7 - 7 7b - a The C. dubia test acceptability criteria state that the test is complete when 60% of controls have 3 broods (approximately 7 days); for purposes of this study, all tests continued for 8 days and each laboratory was requested to carefully distinguish and carefully record the number of broods. b C. parvula were exposed to test samples for 2 days, followed by a 5-day recovery period in control water. 2.2.3 Laboratories A referee laboratory was selected to provide support for each test method evaluated in the WET Variability Study. The referee laboratory was responsible for conducting preliminary testing on sample types to ensure that samples used in the WET Variability Study were stable and provided the desired level of effect. Referee laboratories also collected, prepared, and distributed test samples for interlaboratory testing, and conducted testing simultaneously with participant laboratory testing to provide further information on sample consistency. Interlaboratory testing was conducted by participant laboratories selected for each test method. Each participating laboratory was required to prequalify for the study by documenting that their capacity and capabilities, experience and proficiency, and quality assurance and quality control systems met the needs of the study (see Section 3). The number of participant laboratories conducting each method was ------- dependent upon the number of laboratories responding to the solicitation, the number of laboratories qualified to conduct the given test method, and the number of laboratories that were sponsored by parties external to EPA. The WET Variability Study plan called for a minimum of 9 and a maximum of 20 participant laboratories to constitute a base study design for each method. Additional laboratories (greater than 20) participated on a more limited basis as part of an extended study design. In deviation from this design, only seven qualified laboratories could be procured to participate in the sheepshead acute and chronic test methods. In addition, interlaboratory testing for the Champia chronic and Holmesimysis acute test methods was canceled due to insufficient participant laboratory support. See Section 3 for a detailed description of referee and participant laboratory selection, and the base and extended study designs. 2.2.4 Samples For each test method, four test sample types were prepared in bulk by the referee laboratory, divided, and distributed to participant laboratories for testing. The four sample types included: 1) blank sample, 2) reference toxicant sample, 3) effluent sample, and 4) receiving water sample. Blank and reference toxicant samples were distributed to participant laboratories as liquid ampule samples (to mix and dilute to the required volume at the participant laboratory), while effluent and receiving water samples were distributed as whole-volume samples (consisting of the full volume necessary to conduct the test). The blank sample was a non-toxic sample prepared as the typical synthetic control dilution water for each test method. Testing of the blank sample provided a means of determining the false positive rate for each test method. Interlaboratory precision was evaluated through testing of the reference toxicant, effluent, and receiving water sample types. These sample matrix types (with the exception of the effluent sample type for the Champia chronic and Holmesimysis acute test methods) were spiked with a toxicant to achieve a desired level of effect and facilitate the evaluation of precision. Laboratories participating in the base study design received four blind test samples, which included some combination of the four test sample types (blank, reference toxicant, effluent, and receiving water). Laboratories that participated in the extended study design received three blind ampule samples, which included some combination of blank and reference toxicant samples. The combination of blind test samples received by individual laboratories included replicate (i.e., duplicate) test samples for each test method except for methods with fewer than nine participant laboratories. Replicate samples were always shipped together and tested simultaneously, with the exception of the Selenastrum chronic test, where only one sample was tested per week. See Sections 5 and 6 for a detailed description of sample preparation and distribution, respectively. 2.2.5 Schedule Interlaboratory testing in the WET Variability Study was conducted from September 1999 through April 2000. The final participant laboratory schedule for the WET Variability Study is provided in Table 2.4. For each method (with the exception of the Selenastrum chronic method), testing was conducted in two ------- testing periods. Each participant laboratory simultaneously tested two samples during each testing period. For the Selenastrum chronic method, testing was conducted in four testing periods, with a single sample tested with and without EDTA during each testing period. Table 2.4. Final participant laboratory schedule for interlaboratory testing during the WET Variability Study. Activity Fathead chronic testing period #1 Fathead chronic testing period #2 Ceriodaphnia chronic testing period #1 Silverside chronic testing period #1 Fathead acute testing period #1 Silverside chronic testing period #2 Ceriodaphnia chronic testing period #2 Silverside acute testing period #1 Fathead acute testing period #2 Ceriodaphnia acute testing period #1 Silverside acute testing period #2 Ceriodaphnia acute testing period #2 Mysidopsis chronic testing period #1 Mysidopsis chronic testing period #2 Sheepshead acute testing period #1 Selenastrum chronic testing period #1 Sheepshead acute testing period #2 Selenastrum chronic testing period #2 Sheepshead chronic testing period #1 Selenastrum chronic testing period #3 Sheepshead chronic testing period #2 Selenastrum chronic testing period #4 Testing dates3 (start date - finish date) 9/28/99 - 10/5/99 10/5/99 - 10/12/99 10/12/99 - 10/20/99 10/19/99 - 10/26/99 10/21/99 - 10/25/99 10/26/99- 11/2/99 10/26/99- 11/3/99 11/2/99-11/6/99 11/4/99- 11/8/99 11/9/99- 11/11/99 11/9/99-11/13/99 11/11/99-11/13/99 2/22/00 - 2/29/00 2/29/00 - 3/7/00 3/7/00-3/11/00 3/9/00 - 3/13/00 3/14/00 - 3/18/00 3/16/00 - 3/20/00 3/21/00-3/28/00 3/23/00 - 3/27/00 3/28/00 - 4/4/00 3/30/00 - 4/3/00 a Samples were shipped to arrive at participant laboratories on the indicated start date. Tests were initiated on the same day as sample receipt. Samples were shipped on ice overnight to arrive at each participant laboratory on the day of scheduled testing. Sample distribution was organized such that all effluent samples were tested on the same day in all participant laboratories, and all receiving water samples were tested on the same day (with the exception of delays due to sample or organism shipments). Since the synthetic matrix and ampule form ------- of blank and reference toxicant samples provided more sample stability, these samples were distributed for testing in either the first or second testing periods. As a result, tests conducted on reference toxicant and blank samples were initiated on one of two testing dates. Replicate samples were always distributed together for simultaneous testing. 10 ------- 3.0 LABORATORY PROCUREMENT 3.1 Identification and Solicitation of Potential Referee Laboratories Since the responsibilities of the referee laboratory would be demanding and critical to successful implementation of the WET Variability Study, SCC solicited a select list of laboratories that possessed exceptional qualifications, based on EPA technical staff recommendations. Solicitation bid packages included the following documents: (1) a referee laboratory prequalification document; (2) a referee laboratory statement of work (SOW), including a preliminary study schedule; and (3) a referee laboratory bid sheet. Referee laboratories were first solicited in December 1998. Additional solicitations to an expanded number of laboratories were necessary to fill referee laboratory positions for all of the test methods (Table 3.1). Table 3.1. Summary of referee laboratory solicitation. Solicitation date 12/29/98 1/22/99 2/19/99 Response deadline 1/13/99 2/4/99 3/5/99 Methods solicited All twelve methods All seven marine methods Silverside acute and chronic; Holmesimysis acute Laboratories solicited 7 12 19 Number of qualified responses 2 2 3 3.2 Identification and Solicitation of Potential Participant Laboratories EPA attempted to maximize the number of qualified laboratories participating in the WET Variability Study and select laboratories that were representative of laboratories throughout the United States that routinely conduct WET tests for permittees. SCC and EPA staff identified a list of potential participant laboratories from a variety of sources, including EPA and State environmental agencies, the Society of Environmental Toxicology and Chemistry (SETAC), reviews of the public literature, the Directory of Environmental Laboratories (DynCorp, 1996), and the list of laboratories conducting testing for EPA's Discharge Monitoring Report Quality Assurance (DMRQA) program. In addition, the petitioners provided a list of laboratories interested in participating without EPA sponsorship and a list of potential non-EPA sponsors. On July 9, 1999, SCC solicited all 319 laboratories included in the compiled potential laboratory list. Laboratories solicited included state, academic, municipal, industrial and private laboratories. Solicitation bid packages included the following documents: (1) a detailed cover letter describing the solicitation; (2) a participant laboratory prequalification document; (3) a participant laboratory SOW, including a preliminary study schedule; and (4) a participant laboratory bid sheet. 11 ------- 3.3 Prequalification Requirements for Participant Laboratories To ensure that laboratories participating in the WET Variability Study possessed the capacity and capabilities, experience and proficiency, and quality assurance and quality control systems necessary to meet the needs of the study, EPA required all laboratories to meet specific prequalification requirements. 3.3.1 Prequalification Documentation To demonstrate its qualifications, each potential participant laboratory was required to provide the 16 prequalification items listed below. Prequalification was conducted independently for each test method, and laboratories could submit prequalification materials for any or all methods evaluated in the WET Variability Study. General information (1) A cover page with the laboratory name, address, telephone number, fax number, e-mail address, contact person, and additional contacts for day-to-day sample tracking and technical issues, if different from primary contact. (2) A statement on the number of tests that the laboratory can conduct at one time with the proposed staff, including the number of tests using a single test method and the number of tests using multiple test methods (e.g., three C. dubia survival and reproduction tests, three fathead minnow survival and growth tests, and two of each simultaneously). Capacity and capabilities (3) A statement that the combination of facilities, equipment, staff, and laboratory capabilities are sufficient to meet study needs. (4) Detailed information on the type and size of laboratory and test equipment used for conducting each test method, including information on temperature control, sample storage, water purification devices (i.e., Millipore Milli-Q® filtration and ion exchange), and dilution water sources. Laboratories were required to provide summaries of routine water quality monitoring data on dilution water and water used for culturing or maintaining each species (e.g., 3-4 months of pH, alkalinity, hardness, and salinity measurements on dilution and culture waters). (5) A statement that the laboratory can receive next day deliveries (including Saturday deliveries) via overnight carriers (i.e., Fed Ex, UPS, etc.) and initiate a test on the same day as receipt. (6) A list of laboratory staff able to participate in the study, including resumes and titles. (7) Information on the source of organisms, including whether organisms are cultured in-house or obtained externally. If organisms are cultured in-house, the laboratory was required to provide 12 ------- standard operating procedures for organism culturing (as required in number 9 below), a summary of how culture performance is assessed, and data on culture performance (e.g., Ceriodaphnia dubia brood board monitoring data for the past six months or records of Pimephales promelas egg production). If organisms are obtained from an external source, the laboratory was required to specify the source, number of organisms that can be obtained from this source on a given day, age of obtained organisms, and organism holding and maintenance conditions. Experience and proficiency (8) Copies of internal standard operating procedures (SOPs) for conducting each of the test methods. Internal laboratory SOPs for each test method were required to be in place with dates of SOP origination. (9) Copies of supporting internal laboratory SOPs for organism culturing, food preparation, and dilution water preparation for each species and each method. (10) A statement on the number of effluent tests conducted in the last year using each of the WET test methods. The laboratory was required to specify the frequency with which test acceptability criteria were met in these tests and the average control response measured in these tests. (11) A statement regarding State or regional certifications and documentation of current certifications (if applicable). Quality assurance/quality control (12) Evidence that the laboratory maintains control (cusum) charts for reference toxicant tests for each method. The laboratory was required to submit the most current control chart for each test method, covering at least 12-24 data points and showing control limits. The raw data (actual data sheets and summarized data) for each data point also were required. Data charts with NOEC and/or IC25 for the same test values were requested as well as an explanation of why one is used rather than the other. Explanations were required if methods used to develop reference toxicant control charts deviated from promulgated methods or were from the previous edition of a relevant test protocol. (13) Evidence that reference toxicant tests are conducted at the appropriate frequency (e.g., monthly for tests that are routinely conducted for NPDES permits). Along with control chart information described above, the laboratory was required to provide a statement on the frequency of reference toxicant testing. If control charts were composed of fewer than 12-24 data points, an explanation was required. (14) Copies of internal laboratory SOPs for conducting reference toxicant tests and constructing control charts. This information had to include a narrative explanation of the width of the control 13 ------- limits for the laboratory and a statement of corrective action for any toxicity test result that falls outside the control limits. (15) Results of the most recent DMRQA study, if the laboratory participated. The laboratory was required to provide data point(s) for each method performed for the previous year's DMRQA study. If the laboratory did not participate, a narrative statement to that effect was required. (16) A signed statement of accuracy and completeness of submitted prequalification materials. The following statement, signed and dated by an authorized representative of the laboratory, was required: "I certify that the information provided in this prequalification package is complete and accurate to the best of my knowledge." 3.3.2 Prequalification Determination SCC evaluated prequalification materials and recommended laboratory rejection based on the following criteria: (1) Combination of facilities, equipment, staff and lab capacity and capabilities were insufficient to meet study needs. (2) Organism source information was not provided, culture and/or collection information was severely lacking, or source information was inadequate to assess the health of the organisms routinely used. (3) Internal laboratory SOPs for each method were vague and could not be discerned and/or were generally insufficient to support performance of the methods in accordance with specific instructions provided by EPA. (4) Statements regarding the number of effluent tests conducted per year, test acceptability rates, average control response, and/or State certifications were not provided, did not adequately demonstrate proficiency in the test method, or did not adequately demonstrate that the laboratory is representative of laboratories throughout the United States that routinely conduct WET testing for permittees. (5) Control charts were not adequately maintained for reference toxicant tests, or data were not provided (cusum chart for each endpoint and raw data for each data point). Control charts did not cover 12-24 data points for each species and test method, and an acceptable explanation was not provided. (6) Reference toxicant tests were not conducted at the appropriate frequency (monthly for tests that are routinely conducted for permits) and a satisfactory explanation was not provided. 14 ------- (7) No acceptable explanation or evidence of corrective action was provided for any control chart value falling outside the control limits. (8) The laboratory did not provide the most recent DMRQA study results, or an acceptable explanation for non-passing results was not provided. If the laboratory did not participate in the DMRQA study, the laboratory did not include an acceptable explanation as to why they did not participate. (9) No signed statement of accuracy and completeness of prequalification materials was included. Laboratories that failed to meet the prequalification criteria due to an incomplete submission were notified via fax of their deficiency and allowed an opportunity to submit the missing information. Laboratories that did not respond to this notification or were unable to provide the missing information failed to prequalify for the WET Variability Study. Only eight laboratories failed to prequalify for the WET Variability Study. Two of these laboratories failed to prequalify for more than one test method. 3.4 Prequalification and Selection of Referee Laboratories Referee laboratories were required to submit the same prequalification materials listed in Section 3.3.1 for participant laboratories. In addition, referee laboratories were required to submit three client references and provide background information on potential effluent and receiving water sample sources. Referee laboratory prequalification materials were evaluated based on the rejection criteria listed in Section 3.3.2 and the additional reference and sample source requirements. The capacity and capabilities of potential referee laboratories were highly scrutinized to ensure that the laboratory could meet the sample collection, preparation, distribution, and testing requirements of the study. All potential referee laboratories that met prequalification criteria and were determined to possess the capacity and capabilities to meet the needs of the study were considered for the referee laboratory position. From this pool of qualified laboratories, the lowest bidder was selected as the referee laboratory for each bid lot of test methods. The referee laboratories selected for each test method are listed in Table 3.2. 3.5 Selection of Participant Laboratories Laboratories that met all prequalification criteria (see Section 3.3) were eligible for participation in the WET Variability Study. From the pool of prequalified laboratories, those laboratories with the nine lowest bids per method were selected for EPA sponsorship and participation in the base study design (see Figure 3.1). If a tie for the ninth lowest bid was encountered, selection among equal bid prices from prequalified laboratories was determined based on business classification, with preferences granted in the order of small company, minority-owned, woman-owned, or large company. From the remaining prequalified laboratories, those that had identified a willing sponsor outside of EPA (non-EPA sponsorship) were also selected for participation in the WET Variability Study. From this group of prequalified, non-EPA sponsored laboratories, a maximum of 11 laboratories (for each WET test method) were randomly selected for participation in the base study design. The 9 EPA-sponsored laboratories and 15 ------- the 11 randomly chosen non-EPA-sponsored laboratories constituted the 20 laboratories included in the base study design for each WET test method. All remaining prequalified laboratories with non-EPA sponsorship were selected for participation in the extended study design. Figure 3.1 displays the complete process of participant laboratory selection. SCC formally notified all laboratories on September 8, 1999, of their selection, sponsor for each method (EPA or non-EPA), and level of participation (base or extended design). Table 3.2. Referee laboratories selected for the WET Variability Study. Referee laboratory EA Engineering, Science and Technology, Inc. EA Engineering, Science and Technology, Inc. Ogden Environmental and Energy Services, Inc. MEC Analytical, Inc. EnviroSystems, Inc. Method(s) supported Ceriodaphnia acute and chronic; Fathead acute and chronic; Selenastrum chronic Mysidopsis chronic; Sheepshead acute and chronic Silverside acute and chronic Holmesimysis acute Champia chronic Date awarded 03/10/1999 04/01/1999 06/08/1999 06/02/1999 04/14/1999 The results of laboratory prequalification and selection are displayed in Table 3.3 for each test method. Because only one and two laboratories were procured for the Champia chronic and Holmesimysis acute test methods respectively, interlaboratory testing was canceled for these two methods. For all other methods, 7 to 35 laboratories prequalified and were selected for participation. A total of 55 laboratories were selected to participate in the WET Variability Study, with many laboratories participating in multiple methods. See Appendix C for an alphabetical list of the participant laboratories. 3.6 Participant Laboratory Meeting EPA invited all laboratories that submitted prequalification materials to attend a participant laboratory meeting held in Chicago, Illinois, on September 16, 1999. At the participant laboratory meeting, EPA and SCC staff presented the study plan for the WET Variability Study and highlighted participant laboratory tasks and requirements. Participant laboratories had the opportunity to meet EPA and SCC staff and to ask questions regarding the WET Variability Study and their responsibilities. At the meeting, EPA and SCC staff did not release any information regarding sample contents, sample descriptions, or sample distribution schemes that would jeopardize the blind aspect of the study. Following the meeting, SCC distributed meeting notes to all participant laboratories that were unable to attend. 16 ------- PREQUALIFICATION BID COST EVALUATION RANDOM SELECTION LAB INVOLVEMENT ASSIGNMENT Did lab pass prequalification criteria? Yes Was bid among 9 lowest for the method? Yes Base Design No No No involvement Was external sponsorship available? No involvement Yes Was lab randomly selected for base design? Yes No f Extended Design j Figure 3.1 Participant laboratory selection process. 17 ------- Table 3.3. Results of participant laboratory prequalification and selection." Test method Ceriodaphnia acute Ceriodaphnia chronic Fathead acute Fathead chronic Selenastrum chronic Mysidopsis chronic Sheepshead acute Sheepshead chronic Silverside acute Silverside chronic Champia chronicb Holmesimysis acuteb No. of labs solicited 319 319 319 319 319 319 319 319 319 319 319 319 No. of labs responding 48 55 50 49 16 16 7 9 13 15 2 2 No. of labs failing to prequalify 2 0 1 0 2 3 0 2 1 2 1 0 No. of labs selected for base study design EPA sponsored labs 9 9 9 9 9 9 7 7 9 9 1 2 Non-EPA sponsored labs 11 11 11 11 2 2 0 0 0 1 0 0 No. of labs selected for extended study design Non-EPA sponsored labs 8 15 9 7 0 0 0 0 0 0 0 0 Total no. of participant labs 28 35 29 27 11 11 7 7 9 10 1 2 a Appendix C contains an alphabetical list of the participant laboratories. b Intel-laboratory testing was canceled for this method due to insufficient participant laboratory support. Tests were performed by the referee laboratory only. 18 ------- 4.0 PRELIMINARY TESTING Referee laboratories contracted to support each test method during the WET Variability Study were responsible for conducting preliminary testing of each sample type prior to interlaboratory testing. Preliminary testing was single-laboratory testing conducted at the referee laboratory to determine the appropriate composition of samples for use in the interlaboratory testing phase. A four-part preliminary testing plan was developed and instituted to provide background information on sample toxicity, necessary spiking concentrations, and the persistence of sample toxicity. Based on the results of preliminary testing, SCC selected appropriate sample sources and spiking levels for use in interlaboratory testing. The four parts of the preliminary testing plan consisted of the tests listed in Table 4.1 and accomplished the following specific objectives. • Part 1 - Background testing: determine the suitability of effluent and receiving water sample matrices for use in the WET Variability Study through characterization of physical, chemical, and toxicological properties • Part 2 - Range-finding: determine the appropriate spiking concentrations for the preparation of spiked effluent, receiving water, and reference toxicant sample types • Part 3 - Holding time testing: determine the persistence of toxicity in spiked effluent and receiving water samples • Part 4 - Final preliminary testing: assess whether test samples provided the desired range of toxicity following sample preparation, shipping, and handling The specific requirements for each part of preliminary testing are described in Sections 4.1 to 4.4 below. These preliminary testing requirements were modified for the Champia chronic and Holmesimysis acute test methods after interlaboratory testing of these methods was canceled. With this cancellation, the objectives of any uncompleted preliminary tests were adjusted to better direct the use of preliminary test data toward assessing single-laboratory test precision rather than preparation for interlaboratory testing. Preliminary test results for these two methods are reported in Appendix D and summarized in Section 9 in lieu of interlaboratory test data. For all other methods, preliminary testing results (reported in Appendix D) were not used to assess test precision, but were used to support decisions regarding the selection of test samples for use in interlaboratory testing. 4.1 Part 1 - Background Testing Part 1 of preliminary testing verified that selected effluent and receiving water sample matrix sources were acceptable for study use by assessing the physical, chemical, and toxicological characteristics of the samples. Referee laboratories were required to submit information on potential effluent and receiving water sample sources as part of their prequalification materials. EPA and SCC reviewed these materials, including historical information from the source, and made a preliminary selection of the effluent and receiving water sample sources for each test method. Following this determination, the referee laboratory initiated Part 1 of preliminary testing. 19 ------- Table 4.1. Summary of preliminary testing requirements. Test method Ceriodaphnia acute Ceriodaphnia chronic Fathead acute Fathead chronic Selenastmm chronic Mysidopsis chronic Sheepshead acute Sheepshead chronic Silverside acute Silverside chronic Champia chronic Holmesimysis acute Part 1 - Background testing Unspiked effluent • • •a • • • • • Unspiked receiving water • • •a • • • • • Part 2 - Range-finding Spiked effluent • • • • •a • • • • • •c •c Spiked receiving water • • • • •a • • • • • • • Reference toxicant • • • • •a • • • • • • • Part 3 - Holding time testing Spiked effluent • • •a • • • •c •c Spiked receiving water • • •a • • • • • Part 4 - Final preliminary testing Spiked effluent • • • • •b • • • • • •c •c Spiked receiving water • • • • •b • • • • • • • Reference toxicant • • • • •b • • • • • • • a Test conducted with EDTA. b Tests conducted with and without EDTA. c Tests were conducted only on unspiked effluent sample 20 ------- Referee laboratories collected preliminary test samples according to Section 8 of the method manuals (USEPA, 1993; USEPA, 1994a; USEPA, 1994b). During Part 1 of preliminary testing, referee laboratories conducted physical and chemical analyses of both the effluent sample and the receiving water sample, including alkalinity, hardness, pH, temperature, total residual chlorine, total ammonia, dissolved oxygen, total dissolved solids, total suspended solids, total organic carbon, biological oxygen demand, and chemical oxygen demand. For samples that were to be used in marine tests, salinity and copper also were measured. A smaller subset of parameters was analyzed for samples collected during Parts 2-4 of preliminary testing. Following chemical and physical characterization of the sample matrices, a single background toxicity test using each of the test species was conducted on a sample from each effluent and receiving water source. If historical information (chemical analysis or toxicological analysis) on the effluent and receiving water matrix source was available, this information was evaluated along with results of background testing. Following completion of analysis and historical data gathering, a final determination of effluent and receiving water sample sources was made. The selection criteria for effluent and receiving water sample sources included the following elements. • Accessibility - Selected effluent and receiving water sample sources were readily accessible to the referee laboratory. This included logistical accessibility as well as permission from the source provider to collect, test, and use the sample source in the WET Variability Study • Historic testing and experience - It was important for referee laboratories to have significant experience in collecting and testing the selected effluent and receiving water samples. This experience and knowledge of historic testing allowed referee laboratories to identify conditions or characteristics of the source that could potentially pose problems in the WET Variability Study • Characterization - Selected effluent and receiving water sample sources were well characterized by the referee laboratory through historic physical, chemical, and toxicological testing, as well as Part 1 of preliminary testing • Consistency - Selected effluent and receiving water sample sources either provided a consistent level of toxicity or consistently produced no toxicity (in which case the sample could be spiked to achieve the desired effect level) 4.2 Part 2 - Range-finding Part 2 of preliminary testing determined the range of spiking concentrations necessary to achieve a specific level of toxic effect for each sample type. This determination was critical to insuring that test concentrations used in interlaboratory testing bracketed the effect concentrations (LC50, IC25, IC50, and NOEC) evaluated in the WET Variability Study. During interlaboratory testing, participant laboratories were instructed to test each sample using a test concentration range of 6.25 - 100%, so it was important for test samples to produce measurable toxic effects within that test concentration range. Effluent and receiving water samples were spiked to produce target effect concentrations (LCSOs for acute test methods and IC25s for chronic test methods) of 25% sample during interlaboratory testing. The reference toxicant sample was spiked to produce target effect concentrations of 50% sample during interlaboratory 21 ------- testing. Spiking levels for the Selenastrum chronic test method were targeted to produce an IC50 of 38% sample during interlaboratory testing. During Part 2 of preliminary testing, each matrix (effluent for the effluent sample, receiving water for the receiving water sample, and synthetic dilution water for the reference toxicant sample) was spiked at a range of concentrations estimated to encompass the desired effect concentrations (LCSOs for acute test methods and IC25s for chronic test methods). Effluents and receiving water samples were not spiked if they possessed persistent toxicity and produced effect concentrations near the target effect level. Potassium chloride (KC1) was used as the spiking agent for freshwater methods, sheepshead acute and chronic methods, and the Mysidopsis chronic test method; copper sulfate (CuSO4) was used as the spiking agent for silverside acute and chronic test methods, Champia chronic, and Holmesimysis acute test methods. Preliminary spiking levels for Part 2 testing were determined from referee laboratory reference toxicant testing databases, literature values, or range-finding tests conducted just prior to Part 2 testing. Part 2 preliminary tests on spiked matrices were conducted as definitive tests according to the WET method manuals and specific requirements of the WET Variability Study plan (see Section 7). Part 2 testing was conducted for each test method and each sample matrix. If the results of Part 2 testing were not conclusive or if they differed greatly from historic reference toxicant testing conducted in the referee laboratory, Part 2 tests were repeated. Following Part 2 testing, appropriate spiking concentrations for interlaboratory testing were estimated. These spiking levels were estimated as the effect concentration determined in Part 2 testing divided by 25% (0.25) for effluent and receiving water samples and divided by 50% (0.5) for the reference toxicant sample. For example, if Part 2 testing for a given method determined an IC25 of 100 mg KC1/L in the effluent matrix, the final spiking concentration for the effluent sample type should be 400 mg KC1/L. This final spiking concentration was determined by dividing the Part 2 IC25 value (100 mg/L) by the target effect level of 25% (0.25) sample to obtain 400 mg KC1/L. When the final sample spiked at 400 mg KC1/L is diluted during interlaboratory testing using the standard 0.5 dilution factor (i.e., 100%, 50%, 25%, 12.5%, and 6.25% sample), the 25% sample test concentration should contain 100 mg KC1/L, and the test IC25 should be near the targeted 25% sample range. 4.3 Part 3 - Holding Time Testing Part 3 of preliminary testing determined the persistence of toxicity in the effluent and receiving water samples. Excess volume of the spiked effluent and receiving water samples was retained from Part 2 testing and stored in the dark at 4°C. Following storage for seven days, a second test was conducted using the stored sample, and results were compared to that of the initial test. If effluent or receiving water samples were not spiked, Part 3 testing was conducted on excess unspiked effluent and receiving water sample collected in Part 1. For acute and chronic test methods using the same species, Part 3 testing was conducted using only the acute test method. The results of holding time testing provided valuable information on the persistence of sample toxicity. This information was useful in the timing and scheduling of referee laboratory sample preparation for interlaboratory testing. This information also 22 ------- supported the assumption that sample toxicity remained constant when testing at a given participant laboratory was delayed due to problems with sample shipment or organism availability. 4.4 Part 4 - Final Preliminary Testing Part 4 of preliminary testing validated that the samples and spiking concentrations (if applicable) selected for the WET Variability Study achieved the desired range of effect following sample preparation, shipping, and handling. Spiking concentrations for Part 4 testing were determined from the results of Part 2 testing with necessary adjustments to meet the target effect levels. Part 4 testing also served as a trial run for sample collection, preparation, packaging, and shipment in the interlaboratory testing phase. Each sample type that was used in final preliminary testing was collected, prepared, packaged, and shipped exactly as described for interlaboratory testing (see Sections 5 and 6). Final preliminary testing samples were shipped by the referee laboratory round-trip back to the referee laboratory (e.g., sent from the referee laboratory on one day for delivery back to that same facility the next day). Upon receipt, the referee laboratory conducted final preliminary tests as described for interlaboratory testing (see Section 7). Part 4 testing was used to determine if the selected spiking concentrations achieved the targeted effect levels following sample preparation and shipment. Spiking levels that produced effect concentrations within the range of 10-35% for effluent and receiving water samples or 35-60% for reference toxicant samples were considered appropriate for use in the interlaboratory testing phase. Based on the results from Part 4 preliminary testing, SCC selected the sample composition (matrix, spiking agent, and spiking levels) for use in interlaboratory testing. Referee laboratories then prepared (see Section 5) and shipped (see Section 6) samples to participant laboratories for interlaboratory testing (see Section 7). 23 ------- 5.0 SAMPLE PREPARATION For each method, four test sample types were prepared in bulk by the referee laboratory, divided, and distributed to participant laboratories for testing. The four sample types included a blank sample, a reference toxicant sample, an effluent sample, and a receiving water sample. This section describes the preparation of these test samples for use in interlaboratory testing. The test sample types and appropriate spiking concentrations were selected based on preliminary testing conducted by the referee laboratories (see Section 4 and Appendix D). 5.1 Freshwater Methods For the freshwater methods, Table 5.1 describes each of the sample types that were prepared and distributed for interlaboratory testing by the referee laboratory (EA Engineering, Science and Technology, Inc.). The blank sample type for all freshwater test methods, with the exception of the Selenastrum chronic method, consisted of moderately hard synthetic freshwater prepared according to Section 7 of the WET method manual (USEPA, 1994a). The blank sample type was prepared by adding the appropriate amounts of reagents (Section 7, USEPA, 1994a) to deionized water in cleaned and rinsed 5-gallon or 3-gallon (depending on the volume needed for interlaboratory testing) high-density polypropylene (HDPP) carboys. Following preparation, the bulk blank sample was thoroughly mixed and aerated for at least 24 hours (as required in Section 7, USEPA, 1994a) prior to removing aliquots for packaging and distribution to participant laboratories. For the Selenastrum chronic test method, the blank sample was prepared as deionized water. The reference toxicant sample type for all freshwater test methods consisted of the blank sample matrix (moderately hard synthetic freshwater for Ceriodaphnia and fathead test methods and deionized water for the Selenastrum chronic method) spiked with KC1. Moderately hard synthetic freshwater was prepared by adding the appropriate amounts of reagents (Section 7, USEPA, 1994a) to deionized water in cleaned and rinsed 5-gallon or 3-gallon (depending on the volume needed for interlaboratory testing) HDPP carboys. The appropriate amount of reagent grade KC1 was then added to the moderately hard synthetic freshwater (or deionized water for the Selenastrum chronic method) to achieve the spiking concentrations listed in Table 5.2. When ampule samples were reconstituted according to instructions provided in the participant laboratory SOPs (Appendix B), the resulting reconstituted sample yielded the spiking concentrations listed in Table 5.1. 24 ------- Table 5.1. Description of samples used for freshwater methods in the WET Variability Study. Test method Ceriodaphnia acute Ceriodaphnia chronic Fathead acute Fathead chronic Selenastrum chronic Sample type Blank Reference toxicant Effluent Receiving water Blank Reference toxicant Effluent Receiving water Blank Reference toxicant Effluent Receiving water Blank Reference toxicant Effluent Receiving water Blank Reference toxicant Effluent Receiving water Sample matrix MHSFb MHSFb Municipal effluent River water MHSFb MHSFb Municipal effluent River water MHSFb MHSFb Municipal effluent River water MHSFb MHSFb Municipal effluent River water Deionized water Deionized water Municipal effluent River water Spiking agent None KC1 KC1 KC1 None KC1 KC1 KC1 None KC1 KC1 KC1 None KC1 KC1 KC1 None KC1 KC1 KC1 Spiked concentration3 (mg/L as KC1) - 1000 2680 1800 - 270 2100 1200 - 2200 5328 5000 - 1220 3600 2400 - 5655 11540 11713 Collection date - - 11/08/99 11/05/99 - - 10/22/99 10/08/99 - - 11/01/99 10/18/99 - - 10/01/99 09/24/99 - - 03/06/00 03/20/00 Preparation date 11/06/99 11/06/99 11/09/99 11/06/99 10/09/99 10/09/99 10/24/99 10/10/99 10/19/99 10/19/99 11/02/99 10/19/99 09/24/99 09/24/99 10/03/99 09/26/99 03/14/00 03/07/00 03/07/00 03/21/00 Shipment date 11/08/99, 11/10/99 11/08/99, 11/10/99 11/10/99 11/08/99 10/11/99, 10/25/99 10/11/99, 10/25/99 10/25/99 10/11/99 10/20/99, 11/03/99 10/20/99, 11/03/99 11/03/99 10/20/99 09/27/99, 10/04/99 09/27/99, 10/04/99 10/04/99 09/27/99 03/15/00, 03/22/00, 03/29/00 03/08/00, 03/15/00, 03/22/00, 03/29/00 03/08/00, 03/15/00 03/22/00 a Spiking concentrations are nominal values. Spiked concentrations for reference toxicant samples represent concentrations after proper reconstitution of the ampule sample. See Table 5.2 for a description of ampule sample contents. b Moderately hard synthetic freshwater prepared according to Section 7 of the WET method manual (USEPA, 1994a). 25 ------- Table 5.2. Spiking concentrations in reference toxicant ampule samples. Test method3 Ceriodaphnia acute Ceriodaphnia chronic Fathead acute Fathead chronic Selenastmm chronic Mysidopsis chronic Sheepshead acute Sheepshead chronic Silverside acute Silverside chronic Reference toxicant KC1 KC1 KC1 KC1 KC1 KC1 KC1 KC1 CuS04 CuSO, Spiked cone. in prepared ampule sample (mg/L)b 10,000 8,100 88,000 97,600 113,100 108,000 25,280 90,000 40.0 210 Volume of ampule sample added to reconstituted sample (mL) 100 100 100 100 200 100 500 500 100 100 Reconstituted sample volume (L) 1 3 4 8 4 9 4 15 4 21 Resulting cone, in reconstituted sample (mg/L)b 1,000 270 2,200 1,220 5,655 1,200 3,160 3,000 1.0 1.0 a Intel-laboratory testing was not conducted for the Champia chronic and Holmesimysis acute test methods. b Spiking concentrations are nominal values. The effluent sample type for all freshwater test methods consisted of a freshwater municipal effluent spiked with KC1. The effluent was collected from a municipal wastewater treatment plant that is designed to treat 180 mgd, is able to handle peak flows of 400 mgd, and currently treats 140 to 150 mgd. The facility employs tertiary treatment for biological nutrient removal including single-stage nitrification/ denitrification, sand filtration, chlorination/dechlorination, and anaerobic digestion. Effluent was collected at this site from a small access pipe through which effluent was pumped from the main discharge outfall. Sample was collected from this access pipe using a funnel and tubing to fill 5-gallon HDPP carboys. Sample was then immediately transported to the referee laboratory at ambient temperature. Upon arrival, the referee laboratory stored the sample in the dark at <4°C. Table 5.3 lists the volume of effluent collected for each test method. The effluent sample was then homogenized in 50- gallon HDPP containers. If the volume of the sample collected for interlaboratory testing was greater than 50 gallons for a given test method, the sample was stored and homogenized in multiple 50-gallon containers connected with piping. Submersible pumps were used to circulate and homogenize the sample among individual containers. Following homogenization, the effluent sample for each test method was spiked at the appropriate concentration with KC1 (Table 5.1). The appropriate amount of KC1 was initially dissolved in a small volume of effluent and then added to the bulk effluent sample. When multiple containers were used, equal amounts of KC1 were added to each container. The receiving water sample type for all freshwater test methods consisted of a natural river water spiked with KC1. The receiving water was collected from the Gunpowder River near Falls Road, in Baltimore County, Maryland. Receiving water was collected by filling 1-gallon HDPP containers under the surface of the water. Receiving water sample was then transferred to 5-gallon HDPP containers and immediately transported to the referee laboratory at ambient temperature. Upon arrival at the referee laboratory, the sample was stored at <4°C in the dark. Table 5.4 lists the volume of receiving water collected for each 26 ------- test method. The receiving water sample was then homogenized and spiked as described in the preceding paragraph for the effluent sample type. Bulk samples for all sample types were prepared in cleaned and rinsed containers. Containers were cleaned with detergent, hydrochloric acid, acetone, and rinsed with deionized water. All bulk sample preparations were mixed thoroughly prior to spiking, following spiking, and prior to removing aliquots for distribution to participant laboratories. All bulk samples were stored in the dark at <4°C prior to shipment to participant laboratories. Table 5.3. Effluent sample volumes collected for interlaboratory testing. Test method3 Ceriodaphnia acute Ceriodaphnia chronic Fathead acute Fathead chronic Selenastrum chronic Mysidopsis chronic Sheepshead acute Sheepshead chronic Silverside acute Silverside chronic Volume required per sample (L) 1 7 4 17.5 4 21 4 35 4 49 Number of participant lab samples required 28 28 28 28 16 16 8 8 13 14 Minimum required volume (L)b 28 196 112 490 64 336 32 280 52 686 Collected volume (L) 45 270 150 660 100 450 60 375 60 1000 a Interlaboratory testing was not conducted for the Champia chronic and Holmesimysis acute test methods. b Minimum required volume = Volume per sample X number of samples. Table 5.4. Receiving water sample volumes collected for interlaboratory testing. Test method3 Ceriodaphnia acute Ceriodaphnia chronic Fathead acute Fathead chronic Selenastrum chronic Mysidopsis chronic Sheepshead acute Sheepshead chronic Silverside acute Silverside chronic Volume required per sample (L) 1 7 4 17.5 4 21 4 35 4 49 Number of participant lab samples required 14 14 14 14 9 9 8 8 7 8 Minimum required volume (L)b 14 98 56 245 36 189 32 280 28 392 Collected volume (L) 19 135 90 345 60 278 75 389 36 540 a Interlaboratory testing was not conducted for the Champia chronic and Holmesimysis acute test methods. b Minimum required volume = Volume per sample X number of samples. 27 ------- 5.2 Marine Methods For the marine methods, Table 5.5 describes each of the sample types that were prepared and distributed for interlaboratory testing by the referee laboratory (EA Engineering, Science and Technology, Inc. for the Mysidopsis chronic and sheepshead acute and chronic test methods; Ogden Environmental and Energy Services, Inc. for the silverside acute and chronic test methods). 5.2.1 Mysidopsis Chronic and Sheepshead Acute and Chronic Test Methods The blank sample type for the Mysidopsis chronic, sheepshead acute, and sheepshead chronic test methods consisted of synthetic seawater prepared at a salinity of 25 ppt. The blank sample type was prepared by adding the appropriate amounts of bioassay grade Forty Fathoms® artificial sea salts to deionized water in cleaned and rinsed 5-gallon or 3-gallon (depending on the volume needed for interlaboratory testing) HDPP carboys. Following preparation, the bulk blank sample was thoroughly mixed to dissolve the added reagents. The reference toxicant sample type for the Mysidopsis chronic, sheepshead acute, and sheepshead chronic test methods consisted of deionized water spiked with KC1. For these test methods, deionized water rather than artificial seawater was spiked with KC1 to improve the solubility of KC1 in the highly concentrated reference toxicant ampule samples. The appropriate amount of reagent grade KC1 was added to deionized water in cleaned and rinsed 5-gallon or 3-gallon (depending on the volume needed for interlaboratory testing) HDPP carboys to achieve the spiking concentrations listed in Table 5.2 for the reference toxicant samples. When ampule samples were reconstituted according to instructions provided in the participant laboratory SOPs (see Appendix B), the resulting reconstituted sample yielded the spiking concentrations listed in Table 5.5. The effluent sample type for the Mysidopsis chronic, sheepshead acute, and sheepshead chronic test methods consisted of a municipal effluent spiked with KC1. The municipal effluent described previously for freshwater methods (see Section 5.1) also was used for the Mysidopsis chronic, sheepshead acute, and sheepshead chronic test methods. The effluent was collected and transported to the referee laboratory as previously described for the freshwater methods (see Section 5.1). Table 5.3 lists the volume of effluent collected for each test method. The effluent sample was then homogenized in 50-gallon HDPP containers. If the volume collected and necessary for interlaboratory testing was greater than 50 gallons for a given test method, the sample was stored and homogenized in multiple 50-gallon containers connected with piping. Submersible pumps were used to circulate and homogenize sample among individual containers. Following homogenization, the salinity of the effluent sample was adjusted by the addition of bioassay grade Forty Fathoms® artificial sea salts. Salinity of the effluent sample was adjusted to 23 ppt, 20 ppt, and 21 ppt for the Mysidopsis chronic, sheepshead acute, and sheepshead chronic test methods, respectively. Since the addition of the KC1 spike increased salinity, the initial salinity levels were selected to achieve a final salinity of 25 ppt in the effluent sample following spiking with KC1. Following salinity adjustment, the effluent sample was spiked at the appropriate concentration with KC1 (Table 5.5). The appropriate amount of KC1 was initially dissolved in a small volume of 28 ------- Table 5.5. Description of samples used for marine methods in the WET Variability Study. Test method3 Mysidopsis chronic Sheepshead acute Sheepshead chronic Silverside acute Silverside chronic Sample type Blank Reference toxicant Effluent Receiving water Blank Reference toxicant Effluent Receiving water Blank Reference toxicant Effluent Receiving water Blank Reference toxicant Effluent Receiving water Blank Reference toxicant Effluent Receiving water Sample matrix synthetic seawater0 synthetic seawater0 municipal effluent natural seawater synthetic seawater0 synthetic seawater0 municipal effluent natural seawater synthetic seawater0 synthetic seawater0 municipal effluent natural seawater synthetic seawater0 synthetic seawater0 industrial effluent natural seawater synthetic seawater0 synthetic seawater0 industrial effluent natural seawater Spiking agent None KC1 KC1 KC1 None KC1 KC1 KC1 None KC1 KC1 KC1 None CuSO4 CuSO4 CuSO4 None CuSO4 CuSO4 CuSO4 Spiked concentrationb (mg/L as KC1 and as Cu) - 1200 2000 2400 - 3160 5200 5600 - 3000 4880 4400 - 1.000 0.922 0.565 - 1.000 0.800 0.494 Collection date - - 02/16/00 02/24/00 - - 03/06/00 02/24/00 - - 03/23/00 03/17/00 - - 11/03/99 10/30/99 - - 10/20/99 10/16/99 Preparation date 02/18/00 02/18/00 02/18/00 02/25/00 03/03/00 03/10/00 03/10/00 03/03/00 03/17/00 03/25/00 03/24/00 03/17/00 10/31/99 10/31/99 11/07/99 10/31/99 10/17/99 10/17/99 10/23/99 10/17/99 Shipment date 02/21/00, 02/28/00 02/21/00, 02/28/00 02/21/00 02/28/00 03/06/00 03/13/00 03/13/00 03/06/00 03/20/00 03/27/00 03/27/00 03/20/00 11/01/99, 11/08/99 11/01/99, 11/08/99 11/08/99 11/01/99 10/18/99, 10/25/99 10/18/99 10/25/99 10/18/99 a Intel-laboratory testing was not conducted for the Champia chronic and Holmesimysis acute test methods. b Spiking concentrations are nominal values. Spiked concentrations for reference toxicant samples represent concentrations after proper reconstitution of the ampule sample. See Table 5.2 for a description of ampule sample contents. c Synthetic seawater was prepared using bioassay grade Forty Fathoms® added to deionized water. 29 ------- effluent and then added to the bulk effluent sample. When multiple containers were used, equal amounts of KC1 were added to each container. The receiving water sample type for the Mysldopsls chronic, sheepshead acute, and sheepshead chronic test methods consisted of a natural seawater spiked with KC1. Receiving water was collected from Manasquan Inlet, in Manasquan, Monmouth County, New Jersey. Seawater was collected within approximately 1 hour of high tide using a submersible pump, and transferred into 5-gallon HDPP carboys for transport. The seawater sample collected for each test method was filtered through a 5-(im filter either at the time of collection or upon receipt at the referee laboratory. The receiving water sample was transported to the referee laboratory at ambient temperature. Upon arrival, the referee laboratory stored the sample in the dark at <4°C. Table 5.4 lists the volume of receiving water collected for each test method. The receiving water sample was then homogenized and spiked as described in the preceding paragraph for the effluent sample type. Bulk samples for all sample types were prepared in cleaned and rinsed containers. Containers were cleaned with detergent, hydrochloric acid, acetone, and rinsed with deionized water. All bulk sample preparations were mixed thoroughly prior to spiking, following spiking, and prior to removing aliquots for distribution to participant laboratories. All bulk samples were stored in the dark at <4°C prior to shipment to participant laboratories. 5.2.2 Silverside Acute and Chronic Test Methods The blank sample type for the silverside acute and chronic test methods consisted of synthetic seawater prepared at a salinity of 25 ppt. The blank sample type was prepared by adding the appropriate amounts of bioassay grade Forty Fathoms® artificial sea salts to deionized water in polycarbonate carboys. Following preparation, the bulk blank sample was thoroughly mixed to dissolve the added reagents. The reference toxicant sample type for the silverside acute and chronic test methods consisted of the blank sample matrix (synthetic seawater at 25 ppt salinity) spiked with CuSO4. The appropriate volume of a concentrated CuSO4 stock solution was added to synthetic seawater in 10-L polycarbonate containers to achieve the spiking concentrations listed in Table 5.2 for the reference toxicant ampule sample. When ampule samples were reconstituted according to instructions provided in the participant laboratory SOPs, the resulting reconstituted sample yielded the spiking concentrations listed in Table 5.5. The effluent sample type for the silverside acute and chronic test methods consisted of an industrial effluent spiked with CuSO4. The effluent was collected from an industrial wastewater treatment facility that treats wastes from an oil refinery. Effluent was collected using an automatic sampler set to collect a single grab sample. Sample was collected and stored in 5-gallon buckets with 4mm polyethylene liners and plastic lids for transport to the referee laboratory. Upon arrival, the referee laboratory stored the sample in the dark at <4°C. Table 5.3 lists the volume of effluent collected for each test method. The effluent sample was then homogenized in 250-L polycarbonate containers using a mechanical mixer. The volume required for the silverside chronic test method necessitated the preparation of separate batches of 30 ------- effluent sample prior to shipment of test samples for test initiation, the first renewal shipment, and the second renewal shipment. For each batch, it was necessary to mix and hold sample in two 250-L containers. The sample was homogenized among the two containers by repeatedly (six times) adding 40% of one container to the other and mixing each time. For the silverside acute test method, all effluent sample was mixed and prepared in a single 250-L container. Following homogenization, the salinity of the effluent sample was adjusted to 25 ppt by the addition of bioassay grade Forty Fathoms® artificial sea salts. The effluent sample was then spiked by adding the appropriate volume of a concentrated CuSO4 stock solution to achieve the spiking concentrations listed in Table 5.5. The receiving water sample type for the silverside acute and chronic test methods consisted of a natural seawater spiked with CuSO4. Natural seawater was collected from the Scripps Institution of Oceanography seawater system in La Jolla, CA. Seawater was pumped from a fixed collection site 320m offshore of La Jolla, CA, filtered through a sand filter, and trucked to the referee laboratory. At the referee laboratory, the seawater was incorporated into the laboratory's flow-through seawater system that includes two 2,200-gallon storage tanks, an in-line 20-(im filter, and an in-line heater/chiller unit. The salinity of receiving water (initially 34 ppt) was adjusted to 25 ppt with the addition of deionized water. The bulk receiving water sample then was homogenized in 250-L polycarbonate containers and spiked by adding the appropriate volume of a concentrated CuSO4 stock solution to achieve the spiking concentrations listed in Table 5.5. Bulk samples for all sample types were prepared in cleaned and rinsed containers. Containers were cleaned with detergent, rinsed with tap water, then deionized water, and rinsed again with sample. All bulk sample preparations were mixed thoroughly prior to spiking, following spiking, and prior to removing aliquots for distribution to participant laboratories. All bulk samples were stored in the dark at <4°C prior to shipment to participant laboratories. 5.3 Problems Encountered in Sample Preparation The reference toxicant sample prepared for the Ceriodaphnia chronic test method produced toxicity in only some of the participant laboratories (see Section 9.3). The referee laboratory also did not detect toxicity in this sample during the interlaboratory testing phase. It was determined that the reference toxicant sample prepared for interlaboratory testing was spiked at a level that was only slightly toxic and very near the minimum detection level (100% sample). Depending on the sensitivity of test organisms at individual laboratories, some laboratories identified the sample as toxic, while other laboratories did not. The spiking level used for this sample was based on preliminary testing results from three tests that indicated an IC25 of 138, 132, and 134 mg KC1/L. The average of these tests (135 mg KC1/L) was multiplied by two to obtain the target spiking level of 270 mg KC1/L. It should be noted that one of the preliminary tests produced an IC25 of 320 mg/L KC1, indicating that 135 mg KC1/L may have been a low estimate of the IC25 for KC1. The reference toxicant sample for the silverside acute test also did not produce a toxic response in interlaboratory testing. This was caused by precipitation of the spiked copper in the liquid ampule 31 ------- sample. Precipitation of copper in seawater samples also was observed in preliminary testing for the Mysidopsis and sheepshead test methods, so the spiking agent for these methods was changed (see Appendix D, Section D.9). For the silverside methods, this potential problem was not identified during preliminary testing so no change in the spiking agent was made prior to interlaboratory testing. Part 4 preliminary testing of the reference toxicant sample did not reveal copper precipitation and produced an LC50 of 0.29 mg Cu/L. The same spiking levels used in Part 4 testing were used for the interlaboratory sample, yet results for that sample showed precipitation and no toxicity. The referee laboratory discovered that in Part 4 preliminary testing, the reference toxicant ampule sample was prepared using deionized water, but the interlaboratory sample was prepared using synthetic seawater. To confirm that this difference in sample matrix resulted in the non-toxic interlaboratory sample, the referee laboratory prepared two additional reference toxicant samples at the same spiking level used for interlaboratory testing. One sample was prepared using deionized water and one was prepared using synthetic seawater. Testing results from the sample prepared using deionized water were consistent with Part 4 preliminary testing results (LC50 of 25.5% sample), and results from the ampule sample prepared using synthetic seawater were consistent with interlaboratory testing results (LC50 >100% sample). These additional tests confirmed that preparation of the reference toxicant ampule sample using synthetic seawater caused the precipitation of the spiked copper and produced a nontoxic sample for interlaboratory testing. 32 ------- 6.0 PACKAGING AND DISTRIBUTION OF TEST SAMPLES 6.1 Sample Distribution Scheme Laboratories participating in the base study design (see Sections 2.2.3 and 3.5) each received four blind test samples; laboratories participating in the extended study design each received three blind test samples. The sample distribution scheme in Table 6.1 shows the number and type of samples distributed to each laboratory according to the laboratory's assigned position in the study. EPA-sponsored laboratories were randomly assigned among positions 1-9, non-EPA sponsored laboratories in the base study design were randomly assigned among positions 10-20, and laboratories participating in the extended study design were randomly ordered in positions 21- the total number of laboratories. As noted in Table 6.1, an alternate sample distribution scheme was used for the sheepshead acute and chronic test methods. Since fewer than nine laboratories participated in these methods, the sample distribution scheme would not have met the data quality objective of obtaining six data sets for each test method and sample type. For these test methods, one of each of the sample types was distributed to each of the seven participant laboratories. 6.2 Packaging and Shipment of Samples After bulk test samples were prepared according to Section 5, each bulk test sample was divided into individual test sample aliquots for shipment to participant laboratories. Test sample aliquots were divided into containers appropriate for the individual test sample volumes. High-density polypropylene sample containers, of the appropriate size (refer to Tables 6.2 and 6.3) were pre-rinsed, filled with the sample, and sealed with zero head-space. All blank and reference toxicant samples were prepared and packaged as ampule samples. Ampule samples were small volume (generally 100 mL) liquid samples that were reconstituted at participant laboratories to provide the necessary test sample volume. For the Selenastrum chronic test and the sheepshead acute and chronic tests, larger volumes (200 mL, 500 mL, and 500 mL, respectively) were used for ampule samples to reduce the concentration of KC1 in the reference toxicant ampule sample and avoid potential solubility problems. All ampule samples (blank and reference toxicant sample types) for a given test method were shipped in the same container style and size. Effluent and receiving water samples were prepared and packaged as whole volume samples in HDPP cubitainers. All whole volume samples for a given test method were shipped in the same container style and size. Tables 6.2 and 6.3 show the volumes and numbers of samples prepared for freshwater and marine test methods, respectively. For chronic test methods that required daily renewal, samples were packaged and shipped in three separate aliquots. The first aliquot (Initiation) was received by the participant laboratory on test Day 0 and was used for test initiation on Day 0 and test renewal on Day 1. The second aliquot (Renewal 1) was received on test Day 2 and used for test renewals on Day 2 and Day 3. The final aliquot (Renewal 2) was received on test Day 4 and used for test renewal on Day 4, Day 5, Day 6, and Day 7 if necessary (for the Ceriodaphnia chronic test method). For ampule samples, all three separate aliquots were received on test Day 0, and participant laboratories were instructed to reconstitute the samples following the scheme 33 ------- described above (separate aliquots reconstituted on test Days 0, 2, and 4). For all acute test methods, a single aliquot of sample was received and used for test initiation and any required renewals. Table 6.1. Sample distribution scheme for the WET Variability Study." Participant laboratory assigned position Base study design EPA- sponsored Base study design non-EPA sponsored Extended study design non-EPA sponsored 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 -up (odd #s) 21 -up (even #s) Number of samples of each type received Blank 1 1 1 1 0 1 0 1 0 1 1 1 0 0 1 1 0 0 1 1 2 1 Effluent 1 2 0 1 2 1 2 1 2 1 2 0 2 2 2 0 2 2 2 0 0 0 Receiving water 1 1 1 0 1 0 1 0 1 1 1 1 0 0 1 1 0 0 1 1 0 0 Reference toxicant 1 0 2 2 1 2 1 2 1 1 0 2 2 2 0 2 2 2 0 2 1 2 a This sample distribution scheme was used for interlaboratory testing of all test methods except the sheepshead acute and chronic test methods. For these test methods, one of each of the sample types was distributed to each of the seven participant laboratories. Samples were cooled to <4°C prior to shipment and then packed in coolers (e.g., 28, 48, 54-qt) containing wet ice. Depending on the test method performed by an individual participant laboratory, multiple test samples were shipped in one cooler if possible to reduce the number of coolers shipped. Test sample volumes that exceeded the maximum weight limit for overnight shipping were divided into separate coolers for shipment. Duplicate test sample aliquots were shipped in the same cooler whenever possible; if test sample volume prohibited shipping duplicates in the same cooler, they were shipped under the same airbill to ensure they were shipped together. All samples were shipped FedEx Priority Overnight for delivery on the day of scheduled testing (see Section 2.2.4). Referee laboratories conducted testing 34 ------- simultaneously with participant laboratories on samples prepared identically to those for interlaboratory testing and shipped round-trip back to the referee laboratory. 6.3 Sample Tracking Each WET test method received an EPA episode number to designate samples prepared for that test method. Each sample aliquot that was prepared and shipped was assigned a unique sample number and was accompanied by an EPA traffic report form. Duplicate samples received different sample numbers to retain the blind sample aspect of the study design. For chronic test methods that required additional shipments for sample renewal, the sample number remained the same for each initiation and renewal shipment with the addition of a letter (A, B, and C) after the sample number to designate the sample for use as Initiation (A), Renewal 1 (B), or Renewal 2 (C). The sample number was clearly and permanently marked on each container and the accompanying EPA traffic report form. Sample numbers for each test method are given in Table 6.4. Following completion of the study, each test sample was assigned a sample code (in addition to the sample number previously assigned) as an alternate unique identifier. For the results section of this report, samples are identified by sample codes to aid in blinding the identity of individual participant laboratories. Referee laboratories included an EPA traffic report form with each sample that was shipped to document the chain-of-custody for that sample. The traffic report form (see Appendix B) identified the episode number, sample number, name and address of the referee laboratory, name and address of the participant laboratory, date shipped, airbill number, tests requested, and pre-shipment sample information (sample preparation date and initial water chemistry). A traffic report form specific to each sample was placed in a waterproof enclosure (e.g., Ziploc® bag) and packed in the cooler with the respective sample. Each cooler used in the study was permanently numbered and labeled (with the referee laboratory name and address) to assist in locating lost coolers and to assist in retrieving coolers from participant laboratories. For each shipment event, the referee laboratory also completed a sample shipment documentation form. The referee laboratory faxed this form to SCC immediately after sample pickup by FedEx. The sample shipment form documented the following information for each shipping event: • Sample number - the unique identifying number for each sample aliquot Sample description - identified the sample as either blank, spiked effluent, spiked receiving water, or reference toxicant • Participant laboratory name - name of the laboratory to which the sample was shipped • Airbill number - the overnight shipping service's number that identified each individual shipment Size of test containers - the volume of the test container in which the sample was shipped Cooler number - a unique identifying number for the cooler in which the sample was shipped. • Comments - any miscellaneous information related to sample shipment. 35 ------- Table 6.2. Number and volume of samples required for freshwater methods in the WET Variability Study. Test method Ceriodaphnia acute Ceriodaphnia chronic Fathead acute Fathead chronic Selenastrum chronic Sample type Blank Reference toxicant Effluent Receiving water Blank Reference toxicant Effluent Receiving water Blank Reference toxicant Effluent Receiving water Blank Reference toxicant Effluent Receiving water Blank Reference toxicant Effluent Receiving water Ampule or whole volume sample Ampule Ampule Whole volume Whole volume Ampule Ampule Whole volume Whole volume Ampule Ampule Whole volume Whole volume Ampule Ampule Whole volume Whole volume Ampule Ampule Whole volume Whole volume Sample volume3 100 mL 100 mL 1L 1L 100 mL 100 mL 2-3L 2-3L 100 mL 100 mL 4L 4L 100 mL 100 mL 5-7.5L 5-7.5L 200 mL 200 mL 4L 4L Number of samples requiredb Initiation 34 32 28 14 35 49 28 14 28 39 30 14 25 38 28 14 9 14 16 9 Renewal 1 - - - - 35 49 28 14 - - - - 25 38 28 14 - - - - Renewal 2 - - - - 35 49 28 14 - - - - 25 38 28 14 - - - - Total 34 32 28 14 105 147 84 42 28 39 30 14 75 114 84 42 9 14 16 9 a For chronic test methods, sample volumes ranged depending upon the aliquot. More volume was required for the Renewal 2 aliquot than the Initiation or Renewal 1 aliquots. b Number of samples includes samples shipped round-trip back to referee laboratory. 36 ------- Table 6.3. Number and volume of samples required for marine methods in the WET Variability Study. Test method3 Mysidopsis chronic Sheepshead acute Sheepshead chronic Silverside acute Silverside chronic Sample type Blank Reference toxicant Effluent Receiving water Blank Reference toxicant Effluent Receiving water Blank Reference toxicant Effluent Receiving water Blank Reference toxicant Effluent Receiving water Blank Reference toxicant Effluent Receiving water Ampule or whole volume sample Ampule Ampule Whole volume Whole volume Ampule Ampule Whole volume Whole volume Ampule Ampule Whole volume Whole volume Ampule Ampule Whole volume Whole volume Ampule Ampule Whole volume Whole volume Sample volume1" 100 mL 100 mL 6-9L 6-9L 500 mL 500 mL 4L 4L 500 mL 500 mL 10-15L 10-15L 100 mL 100 mL 4L 4L 100 mL 100 mL 14-21L 14-21L Number of samples required0 Initiation 9 14 16 9 8 8 8 8 8 8 8 8 7 13 13 7 8 14 14 8 Renewal 1 9 14 16 9 - - - - 8 8 8 8 - - - - 8 14 14 8 Renewal 2 9 14 16 9 - - - - 8 8 8 8 - - - - 8 14 14 8 Total 27 42 48 27 8 8 8 8 24 24 24 24 7 13 13 7 24 42 42 24 a Interlaboratory testing was not conducted for the Champia chronic and Holmesimysis acute test methods. b For chronic test methods, sample volumes ranged depending upon the aliquot. More volume was required for the Renewal 2 aliquot than the Initiation or Renewal 1 aliquots. c Number of samples includes samples shipped round-trip back to referee laboratory. 37 ------- SCC entered shipment information into a database and used this information on the day of expected sample arrival to track the delivery of each sample through the FedEx automated shipment tracking system. If sample shipment problems were encountered, SCC notified participant laboratories of the problem and instructed the laboratory how to proceed. The instructions provided by SCC were specific to the individual case, but laboratories were generally instructed to initiate testing on the day of sample arrival if the sample was delivered prior to the close of business. If the delivery of renewal samples was delayed, laboratories were instructed to renew the test on time with remaining sample from the previous shipment. If sample shipments could not be located or if tracking did not indicate progress of the sample, the referee laboratory was instructed to resend the sample (prepared from the remaining bulk sample) for delivery the following day. Upon receipt of each sample, participant laboratories were responsible for determining that the sample arrived in satisfactory condition and for documenting receipt of the sample, post-shipment sample water quality (temperature, pH, dissolved oxygen, and conductivity or salinity), and any problems on the EPA traffic report form. Laboratories faxed the completed traffic report form to SCC immediately upon sample receipt and retained a copy for inclusion in the data report. SCC's faxed receipt of the completed traffic report form from the participant laboratory served as the notification that the sample had arrived in good condition at the participant laboratory. For ampule samples, participant laboratories were not required to measure post-shipment sample water quality. To avoid possible contamination between the highly concentrated reference toxicant ampule samples and blank samples, no direct measurements were made on the ampule samples. Temperature was measured in a temperature check sample that was included with each cooler containing ampule samples. This sample, which was clearly marked as a temperature check, contained tap water in the same volume and container as ampule samples. Table 6.4. Episode numbers and sample numbers used in the WET Variability Study. Test method Ceriodaphnia acute Ceriodaphnia chronic Fathead acute Fathead chronic Selenastrum growth Mysidopsis chronic Sheepshead acute Sheepshead chronic Silverside acute Silverside chronic Episode number 6207 6208 6205 6206 6209 6216 6214 6215 6210 6211 Sample number range 03000-03107 04000-04128 01000-01110 02000-02104 05000 - 05047 12000 - 12047 10000- 10031 11000-11031 06000-06041 07000 - 07043 Sample code range 9217 - 9324 9325 - 9453 9001-9111 9112-9216 9454 . 9501 9650 - 9697 9586 - 9617 9618-9649 9502 - 9541 9542 - 9585 38 ------- 6.4 Problems Encountered in Sample Distribution On the first shipment day for the Ceriodaphnia acute test method (11/08/99), the referee laboratory inadvertently switched the blank and reference toxicant ampule samples. All participant laboratories that were intended to receive blank samples received reference toxicant samples, and all laboratories that were intended to receive reference toxicant samples received blank samples. This error was not noticed at the time, and the referee laboratory incorrectly reported sample types on the sample shipment documentation form faxed to SCC. Since all samples were received by participant laboratories as blind test samples, this error had no effect on the participant laboratory testing of samples. This error did alter the intended sample distribution scheme (Table 6.1); however, the number of blank and reference toxicant samples were approximately equal so effects on the study design were minimal. The error was first identified by SCC as a result of participant laboratory data reports. Test results and conductivity measurements on the two samples indicated that the samples were switched. Since KC1 was used as the reference toxicant, conductivity measurements were used to properly identify the blank and reference toxicant samples. Conductivity of the reconstituted blank sample was approximately 300 //mhos/cm, and conductivity of the reconstituted reference toxicant sample was approximately 2900 ^mhos/cm. The referee laboratory determined that the error was caused by inadvertently filling ampules intended for blank samples with the reference toxicant bulk sample and filling ampules intended for reference toxicant samples with the blank bulk sample. Due to weather or other circumstances, sample shipments occasionally failed to arrive at the participant laboratory on time. Of the 1438 sample aliquots shipped in the WET Variability Study, 1412 (or 98%) successfully arrived on the intended delivery date. Tests that were initiated on samples greater than 36 hours old are identified with a data qualifier flag in the "Results" section of this report (Section 9). No participant laboratory tested samples that were greater than 72 hours old. Samples also occasionally arrived at participant laboratories at temperatures above 4°C. Only 7.4% of the 1438 sample aliquots arrived at above the recommended sample shipment temperature. This also had little effect on test results in the WET Variability Study since sample characteristics were known and selected toxicants were not likely to be altered as a result of slight temperature fluctuations. On seven occasions, participant laboratories noted that sample labels were smeared by melting ice in the coolers and were difficult to read. This problem was addressed by placing the ice in double-lined plastic bags within the coolers. On all occasions the sample numbers were identified using the accompanying traffic report form. Participant laboratories then re-labeled the sample. 39 ------- 7.0 INTERLABORATORY TESTING Interlaboratory testing was conducted to obtain data from multiple laboratories on the same test sample. These data were used to evaluate the performance of the WET test methods. Prior to interlaboratory testing of each test method, SCC provided participant laboratories with method-specific SOPs documenting participant laboratory requirements (see Appendix B). These SOPs described the shipment and tracking of test samples, provided instructions for any necessary pre-test sample adjustments (including preparation of reconstituted ampule samples), provided general and method-specific testing requirements as described in Section 7.1 and 7.2, and described data reporting requirements. Participant laboratories were then provided with samples (prepared according to Section 5) for immediate testing. Interlaboratory testing was conducted according to the schedule provided in Section 2.2.4 for the Ceriodaphnia acute and chronic, fathead acute and chronic, Selenastrum chronic, Mysidopsis chronic, sheepshead acute and chronic, and silverside acute and chronic test methods. Interlaboratory testing was canceled for the Champia chronic and Holmesimysis acute test methods (see Section 2.1). 7.1 General Testing Requirements Except where indicated in the SOPs provided to participant laboratories, each test was conducted in accordance with the general guidance and method-specific requirements for effluent testing included in the WET methods manuals. Additional general WET test requirements that were listed in participant laboratory SOPs are provided below: (1) Tests were required to be conducted by the same laboratory personnel that routinely conduct WET tests at that laboratory facility and who were identified in the prequalification materials. The laboratory was required to contact SCC if these individuals could not be available during any part of the study. Personnel conducting the tests were to be identified clearly and consistently in records. (2) To coordinate testing at participant laboratories, testing of each sample with each method was required to be initiated on the precise day specified in the study schedule. The study schedule was distributed to participating laboratories prior to commencement of each study round and in ample time to prepare for testing. Laboratories were required to test samples within 36 hours from the time of sample preparation (determined in the WET Variability Study as the time at which individual sample aliquots were divided from the bulk test sample for distribution to participant laboratories). Laboratories were required to report deviations from the study schedule to SCC immediately for approval. (3) Laboratories were required to conduct tests within the physical and chemical water quality ranges specified in the study plan, the SOW, specific instructions, and the methods manuals. Method- specific instructions for any adjustments to the test samples prior to sample use (such as reconstitution of ampule samples or salinity adjustments) were provided to the testing laboratories prior to test initiation. Laboratories were required to refrigerate (at 4°C ± 2°C) test samples 40 ------- immediately upon receipt and throughout the period of testing. Routine or continuous monitoring of refrigeration temperature was recommended to ensure that these sample holding requirements were met. (4) Laboratories were required to measure test conditions (pH, conductivity or salinity, total alkalinity, total hardness, and dissolved oxygen) in each test according to guidance in the WET method manuals. (5) Laboratories were required to use the dilution and control waters specified in Tables 7.1 -7.12 for each test method. Laboratories were required to prepare these dilution waters according to instructions in Section 7 of the method manuals. For marine test methods, laboratories were required to prepare dilution waters that meet the salinity ranges specified in Tables 7.6-7.12. (6) Laboratories were required to conduct all tests as definitive tests with a control and a minimum of five test concentrations prepared using a dilution factor of 0.5. (7) Laboratories were required to conduct all tests using the number of replicates and number of test containers per concentration as specified in Tables 7.1 - 7.12. (8) For a given test method, laboratories were required to use the same type, size, shape, and material for all test chambers. The test chamber material used had to be allowed by the WET method manuals. (9) Laboratories were required to randomize test vessels in accordance with the WET method manuals. In addition, block randomization and use of known parentage were required for the Ceriodaphnia chronic test method as described in the method manual. The Agency plans to amend Method 1002.0 (Ceriodaphnia chronic test method) to require that test organisms be allocated among test replicates so that offspring of each female are evenly distributed among test replicates ("blocking by known parentage"). (10) While the method manual requires the termination of the Ceriodaphnia chronic test after the production of three broods in 60% of the controls, laboratories were required to conduct the Ceriodaphnia chronic test for eight days. Laboratories were required to record the survival, number of young per day, and number of broods at the end of Day 6, 7, and 8 (specifically, at 144 hours, at 168 hours, and at 192 hours, respectively, from test initiation). This was done to assess the effect of the three brood test acceptance criterion on test results. No test was allowed to be terminated prior to Day 8 for any reason, including a failure to meet test acceptance criteria. Laboratories were required to include the additional measurements on Days 6, 7, and 8 as raw data in the final data report. However, laboratories were required to analyze data from the Ceriodaphnia chronic test using the three brood approach as specified in the method manual. 41 ------- (11) Laboratories were required to conduct the Selenastrum chronic test simultaneously with and without EDTA for each sample. For laboratories participating in the base study design (refer to Sections 2.2.3 and 3.5), four samples were tested with and without EDTA for a total of eight analyses. (12) Laboratories were required to observe mortality and remove dead organisms in each test daily, except for the Selenastrum chronic and Champia chronic test methods. For the Ceriodaphnia chronic test method, laboratories were required to count young daily and determine the number of broods at each count. (13) Laboratories were required to contact SCC immediately if test results indicated extreme toxicity (i.e., control mortalities, or complete mortality in all concentrations). Laboratories were then required to investigate possible causes, first by checking for transcription and calculation mistakes, and then by investigating possible contamination in dilution waters, organism cultures, equipment, or other procedural steps. (14) If any initiated test failed to be completed for any reason, the laboratory was required to contact SCC immediately for problem resolution and scheduling of additional testing. In this case, laboratories were required to report the incomplete test data and fully document the reason for not completing the test. (15) Laboratories were required to report all data obtained during the course of testing, including the response of organisms in control treatments. (16) Laboratories were required to perform all QA/QC tests listed in Section 4 of the WET method manuals. Laboratories that purchased organisms were required to supply QA/QC from the test organism supplier and follow WET method manuals for the appropriate QA/QC for purchasing organisms. (17) Laboratories were required to perform a reference toxicant QC test for each test method in the month that interlaboratory testing occurred. Results of this test had to be submitted with the final data package. (18) Laboratories were required to submit hard copies of all data from laboratory bench sheets and statistical analyses, including but not limited to all bench sheets, raw data, sample tracking forms, and chemical analysis data. Laboratories also were required to submit data electronically according to the electronic template (Microsoft Excel® spreadsheet) that was provided by SCC prior to test initiation. (19) Laboratories were required to analyze data in accordance with the statistical programs specified in the WET method manuals. Statistical methods and programs used had to be reported along with sample calculations. 42 ------- (20) Laboratories were required to report a LC50 for each acute test. A NOEC and LC50 for survival, and a NOEC and IC25 for growth/reproduction were required as appropriate for each short-term chronic test as described in the method manuals and Table 2.3 of this report. Laboratories were required to report individual toxicity results and were not allowed to average or perform other data manipulations unless required by the WET method manual. 7.2 Method-Specific Requirements EPA acknowledges that the promulgated WET methods distinguish between requirements (indicated by the compulsory terms "must" and "shall") and recommendations and guidance (indicated by discretionary terms "should" and "may"). The latter terms indicate that the analyst has flexibility to optimize successful test completion. Additionally, the WET method manuals allow variations of the methods that are typically fixed in the permit; therefore, for the purposes of this study, a set of test condition variables were defined by EPA (for example, dilution water, salinity, and acute test duration). The summary of test conditions for the 12 WET methods evaluated in the WET Variability Study are provided in Tables 7.1 -7.12. These tables were extracted from the summary test condition tables in the WET method manuals and modified to fit the scope of this study. Items shown in bold italic in these tables represent conditions standardized for the purposes of this study where WET method manuals provide a range. These test conditions were reiterated in participant laboratory SOPs (Appendix B). 43 ------- Table 7.1. Summary of test conditions and test acceptability criteria for the Ceriodaphnia acute test method. 1. Test type: 2. Test duration: 3. Temperature: 4. Light quality: 5. Light intensity: 6. Photoperiod: 7. Test chamber size: 8. Test solution volume: 9. Renewal of test solutions: 10. Age of test organisms: 11. No. organisms per test chamber: 12. No. replicate chambers per concentration: 13. No. organisms per concentration: 14. Feeding regime: 15. Test chamber cleaning: 16. Test chamber aeration: 17. Dilution water: 18. Test concentrations: 19. Dilution factor: 20. Endpoint: 21. Sample handling and holding requirements: 22. Sample volume required: 23. Test acceptability criterion: Static non-renewal 48 h 25±1°C Ambient laboratory illumination 10-20 uE/m2/s (50-100 ft-c) (ambient laboratory levels) 16 h light, 8 h darkness 30 mL 15 mL None Less than 24-h old 5 4 20 Feed YCT and Selenastrum while holding prior to the test; newly-released young should have food available a minimum of 2 h prior to use in a test. Cleaning not required None Moderately hard synthetic water prepared using MILLIPORE MILLI-Q® or equivalent deionized water and reagent grade chemicals (see Methods Manual Section 7, Dilution Water) Five concentrations and a control 0.5 Mortality (LC50) Samples treated as effluent samples for NPDES monitoring. Samples are to be used on the day specified by the interlaboratory study testing schedule 1L 90% or greater survival in controls NOTE: Test vessels shall be randomized in accordance with the WET method manuals. 44 ------- Table 7.2. Summary of test conditions and test acceptability criteria for the Ceriodaphnia chronic test method. 1. Test type: 2. Temperature: 3. Light quality: 4. Light intensity: 5. Photoperiod: 6. Test chamber size: 7. Test solution volume: 8. Renewal of test solutions: 9. Age of test organisms: 10. No. neonates per test chamber:1 11. No. replicate test chambers per concentration: 12. No. neonates per test concentration: 13. Feeding regime: 14. Cleaning: 15. Aeration: 16. Dilution water: 17. Test concentrations: 18. Dilution factor: 19. Test duration:2 20. Endpoints: 21. Test acceptability criteria: 22. Sample handling and holding requirements: 23. Sample volume required: Static renewal 25±1°C Ambient laboratory illumination 10-20 uE/m2/s, or 50-100 ft-c (ambient laboratory levels) 16 h light, 8 h darkness 30 mL 15 mL Daily Less than 24 h; and all released within a 8-h period 1 10 10 Feed 0.1 mL each of YCT and algal suspension per test chamber daily. Use freshly cleaned glass beakers or new plastic cups daily None Moderately hard synthetic water prepared using MILLIPORE MILLI-Q® or equivalent deionized water and reagent grade chemicals (see Methods Manual Section 7, Dilution Water) Five concentrations and a control 0.5 8 days Survival and reproduction 80% or greater survival and an average of 15 or more young per surviving female in the control solutions. 60% of surviving control organisms must produce three broods. Samples treated as effluent samples for NPDES monitoring. Samples are to be used on the day specified by the interlaboratory study testing schedule 1 L/day 1 Test vessels shall be randomized in accordance with the WET method manuals. In addition, block randomization and use of known parentage will be required for the Ceriodaphnia survival and reproduction test as described in the manual and guidance will be reiterated in the specific instructions provided to the laboratories. 2The Ceriodaphnia dubia test, which would otherwise be terminated after 3 broods according to methods manual Section 13.12.1 of that Method, must be conducted for 8 days, with endpoints (survival and number of young per day and number of broods at each recording interval) recorded at the end of the 6th, 7th and 8th day (specifically, at 144, 168, and 192 hours, respectively, from test initiation). No test shall be terminated prior to the 8th day for any reason, including a failure to meet test acceptance criteria. 45 ------- Table 7.3. Summary of test conditions and test acceptability criteria for the fathead acute test method. 1. Test type: 2. Test duration: 3. Temperature: 4. Light quality: 5. Light intensity: 6. Photoperiod: 7. Test chamber size: 8. Test solution volume: 9. Renewal of test solutions: 10. Age of test organisms: 11. No. organisms per test chamber: 12. No. replicate chambers per concentration: 13. No. organisms per concentration: 14. Feeding regime: 15. Test chamber cleaning: 16. Test solution aeration: 17. Dilution water: 18. Test concentrations: 19. Dilution factor: 20. Endpoint: 21. Sample handling and holding requirements: 22. Sample volume required: 23. Test acceptability criterion: Static-renewal 96 h 25±1°C Ambient laboratory illumination 10-20 uE/m2/s (50-100 ft-c) (ambient laboratory levels) 16 h light, 8 h darkness 250 mL 200 mL At48h 1-14 days; 24-h range in age 10 2 20 Artemia nauplii are made available while holding prior to the test; add 0.2 mLArtemia nauplii concentrate 2 h prior to test solution renewal at 48 h Cleaning not required None, unless DO concentration falls below 4.0 mg/L; rate should not exceed 100 bubbles/min Moderately hard synthetic water prepared using MILLIPORE MILLI-Q® or equivalent deionized water and reagent grade chemicals (see Methods Manual Section 7, Dilution Water) Five concentrations and a control 0.5 Mortality (LC50) Samples treated as effluent samples for NPDES monitoring. Samples are to be used on the day specified by the interlaboratory study testing schedule 2 Lfor effluents 90% or greater survival in controls NOTE: Test vessels shall be randomized in accordance with the WET method manuals. 46 ------- Table 7.4. Summary of test conditions and test acceptability criteria for the fathead chronic test method. 1. Test type: 2. Temperature: 3. Light quality: 4. Light intensity: 5. Photoperiod: 6. Test chamber size: 7. Test solution volume: 8. Renewal of test solutions: 9. Age of test organisms: 10. No. larvae per test chamber: 11. No. replicate chambers per concentration: 12. No. larvae per concentration: 13. Source of food: 14. Feeding regime: 15. Cleaning: 16. Aeration: 17. Dilution water: 18. Test concentrations: 19. Dilution factor: 20. Test duration: 21. Endpoints: 22. Test acceptability criteria: 23. Sample handling and holding requirements: 24. Sample volume required: Static renewal 25±1°C Ambient laboratory illumination 10-20 uE/m2/s (50-100 ft-c) (ambient laboratory levels) 16 h light, 8 h darkness 500 mL 250 mL Daily Newly hatched larvae less than 24h old. If shipped, not more than 48h old, 24h range in age 10 4 40 Newly hatched Artemia nauplii (less than 24 h old) Feed 0.1 g newly hatched (less than 24-h old) brine shrimp nauplii three times daily at 4-h intervals or, as a minimum, 0.15 g twice daily, 6 h between feedings (at the beginning of the work day prior to renewal, and at the end of the work day following renewal). Sufficient nauplii are added to provide an excess. Larvae fish are not fed during the final 12 h of the test Siphon daily, immediately before test solution renewal None, unless DO concentration falls below 4.0 mg/L. Rate should not exceed 100 bubbles/min Moderately hard synthetic water prepared using MILLIPORE MILLI-Q® or equivalent deionized water and reagent grade chemicals (see Methods Manual Section 7, Dilution Water) Five concentrations and a control 0.5 1 days Survival and growth (weight as mean per original) 80% or greater survival in controls; average dry weight per surviving organism in control chambers equals or exceeds 0.25 mg/surviving Samples treated as effluent samples for NPDES monitoring. Samples are to be used on the day specified by the interlaboratory study testing schedule 2.5 L/day NOTE: Test vessels shall be randomized in accordance with the WET method manuals. 47 ------- Table 7.5. Summary of test conditions and test acceptability criteria for the Selenastrum chronic test method. 1. Test type: 2. Temperature: 3. Light quality: 4. Light intensity: 5. Photoperiod: 6. Test chamber size: 7. Test solution volume: 8. Renewal of test solutions: 9. Age of test organisms: 10. Initial cell density in test chambers: 11. No. replicate chambers per concentration: 12. Shaking rate: 13. Aeration: 14. Dilution water: 15. Test concentrations: 16. Test dilution factor: 17. Test duration: 18. Endpoint: 19. Test acceptability criteria: 20. Sample handling and holding requirements: 21. Sample volume required: Static non-renewal 25±1°C "Cool white" fluorescent lighting 86 ± 8.6 uE/m2/s (400 ± 40 ft-c or 4306 lux) Continuous illumination 250 mL 100 mL None 4 to 7 days 10,000 cells/mL 4 100 cpm continuous None Algal stock culture medium, moderately hard synthetic water prepared using MILLIPORE MILLI-Q® or equivalent deionized water and reagent grade chemicals(see Methods Manual Section 7, Dilution Water) Five concentrations and a control 0.5 96 h Growth (cell counts) 1 X 106 cells/mL with EDTA or 2 X 105 cells/mL without EOT A in the controls; Variability of controls should not exceed 20% Samples treated as effluent samples for NPDES monitoring. Samples are to be used on the day specified by the interlaboratory study testing schedule 2L NOTE: Test vessels shall be randomized in accordance with the WET method manuals. 48 ------- Table 7.6. Summary of test conditions and test acceptability criteria for the Mysidopsis chronic test method. 1. Test type: 2. Salinity: 3. Temperature: 4. Light quality: 5. Light intensity: 6. Photoperiod: 7. Test chamber: 8. Test solution volume: 9. Renewal of test solutions: 10. Age of test organisms: 11. No. organisms per test chamber: 12. No. replicate chambers per concentration: 13. No. larvae per concentration: 14. Source of food: 15. Feeding regime: 16. Cleaning: 17. Aeration: 18. Dilution water: 19. Test concentrations: 20. Dilution factor: 21. Test duration: 22. Endpoints: 23. Test acceptability criteria: 24. Sample handling and holding requirements: 25. Sample volume required: Static renewal 25%0(±2%<) 26±1°C Ambient laboratory illumination 10-20 uE/m2/s (50-100 ft-c.) (ambient laboratory levels) 16 h light, 8 h darkness, with phase in/out period 8 oz plastic disposable cups, or 400 mL glass beakers 150 mL per replicate Daily 7 days 5 8 40 Newly hatched Artemia nauplii (less than 24 h old) Feed 150 24 h old nauplii per Mysidopsis daily, half after test solution renewal and half after 8-12 h. Pipette excess food from cups daily immediately before test solution renewal and feeding. None unless DO falls below 4.0 mg/L, then gently aerate in all cups 25%o(± 2%<) salinity synthetic seawater prepared with Bioassay Grade Forty Fathoms® artificial sea salts and MILLIPORE MILLI-Q® or equivalent deionized water (see WET Methods Manual Section 7, Dilution Water) Five concentrations and a control 0.5 7 days Survival, growth, and egg development 80% or greater survival, average dry weight 0.20 mg or greater in controls; fecundity may be used if 50% or more of females in controls produce eggs Samples treated as effluent samples for NPDES monitoring. Samples are to be used on the day specified by the interlaboratory study testing schedule 3 L/day NOTE: Test vessels shall be randomized in accordance with the WET method manuals. 49 ------- Table 7.7. Summary of test conditions and test acceptability criteria for the sheepshead acute test method. 1. Test type: 2. Test duration: 3. Temperature: 4. Light quality: 5. Light intensity: 6. Photoperiod: 7. Test chamber size: 8. Test solution volume: 9. Renewal of test solutions: 10. Age of test organisms: 11. No. organisms per test chamber: 12. No. replicate chambers per concentration: 13. No. organisms per concentration: 14. Feeding regime: 15. Test chamber cleaning: 16. Test solution aeration: 17. Dilution water: 18. Test concentrations: 19. Dilution factor: 20. Endpoint: 21. Sample handling and holding requirements: 22. Sample volume required: 23. Test acceptability criterion: 24. Salinity: Static renewal 96 h 25°C±1°C Ambient laboratory illumination 10-20 uE/m2/s or (50-100 ft-c) (ambient laboratory levels) 16 h light, 8 h darkness 250 mL 200 mL At48h 1-14 days; 24-h range in age 10 2 20 Artemia nauplii are made available while holding prior to the test; add 0.2 mLArtemia nauplii concentrate 2 h prior to test solution renewal at 48 h Cleaning not required None, unless DO concentration falls below 4.0 mg/L; rate should not exceed 100 bubbles/min 25 %o±2%osalinity synthetic seawater prepared with Bioassay Grade Forty Fathoms® artificial sea salts andMILLIPORE MILLI-Q® or equivalent deionized water (see Methods Manual Section 7, Dilution Water) Five concentrations and a control 0.5 Mortality (LC50) Samples treated as effluent samples for NPDES monitoring. Samples are to be used on the day specified by the interlaboratory study testing schedule 1 Lfor effluents 90% or greater survival in controls 25%o(±2%j NOTE: Test vessels shall be randomized in accordance with the WET method manuals. 50 ------- Table 7.8. Summary of test conditions and test acceptability criteria for the sheepshead chronic test method. 1. Test type: 2. Salinity: 3. Temperature: 4. Light quality: 5. Light intensity: 6. Photoperiod: 7. Test chamber size: 8. Test solution volume: 9. Renewal of test solutions: 10. Age of test organisms: 11. No. larvae per test chamber: 12. No. replicate chambers per concentration: 13. No. larvae per concentration: 14. Source of food: 15. Feeding regime: 16. Cleaning: 17. Aeration: 18. Dilution water: 19. Test concentrations: 20. Dilution factor: 21. Test duration: 22. Endpoints: 23. Test acceptability criteria: 24. Sample handling and holding requirements: 25. Sample volume required: Static renewal 25%0(±2%0) 25±1°C Ambient laboratory illumination 10-20 uE/m2/s (50-100 ft-c) (ambient laboratory levels) 16 h light, 8 h darkness 600 mL beaker 500 mL/replicate (loading and DO restrictions must be met) Daily Newly hatched larvae (less than 24 h old; 24-h range in age) 10 4 40 Newly hatched Artemia nauplii, (less than 24-h old) Feed once a day 0.10 g wet weight Artemia nauplii per replicate on Days 0-2; Feed 0.15 g wet weight Artemia nauplii per replicate on Days 3-6 Siphon daily, immediately before test solution renewal and feeding None, unless DO falls below 4.0 mg/L, then aerate all chambers. Rate should be less than 100 bubbles/min 25%o (±2%o) salinity synthetic seawater prepared with Bioassay Grade Forty Fathoms® artificial sea salts and MILLIPORE MILLI-Q® or equivalent deionized water (see Methods Manual Section 7, Dilution Water) Five concentrations and a control 0.5 1 days Survival and growth (weight) 80% or greater survival in controls; average dry weight per surviving organism in control chambers should be 0.60 mg or greater, if unpreserved, or 0.50 mg or greater after no more than 7 days in 4% formalin or 70% ethanol Samples treated as effluent samples for NPDES monitoring. Samples are to be used on the day specified by the interlaboratory study testing schedule 6 L/day NOTE: Test vessels shall be randomized in accordance with the WET method manuals. 51 ------- Table 7.9. Summary of test conditions and test acceptability criteria for the silverside acute test method. 1. Test type: 2. Test duration: 3. Temperature: 4. Light quality: 5. Light intensity: 6. Photoperiod: 7. Test chamber size: 8. Test solution volume: 9. Renewal of test solutions: 10. Age of test organisms: 11. No. organisms per test chamber: 12. No. replicate chambers per concentration: 13. No. organisms per concentration: 14. Feeding regime: 15. Test chamber cleaning: 16. Test solution aeration: 17. Dilution water: 18. Test concentrations: 19. Dilution factor: 20. Endpoint: 21. Sample handling and holding requirements: 22. Sample volume required: 23. Test acceptability criterion: 24. Salinity: Static-renewal 96 h 25°C±1°C Ambient laboratory illumination 10-20 AiE/mVs (50-100 ft-c) (ambient laboratory levels) 16 h light, 8 h darkness 250 mL 200 mL At48h 9-14 days; 24-h range in age 10 2 20 Artemia nauplii are made available while holding prior to the test; add 0.2 mLArtemia nauplii concentrate 2 h prior to test solution renewal at 48 h Cleaning not required None, unless DO concentration falls below 4.0 mg/L; rate should not exceed 100 bubbles/min 25%o (± 2%o) salinity synthetic seawater prepared with Bioassay Grade Forty Fathoms® artificial sea salts and MILLIPORE MILLI-Q® or equivalent deionized water (see Methods Manual Section 7, Dilution Water) Five concentrations and a control 0.5 Mortality (LC50) Samples treated as effluent samples for NPDES monitoring. Samples are to be used on the day specified by the interlaboratory study testing schedule 1 Lfor effluents 90% or greater survival in controls 25%0(±2%0) NOTE: Test vessels shall be randomized in accordance with the WET method manuals. 52 ------- Table 7.10. Summary of test conditions and test acceptability criteria for the silverside chronic test method. 1. Test type: 2. Salinity: 3. Temperature: 4. Light quality: 5. Light intensity: 6. Photoperiod: 7. Test chamber size: 8. Test solution volume: 9. Renewal of test solutions: 10. Age of test organisms: 11. No. larvae per test chamber: 12. No. replicate chambers per concentration: 13. No. larvae per concentration: 14. Source of food: 15. Feeding regime: 16. Cleaning: 17. Aeration: 18. Dilution water: 19. Test concentrations: 20. Dilution factor: 21. Test duration: 22. Endpoints: 23. Test acceptability criteria: 24. Sample handling and holding requirements: 25. Sample volume required: Static renewal 25%0(±2%<) 25±1°C Ambient laboratory illumination 10-20 uE/m2/s (50-100 ft-c) (Ambient laboratory levels) 16 h light, 8 h darkness 1 L containers 750 mL/replicate (loading and DO restrictions must be met) Daily 7-11 days post hatch; 24-h range in age 10 4 40 Newly hatched Artemia nauplii (survival of 7-9 days oldMenidia beryllina larvae improved by feeding 24 hold Artemia) Feed 0.10 g wet weight^4rte/w/'a nauplii per replicate on days 0-2; Feed 0.15 g wet weight Artemia nauplii per replicate on days 3-6 Siphon daily, immediately before test solution renewal and feeding None, unless DO concentration falls below 4.0 mg/L, then aerate all chambers. Rate should be less than 100 bubbles/min. 25%o (± 2%o) salinity synthetic seawater prepared with Bioassay Grade Forty Fathoms® artificial sea salts and MILLIPORE MILLI-Q® or equivalent deionized water (see Methods Manual Section 7, Dilution Water) Five concentrations and a control 0.5 1 days Survival and growth (weight) 80% or greater survival in controls, 0.50 mg average dry weight of control larvae when larvae dried immediately after test termination, or 0.43 mg or greater average dry weight of control larvae, preserved not more than 7 days in 4% formalin or 70% ethanol Samples treated as effluent samples for NPDES monitoring. Samples are to be used on the day specified by the interlaboratory study testing schedule 6 L/day NOTE: Test vessels shall be randomized in accordance with the WET method manuals. 53 ------- 16. Endpoints: 17. Test acceptability criteria: 18. Sample handling and holding requirements: 19. Sample volume required: Table 7.11. Summary of test conditions and test acceptability criteria for the Champia chronic test method. 1. Test type: 2. Salinity: 3. Temperature: 4. Photoperiod: 5. Light intensity: 6. Light source: 7. Test chamber size: 8. Test solution volume: 9. No. organisms per test chamber: 10. No. replicate chambers per concentration: 11. No. organisms per concentrations: 12. Dilution water: 13. Test concentrations: 14. Test dilution factor: 15. Test duration: Static non-renewal 30%« (± 2%«) 23±1°C 16 h light, 8 h darkness 75 uE/m2/s (500 ft-c) Cool-white fluorescent lights 200 mL polystyrene cups, or 250 mL Erlenmeyer flasks 100 ml 5 female branch tips and 1 male plant 4 24 30%o salinity natural seawater Five concentrations and a control 0.5 2 day exposure to effluent, followed by 5 to 7-day recovery period in control medium for cystocarp development Reduction in cystocarp production compared to controls 80% or greater survival, and an average of 10 cystocarps per plant in controls Samples treated as effluent samples for NPDES monitoring. Samples are to be used on the day specified by the inter lab oratory study testing schedule 1L per test NOTE: Test vessels shall be randomized in accordance with the WET method manuals. 54 ------- Table 7.12. Summary of test conditions and test acceptability criteria for the Holmesimysis acute test method. The acute test procedure described in the Acute Methods Manual for Mysidopsis bahia will be used for this test with a salinity of 32%o (± 2%o) and a temperature of 12 °C ± 1 °C. 1. Test type: 2. Test duration: 3. Temperature: 4. Light quality: 5. Light intensity: 6. Photoperiod: 7. Test chamber size: 8. Test solution volume: 9. Renewal of test solutions: 10. Age of test organisms: 11. No. organisms per test chamber: 12. No. replicate chambers per concentration: 13. No. organisms per concentration: 14. Feeding regime: 15. Test chamber cleaning: 16. Test solution aeration: 17. Dilution water: 18. Test concentrations: 19. Dilution factor: 20. Endpoint: 21. Sample handling and holding requirements: 22. Sample volume required: 23. Test acceptability criterion: 24. Salinity: Static-renewal 96 h 12°C±1°C Ambient laboratory illumination 10-20 AiE/mVs (50-100 ft-c) (ambient laboratory levels) 16 h light, 8 h darkness 250 mL 200 mL At48h 1-5 days; 24-h range in age 10 4 40 Artemia nauplii are made available while holding prior to the test; feed 0.2 mL of concentrated suspension of Artemia nauplii < 24-h old, daily (approximately 100 nauplii per Mysidopsis) Cleaning not required None, unless DO concentration falls below 4.0 mg/L; rate should not exceed 100 bubbles/min 32%o salinity natural seawater Five concentrations and a control 0.5 Mortality (LC50) Samples treated as effluent samples for NPDES monitoring. Samples are to be used on the day specified by the interlaboratory study testing schedule 1 Lfor effluents 90% or greater survival in controls 32%0(±2%,) NOTE: Test vessels shall be randomized in accordance with the WET method manuals. 55 ------- 8.0 DATA REPORTING AND EVALUATION 8.1 Report Submission Within 30 days following the completion of interlaboratory testing for a given method, each laboratory was required to submit a data report detailing the conduct and results of WET testing completed on each sample. Table 8.1 lists the report due dates and number of reports received for each test method. Reports were received for all samples from all participating laboratories with the exception of one participant laboratory (laboratory ID #3) for the Ceriodaphnia chronic method. Participant laboratory #3 received samples for the Ceriodaphnia chronic method and initiated testing. During testing of the final two of three samples received, an overnight power failure caused the malfunction of water baths containing test chambers. The malfunctioning water baths over-heated test chambers, killing all test organisms (including controls) and terminating the test prematurely. The laboratory was unable to retest within reasonable sample holding times, so all further testing of this method was canceled at this laboratory. The laboratory was non-EPA-sponsored, and the sponsor declined to reimburse the laboratory for testing costs since all tests were not completed. In turn, the laboratory declined to submit a data report for this test method. As a result, interlaboratory data for the Ceriodaphnia chronic method include results from one referee laboratory and 34 of the 35 participant laboratories. According to the Participant Laboratory Statement of Work and SOP (see Appendix B), each data report was required to consist of: Narrative summary of testing - The narrative summary was intended to quickly and clearly identify the laboratory, test method, samples tested, summarized test results, any problems associated with the samples or conduct of the tests, any modifications from approved procedures, and any laboratory comment on the performance of the method. Hardcopy results - This deliverable consisted of the items outlined in Table 8.2, all raw data (biological and chemical), all laboratory benchsheets, all pertinent sample information including copies of all completed EPA traffic report forms, and all pertinent quality assurance information including results of the monthly QA/QC reference toxicant tests. • Electronic results - Laboratories also were required to submit selected raw and summarized data electronically using method-specific Microsoft Excel® spreadsheet templates that were provided to participant laboratories by SCC. Electronic data included general information, sample collection/receipt information, test condition information, raw biological data, raw water quality data, and summarized test results. Electronic data submission facilitated automated review and statistical analysis of study results. 56 ------- Table 8.1. Report due dates. Due date 11/12/99 12/02/99 12/03/99 12/08/99 12/13/99 12/13/99 04/06/00 04/17/00 05/03/00 05/04/00 Test method Fathead chronic Silverside chronic Ceriodaphnia chronic Fathead acute Ceriodaphnia acute Silverside acute Mysidopsis chronic Sheepshead acute Selenastmm chronic Sheepshead chronic Number of reports received3 28 11 35 30 29 10 12 8 12 8 aThe number of laboratory reports received includes participant and referee laboratories. One referee laboratory report was received for each method. Table 8.2. Data reporting elements. Section 1 - Summary Page 1.1 Laboratory name 1.2 Laboratory address and phone number 1.3 Name and signature of laboratory QA Officer, certifying that data have been internally reviewed and that personnel meticulously followed the methods, and the procedures are deemed to be compliant with the methods and acceptable for reporting purposes 1.4 Laboratory contact responsible for study 1.5 Analyst(s) who performed WET tests (full name) 1.6 Toxicity tests performed 1.7 Detailed explanations of any difficulties encountered and any approved modifications to the techniques specified in the SOW, specific instructions, or the methods manuals 1.8 Number of successful tests completed Section 2 - Sample Information 2.1 Number of samples received and EPA sample number assigned to each sample 2.2 Dates of sample receipt 2.3 Sample temperature when received at laboratory 2.4 Physical and chemical data of sample contents (as required in appropriate method) 2.5 Dilution water 2.5.1 Source and time frame water is used or how maintained 2.5.2 Collection or preparation date(s), where applicable 2.5.3 Pretreatment information 2.5.4 Physical and chemical characteristics (pH, hardness, conductivity, salinity, etc.) 2.6 Sample storage information 2.7 Sample preparation for testing information 57 ------- Table 8.2. Data reporting elements, (continued) Section 3 - Test Conditions 3.1 Toxicity test method used (title, number, source) 3.2 Endpoint(s) of test(s) 3.3 Deviations from reference method(s), if any, and reason(s) 3.4 Date and time test(s) started, date and time samples were prepared and solutions transferred for renewals 3.5 Date and time test(s) terminated 3.6 Type and volume of test chambers 3.7 Volume of solution used per chamber 3.8 Number of organisms per test chamber 3.9 Number of replicate test chambers per treatment 3.10 Feeding frequency and amount and type of food (be specific with sources, concentrations of foods (i.e., algae concentration, YCT solids level, preparation dates) 3.11 Acclimation of test organisms (temperature mean and range and, where applicable, salinity mean and range) 3.12 Test temperature (mean and range) 3.13 Test salinity, where applicable (mean and range) 3.14 Specify if aeration was needed 3.15 Specify if organisms were dried immediately for weighing or preserved prior to drying 3.16 Specify how food was prepared and sources of food. Include test results that validate the quality of batch food preparations (i.e., Ceriodaphnia dubia tests on YCT preparation) 3.17 Describe how routine chemistries on new solutions were made (in actual test chamber or in beakers after dispensing) 3.18 Describe how randomization was conducted, especially blocking and known parentage; report how brood distinctions were made and male (if any) identification was made Section 4 - Test Organisms 4.1 Scientific name of test species, verification of species documented 4.2 Age (life stage) of test species (be specific for all species); age at time of test initiation (for example, for C. dubia be sure to clarify the window of age of the neonates as well as the overall age of the animals) 4.3 Mean length and weight (where applicable) 4.4 Source and QA/QC test conditions 4.5 Holding Conditions 4.6 Diseases and treatment (where applicable) 4.7 Taxonomic key used for species identification Section 5 - Quality Assurance 5.1 Reference toxicant used routinely; source; date received; lot number 5.2 Date and time of most recent reference toxicant test; test results and current control (cusum) chart including 20 most recent data points 5.3 Dilution water used in reference toxicant tests (with characteristics provided) 5.4 Physical and chemical methods used 5.5 Reference toxicant results (NOEC, IC25, or LC50 where applicable, LOEC or EC50) Section 6 - Results 6.1 Copies of all bench sheets. Be sure to count and note broods for reproduction test with Ceriodaphnia 6.2 Raw toxicity data in tabular form, including daily records of affected organisms in each replicate at each concentration (including controls) and plots of toxicity data 6.3 Table of results (LC50, IC25, NOEC for each endpoint) and confidence limits (where applicable) 6.4 Statistical methods and software used to calculate results 6.5 Summary table of physical and chemical data 58 ------- 8.2 Data Review 8.2.1 Data Package Receipt and Gross Completeness Check Data reports from all laboratories were submitted to SCC for review and verification of test results. Upon receipt, SCC personnel date-stamped data packages and performed an initial review to ensure that all required information was provided. Most laboratories reported results using their standard reporting formats rather than the suggested format in Table 8.2. This was acceptable provided that all pertinent information was included. If necessary information was not provided in the data report, SCC personnel contacted the laboratory and asked them to supply the additional information. 8.2.2 Data Accuracy and Quality Check Following initial review of data packages for completeness, SCC personnel performed a detailed review of data reports to ensure that data were accurate and generated in accordance with the required procedures. The following review steps were completed for each data report: Cross reference of raw data - Raw data on submitted electronic benchsheets were compared to hardcopy laboratory benchsheets to ensure that no transcription or data entry errors occurred. Every entry of biological raw data (including individual replicate values of daily survival, daily reproduction, and weight) was cross-referenced against hardcopy benchsheets to ensure accuracy. Sample collection, test condition, and water quality raw data entered on electronic benchsheets were spot-checked against hardcopy benchsheets. If errors were found during the spot-check, more intensive review was initiated. All data fields that triggered automated data qualifier flags in the electronic benchsheet were also individually compared to hardcopy benchsheets to ensure accuracy. When errors or inconsistencies were identified, the electronic benchsheet was corrected to match the hardcopy laboratory benchsheets. Narrative summary and hardcopy report review - SCC reviewed the narrative summary and hardcopy report to ensure that tests were conducted in accordance with the WET method manuals, the WET Variability Study plan, and guidance provided in method-specific SOPs. If any deviations from the required or recommended procedures were identified, SCC data reviewers verified that those deviations were captured by automated or manual data qualifier flags in the electronic benchsheet. Data qualifier flags - The electronic benchsheets for each method were programmed to automatically identify and flag deviations in test conditions, sample holding times, sample temperatures, test acceptability criteria, or test water quality. For all flags that were automatically identified, SCC evaluated the electronic benchsheet and hardcopy report to ensure that the flag was warranted. Additional parameters, test conditions, and comments that were unable to be programmed for automated review were checked and flagged manually if necessary. Table 8.3 lists and describes the categories of data qualifier flags used in the WET Variability Study. Tests that were flagged for a failure to meet test acceptability criteria (flags a1; aj, a3, a4, a6, and a7) were considered invalid and were not included in the analysis of method performance (see Section 9.1.1). The large number of other flags (bl - g12) described in Table 8.3 and identified for particular tests in 59 ------- Section 9, emphasizes the extensive nature of test review rather than deficiencies in test conduct or quality control. While the presence of these flags may indicate a deviation from optimal test conditions, their presence alone was not used to invalidate and exclude test data from the analysis of method performance. Rather, these flags were used to identify potential causes for aberrant test results and to support inclusion or exclusion of data in outlier analyses (see Section 9.1.4). Reference toxicant test review - Laboratories were required to conduct a reference toxicant test within the month of testing for the WET Variability Study and provide the test results and current control chart with the data report. SCC data review staff verified that a reference toxicant test was conducted during the required time frame and reviewed the laboratory's control chart to ensure that the current test result fell within the control chart limits (2 standard deviations for point estimates and 1 concentration interval for hypothesis testing results). If the reference toxicant test fell outside of the control chart limits or was not conducted during the required time frame, a data qualifier flag for reference toxicant testing was associated (in the results database) with all test results from this laboratory for the given method. Table 8.3. Test data qualifier flags. Flag code Flag Description Test acceptability &i a2 a3 a4 as a6 a. Survival of control organisms failed to meet the minimum test acceptability criteria for the method. Growth of control organisms (measured as the mean weight of surviving control organisms) failed to meet the minimum test acceptability criteria for the method. Reproduction of control organisms in the Ceriodaphnia chronic test failed to meet the minimum test acceptability criteria requiring that 60% of surviving control organisms have 3 broods prior to test termination at 8 days. Mean reproduction of surviving control organisms in the Ceriodaphnia chronic test failed to meet the minimum test acceptability criteria of 1 5 neonates. Fecundity endpoints were not generated because less than 50% of control females in the Mysidopsis chronic test produced eggs. Mean algal growth of control organisms in the Selenastrum chronic test did not meet the minimum test acceptability criteria requiring a mean cell density of IxlO6 cells/mL with EDTA or 2xl05 cells/mL without EDTA. Algal growth variability between control replicates in the Selenastrum chronic test did not meet the minimum test acceptability criteria requiring less than 20% variability (measured as % CV). Sample Receipt b, b2 b3 b4 b5 Sample temperature was >4°C upon arrival at the participant lab. Sample was >36 hr old at the time of test initiation. Sample was >72 hr old at the time of test initiation. Sample was aerated upon receipt due to over saturation of dissolved oxygen. Sample was inadvertently delivered to and opened by a laboratory not involved in the study. The sample was quickly rerouted to the correct laboratory and tests were initiated on time. 60 ------- Table 8.3. Test data qualifier flags (continued) Flag code Flag Description Dilution Water Cl C2 C3 C4 C5 Dilution water used for the test was different from that required in the study. Alkalinity of dilution water was >10% outside of recommended ranges for moderately hard synthetic water. Hardness of dilution water was >10% outside of recommended ranges for moderately hard synthetic water. pH of dilution water was >10% outside of recommended ranges for moderately hard synthetic water. Salinity of dilution water was outside of the range required in the study. Water Quality d, d2 d3 d4 d5 d6 d7 Temperature of one or more test concentrations was outside of range required in the study. Dissolved oxygen (DO) was less than 4 mg/L in one or more test concentrations. Aeration was not provided in test when DO was <4 mg/L. pH was <6 or >9 in one or more test concentrations. Salinity of one or more test concentrations was outside of the range required in the study. Salinity was adjusted during the test to compensate for evaporation due to test aeration. Total ammonia was >5 mg/L in one or more test concentrations. Test Conditions ei e2 e3 e4 e5 e6 e7 e8 e9 610 en Number of organisms per test container differed from the required number of organisms due to accidental loss of one or more test organisms. Test chamber size was different from that required in the study. Test solution volume in test containers was outside of the range required in the study. Test renewals were conducted more than 2 hours outside of the required time for test renewal. Feeding schedule used during the test differed from feeding schedule recommended in the method manuals. Test termination was conducted more than 2 hours prior to proper test termination time. Dilution series used for testing was 12.5, 25, 50, 75, and 100% instead of the standard dilution series required in the study (6.25, 12.5, 25, 50, and 100%). Lighting cycle was interrupted for 2 hours during power outage. Continuous shaking rate of 100 cpm was not used. Initial cell density average was more than 10% outside of required 10,000 cell/mL inoculation level. Initial cell density variability among replicates was greater than a CV of 10%. Organisms fi f2 f3 f4 f5 Age of organism was outside of range required in the study. Organism culture contaminated with rotifers. Organism culture crashed just prior to testing. Males were identified in the test. Organisms were cultured at 20°C and directly transferred to test temperature of 25°C. 61 ------- Table 8.3. Test data qualifier flags (continued) Flag code Flag Description Quality control gi g2 S3 g4 g5 g6 g7 gs g9 gio gll gl2 Reference toxicant test conducted during the month of study testing was outside of the 2 standard deviation control chart limits or the test was not conducted. Percent minimum significant difference (PMSD) was greater than the recommended criteria for this method (USEPA, 2000d). ASTM h statistic for this test was greater than the recommended criteria, indicating that this test may be an outlier. ASTM k statistic for this test was greater than the recommended criteria, indicating that this test may be an outlier. Interrupted concentration - response relationship was observed (i.e., a test concentration was determined to be significantly different from the control, while one or more higher test concentrations were not significantly different from the control). One or multiple test replicates were lost due to laboratory error. Two cell count methods were used. Test failed test acceptability criteria for growth using coulter counter method, but passed test acceptability criteria for growth using Hemacytometer method. Test was repeated due to laboratory error. Initial test was incorrectly terminated at 48 hours. Test was repeated due to laboratory error. Initial test was incorrectly renewed with the wrong sample. Two sets of controls were conducted for this test, and one did not meet test acceptability criteria for reproduction. Cell density was measured using turbidity technique instead of cell counts. Referee laboratory test repeated. Initial test failed test acceptability criteria for survival. 8.2.3 Effect Concentration Recalculation and Verification To confirm that test results were calculated correctly and according to WET method manual requirements for statistical data analysis, all test results were recalculated by SCC using reviewed raw data from the electronic benchsheets. SCC conducted statistical analysis of WET test data using ToxCalc version 5.0 (Tidepool Scientific, 1996). Biological test data were electronically copied from electronic benchsheets directly into the ToxCalc software to avoid additional data transcription or data entry errors. Statistical methods for analysis were selected according to the WET method manuals using the EPA flowchart option in the ToxCalc software. Test results for the endpoints listed in Table 2.3 were calculated for the respective test methods. Any error messages that were produced by the software were noted and evaluated to ensure that the software defaulted to the correct alternate statistical method. As part of the effect concentration recalculation and verification process, SCC reviewed the concentration- response curve generated for each sample and endpoint. SCC reported test results for each sample in accordance with EPA's guidance on the evaluation of concentration-response relationships (USEPA, 2000a). When SCC observed unexpected concentration-response patterns, EPA's guidance (USEPA, 2000a) was followed for determining whether the derived effect concentration was reliable and should be reported, the effect concentration was anomalous and should be explained, or the test was inconclusive and the sample should be retested. When EPA's concentration-response relationship guidance recommended retesting of a sample, 62 ------- SCC reported the result for this sample as inconclusive. Table 8.4 lists the samples that produced unexpected concentration-response curves and were affected by the concentration-response guidance. After test results were recalculated by SCC, these results were compared to results as reported by the participant laboratory. If recalculated and laboratory-reported results differed, the summarized test data and statistical analyses used for calculation were reviewed to isolate the source of the deviation. Following recalculation and review, test results were incorporated into a results database for the WET Variability Study. Table 8.4. Sample results affected by EPA guidance on concentration-response relationships (USEPA, 2000a). Method Ceriodaphnia chronic Fathead chronic Sample code 9328 9332 9333 9341 9343 9379 9380 9392 9408 9415 9122 9129 9145 9161 9168 9193 Concentration -response pattern observed3 4 4 5 5 6 5 6 6 4 5 5 6 5 6 5 6 Effect on reported test result Reproduction NOEC and IC25 were reported as inconclusive since test percent minimum significant difference (PMSD) was greater than recommended criterion Calculated IC25 was determined to be anomalous due to ICp smoothing; IC25 was adjusted to >100% since mean response in the 100% treatment was within 25% of control mean 6.25% treatment was determined to be anomalous and reproduction NOEC was reported as highest concentration not significantly different from control 6.25% and 12.5% treatments were determined to be anomalous and reproduction NOEC was reported as highest concentration not significantly different from control Reproduction NOEC was reported as concentration below the LOEC 6.25% treatment was determined to be anomalous and survival NOEC was reported as highest concentration not significantly different from control Survival NOEC was reported as inconclusive since test PMSD was greater than recommended criterion Reproduction NOEC was reported as concentration below the LOEC Calculated IC25 was determined to be anomalous due to ICp smoothing; IC25 was adjusted to >100% since mean response in the 100% treatment was within 25% of control mean Survival NOEC was reported as inconclusive since test PMSD was greater than recommended criterion 6.25% treatment was determined to be anomalous and growth NOEC was reported as highest concentration not significantly different from control Growth NOEC was reported as concentration below the LOEC 25% treatment was determined to be anomalous and survival NOEC was reported as highest concentration not significantly different from control Growth NOEC was reported as concentration below the LOEC 6.25% treatment was determined to be anomalous and growth NOEC was reported as highest concentration not significantly different from control Growth NOEC was reported as concentration below the LOEC 63 ------- Table 8.4. Sample results affected by EPA guidance on concentration-response relationships (USEPA, 2000a). (continued) Method Fathead chronic (continued) Selenastrum chronic Mysidopsis chronic Silverside chronic Sample code 9194 9209 9212 9454 (w/o EDTA) 9455 (w/ EDTA) 9455 (w/o EDTA 9468 (w/ EDTA) 9468 (w/o EDTA) 9473 (w/o EDTA) 9682 9694 9696 9545 9556 Concentration -response pattern observed3 6 5 6 4 4 4 5 5 5 5 5 5 4 5 Effect on reported test result Growth NOEC was reported as concentration below the LOEC Growth NOEC was reported as inconclusive since test PMSD was greater than recommended criterion Survival and growth NOECs were reported as concentration below the LOEC Growth NOEC, IC25, and IC50 were reported as inconclusive since test PMSD was greater than recommended criterion Growth NOEC, IC25, and IC50 were reported as inconclusive since control response was marginal and below laboratory's normal range of control performance Calculated IC25 was determined to be anomalous due to ICp smoothing; IC25 was adjusted to >100% since mean response in the 100% treatment was within 25% of control mean 50% treatment was determined to be anomalous and growth NOEC was reported as highest concentration not significantly different from control Growth NOEC was reported as inconclusive since test PMSD was greater than recommended criterion Growth NOEC was reported as inconclusive since test PMSD was greater than recommended criterion 12.5% treatment was determined to be anomalous and fecundity NOEC was reported as highest concentration not significantly different from control 12.5% treatment was determined to be anomalous and fecundity NOEC was reported as highest concentration not significantly different from control 12.5% treatment was determined to be anomalous and growth NOEC was reported as highest concentration not significantly different from control Growth NOEC and IC25 were reported as inconclusive since test PMSD was greater than recommended criterion 12.5% treatment was determined to be anomalous and growth NOEC was reported as highest concentration not significantly different from control 1 Concentration-response patterns are numbered as identified in Chapter 4 of USEPA, 2000a. 64 ------- 9.0 RESULTS 9.1 Analysis of Results SCC personnel entered recalculated and verified test results for each sample tested into a results database along with associated data qualifier flags, sample information, and summary test information (e.g., control mean, control CV, test minimum significant difference, etc.). Information in the results database was used to evaluate the test completion rate, false positive rate, and precision for each test method. All calculated test results presented in this section were rounded to three significant figures for consistency. A formal analysis of measurement error for each data type was not conducted; however, three significant figures is believed to be consistent with most WET test measurements (e.g., test concentrations, weights, counts). Summary statistics presented in this section (e.g., mean, standard deviation, CV) were calculated from test results prior to rounding (so summary statistics of rounded results may differ slightly). 9.1.1 Valid Tests Only valid tests were used in the determination of false positive rates and precision. A valid test was defined as a test that met the required test acceptability criteria for the method as stated in the WET method manuals. Tests that deviated from specified test conditions were identified with data qualifier flags (see Section 8.2.2) but were not excluded as invalid tests. The WET method manuals state that tests that deviate from specified test conditions may be conditionally acceptable depending on the degree of the departure and the objectives of the test. Based on the study objectives of assessing the performance of WET test methods, these tests were included in the analysis of false positive rates and precision unless the combined results of test review (see Section 8.2) and outlier analysis (see Section 9.1.4) indicated that the test condition deviations significantly affected test results. When EPA guidance on the evaluation of concentration-response relationships (USEPA, 2000a) recommended retesting of a sample, the test result in question was identified as inconclusive and was not included in the analysis of false positive rates and precision for the methods. Also, test results from referee laboratories were excluded from determinations of successful test completion rates, false positive rates, and precision for the methods. While referee laboratory testing was conducted similarly to and simultaneously with participant laboratory testing, the identity of samples was not blinded to the referee laboratory. Appendix F summarizes study results when referee laboratory data are included in the analysis of successful test completion rates, false positive rates, and precision. 9.1.2 Successful Test Completion Rate The successful test completion rate was calculated independently for each test method as the percentage of initiated and properly terminated tests that met the test acceptability criteria as specified in the WET method manuals. Participant laboratories that failed to complete tests due to reasons unrelated to the test methods themselves (i.e., laboratory error) were not included in the test completion rate calculations or statistical analyses. This occurred for only four samples (9586, 9587, 9589, and 9618). In three cases a 96-hour test was incorrectly terminated at 48 hours, and in one case the test was renewed using the wrong sample. In each case, the referee laboratory sent a new sample aliquot from the original bulk sample preparation for retesting at the 65 ------- participant laboratory. Results from the repeated tests are presented in this report and were used in the determination of successful test completion rates, false positive rates, and precision. 9.1.3 False Positive Rate The false positive rate was calculated independently for each test method and for each endpoint and effect concentration reported (LC50, survival NOEC, IC25 for growth, IC25 for reproduction, NOEC for growth, and NOEC for reproduction). The false positive rate was determined as the rate at which test results indicated toxicity (i.e., a calculated effect concentration <100% sample) in blank samples, and was calculated as: Number of valid tests indicating toxicity in blank samples Total number of valid tests conducted on blank samples •XlOO% 9.1.4 Precision Precision estimates were generated independently for each test method, point estimate, and sample type tested (except for the blank sample type). For sample types that were tested using within-laboratory replication, estimates were provided for within-laboratory precision (based on the within-laboratory variance component), between-laboratory precision (based on the between-laboratory variance component), and total precision (based on the total variance). For sample types that were not tested using within-laboratory replication, a single precision estimate was generated based on the total variance. When test results were calculated as outside of the test concentration range (i.e., >100% or <6.25%), these censored values were set to the limits of the test concentration range for the purposes of calculating summary statistics and estimating precision. Censored values of > 100% were set to 100%, and censored values of <6.25% were set to 6.25%. Censored IC25 values of >12.5%, >25%, and >50% also were possible for the Mysidopsis chronic fecundity endpoint. These censored values were set to 12.5%, 25%, and 50%, respectively. A large proportion of censored values (e.g., >100% or <6.25%) within a data set, was evidence that the sample type failed to produce toxicity that could be definitively measured within the test concentration range. Because the study was designed to characterize the precision WET methods within their measurement range, precision estimates were not calculated for sample types that failed to produce toxicity that could be definitively measured within the test concentration range. As a result, precision estimates were not calculated for the blank sample type for all test methods, the reference toxicant sample type for the Ceriodaphnia chronic and silverside acute test methods (see Section 5.3), and all sample types for the Mysidopsis chronic fecundity endpoint. Only participant laboratory results from valid tests (see Section 9.1.1) were included in the calculation of precision estimates; invalid tests were excluded. SCC conducted an outlier analysis of valid test results to determine if any additional test results should be excluded from the analysis of precision. SCC used the calculation of ASTM's h and k statistics (ASTM, 1997) to evaluate data consistency and identify potential outliers. ASTM h statistics were used to examine the consistency of test results from laboratory to laboratory. ASTM k statistics were used to examine the consistency of within-laboratory precision from laboratory to laboratory. For each test method, ASTM h statistics were calculated for each laboratory and each sample type (except the blank sample type) using the equation below. 66 ------- x-x where, X = a laboratory's average test result for a given sample type (if the laboratory only tested one sample of a given sample type, that individual result was used) = p _ X = the average of individual laboratory averages (S.X/P) i p = number of laboratories testing a given sample type For each test method, ASTM k statistics were calculated for each laboratory and each sample type (except the blank sample type) using the equation below. The k statistic was not calculable for laboratories that did not test replicate samples of the same sample type. c k = J-^-i / —\2 /, , > \x - x) (n - 1) p = number of laboratories testing a given sample type x = an individual test result X = a laboratory's average test result for a given sample type n = number of test results for a given sample type from a single laboratory For each test method and sample type (excluding the blank sample type), SCC compared the h and k statistics calculated for each laboratory to critical values of h and k statistics at the recommended 0.5% significance level (see ASTM, 1997 for table of critical values). Test results from laboratories with a calculated h statistic above the critical value were significantly different (at the 0.5% significance level) from results reported by other laboratories for the same sample type. These inconsistent test results were flagged (see Table 8.3) and identified as potential outliers. Laboratories with a calculated k statistic above the critical value experienced greater within-laboratory variability than other laboratories testing the same sample type. These inconsistent test results also were flagged (see Table 8.3) for further investigation. Since estimates of coefficients of variation can be biased by extreme values and by small data sets, it was important to closely investigate individual data points before discarding them as outliers. An individual test result was only discarded as an outlier if the laboratory was identified by ASTM h statistics as an outlier and a reasonable cause for producing the aberrant result could be determined. The data qualifier flags associated with each test result were useful in this determination. In general, a very conservative approach to excluding outliers was taken. Only 15 tests in the entire study of 698 tests were identified by ASTM h statistics as potential 67 ------- outliers. Two of these tests were also flagged for extreme ASTM k statistics. Of the 15 tests identified as potential outliers, only 8 were excluded from precision estimates based on the determination of a cause for inconsistent results. Table 9.1 shows the test results that were identified as potential outliers and provides a rationale for the inclusion or exclusion of these data points from precision estimates. Precision was estimated by the coefficient of variation (CV) for point estimates. For NOEC values, precision was simply described by the range and distribution of NOEC values and the percentage of values falling within one concentration of the median (as described in the WET method manuals for evaluating routine reference toxicant test results using NOECs). The CV for point estimates was calculated as: where, S2 = variance (S = the standard deviation) x= mean of valid test results for a given method, endpoint, and sample type For test methods and sample types that included within-laboratory replication (i.e., multiple tests on the same sample type from a given laboratory), the variance identified in the above equation was obtained by maximum likelihood estimation using the PROC MIXED procedure in SAS version 8 (SAS Institute, 2000). This procedure estimated the within-laboratory, between-laboratory, and total variance components. Each of these variance components were individually used to calculate within-laboratory, between-laboratory, and total CVs. The total CVs express the total interlaboratory variability of the results, including both within-laboratory and between-laboratory components of variability. For test methods and sample types that did not include within-laboratory replication, the variance identified in the above equation was obtained by the following equation using the PROC MEANS procedure in SAS version 8 (SAS Institute, 2000). " x 1=1 where, n = number of valid test results for a given method, endpoint, and sample type Xt = individual result /' (/' ranging from 1 to «) S2 = variance of the n test results (S = the standard deviation) X= mean of the n test results 68 ------- Table 9.1. Test results identified as potential outliers by ASTM h statistics. Test method Ceriodaphnia chronic Fathead acute Fathead chronic Selenastrum chronic Sheepshead acute Silverside chronic Lab ID 42 42 62 62 205 205 205 125 125 39 125 425 101 421 421 Sample code 9347 9348 9033 9034 9065 9066 9177 9162 9163 9468 (w/ EDTA) 9476 (w/o EDTA) 9617 9600 9582 9583 Test endpoint survival and reproduction survival and reproduction survival survival survival survival growth growth growth growth growth survival survival survival and growth survival and growth Included/e xcluded excluded excluded included included included included included excluded excluded excluded excluded included included excluded excluded Rationale These duplicate samples were flagged for both h and k statistics, indicating that the laboratory's mean test result was significantly different from other laboratories and that within-lab variability was also significantly higher for this laboratory; this high within-lab variability could explain the inconsistent mean test result from this laboratory No data qualifier flags Only data qualifier flag was for test chamber size No data qualifier flags Dilution water quality was highly variable for this laboratory. For these two tests, dilution water hardness at test initiation was 111 mg/L, compared to the expected hardness range of 80- 100 mg/L for moderately hard reconstituted water. For the remaining two fathead chronic tests conducted in this laboratory, dilution water alkalinity at test initiation was 45 mg/L, compared to the expected alkalinity range of 60-70 mg/L. Cell growth was marginal and failed test acceptability criteria when measured using coulter counter; other flags included sample temperature, test temperature, test pH, continuous shaking rate, and interrupted concentration-response relationship Reference toxicant test was not conducted concurrently with tests or within the month of testing Data qualifier flags were observed for salinity of 30 ppt on test day 4 and for test termination at 93.7 hours (rather than 96); however, the data point was included due to the small size of the data set for this method (this result represented 14% of the data set) No data qualifier flags Reference toxicant test conducted concurrently with samples was outside of control chart limits 69 ------- 9.2 Ceriodaphnia Acute Test Method Results A total of 28 participant laboratories conducted the Ceriodaphnia acute test method in the WET Variability Study. These laboratories tested a total of 34 blank samples, 30 reference toxicant samples, 27 effluent samples, and 13 receiving water samples. For each sample tested, a 48-hour LC50 was generated as a test result. Results of Ceriodaphnia acute testing are shown in Tables 9.2 - 9.5 for each sample type. The sample distribution scheme used for the Ceriodaphnia acute test method was inadvertently altered from the original study design (see Section 6.4) due to an error in sample distribution. On the first week of testing, all laboratories that were intended to receive blank samples were shipped reference toxicant samples, and all laboratories that were intended to receive reference toxicant samples were shipped blank samples. This error was not identified prior to the second week of testing, so shipments during the second week were conducted as planned. This caused some laboratories to receive three blank samples and no reference toxicant samples and other laboratories to receive three reference toxicant samples and no blank samples. This sample distribution error should not affect the evaluation of the Ceriodaphnia acute test method. 9.2.1 Successful Test Completion Rate A total of 104 Ceriodaphnia acute tests were initiated by 28 participant laboratories. All 104 tests were completed; however, tests conducted on five samples (9232, 9222, 9233, 9234, and 9231) were invalid due to failure to meet test acceptability criteria for survival. The resulting successful test completion rate calculated in the WET Variability Study for the Ceriodaphnia acute test method was 95.2%. Four of the five invalid tests were conducted in a single laboratory (Lab 29). This laboratory failed to properly culture test organisms at the test temperature of 25 °C. The laboratory transferred organisms cultured at 20°C to the test temperature of 25 °C, causing significant mortality in all test treatments and failure of test acceptability criteria. 9.2.2 False Positive Rate A total of 33 valid tests were completed on blank samples by 14 participant laboratories (Table 9.2). The LC50 calculated for all 33 blank samples was >100%, indicating no toxicity and no false positives. The resulting false positive rate calculated in the WET Variability Study for the Ceriodaphnia acute test method was 0.00%. 9.2.3 Precision Precision of the Ceriodaphnia acute test method was estimated by calculating the CV of LC50 values obtained for the reference toxicant, effluent, and receiving water samples. Within-laboratory, between-laboratory, and total CVs were calculated for the reference toxicant and effluent samples. Only a total CV was calculated for the receiving water samples since no within-laboratory replication was provided for this sample type. All valid participant laboratory test data for the reference toxicant, effluent, and receiving water samples were used in estimating precision. No test results were identified by ASTM h statistics as possible outliers. Table 9.6 summarizes the precision of point estimates from the Ceriodaphnia acute test method. Within- laboratory CVs ranged from 9.68% to 14.6%, between-laboratory CVs ranged from 15.2% to 32.8%, and total CVs ranged from 21.1%to 34.2%. Total CVs were lower for reference toxicant samples (21.1%) than for 70 ------- effluent (34.2%) or receiving water (31.8%) samples. As expected, the majority of variability was due to the between-laboratory component, with within-laboratory CVs averaging 12.1% and between-laboratory CVs averaging 24.0%. Averaging the CVs based on total variance for the three sample types, a total CV of 29.0% was obtained for the Ceriodaphnia acute test method in the WET Variability Study. 71 ------- Table 9.2. Results for Ceriodaphnia acute test method performed on blank samples. LabID 29 6 18 18 18 33 33 33 46 46 46 62 62 69 69 70 70 70 73 73 73 105 157 157 157 251 251 311 311 417 417 452 452 452 Summary Statistics Sample code 9232 9221 9224 9225 9226 9235 9236 9237 9243 9244 9245 9250 9251 9254 9255 9258 9259 9260 9262 9263 9264 9270 9283 9284 9285 9295 9296 9299 9300 9315 9316 9322 9323 9324 N Min Max Median Mean Test date 11/09/99 11/09/99 11/09/99 11/11/99 11/11/99 11/09/99 11/09/99 11/11/99 11/09/99 11/09/99 11/11/99 11/09/99 11/09/99 11/09/99 11/09/99 11/09/99 11/09/99 11/11/99 11/09/99 11/09/99 11/11/99 11/09/99 11/09/99 11/09/99 11/11/99 11/09/99 11/09/99 11/09/99 11/09/99 11/09/99 11/11/99 11/09/99 11/11/99 11/11/99 False positives False positive rate Survival Information LC50 (% sample) Invalid" >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 33 >100 >100 >100 >100 0 0.00% Control mean (%) 5.00 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 90.0 90.0 100 100 100 100 100 100 100 100 100 100 100 100 100 100 95.0 95.0 Control CV (%) 41.8 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 11.2 11.2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 9.26 9.26 Flags3 ai, e2, f5 d,,e2 d,,e2 d,,e2 c4, d4 c4, d4 C4 C2 C2 C2 g5 b, a Data qualifier flags are described in Table 8.3. b Results from invalid tests were excluded from summary statistics. 72 ------- Table 9.3. Results for Ceriodaphnia acute test method performed on reference toxicant samples. LabID Referee Referee 3 3 3 25 25 25 42 42 42 60 101 101 101 113 113 113 125 125 141 141 141 205 238 406 416 416 416 425 425 425 Summary Statistics Sample code 9307 9310 9217 9218 9219 9227 9229 9230 9239 9241 9242 9246 9265 9267 9268 9273 9274 9275 9276 9279 9280 9281 9282 9287 9291 9303 9311 9312 9313 9318 9320 9321 N Min Max Median Mean Within-lab Between-lab Total Test date 11/09/99 11/11/99 11/09/99 11/09/99 11/11/99 11/09/99 11/11/99 11/11/99 11/09/99 11/11/99 11/11/99 11/09/99 11/09/99 11/11/99 11/11/99 11/09/99 11/09/99 11/11/99 11/09/99 11/11/99 11/09/99 11/11/99 11/11/99 11/09/99 11/09/99 11/09/99 11/09/99 11/11/99 11/11/99 11/09/99 11/11/99 11/11/99 STD cv% STD cv% STD cv% Survival Information LC50 (% sample) 40.6b 34.4b 31.9 27.6 18.3 34.4 33.0 33.0 37.9 35.4 50.0 17.1 21.7 25.9 27.7 35.4 32.0 35.4 26.2 24.8 31.2 32.7 35.4 39.2 30.8 31.8 34.2 24.1 26.8 35.4 30.0 28.9 30 17.1 50.0 31.8 30.9 4.52 14.6% 4.70 15.2% 6.52 21.1% Control mean (%) 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 90.0 100 100 100 100 100 100 100 100 100 100 100 100 100 100 Control CV (%) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 18.7 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Flags3 b4 61 63 63 63 b, b, Ci, e3 e2,e3 e2,e3 e2,e3 C2 C2 C2 GI, c2, d; e2!e6 Cl C2, C3 C2, C3 C2, C3 Cl Cl Cl,g5 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 73 ------- Table 9.4. Results for Ceriodaphnia acute test method performed on effluent samples. LabID Referee 6 29 29 6 33 60 60 62 62 69 69 70 105 105 125 157 205 205 238 238 251 251 311 311 406 406 417 Summary Statistics Sample code 9309 9222 9233 9234 9223 9238 9248 9249 9252 9253 9256 9257 9261 9271 9272 9278 9286 9289 9290 9293 9294 9297 9298 9301 9302 9305 9306 9317 N Min Max Median Mean Within-lab Between-lab Total Test date 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 11/11/99 STD cv% STD cv% STD cv% Survival Information LC50 (% sample) 24.6b Invalid0 Invalid0 Invalid0 9.99 13.0 13.2 10.3 28.2 26.8 27.7 33.0 21.2 23.6 23.1 20.3 35.4 36.6 30.8 16.9 16.5 17.1 19.5 24.1 21.8 27.7 27.0 15.0 24 9.99 36.6 22.4 22.5 2.17 9.68% 7.36 32.8% 7.68 34.2% Control mean (%) 100 40.0 0.00 0.00 100 100 95.0 95.0 90.0 100 100 100 90.0 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 Control CV (%) 0.00 59.5 0.00 0.00 0.00 0.00 9.26 9.26 18.7 0.00 0.00 0.00 11.2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Flags3 &i al5 dj, e2, f5 al5 dj, e2, f5 Cl> 63 Cl> 63 C2 b, bj, b4, Cj, c2, d; bj, b4, Cj, c2, d; e2 e2 t>l> Cl, C2 Cl> C2 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from 0 Results from invalid tests were excluded from summary summary statistics. statistics. 74 ------- Table 9.5. Results for Ceriodaphnia acute test method performed on receiving water samples. LabID Referee 29 6 25 42 60 101 105 125 205 238 406 417 425 Summary Statistics Sample code 9308 9231 9220 9228 9240 9247 9266 9269 9277 9288 9292 9304 9314 9319 N Min Max Median Mean STD CV% Test date 11/9/99 11/9/99 11/9/99 11/9/99 11/9/99 11/9/99 11/9/99 11/9/99 11/9/99 11/9/99 11/9/99 11/9/99 11/9/99 11/9/99 Survival Information LC50 (% sample) 30.8b Invalid0 20.8 34.2 26.8 16.5 21.5 11.1 33.2 34.2 23.3 19.9 20.4 17.7 12 11.1 34.2 21.2 23.3 7.40 31.8% Control mean (%) 100 5.00 100 100 100 100 100 100 90.0 100 100 100 100 100 Control CV (%) 0.00 41.8 0.00 0.00 0.00 0.00 0.00 0.00 11.2 0.00 0.00 0.00 0.00 0.00 Flags3 d, ab e2, f5 63 Ci,e3 e2, e3 bb b4, cb c2, dj e2, e6 Ci Ci a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. c Results from invalid tests were excluded from summary statistics. Table 9.6. Precision of point estimates from the Ceriodaphnia acute test method. Sample type Reference toxicant Effluent Receiving water Average CV (%) Within-laba 14.6 9.68 - 12.1 Between-laba 15.2 32.8 - 24.0 Total 21.1 34.2 31.8 29.0 a Within and between-laboratory components of variability were not calculated for the receiving water sample type since no within- laboratory replication was provided for this sample type. 75 ------- 9.3 Ceriodaphnia Chronic Test Method Results A total of 35 participant laboratories conducted the Ceriodaphnia chronic test method in the WET Variability Study. One of the participant laboratories did not submit a data report (see Section 8.1), so summarized results are based on 34 participant laboratories. These laboratories tested a total of 34 blank samples, 48 reference toxicant samples, 27 effluent samples, and 13 receiving water samples. For each sample tested, a survival NOEC, a reproduction NOEC, a survival LC50, and a reproduction IC25 were generated as test results. While all Ceriodaphnia chronic tests performed in the WET Variability Study were conducted for eight days, results presented in this section were determined using the current WET method manual criterion for test termination. This criterion states that tests should be terminated when 60% of the surviving control females have produced their third brood, or at the end of 8 days, whichever occurs first. As described in the WET method manual, reproduction was measured as the total number of young produced per original female at the time of appropriate test termination. Organisms positively identified as males were excluded from the reproduction analysis. Also, test concentrations above the survival NOEC were excluded from hypothesis testing conducted on the reproduction endpoint. Results of Ceriodaphnia chronic testing are shown in Tables 9.7 - 9.10 for each sample type. Precision estimates were not calculated for the reference toxicant sample type because this sample type failed to produce toxicity that could be definitively measured within the test concentration range. For the reference toxicant sample, 97.3% of LCSOs were >100% sample; 72.2% of IC25s were >100% sample. This was caused by a reference toxicant sample that was only moderately toxic. The spiking concentration of KC1 for this sample was selected to achieve an IC25 of approximately 50% sample based on preliminary testing. Despite preliminary testing efforts, the spiking level selected was insufficient to produce this targeted level of effect (see Section 5.3). 9.3.1 Successful Test Completion Rate A total of 122 Ceriodaphnia chronic tests were initiated by 34 participant laboratories. All 122 tests were completed; however, tests conducted on 22 samples were invalid due to failure to meet test acceptability criteria for survival or reproduction. The resulting successful test completion rate calculated in the WET Variability Study for the Ceriodaphnia chronic test method was 82.0%. In addition, the reproduction results for sample 9328 and the survival NOEC results for samples 9415 and 9380 were reported as inconclusive based on an evaluation of the concentration-response relationship (see Table 8.4). If these tests are considered unsuccessful in addition to invalid tests (since the test would be repeated in a regulatory context if the test endpoint required in the permit produced an inconclusive result), the successful test completion rate becomes 79.5%. Of the 34 participant laboratories, 24 produced valid results for all samples tested. The 22 invalid tests were concentrated in the remaining 10 laboratories. Of these 10 laboratories, 8 laboratories performed invalid tests on 50% or more of the samples tested. Two laboratories performed invalid tests on all samples tested. This attributed to the relatively low successful test completion rate achieved for the Ceriodaphnia chronic test method in the WET Variability Study. 76 ------- 9.3.2 False Positive Rate A total of 27 valid tests were conducted on blank samples by 22 participant laboratories (Table 9.7). No false positives were observed for the survival endpoint. The survival NOEC was 100% for all 27 blank samples, and the LC50 was >100% for all 27 blank samples. One false positive was observed for sublethal endpoints. The reproduction NOEC for sample 9450 was 25%, and the reproduction IC25 for this sample was 15.9%. The resulting false positive rate calculated in the WET Variability Study for the Ceriodaphnia chronic test method was 3.70% for the reproduction endpoint. The one false positive that was observed for this method originated from a laboratory that failed all other Ceriodaphnia chronic tests conducted. In addition to the false positive reported above, a participant laboratory reported a reproduction IC25 of less than 100% for sample 9332, indicating a false positive result. Based on EPA guidance for evaluating concentration-response relationships (USEPA, 2000a), this value was determined to be an anomalous result of the ICp (percentage inhibition concentration) smoothing procedure, and the IC25 was corrected to >100% (Table 8.4). A participant laboratory also reported a survival NOEC and reproduction NOEC of less than 100% for sample 9379, indicating a false positive result. This sample exhibited an interrupted concentration-response curve, and based on EPA guidance for evaluating concentration-response relationships, the survival NOEC and growth NOEC were recalculated and reported as 100% (Table 8.4). Sample 9341 also produced an interrupted concentration-response curve, but the reproduction NOEC was similarly recalculated and reported as 100% (Table 8.4). 9.3.3 Precision Precision of the Ceriodaphnia chronic test method was estimated by calculating the CV of LC50 and IC25 values obtained for the effluent and receiving water samples. CVs were not calculated for the reference toxicant sample type because this sample type failed to produce toxicity that could be definitively measured within the test concentration range (see Sections 5.3 and 9.1.4). For the effluent sample type, within-laboratory, between- laboratory, and total CVs were calculated. Only a total CV was calculated for the receiving water samples since no within-laboratory replication was provided for this sample type. Survival and reproduction results from samples 9347 and 9348 were excluded from the analysis of precision. These test samples were identified by ASTM h statistics as possible outliers, and review of data qualifier flags revealed a possible cause (Table 9.1). Table 9.11 summarizes the precision of point estimates from the Ceriodaphnia chronic test method. Within- laboratory and between-laboratory CVs for LC50 values were 7.09% and 21.8%, respectively. Total CVs ranged from 20.0% for the receiving water sample type to 23.0% for the effluent sample type. Averaging the CVs based on total variance for these two sample types, a total CV of 21.5% was obtained for the Ceriodaphnia chronic survival endpoint. CVs for IC25 values were higher than those for LC50 values. Within-laboratory CVs for the IC25 were 17.4%, between-laboratory CVs were 27.6%, and total CVs ranged from 32.6% to 37.4%. Averaging the CVs for the IC25 based on total variance for the two sample types, a total CV of 35.0% was obtained for the Ceriodaphnia chronic reproduction endpoint. 77 ------- The precision of NOEC values was determined by evaluating the range and distribution of NOEC values and the percentage of values falling within and beyond one concentration from the median. Table 9.12 describes the precision of NOEC values for the Ceriodaphnia chronic method. For the survival endpoint, NOEC values spanned two concentrations for the receiving water sample types and three concentrations for the reference toxicant and effluent sample types. The percentage of values within one concentration of the median was 97.2%, 91.3%, and 100% for the reference toxicant, effluent, and receiving water sample types, respectively. Reproduction NOEC values were less precise for the reference toxicant sample type, spanning five concentrations, but were comparable for the effluent and receiving water sample types, spanning three and two concentrations, respectively. The percentage of values within one concentration of the median was 83.3%, 100%, and 100% for the reference toxicant, effluent, and receiving water sample types, respectively. Of the eight results (for samples 9342, 9343, 9360, 9392, 9397, 9415, 9361, and 9362) that were beyond one concentration from the median, three (samples 9343, 9392, and 9415) were the result of interrupted concentration-response curves. Three results that were beyond one concentration from the median were from very statistically sensitive tests, with percent minimum significant differences (PMSDs) below EPA's recommended lower bound of 11% (USEPA, 2000d). The PMSDs for these three samples (9360, 9361, and 9362) were 8.7%, 8.1%, and 9.2%, respectively. 78 ------- Table 9.7. Results for Ceriodaphnia chronic test method performed on blank samples. LabID Referee 46 69 311 311 333 406 406 6 25 27 27 30 30 33 44 44 49 71 73 101 105 113 Sample code 9425 9352 9363 9411 9412 9420 9422 9423 9330 9332 9337 9338 9340 9341 9344 9349 9350 9356 9367 9371 9376 9379 9381 Test date 10/12/99 10/26/99 10/12/99 10/26/99 10/26/99 10/26/99 10/12/99 10/12/99 10/26/99 10/12/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/12/99 10/12/99 10/12/99 10/13/99 10/12/99 10/26/99 10/26/99 10/12/99 Survival Information NOEC (% sample) 100b Invalid0 Invalid" Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 LC50 (% sample) >100b Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 Control mean (%) 87.5 70.0 100 90.0 100 70.0 70.0 77.8 100 80.0 100 100 80.0 100 100 100 100 90.0 100 100 100 100 100 Reproduction Information NOEC (% sample) 100b Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 IC25 (% sample) >100b Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 Control mean (neonates) 17.1 2.40 9.10 9.50 8.40 26.1 21.1 21.6 25.4 19.2 20.1 24.2 16.0 23.5 15.8 30.8 29.5 24.1 23.1 21.5 20.4 24.9 26.5 Control CV (%) 48.6 79.1 53.7 43.9 56.7 69.7 50.4 54.9 26.0 63.0 27.8 12.3 58.6 22.1 51.7 6.46 4.86 18.7 15.1 20.5 17.2 27.4 12.4 Day of proper test termination 7 8 8 8 8 7 6 6 6 7 6 7 7 6 6 7 7 6 7 6 7 6 7 Flags3 db g2, ge ab a3, a4, ^4, db d4, e5, f3, g2 a4, c2, efo f3, g2 a3, a4, c2, g2 a3, a4, c2, g2 ab 62; 63; 66; §2, Ss ab cb c2, g2 ab cb c2, g2 f4 e6 b, g2 g5 g2 e2 e2 C2 c2, efo g5 c2, e6 79 ------- Table 9.7. Results for Ceriodaphnia chronic test method performed on blank samples, (continued) LabID 113 125 231 299 299 416 417 421 425 448 450 452 Summary Statistics Sample code 9382 9384 9402 9409 9410 9429 9432 9436 9439 9445 9446 9450 N Min Max Median Mean Test date 10/12/99 10/12/99 10/27/99 10/27/99 10/27/99 10/12/99 10/12/99 10/12/99 10/12/99 10/26/99 10/12/99 10/12/99 False positives False positive rate Survival Information NOEC (% sample) 100 100 100 100 100 100 100 100 100 100 100 100 27 100 100 100 0 0.00% LC50 (% sample) >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 27 >100 >100 >100 >100 0 0.00% Control mean (%) 100 90.0 100 90.0 100 100 90.0 100 100 100 100 90.0 Reproduction Information NOEC (% sample) 100 100 100 100 100 100 100 100 100 100 100 25 27 25 100 100 1 3.70% IC25 (% sample) >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 15.9 27 15.9 >100 >100 96.9 1 3.70% Control mean (neonates) 27.5 17.3 22.7 22.2 24.8 31.6 18.8 30.0 18.9 23.6 22.2 19.4 Control CV (%) 9.11 53.1 16.6 12.2 15.7 15.6 28.7 32.2 33.8 4.09 47.9 73.9 Day of proper test termination 7 7 8 6 6 7 8 7 8 6 6 7 Flags3 c2, e6 C2, g2 b2 b2,d, b2,d, C2, C3 C2 di,e6 Ci C2, g2 4 gb g2 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. c Results from invalid tests were excluded from summary statistics. 80 ------- Table 9.8. Results for Ceriodaphnia chronic test method performed on reference toxicant samples. LabID Referee 6 46 46 105 105 238 311 333 421 452 452 6 27 30 33 33 42 42 44 49 49 62 62 73 73 Sample code 9428 9329 9354 9355 9377 9378 9405 9413 9419 9438 9452 9453 9328 9336 9339 9342 9343 9345 9346 9351 9357 9358 9359 9360 9372 9373 Test date 10/26/99 10/12/99 10/26/99 10/26/99 10/12/99 10/12/99 10/12/99 10/26/99 10/26/99 10/26/99 10/27/99 10/27/99 10/12/99 10/12/99 10/12/99 10/12/99 10/12/99 10/12/99 10/12/99 10/26/99 10/26/99 10/26/99 10/12/99 10/12/99 10/26/99 10/26/99 Survival Information NOEC (% sample) 100b Invalid' Invalid' Invalid0 Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' 100 100 100 100 100 100 100 100 100 100 100 100 100 100 LC50 (% sample) >100b Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 Control mean (%) 100 100 100 100 100 60.0 30.0 100 50.0 77.8 90.0 80.0 100 100 100 100 100 100 100 100 90.0 100 100 100 100 90.0 Reproduction Information NOEC (% sample) 100b Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Inconclusive* 100 50 25 12.5 100 100 100 100 100 100 6.25 100 100 IC25 (% sample) >100b Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Inconclusive4 >100 <6.25 28.1 24.9 >100 >100 >100 >100 >100 >100 >100 >100 >100 Control mean (neonates) 25.3 10.8 5.60 5.00 14.8 7.30 9.50 6.30 13.1 28.0 7.80 9.00 15.9 24.7 25.8 19.2 21.1 18.0 17.2 32.8 19.8 21.5 25.6 27.4 22.7 23.0 Control CV (%) 11.9 59.0 68.5 21.1 46.7 95.1 16.6 65.3 115 53.7 58.2 35.1 48.6 9.36 38.7 39.0 18.1 13.4 12.2 13.2 40.7 14.7 8.48 7.14 12.8 20.1 Day of proper test termination 6 8 8 8 6 7 8 8 6 7 8 8 6 7 6 6 6 7 7 7 6 6 6 6 6 6 Flags3 d, a4,g2 83, a4, c4, dj, d4, e5, f3, g2 a3, a4, c4, dj, d4, e5, f3, g2 a4, c2, efo g2 &i, a4, c2, e6, g2,g5 aj, a3, a4, e2 a3, a4, c2, g2 ai, a4, e2, e3, e6; g2 ai, c3, dj, e6 a3, a4, b2, f2, g1: g2 a3, a4, b2, f2, g1: g2 g2 b, g2 g5 e2 C3 C3,f4 81 ------- Table 9.8. Results for Ceriodaphnia chronic test method performed on reference toxicant samples, (continued) LabID 101 101 113 125 125 134 141 141 205 231 231 299 313 313 406 416 416 417 417 421 448 448 450 Summary Statistics Sample code 9374 9375 9383 9386 9387 9389 9392 9393 9397 9400 9401 9408 9414 9415 9424 9430 9431 9434 9435 9437 9443 9444 9449 N Min Max Median Mean Test date 10/12/99 10/12/99 10/26/99 10/26/99 10/26/99 10/12/99 10/12/99 10/12/99 10/12/99 10/12/99 10/12/99 10/12/99 10/20/99 10/20/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/12/99 10/12/99 10/26/99 Survival Information NOEC (% sample) 100 100 100 100 100 100 100 100 25 100 100 100 100 Inconclusive4 100 100 100 100 100 100 100 100 100 36 25 100 100 LC50 (% sample) >100 >100 >100 >100 >100 >100 >100 >100 55.8 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 37 55.8 >100 >100 98.8 Control mean (%) 100 100 100 100 100 100 100 100 100 100 90.0 100 100 100 100 100 100 90.0 100 100 100 100 100 Reproduction Information NOEC (% sample) 100 100 100 100 100 100 12.5 50 6.25 50 50 100 100 12.5 100 100 100 100 100 100 100 100 100 36 6.25 100 100 IC25 (% sample) >100 >100 >100 >100 >100 >100 93.6 >100 <6.25 81.2 91.4 >100 85.6 7.67 >100 >100 >100 >100 82.7 >100 >100 >100 >100 36 <6.25 >100 >100 86.3 Control mean (neonates) 22.8 24.2 32.0 22.6 21.5 25.2 20.7 19.1 16.5 24.6 24.9 24.0 23.3 19.4 26.3 24.7 25.8 17.5 19.6 30.1 21.0 21.4 24.3 Control CV (%) 7.40 8.22 13.4 16.7 19.0 29.1 13.3 10.6 21.6 8.40 12.6 33.0 58.0 40.9 15.5 26.0 12.9 43.8 22.8 33.1 21.2 8.87 15.2 Day of proper test termination 6 6 7 6 6 6 7 7 6 6 6 6 7 6 6 6 6 8 8 7 6 6 6 Flags3 e6 C2 C2 f4,g2 c2, c3, d4, g5 C2, C3 bj, Cj, c2 di,e6 g2 82=85 ClJ4 C2, C3 C2, C3 C2,g2 C2J4 c3= dj, e6 C2 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. c Results from invalid tests were excluded from summary statistics. d Results were excluded from summary statistics. Based on EPA guidance for evaluating concentration-response relationships (EPA, 2000a) the test result was inconclusive and the sample should be retested to obtain a reliable result. 82 ------- Table 9.9. Results for Ceriodaphnia chronic test method performed on effluent samples. LabID Referee 6 25 25 42 42 62 62 69 69 71 71 105 134 134 141 141 205 205 231 238 238 313 313 333 425 Sample code 9427 9331 9334 9335 9347 9348 9361 9362 9365 9366 9369 9370 9380 9390 9391 9394 9395 9398 9399 9403 9406 9407 9416 9417 9421 9441 Test date 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/27/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 Survival Information NOEC (% sample) 25b Invalid0 25 25 50e 25e 6.25 6.25 25 25 25 25 Inconclusive4 12.5 25 25 12.5 25 12.5 25 12.5 12.5 25 25 25 25 LC50 (% sample) 33.0b Invalid0 33.4 33.7 70.7e 50.0e 11.8 12.5 35.4 35.4 35.4 35.4 27.9 26.5 26.0 30.9 26.8 28.7 26.8 28.7 16.5 20.2 33.0 28.7 27.3 30.8 Control mean (%) 100 100 100 90.0 100 100 90.0 100 100 100 100 100 90.0 100 100 100 100 100 100 100 100 100 100 100 80.0 100 Reproduction Information NOEC (% sample) 12.5b Invalid0 12.5 25 50e 25e 6.25 6.25 25 12.5 25 25 12.5 12.5 12.5 12.5 12.5 12.5 12.5 12.5 12.5 12.5 12.5 12.5 12.5 12.5 IC25 (% sample) 18.5b Invalid0 21.1 21.9 60.4e 37.8e 8.96 9.38 31.3 23.2 30.2 29.5 14.1 18.6 17.0 16.3 16.2 13.5 14.3 18.2 15.3 17.1 20.9 16.7 17.0 17.0 Control mean (neonates) 24.0 13.1 22.7 20.2 15.7 18.1 18.8 20.2 19.9 29.3 20.5 22.2 32.4 27.0 38.4 15.9 15.8 18.7 17.8 22.3 15.7 17.1 24.9 19.1 20.5 15.7 Control CV (%) 8.10 17.1 13.0 36.7 18.0 19.0 11.2 10.6 14.9 17.8 26.8 24.3 40.3 45.8 9.94 9.58 7.78 35.3 25.2 21.8 25.0 18.2 18.5 33.3 57.0 12.4 Day of proper test termination 6 6 8 8 7 7 6 6 7 6 6 6 8 6 6 6 6 6 6 8 7 8 6 6 6 8 Flags3 a4 §3, §4 §3, §4 b, bbd, C2 C2 c2, e6, g2, g5 4 §2 f4 C2, C3 c2, c3, d4 bb cb c2 bb cb c2 bi,b2 bb c2, e2 bb c2, e2 e2> e3, efo g2 c,,e6 83 ------- Table 9.9. Results for Ceriodaphnia chronic test method performed on effluent samples, (continued) LabID 425 450 Summary Statistics Sample code 9442 9448 N Min Max Median Mean Within-lab Between-lab Total Test date 10/26/99 10/26/99 STD cv% STD cv% STD cv% Survival Information NOEC (% sample) 25 12.5 23 6.25 25 25 LC50 (% sample) 26.0 21.9 24 11.8 35.4 28.3 27.5 1.95 7.09% 6.00 21.8% 6.31 23.0% Control mean (%) 100 88.9 Reproduction Information NOEC (% sample) 6.25 12.5 24 6.25 25 12.5 IC25 (% sample) <6.25 16.4 24 <6.25 31.3 17.0 17.9 3.13 17.4% 4.95 27.6% 5.85 32.6% Control mean (neonates) 16.4 21.8 Control CV (%) 11.9 40.0 Day of proper test termination 8 6 Flags3 Ci, e6, g10 c2, S6 * Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. c Results from invalid tests were excluded from summary statistics. d Results were excluded from summary statistics. Based on EPA guidance for evaluating concentration-response relationships (EPA, 2000a) the test result was inconclusive and the sample should be retested to obtain a reliable result. e Results were identified as outliers, a probable cause was identified, and therefore results were excluded from summary statistics. 84 ------- Table 9.10. Results for Ceriodaphnia chronic test method performed on receiving water samples. LabID Referee 46 238 452 25 69 71 125 134 205 333 417 425 450 Summary Statistics Sample code 9426 9353 9404 9451 9333 9364 9368 9385 9388 9396 9418 9433 9440 9447 N Min Max Median Mean STD CV% Test date 10/12/99 10/26/99 10/12/99 10/12/99 10/12/99 10/12/99 10/13/99 10/12/99 10/12/99 10/12/99 10/12/99 10/12/99 10/12/99 10/12/99 Survival Information NOEC (% sample) 25b Invalid0 Invalid0 Invalid0 50 25 25 25 25 25 25 25 25 25 10 25 50 25 LC50 (% sample) 35.4b Invalid0 Invalid0 Invalid0 59.0 40.5 40.6 35.4 37.9 34.6 53.6 40.8 33.0 44.3 10 33.0 59.0 40.6 42.0 8.38 20.0% Control mean (%) 100 80.0 40.0 90.0 100 100 100 90.0 100 100 100 100 100 80.0 Reproduction Information NOEC (% sample) 25b Invalid0 Invalid0 Invalid0 25 25 12.5 25 25 12.5 25 25 12.5 25 10 12.5 25 25 IC25 (% sample) 31. 3b Invalid0 Invalid0 Invalid0 53.8 27.1 23.0 31.3 31.6 10.5 28.1 32.6 22.3 33.4 10 10.5 53.8 29.7 29.4 11.0 37.4% Control mean (neonates) 20.5 5.50 7.60 10.2 26.6 15.0 20.7 17.3 31.6 17.2 36.2 18.4 19.5 21.0 Control CV (%) 30.3 67.6 29.2 45.0 10.7 37.7 20.1 53.1 39.9 34.1 7.79 40.5 13.7 77.3 Day of proper test termination 6 8 7 8 6 8 6 7 7 6 6 8 8 7 Flags3 a3, a4, bb b2, b3, c4, db d4, 65, 4 g2 ab a4, bb e2, g2 a4, f2, gi, g2 efo g5 C2, 4 g2 C2, g2 f4 bb cb c2 bb e2, e3, e4 c2, 4 g2 Ci c2, db g2 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from c Results from invalid tests were excluded from summary summary statistics. statistics. 85 ------- Table 9.11. Precision of point estimates from the Ceriodaphnia chronic test method. Sample type Reference toxicant b Effluent Receiving water Average CV (%) LC50 Within-lab a - 7.09 - 7.09 Between-lab a - 21.8 - 21.8 Total - 23.0 20.0 21.5 IC25 Within-lab a - 17.4 - 17.4 Between-lab a - 27.6 - 27.6 Total - 32.6 37.4 35.0 a Within- and between-laboratory components of variability were not calculated for the receiving water sample type since no within-laboratory replication was provided for this sample type. b Precision estimates were not calculated for the reference toxicant sample type because this sample type failed to produce toxicity that could be definitively measured within the test concentration range. Table 9.12. Precision of NOEC values from the Ceriodaphnia chronic test method. Sample type Reference toxicant Effluent Receiving water Endpoint Survival Reproduction Survival Reproduction Survival Reproduction NOEC Frequency <6.25% 0 0 0 0 0 0 6.25% 0 2 2 3 0 0 12.5% 0 3 6 17 0 3 25% 1 1 15 4 9 7 50% 0 4 0 0 1 0 100% 35 26 0 0 0 0 Median (% sample) 100 100 25 12.5 25 25 % within 1 cone, of median 97.2 83.3 91.3 100 100 100 % beyond 1 cone, of median 2.78 16.7 8.70 0.00 0.00 0.00 86 ------- 9.4 Fathead Acute Test Method Results A total of 29 participant laboratories conducted the fathead acute test method in the WET Variability Study. These laboratories tested a total of 27 blank samples, 38 reference toxicant samples, 29 effluent samples, and 13 receiving water samples. For each sample tested, a 96-hour LC50 was generated as a test result. Results of fathead acute testing are shown in Tables 9.13-9.16 for each sample type. 9.4.1 Successful Test Completion Rate A total of 107 fathead acute tests were initiated by 29 participant laboratories. All 107 tests were completed and met test acceptability criteria. The resulting successful test completion rate calculated in the WET Variability Study for the fathead acute test method was 100%. Two tests conducted by the referee laboratory were invalid due to control survival of 65%. These two tests were initiated on the same day, so poor health of organisms used for testing on that day is a likely cause. 9.4.2 False Positive Rate A total of 27 blank samples were analyzed by 22 participant laboratories (Table 9.13). The LC50 calculated for all 27 blank samples was >100%, indicating no toxicity and no false positives. The resulting false positive rate calculated in the WET Variability Study for the fathead acute test method was 0.00%. 9.4.3 Precision Precision of the fathead acute test method was estimated by calculating the CV of LC50 values obtained for the reference toxicant, effluent, and receiving water samples. Within-laboratory, between-laboratory, and total CVs were calculated for the reference toxicant and effluent samples. Only a total CV was calculated for the receiving water samples since no within-laboratory replication was provided for this sample type. All participant laboratory test data for the reference toxicant, effluent, and receiving water samples were used in estimating precision. While results on four test samples were identified by ASTM h statistics as possible outliers, these results were not excluded from the analysis of precision since a reasonable cause for the outliers could not be identified (Table 9.1). Table 9.17 summarizes the precision of point estimates from the fathead acute test method. CVs were consistent among sample types, with within-laboratory CVs ranging from 7.62% to 10.3%, between-laboratory CVs ranging from 19.2% to 19.7%, and total CVs ranging from 17.2% to 21.8%. As expected, the majority of variability was due to the between-laboratory component, with within-laboratory CVs averaging 8.96% and between-laboratory CVs averaging 19.4%. Averaging the CVs based on total variance for the three sample types, a total CV of 20.0% was obtained for the fathead acute test method in the WET Variability Study. 87 ------- Table 9.13. Results for fathead acute test method performed on blank samples. LabID Referee 3 3 6 27 33 42 42 46 60 60 70 73 101 146 157 157 205 231 238 244 251 311 313 406 416 452 452 Summary Statistics Sample code 9093 9002 9003 9004 9012 9018 9024 9025 9029 9030 9031 9037 9041 9044 9056 9061 9062 9063 9069 9073 9077 9081 9082 9085 9089 9097 9109 9110 N Min Max Median Mean Test date 10/21/99 11/04/99 11/04/99 10/21/99 10/21/99 11/04/99 10/21/99 10/21/99 11/04/99 10/22/99 10/22/99 10/21/99 10/21/99 10/21/99 10/21/99 11/04/99 11/04/99 10/21/99 11/04/99 11/04/99 11/04/99 11/04/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 False positives False positive rate Survival Information LC50 (% sample) >100b >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 27 >100 >100 >100 >100 0 0.00% Control mean (%) 95.0 90.0 100 100 95.0 100 100 95.0 100 90.0 100 100 100 100 100 100 100 100 100 100 100 100 100 100 95.0 100 100 100 Control CV (%) 8.66 17.1 0.00 0.00 8.66 0.00 0.00 8.66 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 8.66 0.00 0.00 0.00 Flags3 e4 b5 b5 c4, d4 b2, cb d7, e2, e3 b2, cb d7, e2, e3 db e4 bb e2 e2, e3 db e2 db e4 Ci a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 88 ------- Table 9.14. Results for fathead acute test method performed on reference toxicant samples. LabID Referee o 6 6 6 25 25 33 41 41 42 46 46 60 62 62 73 73 101 101 105 141 141 146 157 231 231 238 238 244 244 251 251 311 311 313 313 417 Sample code 9096 9001 9006 9007 9008 9009 9017 9020 9021 9026 9027 9028 9032 9033 9034 9042 9043 9046 9047 9049 9052 9053 9059 9060 9067 9068 9071 9072 9075 9076 9079 9080 9083 9084 9087 9088 9102 Test date 11/04/99 10/21/99 11/04/99 11/04/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 11/04/99 10/21/99 10/21/99 11/04/99 10/21/99 10/21/99 11/04/99 11/04/99 11/04/99 11/04/99 10/21/99 10/21/99 10/21/99 11/04/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 11/04/99 11/04/99 11/04/99 11/04/99 10/21/99 Survival Information LC50 (% sample) Invalid13 41.8 43.6 46.7 35.4 40.1 32.3 26.8 24.1 31.9 23.4 29.6 33.9 63.9 55.5 35.4 34.2 55.4 53.6 36.7 34.0 37.9 32.7 33.0 34.2 38.2 39.2 46.7 33.7 34.3 34.2 34.6 39.2 42.0 45.6 46.8 42.1 Control mean (%) 65.0 95.0 100 100 100 100 100 100 100 100 95.0 95.0 100 100 100 100 100 100 100 95.0 95.0 95.0 100 100 100 100 100 100 100 100 100 100 100 100 95.0 95.0 100 Control CV (%) 7.92 8.66 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 8.66 8.66 0.00 0.00 0.00 0.00 0.00 0.00 0.00 8.66 8.66 8.66 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 8.66 8.66 0.00 Flags3 ai gi gi e2 e2 c4, d4 c4, d4 cb d2, d3, e2, e3 g3 g3 e2 dbe4 e2, e3 e2, e3 db e2 db e2 d, d, e2, e4 89 ------- Table 9.14. Results for fathead acute test method performed on reference toxicant samples, (continued) LabID 425 452 Summary Statistics Sample code 9106 9111 N Min Max Median Mean Within-lab Between-lab Total Test date 10/21/99 11/04/99 STD cv% STD cv% STD cv% Survival Information LC50 (% sample) 43.5 30.9 38 23.4 63.9 36.0 38.6 2.94 7.62% 7.60 19.7% 8.15 21.1% Control mean (%) 100 100 Control CV (%) 0.00 0.00 Flags3 61 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 90 ------- Table 9.15. Results for fathead acute test method performed on effluent samples. LabID Referee 25 25 27 27 33 41 41 62 62 70 70 105 105 141 141 146 205 205 231 238 244 406 406 416 416 417 417 425 425 Summary Statistics Sample code 9095 9010 9011 9014 9015 9019 9022 9023 9035 9036 9039 9040 9050 9051 9054 9055 9058 9065 9066 9070 9074 9078 9091 9092 9099 9100 9103 9104 9107 9108 N Min Max Median Mean Within-lab Between-lab Total Test date 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 11/4/99 STD cv% STD cv% STD cv% Survival Information LC50 (% sample) Invalid" 21.8 22.5 18.9 19.8 19.6 16.2 18.3 27.7 23.3 21.8 21.0 23.9 15.8 17.5 15.7 18.2 37.7 32.7 18.9 19.6 19.6 27.7 26.2 22.3 23.1 18.2 18.3 19.7 19.1 29 15.7 37.7 19.7 21.6 2.22 10.3% 4.14 19.2% 4.70 21.8% Control mean (%) 65.0 100 100 100 100 100 95.0 100 100 100 100 100 100 100 95.0 95.0 100 100 100 100 100 100 100 100 95.0 100 100 100 95.0 90.0 Control CV (%) 24.0 0.00 0.00 0.00 0.00 0.00 8.66 0.00 0.00 0.00 0.00 0.00 0.00 0.00 8.66 8.66 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 8.66 0.00 0.00 0.00 8.66 17.1 Flags3 a^d; gi gi e2 bl5e2 e4 e4 61 d2,e2 d2,e2 dl5e4 e2,g3 e2,g3 e2,e3 dl5e2 Cl=e4 Cl=e4 e2, e4 e2, e4 d, d, a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 91 ------- Table 9.16. Results for fathead acute test method performed on receiving water samples. LabID Referee 6 27 33 70 101 105 146 205 313 406 416 417 425 Summary Statistics Sample code 9094 9005 9013 9016 9038 9045 9048 9057 9064 9086 9090 9098 9101 9105 N Min Max Median Mean STD CV% Test date 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 10/21/99 Survival Information LC50 (% sample) 19.6b 17.7 17.7 16.6 18.9 26.2 16.9 20.3 16.2 23.9 19.1 16.9 21.6 25.0 13 16.2 26.2 18.9 19.8 3.41 17.2% Control mean (%) 95.0 100 100 100 100 95.0 95.0 100 100 100 100 100 100 100 Control CV (%) 8.66 0.00 0.00 0.00 0.00 8.66 8.66 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Flags3 e4 e2 dbe4 bb e2 Ci e2, e4 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. Table 9.17. Precision of point estimates from the fathead acute test method. Sample type Reference toxicant Effluent Receiving water Average CV (%) Within-lab a 7.62 10.3 - 8.96 Between-lab a 19.7 19.2 - 19.4 Total 21.1 21.8 17.2 20.0 a Within- and between-laboratory components of variability were not calculated for the receiving water sample type since no within- laboratory replication was provided for this sample type. 92 ------- 9.5 Fathead Chronic Test Method Results A total of 27 participant laboratories conducted the fathead chronic test method in the WET Variability Study. These laboratories tested a total of 24 blank samples, 37 reference toxicant samples, 27 effluent samples, and 13 receiving water samples. For each sample tested, a 7-day survival NOEC, a 7-day growth NOEC, a 7-day survival LC50, and a 7-day growth IC25 were generated as test results. As described in the WET method manual, growth was measured as the total weight per replicate divided by the number of original organisms in that replicate. This definition provides a combined growth and survival endpoint that is more accurately termed biomass. Also, test concentrations above the survival NOEC were excluded from hypothesis testing conducted on the growth endpoint. Results of fathead chronic testing are shown in Tables 9.18 - 9.21 for each sample type. 9.5.1 Successful Test Completion Rate A total of 101 fathead chronic tests were initiated by 27 participant laboratories. All 101 tests were completed; however, tests conducted on samples 9199 and 9118 were invalid due to failure to meet test acceptability criteria for growth and survival, respectively. The resulting successful test completion rate calculated in the WET Variability Study for the fathead chronic test method was 98.0%. In addition, the growth NOEC result for sample 9209 was reported as inconclusive based on an evaluation of the concentration-response relationship (Table 8.4). If this test is considered unsuccessful (since the test would be repeated in a regulatory context if the test endpoint required in the permit produced an inconclusive result), the successful test completion rate becomes 97.0%. 9.5.2 False Positive Rate A total of 24 blank samples were analyzed by 20 participant laboratories (Table 9.18). No false positives were observed for the survival endpoint. The survival NOEC was 100% for all 24 blank samples, and the LC50 was >100% for all 24 blank samples. One false positive was observed for sublethal endpoints. The growth NOEC for sample 9158 was 50%, and the growth IC25 for this sample was 93.6%. The resulting false positive rate calculated in the WET Variability Study for the fathead chronic test method was 4.35% for the growth NOEC and 4.17% for the growth IC25. The one false positive that was observed was due to poor survival in a single replicate of the 100% test concentration treatment. For this sample, the survival in the 100% test concentration was 90%, 100%, 90%, and 50% for the 4 replicates, respectively. Disregarding replicate 4, the survival for this treatment would be identical to the control survival (95%). In addition to the false positive reported above, participant laboratories reported a NOEC of less than 100% for two additional samples (9145 and 9209), indicating false positive results. These samples exhibited an interrupted concentration-response curve. Based on EPA guidance for evaluating concentration-response relationships, the growth and survival NOEC for sample 9145 was recalculated and reported as 100%, and the growth NOEC for sample 9209 was reported as inconclusive (Table 8.4). 93 ------- 9.5.3 Precision Precision of the fathead chronic test method was estimated by calculating the CV of LC50 and IC25 values obtained for the reference toxicant, effluent, and receiving water samples. Within-laboratory, between- laboratory, and total CVs were calculated for the reference toxicant and effluent samples. Only a total CV was calculated for the receiving water samples since no within-laboratory replication was provided for this sample type. Results from two samples (9162 and 9163) were excluded from the analysis of precision. These test samples were identified by ASTM h statistics as possible outliers, and review of data qualifier flags revealed a possible cause (Table 9.1). One additional sample (9177) was identified as a possible outlier but was not excluded from the analysis of precision since a reasonable cause was not identified. Table 9.22 summarizes the precision of point estimates from the fathead chronic test method. CVs for LC50 values were consistent among sample types, with within-laboratory CVs ranging from 6.59% to 9.16%, between-laboratory CVs ranging from 10.6% to 12.0%, and total CVs ranging from 12.5% to 15.1%. As expected, the majority of variability was due to the between-laboratory component, with within-laboratory CVs averaging 7.87% and between-laboratory CVs averaging 11.3%. Averaging the CVs for the LC50 based on total variance for the three sample types, a total CV of 13.4% was obtained for the fathead chronic survival endpoint. CVs for IC25 values were higher than those for LC50 values. Within-laboratory CVs for the IC25 ranged from 10.0% to 19.1%, between-laboratory CVs ranged from 12.9% to 17.2%, and total CVs ranged from 19.8% to 23.1%. Within-laboratory CVs (averaging 14.6%) were only slightly lower than between-laboratory CVs (averaging 15.0%). Averaging the CVs for the IC25 based on total variance for the three sample types, a total CV of 20.9% was obtained for the fathead chronic growth endpoint. The precision of NOEC values was determined simply by evaluating the range and distribution of NOEC values and the percentage of values falling within and beyond one concentration from the median. Table 9.23 describes the precision of NOEC values for the fathead chronic method. For the survival endpoint, NOEC values spanned four concentrations for the reference toxicant sample type and two concentrations for the effluent and receiving water sample types. The percentage of values within one concentration of the median was 97.2%, 100%, and 100% for the reference toxicant, effluent, and receiving water sample types, respectively. Growth NOEC values were less precise, spanning five concentrations for the reference toxicant sample type and four concentrations for the effluent and receiving water sample types. The percentage of values within one concentration of the median was 86.1%, 91.7%, and 76.9% for the reference toxicant, effluent, and receiving water sample types, respectively. Of the 10 results (for samples 9150, 9177, 9193, 9194, 9212, 9129, 9166, 9161, 9176, and 9180) that were beyond 1 concentration from the median, 5 results (samples 9193, 9194, 9212, 9129, and 9161) were the consequence of interrupted concentration-response curves. For each of these samples, test concentrations higher than the reported NOEC were not significantly different from the control, but the NOEC was reported as the concentration below the LOEC based on EPA's concentration-response guidance (Table 8.4). Three results that were beyond one concentration from the median were from very statistically sensitive tests, with PMSDs near or below EPA's recommended lower bound of 9.4% (USEPA, 2000d). The PMSDs for these three samples (9150, 9166, and 9180) were 9.2%, 9.7%, and 10.3%, respectively. The 94 ------- remaining two samples (9176 and 9177) that were beyond one concentration from the median were from the same laboratory (205) and were conducted on the same day, indicating that test organisms or specific test procedures in this laboratory may have produced conditions that caused greater sensitivity. 95 ------- Table 9.18. Results for fathead chronic test method performed on blank samples. LabID Referee 3 3 6 25 27 33 42 42 46 49 49 62 73 101 105 125 231 238 299 311 333 333 406 425 Sample code 9201 9113 9114 9117 9119 9123 9131 9135 9136 9138 9142 9143 9145 9151 9152 9158 9160 9182 9186 9188 9192 9196 9197 9198 9209 Test date 09/28/99 10/05/99 10/05/99 10/05/99 09/28/99 09/28/99 09/29/99 09/28/99 09/28/99 09/28/99 09/28/99 09/28/99 09/28/99 10/05/99 09/28/99 10/05/99 09/28/99 10/05/99 10/05/99 09/28/99 09/28/99 10/05/99 10/05/99 09/28/99 09/28/99 Survival Information NOEC (% sample) 100b 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 LC50 (% sample) >100b >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 Control mean (%) 100 100 95.0 97.5 100 100 100 90.0 92.5 96.7 95.2 97.5 95.0 100 100 95.0 92.5 97.5 95.0 100 100 97.5 95.0 90.0 100 Control CV (%) 0.00 0.00 11.4 5.94 0.00 0.00 0.00 9.94 6.32 6.93 6.73 5.94 7.07 0.00 0.00 11.4 6.32 5.94 7.07 0.00 0.00 5.94 11.4 9.94 0.00 Growth Information NOEC (% sample) 100b 100 100 100 100 100 100 100 100 100 100 100 100 100 100 50 100 100 100 100 100 100 100 100 Inconclusive0 IC25 (% sample) >100b >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 93.6 >100 >100 >100 >100 >100 >100 >100 >100 >100 Control mean (mg) 0.477 0.381 0.365 0.371 0.511 0.788 0.383 0.471 0.506 0.300 0.360 0.379 0.293 0.807 0.375 0.392 0.382 0.663 0.373 0.770 0.503 0.602 0.693 0.262 0.643 Control CV (%) 5.23 7.59 11.7 8.46 4.90 4.50 16.1 10.6 5.02 12.0 9.88 3.77 18.2 8.70 22.9 12.5 16.0 6.38 13.4 2.93 4.58 5.75 17.2 9.51 6.89 Flags3 db ej e3 b, b2 b, bb d7 c4, d4, fj 61 61 db g5 c4, e2 C2, g2 e2, e3 c2, e2, e6, g2 db eb e2, e6 eb e6 bi, cb fj c,, g,, g, 96 ------- Table 9.18. Results for fathead chronic test method performed on a blank samples, (continued) LabID Summary Statistics Sample code N Min Max Median Mean Test date False positives False positive rate Survival Information NOEC (% sample) 24 100 100 100 0 0.00% LC50 (% sample) 24 >100 >100 >100 >100 0 0.00% Control mean (%) Control CV (%) Growth Information NOEC (% sample) 23 50 100 100 1 4.35% IC25 (% sample) 24 93.6 >100 >100 99.7 1 4.17% Control mean (mg) Control CV (%) Flags3 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. c Results were excluded from summary statistics. Based on EPA guidance for evaluating concentration-response relationships (EPA, 2000a) the test result was inconclusive and the sample should be retested to obtain a reliable result. 97 ------- Table 9.19. Results for fathead chronic test method performed on reference toxicant samples. LabID Referee 406 o J 6 6 25 27 27 30 30 42 49 62 62 73 73 105 105 134 141 141 157 205 231 238 Sample code 9204 9199 9112 9115 9116 9122 9125 9126 9127 9128 9137 9144 9147 9148 9149 9150 9156 9157 9165 9168 9169 9173 9177 9181 9184 Test date 10/05/99 10/05/99 09/29/99 09/29/99 09/29/99 10/05/99 10/05/99 10/05/99 09/28/99 09/28/99 10/05/99 10/05/99 10/05/99 10/05/99 09/28/99 09/28/99 09/29/99 09/29/99 09/28/99 09/28/99 09/28/99 09/28/99 09/28/99 09/28/99 09/28/99 Survival Information NOEC (% sample) 50b Invalid0 50 50 50 50 25 25 25 25 50 50 50 50 50 25 50 50 50 50 50 50 50 50 50 LC50 (% sample) 70.7b Invalid0 83.7 70.7 66.2 51.4 40.6 54.9 53.2 58.4 66.4 68.5 67.9 62.1 64.0 65.0 63.8 65.5 65.0 69.5 73.7 66.8 69.0 68.3 73.7 Control mean (%) 100 95.0 100 97.5 97.5 100 100 100 95.0 100 95.0 85.0 92.5 90.0 97.5 100 92.5 100 100 100 100 95.0 97.5 100 100 Control CV (%) 0.00 7.07 0.00 5.94 5.94 0.00 0.00 0.00 7.07 0.00 7.07 10.9 6.32 14.0 5.94 0.00 6.32 0.00 0.00 0.00 0.00 7.07 5.94 0.00 0.00 Growth Information NOEC (% sample) 50b Invalid0 50 50 50 50 25 25 25 25 50 50 50 50 25 12.5 50 50 25 50 50 25 6.25 50 50 IC25 (% sample) 63. 3b Invalid0 64.2 62.5 56.7 51.4 33.5 47.5 26.8 45.0 57.1 58.4 53.6 54.2 51.5 56.8 58.2 58.9 49.4 57.4 63.2 53.4 20.2 59.1 59.4 Control mean (mg) 0.435 0.230 0.534 0.250 0.274 0.467 0.671 0.692 0.687 0.574 0.506 0.386 0.275 0.303 0.754 0.697 0.246 0.258 0.617 0.433 0.433 0.505 0.478 0.490 0.442 Control CV (%) 14.3 9.51 4.26 15.3 7.97 9.59 7.57 7.40 32.5 14.2 7.33 7.55 4.70 16.9 7.83 6.87 14.1 17.0 14.1 15.6 15.6 11.0 7.13 9.53 18.1 Flags3 db ej a2, cb e6, fb g5 b2, c2, fj b2 b2 e3, g5 b, b, b, d, b2, d2, e2 b2, d2, e2 C2, C3, g5 C2, C3 C3 g3 e2, e3 e2 98 ------- Table 9.19. Results for fathead chronic test method performed on reference toxicant samples, (continued) LabID 238 299 299 311 311 333 406 416 416 425 425 452 452 Summary Statistics Sample code 9185 9190 9191 9193 9194 9195 9200 9205 9206 9211 9212 9213 9214 N Min Max Median Mean Within-lab Between-lab Total Test date 09/28/99 10/05/99 10/05/99 10/05/99 10/05/99 09/28/99 10/05/99 09/28/99 09/28/99 10/05/99 10/05/99 09/29/99 09/29/99 STD cv% STD cv% STD cv% Survival Information NOEC (% sample) 50 25 25 50 50 50 50 50 50 25 6.25 50 50 36 6.25 50 50 LC50 (% sample) 74.5 59.5 49.6 68.8 69.5 70.7 60.9 71.2 70.7 62.7 54.3 69.5 71.2 36 40.6 83.7 66.6 65.0 4.29 6.59% 6.87 10.6% 8.10 12.5% Control mean (%) 100 100 100 100 100 92.5 90.0 100 100 100 100 97.5 100 Control CV (%) 0.00 0.00 0.00 0.00 0.00 11.3 9.94 0.00 0.00 0.00 0.00 5.94 0.00 Growth Information NOEC (% sample) 50 25 25 <6.25 12.5 50 50 50 50 25 6.25 50 50 36 <6.25 50 50 IC25 (% sample) 56.3 35.7 34.0 56.6 58.0 59.8 58.8 58.9 63.1 57.0 50.4 60.9 57.3 36 20.2 64.2 56.9 52.9 5.30 10.0% 9.10 17.2% 10.5 19.9% Control mean (mg) 0.532 0.743 0.686 0.413 0.392 0.632 0.235 0.454 0.414 0.563 0.493 0.588 0.710 Control CV (%) 16.1 2.97 9.18 9.46 3.76 10.3 12.2 2.22 9.56 4.44 9.14 12.8 6.65 Flags3 c3, e2, e6 db e2, e6 db e2, e6 g5 g5 GI, e6, fj C2, C3 C2 C2 C2> §5 b2, e2 b2, e2 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from c Results from invalid tests were excluded from summary summary statistics. statistics. 99 ------- Table 9.20. Results for fathead chronic test method performed on effluent samples. LabID Referee 6 25 30 30 33 33 46 46 101 101 105 125 125 134 134 141 141 157 157 205 205 231 238 416 416 Sample code 9203 9118 9121 9129 9130 9133 9134 9140 9141 9154 9155 9159 9162 9163 9166 9167 9170 9171 9174 9175 9178 9179 9183 9187 9207 9208 Test date 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 10/05/99 Survival Information NOEC (% sample) 25b Invalid0 12.5 12.5 12.5 12.5 12.5 12.5 12.5 25 25 12.5 12.5 25 12.5 12.5 12.5 12.5 12.5 25 12.5 12.5 12.5 12.5 12.5 25 LC50 (% sample) 31. lb Invalid0 27.8 21.6 19.9 24.5 24.5 19.3 20.0 32.1 30.5 26.3 21.1 27.6 30.3 31.3 29.7 27.6 29.6 28.4 27.7 25.4 24.1 24.2 21.0 28.8 Control mean (%) 97.5 70.0 92.5 92.5 85.0 95.0 87.5 95.0 97.5 97.5 97.5 95.0 92.5 92.5 100 100 100 100 97.5 95.0 100 100 100 100 100 97.5 Control CV (%) 5.94 23.8 6.32 11.3 20.2 11.4 20.0 11.4 5.94 5.94 5.94 11.4 11.3 11.3 0.00 0.00 0.00 0.00 5.94 11.4 0.00 0.00 0.00 0.00 0.00 5.94 Growth Information NOEC (%sample) 12.5b Invalid0 12.5 <6.25 12.5 12.5 12.5 12.5 12.5 25 12.5 12.5 <6.25d <6.25d <6.25 6.25 12.5 6.25 12.5 25 <12.5 12.5 12.5 12.5 12.5 12.5 IC25 (%sample) 26.9b Invalid0 20.6 16.1 16.1 14.9 21.2 16.1 16.3 29.2 20.5 18.4 7.04d <6.25d 18.4 26.1 23.8 15.6 29.0 26.8 20.5 17.3 19.0 16.9 14.8 23.7 Control mean (mg) 0.430 0.348 0.448 0.490 0.375 0.686 0.644 0.345 0.395 0.490 0.528 0.366 0.390 0.390 0.607 0.593 0.435 0.435 0.545 0.494 0.403 0.420 0.545 0.451 0.703 0.698 Control CV (%) 10.7 27.4 11.6 10.9 29.9 20.7 16.6 16.8 14.7 14.3 5.66 5.22 8.36 8.36 7.78 3.49 4.79 4.79 12.4 13.7 6.53 15.4 6.81 6.61 9.21 15.2 Flags3 61 ai e3 d2, gs d2 b,,g2 b, c4, d4, fj c4, d4, fj d2, e2 C3, g3 C3, g3 d, c2, c3, d2 c2, c3, d2 bb b4, £7 bb b4, e7 bb e2, e3 e2 C2, C3, fj c2, c,, f, 100 ------- Table 9.20. Results for fathead chronic test method performed on effluent samples, (continued) LabID 452 452 Summary Statistics Sample code 9215 9216 N Min Max Median Mean Within-lab Between- lab Total Test date 10/05/99 10/05/99 STD cv% STD cv% STD cv% Survival Information NOEC (% sample) 12.5 25 26 12.5 25 12.5 LC50 (% sample) 30.4 34.1 26 19.3 34.1 27.6 26.5 2.42 9.16% 3.16 12.0% 3.99 15.1% Control mean (%) 100 95.0 Control CV (%) 0.00 11.4 Growth Information NOEC (%sample) 12.5 25 24 <6.25 25 12.5 IC25 (%sample) 26.2 29.8 24 14.8 29.8 19.8 20.7 3.96 19.1% 2.68 12.9% 4.78 23.1% Control mean (mg) 0.461 0.420 Control CV (%) 10.0 6.90 Flags3 e2 e2 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. c Results from invalid tests were excluded from summary statistics. d Results were identified as outliers, a probable cause was identified, and therefore results were excluded from summary statistics. 101 ------- Table 9.21. Results for fathead chronic test method performed on receiving water samples. LabID Referee 25 27 33 46 62 101 125 134 157 205 231 299 425 Summary Statistics Sample code 9202 9120 9124 9132 9139 9146 9153 9161 9164 9172 9176 9180 9189 9210 N Min Max Median Mean STD CV% Test date 9/28/99 9/28/99 9/28/99 9/29/99 9/28/99 9/28/99 9/28/99 9/28/99 9/28/99 9/28/99 9/28/99 9/28/99 9/28/99 9/28/99 Survival Information NOEC (% sample) 25b 25 12.5 25 12.5 25 12.5 25 25 25 25 25 12.5 25 13 12.5 25 25 LC50 (% sample) 34.9b 31.2 28.7 34.2 28.0 32.5 31.2 37.4 35.4 34.1 41.8 37.2 27.7 37.1 13 27.7 41.8 34.1 33.6 4.22 12.6% Control mean (%) 97.5 95.0 100 100 100 95.0 97.5 92.5 97.5 90.0 97.5 100 100 100 Control CV (%) 5.94 7.07 0.00 0.00 0.00 7.07 5.94 6.32 5.94 14.0 5.94 0.00 0.00 0.00 Growth Information NOEC (% sample) 25b 25 12.5 25 6.25 25 12.5 <6.25 12.5 25 <6.25 <6.25 12.5 25 13 <6.25 25 12.5 IC25 (%sample) 29.5b 29.3 28.0 29.3 19.0 29.2 25.9 28.6 25.4 29.8 16.1 22.6 18.8 32.4 13 16.1 32.4 28.0 25.7 5.08 19.8% Control mean (mg) 0.473 0.510 0.655 0.381 0.563 0.285 0.353 0.382 0.575 0.545 0.433 0.506 0.719 0.675 Control CV (%) 4.34 5.48 8.70 13.9 11.8 8.35 23.2 16.0 5.88 14.4 17.7 8.43 10.2 12.7 Flags3 61 e3 b2 c4, db d4, e6, fi d, C2, §5 bb b4, d2 bb e2, e3 e2, e6 C2 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 102 ------- Table 9.22. Precision of point estimates from the fathead chronic test method. Sample type Reference toxicant Effluent Receiving water Average CV (%) LC50 Within-lab a 6.59 9.16 - 7.87 Between-lab a 10.6 12.0 - 11.3 Total 12.5 15.1 12.6 13.4 IC25 Within-lab a 10.0 19.1 - 14.6 Between-lab a 17.2 12.9 - 15.0 Total 19.9 23.1 19.8 20.9 a Within- and between-laboratory components of variability were not calculated for the receiving water sample type since no within-laboratory replication was provided for this sample type. Table 9.23. Precision of NOEC values from the fathead chronic test method. Sample type Reference toxicant Effluent Receiving water Endpoint Survival Growth Survival Growth Survival Growth NOEC Frequency <6.25% 0 1 0 2 0 3 6.25% 1 2 0 3 0 1 12.5% 0 2 20 16 4 4 25% 8 10 6 3 9 5 50% 27 21 0 0 0 0 100% 0 0 0 0 0 0 Median (% sample) 50 50 12.5 12.5 25 12.5 % within 1 cone, of median 97.2 86.1 100 91.7 100 76.9 % beyond 1 cone, of median 2.78 13.9 0.00 8.33 0.00 23.1 103 ------- 9.6 Selenastrum Chronic Test Method Results A total of 11 participant laboratories conducted the Selenastrum chronic test method in the WET Variability Study. These laboratories tested a total of 8 blank samples, 13 reference toxicant samples, 15 effluent samples, and 8 receiving water samples. Each of these samples was tested with and without the addition of EDTA to the sample and dilution water. For each test conducted, a 96-hour growth NOEC, a 96-hour growth IC25, and a 96- hour growth IC50 were generated as test results. The growth endpoint was measured as cell density (cells per mL). Test acceptability criteria were assessed independently for tests conducted with EDTA (minimum control cell density of IxlO6 cells/mL) and without EDTA (minimum control cell density of 2xl05 cells/mL). The test acceptability criteria for control variability was determined by calculating the CV of growth in control replicates (control CV must be less than 20% to meet test acceptability criteria). Results of Selenastrum chronic testing are shown in Tables 9.24 - 9.31 for each sample type tested with and without the addition of EDTA. 9.6.1 Successful Test Completion Rate A total of 44 Selenastrum chronic tests were initiated with the addition of EDTA, and 44 tests were initiated without the addition of EDTA. All 88 tests were completed; however, 16 tests conducted with EDTA and 15 tests conducted without EDTA were invalid due to failure to meet test acceptability criteria for control growth or control variability. The resulting successful test completion rate calculated in the WET Variability Study for the Selenastrum chronic test method was 63.6% with EDTA and 65.9% without EDTA. In addition, growth endpoints for samples 9473 without EDTA, 9455 with EDTA, 9454 without EDTA, and 9468 without EDTA were reported as inconclusive based on an evaluation of the concentration-response relationship (Table 8.4). If these tests are considered unsuccessful (since the tests would be repeated in a regulatory context if the test endpoint required in the permit produced an inconclusive result), the successful test completion rate becomes 61.4% with EDTA and 59.1% without EDTA. Two distinct patterns of test failures were observed for laboratories conducting the Selenastrum chronic test method. The first pattern involved the failure of both tests (with and without EDTA) conducted on a given sample and a given day. This pattern was observed in laboratories 39, 299, 33, and 459. All but 2 of the 18 test failures from these laboratories were observed to occur for tests conducted with and without EDTA on a given sample. For instance, laboratory 39 failed tests with and without EDTA on samples 9466 and 9467, but passed tests with and without EDTA on samples 9468 and 9469. The second pattern of test failures involved the failure of all tests conducted with a given nutrient type (with or without EDTA). Laboratories 33, 406, and 209 exhibited this pattern. Laboratories 33 and 406 failed all tests (8 of 8) with EDTA and passed all but 1 (7 of 8) test without EDTA. Laboratory 209 failed all tests (4 of 4) without EDTA and passed all tests with EDTA. This pattern was further explained by the culturing methods and general testing procedures used in each laboratory. Laboratories 33 and 406, which failed all tests with EDTA, normally cultured organisms and conducted tests without the addition of EDTA. Laboratory 209, which failed all tests without EDTA, normally cultured organisms and conducted tests with EDTA. 9.6.2 False Positive Rate A total of five valid tests were conducted on blank samples with the addition of EDTA (Table 9.24). No false positives were observed for the growth NOEC, IC25, or IC50. The growth NOEC was 100% for all 5 blank 104 ------- samples tested with EDTA, and the IC25 and IC50 values were >100% for all 5 blank samples tested with EDTA. The resulting false positive rate calculated in the WET Variability Study for the Selenastrum chronic test method conducted with the addition of EDTA was 0.00%. A total of six valid tests were conducted on blank samples without the addition of EDTA (Table 9.25). For the growth NOEC, one false positive was observed and one test result was determined as inconclusive due to evaluation of the concentration-response relationship (Table 8.4). The growth NOEC for sample 9457 tested without EDTA was reported as 6.25%. Two false positives were observed for the IC25. The IC25 for samples 9457 and 9473 were reported as 11.7% and 24.6%, respectively. No false positives were observed for the IC50. The resulting false positive rate calculated in the WET Variability Study for the Selenastrum chronic test method conducted without the addition of EDTA was 20.0%, 33.3%, and 0.00% for the growth NOEC, IC25, and IC50, respectively. 9.6.3 Precision Precision of the Selenastrum chronic test method was estimated by calculating the CV of IC25 and IC50 values obtained for the reference toxicant, effluent, and receiving water samples. Within-laboratory, between- laboratory, and total CVs were calculated for the reference toxicant and effluent samples. Only a total CV was calculated for the receiving water samples since no within-laboratory replication was provided for this sample type. Results from two samples (9468 w/ EDTA and 9476 w/o EDTA) were excluded from the analysis of precision. These test samples were identified by ASTM h statistics as possible outliers, and review of data qualifier flags revealed a possible cause (Table 9.1). Table 9.32 summarizes the precision of point estimates from the Selenastrum chronic test method conducted with EDTA. CVs for IC25 values ranged from 10.9% to 39.5% for the within-laboratory variability component, 8.48% to 20.8% for the between-laboratory variability component, and 23.5% to 40.4% for the total variability component. Uncharacteristically, within-laboratory CVs (averaging 25.2%) were higher than between-laboratory CVs (averaging 14.6%). This is likely due to the unusually large within-laboratory CV for the effluent sample type. Averaging the CVs based on total variance for the three sample types, a total CV of 34.3% was obtained for the IC25 in the Selenastrum chronic test method conducted with EDTA. IC50 values were slightly more precise than IC25 values. For the IC50, CVs averaged 5.82%, 13.2%, and 32.2% for the within-laboratory, between-laboratory, and total variability components, respectively. Table 9.33 summarizes the precision of point estimates from the Selenastrum chronic test method conducted without EDTA. The Selenastrum test method was less precise when conducted without the addition of EDTA. CVs for IC25 values ranged from 21.0% to 25.6% for the within-laboratory variability component, 60.3% to 83.6% for the between-laboratory variability component, and 24.1% to 87.5% for the total variability component. Within-laboratory CVs averaged 23.3%, and between-laboratory CVs averaged 72.0%. Averaging the CVs based on total variance for the three sample types, a total CV of 58.5% was obtained for the IC25 in the Selenastrum chronic test method conducted without EDTA. IC50 values were again slightly more precise than IC25 values. For the IC50, CVs averaged 14.5%, 43.9%, and 58.5% for the within-laboratory, between- laboratory, and total variability components, respectively. 105 ------- The precision of NOEC values was determined by evaluating the range and distribution of NOEC values and the percentage of values falling within and beyond one concentration from the median. Table 9.34 describes the precision of NOEC values for the Selenastrum chronic method. For tests conducted with EDTA, NOEC values spanned three concentrations for the effluent sample type and four concentrations for the reference toxicant and receiving water sample types. The percentage of values within one concentration of the median was 85.7%, 100%, and 85.7% for the reference toxicant, effluent, and receiving water sample types, respectively. For tests conducted without EDTA, NOEC values spanned six concentrations for the reference toxicant sample type, four concentrations for the effluent sample type, and two concentrations for the receiving water sample type. The percentage of values within one concentration of the median was 40%, 50%, and 100% for the reference toxicant, effluent, and receiving water sample types, respectively. 106 ------- Table 9.24. Results for Selenastrum chronic test method performed on blank samples with EDTA. LabID Referee 33 333 406 3 36 62 209 459 Summary Statistics Sample Code 9497 9459 9489 9491 9457 9463 9473 9481 9499 N Min Max Median Mean Nutrient EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA False positives False positive rate Test date 03/30/00 03/16/00 03/30/00 03/23/00 03/30/00 03/16/00 03/30/00 03/30/00 03/16/00 Growth Information NOEC (% sample) Invalid" Invalid0 Invalid0 Invalid0 100 100 100 100 100 5 100 100 100 0 0.00% IC25 (% sample) Invalid" Invalid0 Invalid0 Invalid0 >100 >100 >100 >100 >100 5 >100 >100 >100 >100 0 0.00% IC50 (% sample) Invalid" Invalid0 Invalid0 Invalid0 >100 >100 >100 >100 >100 5 >100 >100 >100 >100 0 0.00% Control Mean (Cells/mL) 1.71e+06 2.28e+05 1.66e+06 6.03e+05 4.79e+06 1.32e+06 2.98e+06 3.51e+06 1.94e+06 Control CV (%) 28.0 3.25 29.6 26.7 9.77 9.53 15.2 12.6 5.50 Flags3 a7, d4, g2 a6 a7, d4, g2 a*, a7, db d4, e9, §1, g2 d4 d4, g2 e9, gn d4, g2 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. c Results from invalid tests were excluded from summary statistics. 107 ------- Table 9.25. Results for Selenastrum chronic test method performed on blank samples without EDTA. LabID Referee 209 333 o J 33 36 62 406 459 Summary Statistics Sample Code 9497 9481 9489 9457 9459 9463 9473 9491 9499 N Min Max Median Mean Nutrient w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA False positives False positive rate Test date 03/30/00 03/30/00 03/30/00 03/30/00 03/16/00 03/16/00 03/30/00 03/23/00 03/16/00 Growth Information NOEC (% sample) Invalidb Invalid0 Invalid0 6.25 100 100 Inconclusive"1 100 100 5 6.25 100 100 1 20.0% IC25 (% sample) Invalidb Invalid0 Invalid0 11.7 >100 >100 24.6 >100 >100 6 11.7 >100 >100 72.7 2 33.3% IC50 (% sample) Invalidb Invalid0 Invalid0 >100 >100 >100 >100 >100 >100 6 >100 >100 >100 >100 0 0.00% Control Mean (Cells/mL) 1.64e+06 4.38e+05 9.59e+05 2.65e+05 4.89e+05 1.44e+06 3.68e+06 5.65e+05 7.75e+05 Control CV (%) 27.8 39.2 28.8 7.29 4.35 17.9 8.72 16.2 8.70 Flags3 a7, db d4 a7, e9, g2, g5, gn a7, g2 d4,§6 d4, §2, g5 db d4, e9, gb g2 d4, g2 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. c Results from invalid tests were excluded from summary statistics. d Results were excluded from summary statistics. Based on EPA guidance for evaluating concentration-response relationships (EPA, 2000a) the test result was inconclusive and the sample should be retested to obtain a reliable result. 108 ------- Table 9.26. Results for Selenastrum chronic test method performed on reference toxicant samples with EDTA. LabID Referee 33 33 299 406 406 3 36 39 125 209 209 459 459 Summary Statistics Sample Code 9495 9460 9461 9485 9492 9493 9455 9465 9469 9477 9478 9479 9500 9501 N Min Max Median Mean Within-lab Between-lab Total Nutrient EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA STD cv% STD cv% STD cv% Test date 03/16/00 03/23/00 03/30/00 03/30/00 03/23/00 03/30/00 03/16/00 03/30/00 03/30/00 03/30/00 03/09/00 03/17/00 03/23/00 03/30/00 Growth Information NOEC (% sample) <6.25b Invalid' Invalid0 Invalid' Invalid0 Invalid0 Inconclusive*1 25 25 6.25 50 25 25 50 7 6.25 50 25 IC25 (% sample) 8.40" Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Inconclusive4 39.4 40.7 19.9 42.8 39.6 43.0 50.8 7 19.9 50.8 40.7 39.5 4.30 10.9% 8.20 20.8% 9.26 23.5% IC50 (% sample) 35.5" Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Inconclusive4 68.4 73.2 54.8 68.1 65.5 75.6 71.7 7 54.8 75.6 68.4 68.2 2.39 3.51% 6.36 9.33% 6.80 9.97% Control Mean (Cells/mL) 1.98e+06 2.70e+05 4.37e+05 1.50e+05 5.68e+05 6.18e+05 1.53e+06 1.29e+06 1.13e+06 2.19e+06 2.87e+06 2.27e+06 2.24e+06 1.43e+06 Control CV (%) 4.43 5.60 3.46 14.4 17.9 13.8 11.1 17.6 7.19 5.07 19.6 2.65 2.70 11.8 Flags3 d,,d4 a6 a6 a6, d1; d4, e3, eg, e9= eicb en, g2 a6, dl5 d4, e9, gl, g2 a6, bj, dj, e9, gl, g2 d4 d4, e9 d4, e2, e3, e9, gl gi e9, g2, gn di, e9, gll d4, g2, g6 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 0 Results from invalid tests were excluded from summary statistics. d Results were excluded from summary statistics. Based on EPA guidance for evaluating concentration-response relationships (EPA, 2000a) the test result was inconclusive and the sample should be retested to obtain a reliable result. 109 ------- Table 9.27. Results for Selenastrum chronic test method performed on reference toxicant samples without EDTA. LabID Referee 209 209 299 3 33 33 36 39 125 406 406 459 459 Summary Statistics Sample Code 9495 9478 9479 9485 9455 9460 9461 9465 9469 9477 9492 9493 9500 9501 N Min Max Median Mean Within-lab Between-lab Total Nutrient w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA STD cv% STD cv% STD cv% Test date 03/16/00 03/09/00 03/17/00 03/30/00 03/16/00 03/23/00 03/30/00 03/30/00 03/30/00 03/30/00 03/23/00 03/30/00 03/23/00 03/30/00 Growth Information NOEC (% sample) 6.25" Invalid0 Invalid0 Invalid0 100 100 100 12.5 12.5 6.25 25 6.25 <6.25 25 10 <6.25 100 18.75 IC25 (% sample) 10.2" Invalid0 Invalid0 Invalid0 >100 >100 >100 15.7 22.7 13.2 19.4 12.9 9.69 35.9 10 9.69 >100 21.0 42.9 11.0 25.6% 35.9 83.6% 37.6 87.5% IC50 (% sample) 37.6" Invalid0 Invalid0 Invalid0 >100 >100 >100 21.5 57.3 23.7 33.4 22.0 24.7 53.9 10 21.5 >100 43.7 53.7 12.8 23.8% 29.9 55.6% 32.5 60.5% Control Mean (Cells/mL) 1.85e+06 1.48e+06 l.OOe+04 1.71e+05 2.71e+05 2.25e+05 4.77e+05 2.02e+06 1.93e+06 1.18e+06 3.33e+05 4.48e+05 2.27e+06 1.27e+06 Control CV (%) 9.08 102 0.00 22.3 3.13 0.915 7.23 6.99 11.4 5.49 6.67 9.19 4.18 14.7 Flags3 di, d4, g6 a7, e9, g2, gu afo dj, e9, g2, gll a6, a7, d1; e3, e8, 69= eil= g2 d4, e9 d4, e2, e3, e9, gl gi di, d4, e9, gl5 g2 bi= dj, e9, g; g2 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 0 Results from invalid tests were excluded from summary statistics. 110 ------- Table 9.28. Results for Selenastrum chronic test method performed on effluent samples with EDTA. LabID Referee 33 39 39 299 299 333 406 459 3 36 62 62 125 125 333 Summary Statistics Sample Code 9494 9458 9466 9467 9482 9483 9487 9490 9498 9454 9462 9470 9471 9474 9475 9486 N Min Max Median Mean Within-lab Between-lab Total Nutrient EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA STD cv% STD cv% STD cv% Test date 03/09/00 03/09/00 03/09/00 03/16/00 03/09/00 03/16/00 03/16/00 03/09/00 03/10/00 03/09/00 03/09/00 03/09/00 03/16/00 03/09/00 03/16/00 03/09/00 Growth Information NOEC (% sample) <6.25b Invalid0 Invalid0 Invalid' Invalid1 Invalid' Invalid' Invalid' Invalid' <6.25 6.25 12.5 12.5 12.5 6.25 6.25 7 <6.25 12.5 6.25 IC25 (% sample) 10.7b Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' 17.7 13.3 27.2 16.7 36.5 22.6 10.3 7 10.3 36.5 17.7 20.6 8.15 39.5% 1.75 8.48% 8.33 40.4% IC50 (% sample) 29.9b Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' Invalid' 45.6 34.2 54.9 46.7 62.6 60.9 58.3 7 34.2 62.6 54.9 51.9 4.22 8.14% 8.91 17.2% 9.86 19.0% Control Mean (Cells/mL) 2.18e+06 2.52e+05 1.27e+06 5.02e+05 2.75e+06 4.35e+05 2.64e+06 7.58e+05 5.90e+05 4.66e+06 1.94e+06 3.37e+06 3.90e+06 2.30e+06 2.33e+06 3.54e+06 Control CV (%) 4.45 6.19 28.9 21.2 45.4 30.4 27.4 21.2 17.0 3.87 18.9 4.29 8.29 10.0 7.95 8.28 Flags3 d, a6,g5 a7, bj, d4, e9, g2 a6, a,, dj, d4, e9, g2 a7, d4, e3, e9, e10, en, g2 a6, a,, dj, d4, e3, e9, e10, eu, g2 a7, bl5 dj, d4, g2 a6, a,, bj, dl5 d4, 69= gi= g2 a6, bj, b2, g2 d4 dj, d4, e6 d4 d4 gi gi dl5d4 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. ' Results from invalid tests were excluded from summary statistics. Ill ------- Table 9.29. Results for Selenastrum chronic test method performed on effluent samples without EDTA. LabID Referee 39 39 299 299 333 333 406 459 3 33 36 62 62 125 125 Summary Statistics Sample Code 9494 9466 9467 9482 9483 9486 9487 9490 9498 9454 9458 9462 9470 9471 9474 9475 N Min Max Median Mean Within-lab Between-lab Total Nutrient w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA STD cv% STD cv% STD cv% Test date 03/09/00 03/09/00 03/16/00 03/09/00 03/16/00 03/09/00 03/16/00 03/09/00 03/10/00 03/09/00 03/09/00 03/09/00 03/09/00 03/16/00 03/09/00 03/16/00 Growth Information NOEC (% sample) 6.25" Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Inconclusive4 50 12.5 6.25 12.5 50 25 6 6.25 50 18.75 IC25 (% sample) 8.85" Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Inconclusive* 65.1 20.1 11.9 10.3 57.6 43.2 6 10.3 65.1 31.7 34.7 7.27 21.0% 20.9 60.3% 22.2 63.9% IC50 (% sample) 19.7" Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Invalid0 Inconclusive4 80.2 43.6 35.1 37.5 74.0 68.6 6 35.1 80.2 56.1 56.5 2.97 5.25% 18.1 32.1% 18.4 32.5% Control Mean (Cells/mL) 1.44e+06 1.07e+05 3.10e+05 6.28e+04 4.00e+04 2.61e+06 1.44e+06 5.43e+05 6.23e+05 5.11e+05 2.53e+05 2.23e+06 4.11e+06 3.93e+06 8.38e+05 1.03e+06 Control CV (%) 18.2 75.0 29.0 62.9 41.7 20.9 53.8 26.9 26.2 8.03 6.51 10.2 14.1 16.4 18.5 12.3 Flags3 d, a6, a,, bl5 d4, e9, g2 a,, dj, d4, e,,, g2, g5 a6, a7, d1; d4, e3, e9, g2 afo a7, dl5 d4, e3, e9, eu, g2 a7, dls d4 a7, bj, dj, d4, g2 a7, bj, d1; d4, e9, g1; g2 a?, bi= b2, 82 d4, g2 d,,d4 d4 d4 gl=g2 gl a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 0 Results from invalid tests were excluded from summary statistics. d Results were excluded from summary statistics. Based on EPA guidance for evaluating concentration-response relationships (EPA, 2000a) the test result was inconclusive and the sample should be retested to obtain a reliable result. 112 ------- Table 9.30. Results for Selenastrum chronic test method performed on receiving water samples with EDTA. LabID Referee 3 36 39 62 125 209 299 333 Summary Statistics Sample Code 9496 9456 9464 9468 9472 9476 9480 9484 9488 N Min Max Median Mean STD CV% Nutrient EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA EDTA Test date 03/23/00 03/23/00 03/23/00 03/23/00 03/24/00 03/23/00 03/23/00 03/23/00 03/23/00 Growth Information NOEC (% sample) <6.25b 6.25 25 100° 6.25 12.5 12.5 <6.25 25 7 <6.25 25 12.5 IC25 (% sample) <6.25b 17.2 24.7 >100C 20.2 24.7 11.1 <6.25 19.7 7 <6.25 24.7 19.7 17.7 6.88 38.9% IC50 (% sample) 26.8b 35.8 38.0 >100 36.8 64.5 18.1 <6.25 41.2 8 <6.25 >100 37.4 42.6 28.8 67.6% Control Mean (Cells/mL) 1.94e+06 4.00e+06 2.33e+06 1.35e+06 3.39e+06 2.46e+06 2.28e+06 2.53e+06 2.79e+06 Control CV (%) 8.48 2.65 17.1 9.47 3.66 5.59 15.6 19.8 18.7 Flags3 dbd4 d4 d4, e9 bb db d4, e9, §3= §5, g7 b2, d4, f, gi 69, §2, §11 db d4, e3, e9, 610, g« bj, db d4 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. c Results were identified as outliers, a probable cause was identified, and therefore results were excluded from summary statistics. 113 ------- Table 9.31. Results for Selenastrum chronic test method performed on receiving water samples without EDTA. LabID Referee 209 299 3 36 39 62 125 333 Summary Statistics Sample Code 9496 9480 9484 9456 9464 9468 9472 9476 9488 N Min Max Median Mean STD CV% Nutrient w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA w/o EDTA Test date 03/23/00 03/23/00 03/23/00 03/23/00 03/23/00 03/23/00 03/24/00 03/23/00 03/23/00 Growth Information NOEC (% sample) 12.5b Invalid0 Invalid0 <6.25 6.25 Inconclusive4 6.25 <6.25e 6.25 4 <6.25 6.25 6.25 IC25 (% sample) 15.8b Invalid0 Invalid0 <6.25 10.1 10.3 9.76 32. P 6.34 5 <6.25 10.3 9.76 8.54 2.05 24.1% IC50 (% sample) 39.2b Invalid0 Invalid0 39.1 18.4 >100 15.0 55.9 14.6 6 14.6 >100 28.8 40.5 33.4 82.5% Control Mean (Cells/mL) 1.36e+06 2.07e+05 1.12e+05 2.88e+05 1.50e+06 1.06e+06 3.49e+06 1.38e+06 1.14e+06 Control CV (%) 14.3 23.0 39.0 12.2 13.8 2.89 11.1 6.96 7.77 Flags3 dbd4 a7, e9, g2, gn cif.^, ™-Jr> d-1 ; ^3? ^9? eio d4, e9 bb db d4, e9, §2, S5, §6 b2, d4, f, gl, §3 bb db g2 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. c Results from invalid tests were excluded from summary statistics. d Results were excluded from summary statistics. Based on EPA guidance for evaluating concentration-response relationships (EPA, 2000a) the test result was inconclusive and the sample should be retested to obtain a reliable result. e Results were identified as outliers, a probable cause was identified, and therefore results were excluded from summary statistics. 114 ------- Table 9.32. Precision of point estimates from the Selenastrum chronic test method conducted with EDTA. Sample type Reference toxicant Effluent Receiving water Average CV (%) IC25 Within-lab a 10.9 39.5 - 25.2 Between-lab a 20.8 8.48 - 14.6 Total 23.5 40.4 38.9 34.3 IC50 Within-lab a 3.51 8.14 - 5.82 Between-lab a 9.33 17.2 - 13.2 Total 9.97 19.0 67.6 32.2 a Within and between-laboratory components of variability were not calculated for the receiving water sample type since no within-laboratory replication was provided for this sample type. Table 9.33. Precision of point estimates from the Selenastrum chronic test method conducted without EDTA. Sample type Reference toxicant Effluent Receiving water Average CV (%) IC25 Within-lab a 25.6 21.0 - 23.3 Between-lab a 83.6 60.3 - 72.0 Total 87.5 63.9 24.1 58.5 IC50 Within-lab a 23.8 5.25 - 14.5 Between-lab a 55.6 32.1 - 43.9 Total 60.5 32.5 82.5 58.5 a Within- and between-laboratory components of variability were not calculated for the receiving water sample type since no within-laboratory replication was provided for this sample type. 115 ------- Table 9.34. Precision of NOEC values from the Selenastrum chronic test method. Nutrient WithEDTA Without EDTA Sample type Reference toxicant Effluent Receiving water Reference toxicant Effluent Receiving water NOEC Frequency <6.25% 0 1 1 1 0 1 6.25% 1 o 5 2 2 1 3 12.5% 0 o 5 2 2 2 0 25% 4 0 2 2 1 0 50% 2 0 0 0 2 0 100% 0 0 0 3 0 0 Median 25 6.25 12.5 18.75 18.75 6.25 % within 1 cone, of median 85.7 100 85.7 40.0 50.0 100 % beyond 1 cone, of median 14.3 0.00 14.3 60.0 50.0 0.00 116 ------- 9.7 Mysidopsis Chronic Test Method Results A total of 11 participant laboratories conducted the Mysidopsis chronic test method in the WET Variability Study. These laboratories tested a total of 8 blank samples, 13 reference toxicant samples, 15 effluent samples, and 8 receiving water samples. For each sample tested, a 7-day survival NOEC, a 7-day growth NOEC, a 7-day fecundity NOEC, a 7-day LC50, a 7-day growth IC25, and a 7-day fecundity IC25 were generated as test results. As described in the WET method manual, growth was measured as the total weight per replicate divided by the number of original organisms in that replicate. This definition provides a combined growth and survival endpoint that is more accurately termed biomass. The fecundity endpoint was measured as the percentage of females with eggs. Replicates without identified females were excluded from the analysis of fecundity. Results for the fecundity endpoint were not calculated if less than 50% of control females produced eggs. Also, test concentrations above the survival NOEC were excluded from hypothesis testing conducted on growth and fecundity endpoints. Results of Mysidopsis chronic testing are shown in Tables 9.35 - 9.38 for each sample type. 9.7.1 Successful Test Completion Rate A total of 44 Mysidopsis chronic tests were initiated by 11 participant laboratories. All 44 tests were completed; however, the test conducted on sample 9690 was invalid due to failure to meet test acceptability criteria for survival. The resulting successful test completion rate calculated in the WET Variability Study for the Mysidopsis chronic test method was 97.7%. One test in the referee laboratory was also invalid due to failure to meet test acceptability criteria for survival. Of the 44 Mysidopsis chronic tests initiated, laboratories were able to report results for the fecundity endpoint in only 22 (or 50% of) tests. The remaining 22 tests did not meet the minimum control fecundity necessary to report fecundity results. The fecundity endpoint is an optional endpoint in the Mysidopsis chronic test, so failure to generate fecundity data does not invalidate a test. 9.7.2 False Positive Rate A total of seven valid tests were conducted by seven participant laboratories on blank samples (Table 9.35). No false positives were observed for survival, growth, or fecundity endpoints. The survival NOEC was 100% for all 7 blank samples, and the LC50 was >100% for all 7 blank samples. The NOEC for growth was 100% for all 7 blank samples, and the IC25 was >100% for all 7 blank samples. The fecundity endpoint was only calculable for four samples. The NOEC for fecundity was 100% for all 4 samples, and the IC25 for fecundity was >100% for all 4 samples. The resulting false positive rate calculated in the WET Variability Study for the Mysidopsis chronic test method was 0.00%. A false positive growth NOEC result was reported by the participant laboratory for sample 9658. Upon SCC recalculation and verification, the growth NOEC was properly reported as 100%. SCC test review discovered that the participant laboratory had calculated growth based on the weight per surviving Mysidopsis rather than per original Mysidopsis. Also, sample 9696 produced an interrupted concentration-response curve; however, based on EPA guidance for evaluating concentration-response relationships (USEPA, 2000a), the NOEC for this sample should be reported as 100%. 117 ------- 9.7.3 Precision Precision of the Mysidopsis chronic test method was estimated by calculating the CV of LC50 and growth IC25 values obtained for the reference toxicant, effluent, and receiving water samples. Precision estimates were not calculated for fecundity IC25 values because results for this endpoint could not be consistently and definitively measured within the test concentration range (as evidenced by a large proportion of censored data (e.g., >12.5%, >25%, or >50%) for this endpoint). Within-laboratory, between-laboratory, and total CVs were calculated for the reference toxicant and effluent samples. Only a total CV was calculated for the receiving water samples since no within-laboratory replication was provided for this sample type. All participant laboratory test data for the reference toxicant, effluent, and receiving water samples were used in estimating precision. No test results were identified by ASTM h statistics as possible outliers. Table 9.39 summarizes the precision of point estimates from the Mysidopsis chronic test method. CVs for LC50 values were consistent among sample types; within-laboratory CVs ranged from 6.09% to 7.06%, between-laboratory CVs ranged from 24.6% to 30.0%, and total CVs ranged from 25.6% to 37.4%. As expected, the majority of variability was due to the between-laboratory component, with within-laboratory CVs averaging 6.57% and between-laboratory CVs averaging 27.3%. Averaging the CVs for the LC50 based on total variance for the three sample types, a total CV of 31.2% was obtained for the Mysidopsis chronic survival endpoint. CVs for growth IC25 values were higher than those for LC50 values. Within-laboratory CVs for the growth IC25 ranged from 5.26% to 8.69%, between-laboratory CVs ranged from 36.6% to 40.0%, and total CVs ranged from 37.0% to 45.9%. Within-laboratory CVs (averaging 6.98%) were much lower than between- laboratory CVs (averaging 38.3%). Averaging the CVs for the IC25 based on total variance for the three sample types, a total CV of 41.3% was obtained for the Mysidopsis chronic growth endpoint. The precision of NOEC values was determined by evaluating the range and distribution of NOEC values and the percentage of values falling within and beyond one concentration from the median. Table 9.40 describes the precision of NOEC values for the Mysidopsis chronic test method. For the survival endpoint, NOEC values spanned three concentrations for all three sample types, and 100% of NOEC values were within one concentration of the median. Growth endpoints performed similarly, with the exception of the reference toxicant sample. Growth NOEC values spanned four concentrations for the reference toxicant sample type, and 7.69% were beyond one concentration from the median. The data set for the fecundity endpoint was much smaller than that for the survival and growth endpoints, since only 50% of tests achieved the necessary fecundity in controls (egg production in 50% of control females). In this reduced data set, fecundity NOEC values spanned three concentrations for each sample type and 75.0%, 87.5%, and 66.7% of values were within one concentration of the median for the reference toxicant, effluent, and receiving water sample types, respectively. 118 ------- Table 9.35. Results for Mysidopsis chronic test method performed on blank samples. LabID Referee 417 22 39 73 77 101 125 420 Summary Statistics Sample Code 9687 9690 9650 9658 9664 9666 9672 9676 9696 N Min Max Median Mean Test date 02/29/00 02/22/00 02/22/00 02/22/00 02/29/00 02/22/00 02/29/00 02/29/00 02/29/00 False positives False positive rate Survival Information NOEC (% sample) Invalid11 Invalid" 100 100 100 100 100 100 100 7 100 100 100 0 0.00% LC50 (% sample) Invalid" Invalid0 >100 >100 >100 >100 >100 >100 >100 7 >100 >100 >100 >100 0 0.00% Control mean (%) 76.3 78.1 95.0 90.0 100 100 100 92.5 100 Control CV (%) 21.5 20.9 8.57 10.4 0.716 0.00 0.00 9.81 0.00 Growth Information NOEC (% sample) Invalid* Invalid0 100 100 100 100 100 100 100 7 100 100 100 0 0.00% IC25 (% sample) Invalid" Invalid0 >100 >100 >100 >100 >100 >100 >100 7 >100 >100 >100 >100 0 0.00% Control mean (mg) 0.322 0.224 0.194 0.403 0.417 0.278 0.360 0.621 0.354 Control CV (%) 21.9 45.4 16.1 18.1 7.68 8.75 7.86 43.9 10.7 Fecundity Information NOEC (% sample) Invalid" - - - 100 - 100 100 100 4 100 100 100 0 0.00% IC25 (% sample) Invalid" - - - >100 - >100 >100 >100 4 >100 >100 >100 >100 0 0.00% Control mean (%) 85.7 - - - 72.6 - 96.9 77.1 85.4 Control CV (%) 17.2 - - - 31.5 - 7.53 23.4 13.5 Flags3 &i, d2, d,, d5, QI, e2, gi ai, a5, d5, 61, g2 a5, bj, d5, e2, e3, gj a5, d5 61 a5 61 g2 61, e2, g5 a Data qualifier flags are described in Table 8.3. " Results from the referee laboratory were excluded from summary statistics. 0 Results from invalid tests were excluded from summary statistics. 119 ------- Table 9.36. Results for Mysidopsis chronic test method performed on reference toxicant samples. LabID Referee 22 22 36 39 39 77 77 101 101 181 238 417 420 Summary Statistics Sample Code 9689 9652 9653 9657 9660 9661 9668 9669 9670 9671 9681 9685 9693 9695 N Min Max Median Mean Within- lab Between- lab Total Test date 02/22/00 02/29/00 02/29/00 02/29/00 02/29/00 02/29/00 02/29/00 02/29/00 02/22/00 02/22/00 02/29/00 02/29/00 02/29/00 02/22/00 STD cv% STD cv% STD cv% Survival Information NOEC (% sample) 50b 25 25 25 25 25 25 50 12.5 12.5 50 25 50 50 13 12.5 50 25 LC50 (% sample) 66. 9b 42.5 43.9 54.8 35.4 35.4 57.3 63.7 29.2 24.0 70.7 41.6 59.5 70.7 13 24.0 70.7 43.9 48.4 2.94 6.09% 14.5 30.0% 14.8 30.6% Control mean (%) 87.1 97.5 95.0 97.5 82.5 80.0 100 100 100 100 87.5 95.0 100 100 Control CV (%) 13.8 6.40 8.57 6.40 16.5 15.6 0.00 0.00 0.00 0.00 14.4 8.57 0.940 0.00 Growth Information NOEC (% sample) 50b 25 6.25 25 12.5 25 25 25 12.5 12.5 50 25 25 50 13 6.25 50 25 IC25 (% sample) 36.4b 30.8 26.4 38.2 19.5 26.7 39.3 38.4 21.0 21.4 60.9 36.4 27.6 61.8 13 19.5 61.8 30.8 34.5 3.00 8.69% 13.8 40.0% 14.1 40.9% Control mean (mg) 0.180 0.211 0.228 0.304 0.419 0.415 0.259 0.251 0.275 0.255 0.173 0.194 0.283 0.338 Control CV (%) 40.2 23.6 22.4 18.0 32.3 34.4 14.9 21.5 17.3 13.7 11.4 9.81 40.6 6.84 Fecundity Information NOEC (% sample) 50b - - - - - - - 12.5 12.5 - 25 - 50 4 12.5 50 18.75 IC25 (% sample) >50b - - - - - - - 12.7 >25 - 46.4 - >50 4 12.7 >50 35.7 33.5 Control mean (%) 59.5 - - - - - - - 85.4 84.6 - 80.6 - 81.3 Control CV (%) 36.3 - - - - - - - 13.8 14.7 - 19.2 - 24.5 Flags3 di, ej, e2, gl, 82 as, bj, c5, d5, e2= e3, gl5 g2 a5, bl5 c5, d5, e2! e3, gj a5, e1; e6 a5, d,, d,, d5, 61 a5, d2, dj, d5 as a5 a5,gi d5, e2, e3 a5, d5, el5 g2 el5 e2 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 120 ------- Table 9.37. Results for Mysidopsis chronic test method performed on effluent samples. LabID Referee 22 36 36 39 73 73 77 125 125 181 181 238 238 417 420 Summary Statistics Sample Code 9686 9651 9654 9655 9659 9662 9663 9667 9674 9675 9678 9679 9682 9683 9691 9694 N Min Max Median Mean Within- lab Between- lab Total Test date 02/22/00 02/22/00 02/23/00 02/23/00 02/22/00 02/22/00 02/22/00 02/22/00 02/22/00 02/22/00 02/22/00 02/22/00 02/22/00 02/22/00 02/22/00 02/22/00 STD cv% STD cv% STD cv% Survival Information NOEC (% sample) 25b 12.5 25 12.5 12.5 25 25 12.5 6.25 12.5 12.5 12.5 25 25 25 25 15 6.25 25 12.5 LC50 (% sample) 33. 4b 21.0 30.3 27.5 18.4 34.7 35.2 26.2 12.1 15.6 27.2 26.9 35.4 31.3 30.0 35.4 15 12.1 35.4 27.5 27.1 1.92 7.06% 6.67 24.6% 6.94 25.6% Control mean (%) 95.0 87.5 87.5 100 82.5 100 100 100 90.0 90.0 95.0 95.0 100 100 95.0 100 Control CV (%) 8.34 10.3 14.4 0.00 7.41 0.940 0.940 0.00 17.3 17.3 8.57 12.6 0.00 0.00 12.6 0.00 Growth Information NOEC (% sample) 25b 12.5 12.5 12.5 6.25 25 25 12.5 6.25 12.5 12.5 12.5 25 25 25 25 15 6.25 25 12.5 IC25 (% sample) 30.2b 16.5 19.5 16.9 9.32 31.1 30.8 18.3 10.3 8.66 20.2 21.5 31.3 30.1 28.1 30.0 15 8.66 31.3 20.2 21.5 1.13 5.26% 7.88 36.6% 7.96 37.0% Control mean (mg) 0.323 0.175 0.251 0.289 0.408 0.338 0.430 0.292 0.622 0.622 0.193 0.202 0.206 0.205 0.288 0.339 Control CV (%) 22.7 17.4 18.8 17.7 18.5 5.25 6.34 22.0 67.9 67.9 16.2 17.6 11.1 11.7 28.8 11.2 Fecundity Information NOEC (% sample) 25b - - - - 25 25 - 6.25 12.5 - 12.5 25 25 - 25 8 6.25 25 25 IC25 (% sample) >25b - - - - >25 >25 - >12.5 >12.5 - >25 >25 >25 - 8.45 8 8.45 >25 >25 19.8 Control mean (%) 77.1 - - - - 71.9 71.9 - 92.7 92.7 - 69.4 93.8 89.6 - 83.3 Control CV (%) 16.6 - - - - 16.2 25.3 - 11.4 11.4 - 33.1 10.5 14.9 - 19.0 Flags3 di, ej, e2, g; a,, bj, d5, e2, e3, gi a5, b2= ei= e6 a5, b2, e6 a5, t>i, d5 61 61 a5 d4,g2 d4,g2 a5,gi gi d5, e2, e3, gj, g5 d5, e2, e3, g; a5, d5, el e2!g5 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 121 ------- Table 9.38. Results for Mysidopsis chronic test method performed on receiving water samples. LabID Referee 36 73 101 125 181 238 417 420 Summary Statistics Sample Code 9688 9656 9665 9673 9677 9680 9684 9692 9697 N Min Max Median Mean STD CV% Test date 02/29/00 02/29/00 02/29/00 02/29/00 02/29/00 02/29/00 02/29/00 02/29/00 02/29/00 Survival Information NOEC (% samplle) 25b 12.5 25 6.25 6.25 12.5 12.5 25 25 8 6.25 25 12.5 LC50 (% sample) 33.6b 21.2 33.5 13.0 9.69 26.5 30.2 28.2 34.4 8 9.69 34.4 27.3 24.6 9.20 37.4% Control mean (%) 86.9 97.5 100 97.5 92.5 97.5 97.5 100 100 Control CV (%) 14.8 6.40 1.05 6.40 9.81 6.40 6.40 1.83 0.00 Growth Information NOEC (% sample) 12.5b 6.25 12.5 6.25 6.25 12.5 12.5 12.5 25 8 6.25 25 12.5 IC25 (% sample) 23.5b 14.6 25.3 10.2 8.02 23.2 31.3 15.2 31.3 8 8.02 31.3 19.2 19.9 9.12 45.9% Control mean (mg) 0.337 0.323 0.494 0.350 0.621 0.195 0.209 0.235 0.342 Control CV (%) 13.5 9.03 8.76 10.1 43.9 18.3 11.0 23.1 11.3 Fecundity Information NOEC (% sample) - 12.5 25 6.25 6.25 - 6.25 - 25 6 6.25 25 9.375 IC25 (% sample) - >25 >25 >12.5 8.53 - 12.5 - >25 6 8.53 >25 18.8 18.1 Control mean (%) - 50.0 78.6 95.8 77.1 - 100 - 75.0 Control CV (%) - 38.6 32.9 10.5 23.4 - 6.86 - 29.5 Flags3 a5, d2, d3, 61= e2, g; e6 61 82 as=gi bls d5, e2, e3 a5= d5, 6; 61=62 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 122 ------- Table 9.39. Precision of point estimates from the Mysidopsis chronic test method. Sample type Reference toxicant Effluent Receiving water Average CV (%) LC50 Within-lab a 6.09 7.06 - 6.57 Between-lab a 30.0 24.6 - 27.3 Total 30.6 25.6 37.4 31.2 IC25 Within-lab a 8.69 5.26 - 6.98 Between-lab a 40.0 36.6 - 38.3 Total 40.9 37.0 45.9 41.3 a Within and between-laboratory components of variability were not calculated for the receiving water sample type since no within-laboratory replication was provided for this sample type. Table 9.40. Precision of NOEC values from the Mysidopsis chronic test method. Sample type Reference toxicant Effluent Receiving water Endpoint Survival Growth Fecundity Survival Growth Fecundity Survival Growth Fecundity NOEC Frequency <6.25% 0 0 0 0 0 0 0 0 0 6.25% 0 1 0 1 2 1 2 3 o J 12.5% 2 3 2 7 7 2 3 4 1 25% 7 7 1 7 6 5 3 1 2 50% 4 2 1 0 0 0 0 0 0 100% 0 0 0 0 0 0 0 0 0 Median (% sample) 25 25 18.75 12.5 12.5 25 12.5 12.5 9.375 % within 1 cone, of median 100 92.3 75.0 100 100 87.5 100 100 66.7 % beyond 1 cone, of median 0.00 7.69 25.0 0.00 0.00 12.5 0.00 0.00 33.3 123 ------- 9.8 Sheepshead Acute Test Method Results A total of seven participant laboratories conducted the sheepshead acute test method in the WET Variability Study. These laboratories tested a total of seven blank samples, seven reference toxicant samples, seven effluent samples, and seven receiving water samples. For each sample tested, a 96-hour LC50 was generated as atest result. Results of sheepshead acute testing are shown in Tables 9.41 - 9.44 for each sample type. 9.8.1 Successful Test Completion Rate A total of 28 samples were tested at 7 participant laboratories. All 28 tests were completed and met test acceptability criteria. The resulting successful test completion rate calculated in the WET Variability Study for the sheepshead acute test method was 100%. Testing on two samples (9586 and 9589) was repeated due to laboratory error. The laboratory terminated the initial tests after 48 hours rather than the required 96 hours. Due to this error, the laboratory retested samples the following week at their expense. Because this error was not a result of the test method or test method performance, it was not included in the calculation of the successful test completion rate. 9.8.2 False Positive Rate A total of seven blank samples were analyzed by seven participant laboratories (Table 9.41). The LC50 calculated for all 7 blank samples was >100%, indicating no toxicity and no false positives. The resulting false positive rate calculated in the WET Variability Study for the sheepshead acute test method was 0.00%. 9.8.3 Precision Precision of the sheepshead acute test method was estimated by calculating the CV of LC50 values obtained for the reference toxicant, effluent, and receiving water samples. Since no within-laboratory replication was provided for this method, CVs were calculated based on total variability and no estimates for precision were available for within-laboratory and between-laboratory components of variability. All participant laboratory test data for the reference toxicant, effluent, and receiving water samples were used in estimating precision. While results on two test samples (9617 and 9600) were identified by ASTM h statistics as possible outliers, these results were not excluded from the analysis of precision due to the small size of the data set. Exclusion of either result would translate to exclusion of 14% of the respective data set. Table 9.45 summarizes the precision of point estimates from the sheepshead acute test method. CVs based on total variance ranged from 19.4% to 32.5% with an average CV of 26.0% obtained for the sheepshead acute test method in the WET Variability Study. 124 ------- Table 9.41. Results for sheepshead acute test method performed on blank samples. LabID Referee 22 29 73 101 238 420 425 Summary Statistics Sample Code 9606 9586 9590 9594 9598 9602 9610 9614 N Min Max Median Mean Test date 03/07/00 03/17/00 03/07/00 03/07/00 03/07/00 03/07/00 03/07/00 03/07/00 False positives False positive rate Survival Information LC50 (% sample) >100b >100 >100 >100 >100 >100 >100 >100 7 >100 >100 >100 >100 0 0.00% Control mean (%) 100 100 95.0 100 100 100 100 100 Control CV (%) 0.00 0.00 8.66 0.00 0.00 0.00 0.00 0.00 Flags3 d5 "5> ^2> ^3> ^4> ^5> Sg dbe2 e2, e3 e2 d5, e6 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. Table 9.42. Results for sheepshead acute test method performed on reference toxicant samples. LabID Referee 22 29 73 101 238 420 425 Summary Statistics Sample Code 9609 9589 9593 9597 9601 9605 9613 9617 N Min Max Median Mean STD CV% Test date 03/14/00 03/17/00 03/14/00 03/14/00 03/14/00 03/14/00 03/14/00 03/14/00 Survival Information LC50 (% sample) 40.6b 37.6 46.7 37.2 35.4 35.4 39.2 66.0 7 35.4 66.0 37.6 42.5 11.1 26.0% Control mean (%) 100 100 100 100 100 100 100 100 Control CV (%) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Flags3 db d2, ds d5, e2, e3, e4, e5, Ss d2, d3, d5, e2 e2, e3, e4 e2 d5, e6, g3 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 125 ------- Table 9.43. Results for sheepshead acute test method performed on effluent samples. LabID Referee 22 29 73 101 238 420 425 Summary Statistics Sample Code 9608 9588 9592 9596 9600 9604 9612 9616 N Min Max Median Mean STD CV% Test date 03/14/00 03/14/00 03/14/00 03/14/00 03/14/00 03/14/00 03/14/00 03/14/00 Survival Information LC50 (% sample) 35,4b 33.4 35.4 35.4 18.3 35.4 35.4 34.2 7 18.3 35.4 35.4 32.5 6.29 19.4% Control mean (%) 100 100 100 100 100 100 100 100 Control CV (%) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Flags3 d2,d5 d5, e2, e3, e4, e5 d2, e2 g3 ^2; ^3> ^4> ^6 e2 d5, e6 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. Table 9.44. Results for sheepshead acute test method performed on receiving water samples. LabID Referee 22 29 73 101 238 420 425 Summary Statistics Sample Code 9607 9587 9591 9595 9599 9603 9611 9615 N Min Max Median Mean STD CV% Test date 03/07/00 03/17/00 03/07/00 03/07/00 03/07/00 03/07/00 03/07/00 03/07/00 Survival Information LC50 (% sample) 25.9b 20.3 17.7 35.4 15.5 22.5 35.4 27.7 7 15.5 35.4 22.5 24.9 8.10 32.5% Control mean (%) 100 100 100 100 100 100 100 100 Control CV (%) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Flags3 d2,d5 d5, e2, e3, e4, e5, Sg db d2, e2 e2, e3 e2 d5, e6 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 126 ------- Table 9.45. Precision of point estimates from the sheepshead acute test method. Sample type Reference toxicant Effluent Receiving water Average CV (%) Within-laba - - - - Between-laba - - - - Total 26.0 19.4 32.5 26.0 a Within and between-laboratory components of variability were not calculated for this method since no within-laboratory replication was provided. 9.9 Sheepshead Chronic Test Method Results A total of seven participant laboratories conducted the sheepshead chronic test method in the WET Variability Study. These laboratories tested a total of seven blank samples, seven reference toxicant samples, seven effluent samples, and seven receiving water samples. For each sample tested, a 7-day survival NOEC, a 7-day growth NOEC, a 7-day LC50, and a 7-day growth IC25 were generated as test results. As described in the WET method manual, growth was measured as the total weight per replicate divided by the number of original organisms in that replicate. This definition provides a combined growth and survival endpoint that is more accurately termed biomass. Also, test concentrations above the survival NOEC were excluded from hypothesis testing conducted on the growth endpoint. Results of sheepshead chronic testing are shown in Tables 9.46 - 9.49 for each sample type. 9.9.1 Successful Test Completion Rate A total of 28 samples were tested at 7 participant laboratories. All 28 tests were completed and met test acceptability criteria. The resulting successful test completion rate calculated in the WET Variability Study for the sheepshead chronic test method was 100%. Testing on sample 9618 was repeated due to laboratory error. The laboratory inadvertently renewed the test on Day 6 with the wrong sample. Due to this error, the laboratory retested the sample the following week at their expense. Because this error was not a result of the test method or test method performance, it was not included in the calculation of the successful test completion rate. 9.9.2 False Positive Rate A total of seven blank samples were analyzed by seven participant laboratories (Table 9.46). No false positives were observed for survival or growth endpoints. The survival NOEC was 100% for all 7 blank samples, and the LC50 was >100% for all 7 blank samples. The NOEC for growth was 100% for all 7 blank samples, and the IC25 was >100% for all 7 blank samples. The resulting false positive rate calculated in the WET Variability Study for the sheepshead chronic test method was 0.00%. 127 ------- 9.9.3 Precision Precision of the sheepshead chronic test method was estimated by calculating the CV of LC50 and IC25 values obtained for the reference toxicant, effluent, and receiving water samples. Since no within-laboratory replication was provided for this method, CVs were calculated based on total variability and no estimates for precision were available for within-laboratory and between-laboratory components of variability. All participant laboratory test data for the reference toxicant, effluent, and receiving water samples were used in estimating precision. No test results were identified by ASTM h statistics as possible outliers. Table 9.50 summarizes the precision of point estimates from the sheepshead chronic test method. CVs for LC50 values were extremely low for effluent (2.33%) and receiving water (1.63%) sample types and higher for the reference toxicant sample type (22.2%). This trend was also observed for CVs of IC25 values, which were 18.4%, 6.12%, and 7.15% for the reference toxicant, effluent, and receiving water sample types, respectively. The higher variability associated with the reference toxicant sample could be due to the added step of reconstituting the reference toxicant ampule sample in each laboratory, or it could be an anomaly associated with the small data set for this method. Averaging the CVs based on total variance for the three sample types, total CVs of 8.73% for the LC50 and 10.5% for the IC25 were obtained for the sheepshead chronic test method in the WET Variability Study. The precision of NOEC values was determined by evaluating the range and distribution of NOEC values and the percentage of values falling within and beyond one concentration from the median. Table 9.51 describes the precision of NOEC values for the sheepshead chronic method. For survival and growth endpoints, NOEC values spanned only two concentrations or did not vary at all between laboratories. All NOEC values (100%) were within one concentration of the median. 128 ------- Table 9.46. Results for sheepshead chronic test method performed on blank samples. LabID Referee 22 73 101 181 238 273 420 Summary Statistics Sample Code 9642 9618 9622 9626 9630 9634 9638 9646 N Min Max Median Mean Test date 03/21/00 03/29/00 03/21/00 03/21/00 03/21/00 03/21/00 03/21/00 03/21/00 False positives False positive rate Survival Information NOEC (% sample) 100b 100 100 100 100 100 100 100 7 100 100 100 0 0.00% LC50 (% sample) >100b >100 >100 >100 >100 >100 >100 >100 7 >100 >100 >100 >100 0 0.00% Control mean (%) 100 100 100 100 100 100 100 100 Control CV (%) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Growth Information NOEC (% sample) 100b 100 100 100 100 100 100 100 7 100 100 100 0 0.00% IC25 (% sample) >100b >100 >100 >100 >100 >100 >100 >100 7 >100 >100 >100 >100 0 0.00% Control mean (mg) 1.21 0.825 0.751 0.908 0.913 0.669 1.61 1.92 Control CV (%) 3.06 8.30 7.08 5.35 11.5 14.8 6.56 6.78 Flags3 db 65 b2, d5, e2, e3, 65, gl, g9 62,65 65 d5, e2 d2,d3 62,65 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 129 ------- Table 9.47. Results for sheepshead chronic test method performed on reference toxicant samples. LabID Referee 22 73 101 181 238 273 420 Summary Statistics Sample Code 9645 9621 9625 9629 9633 9637 9641 9649 N Min Max Median Mean STD CV% Test date 03/28/00 03/28/00 03/28/00 03/28/00 03/28/00 03/28/00 03/28/00 03/28/00 Survival Information NOEC (% sample) 50b 25 50 25 25 50 50 25 7 25 50 25 LC50 (% sample) 64.4b 45.1 68.9 37.2 51.1 66.0 69.5 61.6 7 37.2 69.5 61.6 57.0 12.7 22.2% Control mean (%) 100 100 97.5 100 100 100 100 100 Control CV (%) 0.00 0.00 5.94 0.00 0.00 0.00 0.00 0.00 Growth Information NOEC (% sample) 50b 25 25 25 25 25 25 25 7 25 25 25 IC25 (% sample) 54.3b 31.1 44.9 29.6 40.6 49.3 43.1 37.3 7 29.6 49.3 40.6 39.4 7.24 18.4% Control mean (mg) 0.889 0.844 0.654 1.26 0.760 0.641 1.60 2.10 Control CV (%) 8.31 3.65 12.9 3.06 9.77 2.73 5.47 2.15 Flags3 db ds, es c5, d5, e2, e3, e5, §1 d2, e2, e5 65 d5, e2 d2 e2,e5 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 130 ------- Table 9.48. Results for sheepshead chronic test method performed on effluent samples. LabID Referee 22 73 101 181 238 273 420 Summary Statistics Sample Code 9644 9620 9624 9628 9632 9636 9640 9648 N Min Max Median Mean STD CV% Test date 03/28/00 03/28/00 03/28/00 03/28/00 03/28/00 03/28/00 03/28/00 03/28/00 Survival Information NOEC (% sample) 25b 25 25 25 25 25 25 25 7 25 25 25 LC50 (% sample) 34. lb 33.6 34.2 33.7 34.3 35.4 35.4 35.4 7 33.6 35.4 34.3 34.5 0.805 2.33% Control mean (%) 100 100 100 100 97.5 100 100 100 Control CV (%) 0.00 0.00 0.00 0.00 5.94 0.00 0.00 0.00 Growth Information NOEC (% sample) 25b 12.5 12.5 25 25 25 12.5 12.5 7 12.5 25 12.5 IC25 (% sample) 29.4b 27.1 26.8 28.7 29.7 30.5 27.9 25.6 7 25.6 30.5 27.9 28.0 1.72 6.12% Control mean (mg) 0.994 0.896 0.653 1.23 0.768 0.626 1.56 2.02 Control CV (%) 4.11 6.51 13.7 10.0 15.5 2.62 2.26 10.1 Flags3 db e5 c5, d5, e2, e3, e5, §1 d2, e2, e5 65 bb d5, e2 d2 d5, e2, e5 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 131 ------- Table 9.49. Results for sheepshead chronic test method performed on receiving water samples. LabID Referee 22 73 101 181 238 273 420 Summary Statistics Sample Code 9643 9619 9623 9627 9631 9635 9639 9647 N Min Max Median Mean STD CV% Test date 03/21/00 03/21/00 03/21/00 03/21/00 03/21/00 03/21/00 03/21/00 03/21/00 Survival Information NOEC (% sample) 25b 25 25 25 25 25 25 25 7 25 25 25 LC50 (% sample) 35.4b 35.4 35.3 34.2 34.2 35.4 35.3 35.4 7 34.2 35.4 35.3 35.0 0.569 1.63% Control mean (%) 100 92.5 100 100 100 100 100 100 Control CV (%) 0.00 11.3 0.00 0.00 0.00 0.00 0.00 0.00 Growth Information NOEC (% sample) 6.25b 25 12.5 25 12.5 12.5 12.5 12.5 7 12.5 25 12.5 IC25 (% sample) 27.5b 28.7 27.9 29.8 27.3 27.8 25.4 24.0 7 24.0 29.8 27.8 27.3 1.95 7.15% Control mean (mg) 1.20 1.08 0.841 0.933 1.05 0.637 1.61 2.08 Control CV (%) 1.12 4.51 6.93 6.63 3.90 5.79 3.44 5.03 Flags3 di, eb e5, g6 c5, d5, e2, e3, e5, §1 e2,e5 65 bb d5, e2 d2,d3 62,65 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 132 ------- Table 9.50. Precision of point estimates from the sheepshead chronic test method. Sample type Reference toxicant Effluent Receiving water Average CV (%) LC50 Within-laba - - - - Between-laba - - - - Total 22.2 2.33 1.63 8.73 IC25 Within-laba - - - - Between-laba - - - - Total 18.4 6.12 7.15 10.5 1 Within- and between-laboratory components of variability were not calculated for this method since no within-laboratory replication was provided. Table 9.51. Precision of NOEC values from the sheepshead chronic test method. Sample type Reference toxicant Effluent Receiving water Endpoint Survival Growth Survival Growth Survival Growth NOEC Frequency <6.25% 0 0 0 0 0 0 6.25% 0 0 0 0 0 0 12.5% 0 0 0 4 0 5 25% 4 7 7 3 7 2 50% o 5 0 0 0 0 0 100% 0 0 0 0 0 0 Median (% sample) 25 25 25 12.5 25 12.5 % within 1 cone, of median 100 100 100 100 100 100 % beyond 1 cone, of median 0.00 0.00 0.00 0.00 0.00 0.00 133 ------- 9.10 Silverside Acute Test Method Results A total of nine participant laboratories conducted the silverside acute test method in the WET Variability Study. These laboratories tested a total of 6 blank samples, 12 reference toxicant samples, 12 effluent samples, and 6 receiving water samples. For each sample tested, a 96-hour LC50 was generated as a test result. Results of silverside acute testing are shown in Tables 9.52 - 9.55 for each sample type. Precision estimates for the silverside acute test were not calculated for the reference toxicant sample type because this sample type failed to produce toxicity that could be definitively measured within the test concentration range. All LC50s for the reference toxicant sample were >100% sample. This was caused by precipitation of the spiked copper in the liquid ampule sample (see Section 5.3). 9.10.1 Successful Test Completion Rate A total of 36 silverside acute tests were initiated by 9 participant laboratories. Of the 36 tests initiated, 2 tests (samples 9507 and 9506) failed to meet test acceptability criteria for control survival. The resulting successful test completion rate calculated in the WET Variability Study for the silverside acute test method was 94.4%. The two invalid tests were initiated in the same laboratory (33) on the same day and were likely due to poor health of test organisms supplied on that day. In addition, the referee laboratory failed to meet the test acceptability criteria for control survival in three tests (samples 9531, 9532, and 9533). The referee laboratory successfully repeated these tests, and results from the repeated tests are shown in Tables 9.52 - 9.54. 9.10.2 False Positive Rate A total of six blank samples were analyzed by six participant laboratories (Table 9.52). The LC50 calculated for all 6 blank samples was >100%, indicating no toxicity and no false positives. The resulting false positive rate calculated in the WET Variability Study for the silverside acute test method was 0.00%. 9.10.3 Precision Precision of the silverside acute test method was estimated by calculating the CV of LC50 values obtained for the effluent and receiving water samples. Precision estimates were not calculated for the reference toxicant sample type because this sample type failed to produce toxicity that could be definitively measured within the test concentration range (see Sections 5.3 and 9.1.4). Within-laboratory, between-laboratory, and total CVs were calculated for the effluent samples. Only a total CV was calculated for the receiving water samples since no within-laboratory replication was provided for this sample type. All participant laboratory test data for the effluent and receiving water samples were used in estimating precision. No results were identified by ASTM h statistics as possible outliers. Table 9.56 summarizes the precision of point estimates from the silverside acute test method. As expected, the majority of variability was due to the between-laboratory component, with a within-laboratory CV of 9.91% and a between-laboratory CV of 49.7%. Total CVs were much higher for the effluent sample type (50.7%) than the receiving water sample type (26.3%). Averaging the CVs based on total variance for the two sample types, a total CV of 38.5% was obtained for the silverside acute test method in the WET Variability Study. 134 ------- Table 9.52. Results for silverside acute test method performed on blank samples. LabID Referee 29 36 209 244 425 459 Summary Statistics Sample code 9532 9504 9512 9518 9528 9536 9538 N Min Max Median Mean Test date 12/03/99 11/09/99 11/09/99 11/02/99 11/09/99 11/09/99 11/03/99 False positives False positive rate Survival Information LC50 (% sample) >100b >100 >100 >100 >100 >100 >100 6 >100 >100 >100 >100 0 0.00% Control mean (%) 100 100 100 100 100 95.0 100 Control CV (%) 0.00 0.00 0.00 0.00 0.00 8.66 0.00 Flags3 b2, b3, e2, e4, gb §12 eb e2 d2, dfo e4 e4 db e6 b2, d2 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 135 ------- Table 9.53. Results for silverside acute test method performed on reference toxicant samples. LabID Referee 33 29 29 36 36 125 209 209 221 244 244 425 Summary Statistics Sample code 9531 9507 9502 9503 9510 9511 9515 9520 9521 9523 9526 9527 9535 N Min Max Median Mean Test date 12/03/99 11/02/99 11/02/99 11/02/99 11/02/99 11/02/99 11/02/99 11/09/99 11/09/99 11/02/99 11/02/99 11/02/99 11/02/99 Survival Information LC50 (% sample) >100b Invalid0 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100 11 >100 >100 >100 >100 Control mean (%) 100 55.0 100 100 95.0 95.0 100 95.0 90.0 95.0 95.0 95.0 100 Control CV (%) 0.00 8.52 0.00 0.00 8.66 8.66 0.00 8.66 0.00 8.66 8.66 8.66 0.00 Flags3 b2, b3, e2, e4, gb §12 ai bb e2 bbe2 cs, d2, d5, d6 d2, d5, d6 6i,e4 64 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from c Results from invalid tests were excluded from summary summary statistics. statistics. 136 ------- Table 9.54. Results for silverside acute test method performed on effluent samples. LabID Referee 29 33 33 36 125 125 221 221 244 425 459 459 Summary Statistics Sample code 9533 9505 9508 9509 9513 9516 9517 9524 9525 9529 9537 9540 9541 N Min Max Median Mean Within-lab Between-lab Total Test date 12/03/99 11/09/99 11/09/99 11/09/99 11/09/99 11/09/99 11/09/99 11/09/99 11/09/99 11/09/99 11/09/99 11/09/99 11/09/99 STD cv% STD cv% STD cv% Survival Information LC50 (% sample) 28.6b 42.0 20.4 17.9 40.8 23.9 18.3 68.2 60.3 84.7 37.7 38.2 33.0 12 17.9 84.7 38.0 40.4 4.01 9.91% 20.1 49.7% 20.5 50.7% Control mean (%) 90.0 100 90.0 90.0 90.0 100 100 90.0 95.0 100 100 100 100 Control CV (%) 0.00 0.00 0.00 17.1 0.00 0.00 0.00 0.00 8.66 0.00 0.00 0.00 0.00 Flags3 b2, b3, e2, e4, gb §12 d2, e2 d2, d7, e4 db e4 db e4 e4 d2 d2 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 137 ------- Table 9.55. Results for silverside acute test method performed on receiving water samples. LabID Referee 33 125 209 221 425 459 Summary Statistics Sample code 9530 9506 9514 9519 9522 9534 9539 N Min Max Median Mean STD CV% Test date 11/02/99 11/02/99 11/02/99 11/02/99 11/02/99 11/02/99 11/03/99 Survival Information LC50 (% sample) 47,3b Invalid0 21.5 28.7 44.3 35.4 39.0 5 21.5 44.3 35.4 33.8 8.90 26.3% Control mean (%) 95.0 65.0 100 100 90.0 100 100 Control CV (%) 8.66 24.0 0.00 0.00 17.1 0.00 0.435 Flags3 e2,gi ai d, d5 b2, d2, el * Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. c Results from invalid tests were excluded from summary statistics. Table 9.56. Precision of point estimates from the silverside acute test method. Sample type Reference toxicant1" Effluent Receiving water Average CV (%) Within-laba - 9.91 - 9.91 Between-laba - 49.7 - 49.7 Total - 50.7 26.3 38.5 a Within- and between-laboratory components of variability were not calculated for the receiving water sample type since no within- laboratory replication was provided for this sample type. b Precision estimates were not calculated for the reference toxicant sample type because this sample type failed to produce toxicity that could be definitively measured within the test concentration range (see Section 5.3 and 9.1.4). 138 ------- 9.11 Silverside Chronic Test Method Results A total of 10 participant laboratories conducted the silverside chronic test method in the WET Variability Study. These laboratories tested a total of 7 blank samples, 13 reference toxicant samples, 13 effluent samples, and 7 receiving water samples. For each sample tested, a 7-day survival NOEC, a 7-day growth NOEC, a 7-day LC50, and a 7-day growth IC25 were generated as test results. As described in the WET method manual, growth was measured as the total weight per replicate divided by the number of original organisms in that replicate. This definition provides a combined growth and survival endpoint that is more accurately termed biomass. Also, test concentrations above the survival NOEC were excluded from hypothesis testing conducted on the growth endpoint. Results of silverside chronic testing are shown in Tables 9.57 - 9.60 for each sample type. 9.11.1 Successful Test Completion Rate A total of 40 samples were tested at 10 participant laboratories. All 40 tests were completed and met test acceptability criteria. The resulting successful test completion rate calculated in the WET Variability Study for the silverside chronic test method was 100%. The growth endpoint results for sample 9545 were reported as inconclusive based on an evaluation of the concentration-response relationship (Table 8.4). If this test is considered unsuccessful (since the test would be repeated in a regulatory context if the test endpoint required in the permit produced an inconclusive result), the successful test completion rate becomes 97.5%. 9.11.2 False Positive Rate A total of seven blank samples were analyzed by seven participant laboratories (Table 9.57). No false positives were observed for survival or growth endpoints. The survival NOEC was 100% for all 7 blank samples, and the LC50 was >100% for all 7 blank samples. The growth NOEC was 100% for all 7 blank samples, and the IC25 was >100% for all 7 blank samples. The resulting false positive rate calculated in the WET Variability Study for the silverside chronic test method was 0.00%. A participant laboratory reported a growth NOEC of less than 100% for sample 9556, indicating a false positive. This sample exhibited an interrupted concentration-response curve, with only the 12.5% treatment producing a significant difference from the control. Based on EPA guidance for evaluating concentration- response relationships, the growth NOEC for sample 9556 was recalculated and reported as 100% (Table 8.4). 9.11.3 Precision Precision of the silverside chronic test method was estimated by calculating the CV of LC50 and IC25 values obtained for the reference toxicant, effluent, and receiving water samples. Within-laboratory, between- laboratory, and total CVs were calculated for the reference toxicant and effluent samples. Only a total CV was calculated for the receiving water samples since no within-laboratory replication was provided for this sample type. Results from two samples (9582 and 9583) were excluded from the analysis of precision. Samples 9582 and 9583 were identified by ASTM h statistics as possible outliers, and review of data qualifier flags revealed a possible cause (Table 9.1). 139 ------- Table 9.61 summarizes the precision of point estimates from the silverside chronic test method. For LC50 values, within-laboratory CVs ranged from 9.17% to 12.2%, between-laboratory CVs ranged from 32.2% to 46.8%, and total CVs ranged from 33.5% to 48.4%. CVs were generally higher for the effluent sample (48.4%) than for reference toxicant (33.5%) or receiving water (40.0%) samples. As expected, the majority of variability was due to the between-laboratory component, with within-laboratory CVs averaging 10.7% and between-laboratory CVs averaging 39.5%. Averaging the CVs for the LC50 based on total variance for the three sample types, a total CV of 40.6% was obtained for the silverside chronic survival endpoint. CVs for IC25 values were slightly higher than those for LC50 values. Within-laboratory CVs for the IC25 ranged from 7.24% to 22.0%, between-laboratory CVs ranged from 29.1% to 55.5%, and total CVs ranged from 36.4% to 56.0%. As expected, the majority of variability was due to the between-laboratory component, with within-laboratory CVs averaging 14.6% and between-laboratory CVs averaging 42.3%. Averaging the CVs for the IC25 based on total variance for the three sample types, a total CV of 43.8% was obtained for the silverside chronic growth endpoint. The precision of NOEC values was determined by evaluating the range and distribution of NOEC values and the percentage of values falling within and beyond one concentration from the median. Table 9.62 describes the precision of NOEC values for the silverside chronic method. For the survival endpoint, NOEC values spanned four concentrations for the reference toxicant sample type, five concentrations for the effluent sample type, and three concentrations for the receiving water sample type. The percentage of values within one concentration of the median was 90.9%, 84.6%, and 85.7% for the reference toxicant, effluent, and receiving water sample types, respectively. Growth NOEC values spanned four concentrations for the reference toxicant and effluent sample types and spanned three concentrations for the receiving water sample type. The percentage of values within one concentration of the median was 90.9%, 91.7%, and 85.7% for the reference toxicant, effluent, and receiving water sample types, respectively. Of the four results (for samples 9561, 9545, 9564, and 9562) that were beyond one concentration from the median, two (samples 9564 and 9562) were from the same laboratory (125). During tests on both samples, the pH in control replicates and lower test concentrations exceeded 9.0. 140 ------- Table 9.57. Results for silverside chronic test method performed on blank samples. LabID Referee 22 36 39 124 221 333 421 Summary Statistics Sample code 9581 9544 9550 9556 9558 9570 9576 9584 N Min Max Median Mean Test date 10/26/99 10/26/99 10/19/99 10/26/99 10/19/99 10/19/99 10/26/99 10/27/99 False positives False positive rate Survival Information NOEC (% sample) 100b 100 100 100 100 100 100 100 7 100 100 100 0 0.00% LC50 (% sample) >100b >100 >100 >100 >100 >100 >100 >100 7 >100 >100 >100 100 0 0.00% Control mean (%) 97.5 97.5 97.5 97.5 100 95.0 95.0 97.5 Control CV (%) 5.94 5.94 5.94 5.94 0.00 7.07 7.07 5.94 Growth Information NOEC (% sample) 100b 100 100 100 100 100 100 100 7 100 100 100 0 0.00% IC25 (% sample) >100b >100 >100 >100 >100 >100 >100 >100 7 >100 >100 >100 100 0 0.00% Control mean (mg) 0.647 0.566 1.28 0.783 2.33 0.896 1.46 1.07 Control CV (%) 5.66 8.50 12.0 5.75 5.48 16.2 7.76 14.4 Flags3 65 bb c5, d5, e5 S5 e5, e6 db d2, e5 db eb efo gj db d5, e6, gj a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 141 ------- Table 9.58. Results for silverside chronic test method performed on reference toxicant samples. LabID Referee 22 33 36 36 39 39 124 125 209 333 333 421 421 Summary Statistics Sample code 9578 9543 9547 9552 9553 9554 9555 9561 9563 9567 9574 9575 9582 9583 N Min Max Median Mean Within-lab Between-lab Total Test date 10/19/99 10/19/99 10/19/99 10/26/99 10/26/99 10/19/99 10/19/99 10/26/99 10/19/99 10/19/99 10/19/99 10/19/99 10/20/99 10/20/99 STD cv% STD cv% STD cv% Survival Information NOEC (% sample) 6.25b 12.5 12.5 12.5 12.5 12.5 12.5 <6.25 6.25 12.5 12.5 25 25C 100C 11 <6.25 25 12.5 LC50 (% sample) 17.1b 28.1 23.5 24.3 21.1 27.3 23.7 11.0 10.0 26.7 34.6 37.0 65.6C >100C 11 9.96 37.0 24.3 24.3 2.23 9.17% 7.82 32.2% 8.13 33.5% Control mean (%) 97.5 97.5 100 100 97.5 95.0 95.0 100 80.0 100 100 100 100 95.0 Control CV (%) 5.94 5.94 0.00 0.00 5.94 7.07 7.07 0.00 9.48 0.00 0.00 0.00 0.00 7.07 Growth Information NOEC (% sample) 6.25b 12.5 12.5 12.5 12.5 12.5 12.5 <6.25 6.25 12.5 12.5 25 25C 100C 11 <6.25 25 12.5 IC25 (% sample) 15.3b 18.1 17.8 26.1 17.3 17.7 16.9 6.95 8.02 20.6 30.5 25.6 53.7C >100C 11 6.95 30.5 17.8 18.7 4.10 22.0% 5.43 29.1% 6.80 36.4% Control mean (mg) 0.603 0.951 0.941 1.41 1.36 1.06 1.06 2.81 1.34 1.57 1.92 2.09 1.25 0.915 Control CV (%) 13.1 28.7 14.2 3.36 4.08 13.9 13.9 10.4 8.50 13.5 6.56 3.06 14.9 8.44 Flags3 t>i, e5, e6 t>i, c5, d5, e5 gi t>i, e6 t>i,e6 e5 d4,e5 gi gi b2, dj, e6, gl, g3 b2, dj, e6, gl, g3 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. c Results were identified as outliers, a probable cause was identified, and therefore results were excluded from summary statistics. 142 ------- Table 9.59. Results for silverside chronic test method performed on effluent samples. LabID Referee 22 33 33 39 124 125 125 209 209 221 221 333 421 Summary Statistics Sample code 9580 9545 9548 9549 9557 9560 9564 9565 9568 9569 9572 9573 9577 9585 N Min Max Median Mean Within-lab Between-lab Total Test date 10/26/99 10/26/99 10/27/99 10/27/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/26/99 10/27/99 STD cv% STD cv% STD cv% Survival Information NOEC (% sample) 12.5b 100 12.5 25 25 25 6.25 12.5 25 12.5 50 50 25 50 13 6.25 100 25 LC50 (% sample) 28.9b >100 57.7 75.1 43.2 48.2 16.4 20.9 35.6 32.0 66.1 70.7 47.2 >100 13 16.4 >100 48.2 54.9 6.70 12.2% 25.7 46.8% 26.5 48.4% Control mean (%) 92.5 97.5 100 97.5 97.5 100 100 100 92.5 100 92.5 100 95.0 100 Control CV (%) 11.3 5.94 0.00 5.94 5.94 0.00 0.00 0.00 6.32 0.00 16.1 0.00 11.4 0.00 Growth Information NOEC (% sample) 12.5b Inconclusive0 12.5 25 25 25 6.25 12.5 25 12.5 50 50 25 50 12 6.25 50 25 IC25 (% sample) 27.8" Inconclusive0 53.7 56.9 32.5 36.8 8.74 16.1 31.1 30.8 63.0 67.3 40.4 >100 12 8.74 >100 38.6 44.8 3.24 7.24% 24.9 55.5% 25.1 56.0% Control mean (mg) 0.561 0.613 0.867 1.02 0.783 2.86 0.984 0.984 1.29 1.19 0.602 0.607 1.58 0.921 Control CV (%) 11.1 24.8 21.7 7.85 5.75 7.31 13.0 13.0 5.27 14.9 7.37 13.3 2.01 21.8 Flags3 e5 t>i= c5, di, e5, g2 bi.bj.g, t>i, b2, g; bj, d7 e5 d4, e5 d4,e5 d7 di,d, di, e5 di,e5 di, gi di, d,, e6, gl, g2 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 0 Results were excluded from summary statistics. Based on EPA guidance for evaluating concentration-response relationships (EPA, 2000a) the test result was inconclusive and the sample should be retested to obtain a reliable result. 143 ------- Table 9.60. Results for silverside chronic test method performed on receiving water samples. LabID Referee 22 33 36 124 125 209 221 Summary Statistics Sample code 9579 9542 9546 9551 9559 9562 9566 9571 N Min Max Median Mean STD CV% Test date 10/19/99 10/19/99 10/19/99 10/19/99 10/19/99 10/19/99 10/19/99 10/19/99 Survival Information NOEC (% sample) 12,5b 25 25 12.5 12.5 6.25 25 25 7 6.25 25 25 LC50 (% sample) 26,2b 52.7 43.8 38.4 26.5 11.5 55.8 39.2 7 11.5 55.8 39.2 38.3 15.3 40.0% Control mean (%) 85.0 95.0 100 100 100 80.0 97.5 100 Control CV (%) 6.95 7.07 0.00 0.00 0.00 9.48 5.94 0.00 Growth Information NOEC (% sample) 12,5b 25 25 12.5 12.5 6.25 25 25 7 6.25 25 25 IC25 (% sample) 19.1b 36.0 36.0 29.3 18.2 9.35 41.6 32.8 7 9.35 41.6 32.8 29.0 11.4 39.1% Control mean (mg) 0.644 0.840 0.941 1.34 2.33 1.34 1.58 1.18 Control CV (%) 9.60 21.3 14.2 8.12 5.48 8.45 9.46 12.2 Flags3 t>i, e5, e6 bb c5, d5, e5 bb gi e5, e6 d4, e5 db d2, e5 a Data qualifier flags are described in Table 8.3. b Results from the referee laboratory were excluded from summary statistics. 144 ------- Table 9.61. Precision of point estimates from the silverside chronic test method. Sample type Reference toxicant Effluent Receiving water Average CV (%) LC50 Within-lab a 9.17 12.2 - 10.7 Between-lab a 32.2 46.8 - 39.5 Total 33.5 48.4 40.0 40.6 IC25 Within-lab a 22.0 7.24 - 14.6 Between-lab a 29.1 55.5 - 42.3 Total 36.4 56.0 39.1 43.8 a Within- and between-laboratory components of variability were not calculated for the receiving water sample type since no within-laboratory replication was provided for this sample type. Table 9.62. Precision of NOEC values from the silverside chronic test method. Sample type Reference toxicant Effluent Receiving water Endpoint Survival Growth Survival Growth Survival Growth NOEC Frequency <6.25% 1 1 0 0 0 0 6.25% 1 1 1 1 1 1 12.5% 8 8 3 3 2 2 25% 1 1 5 5 4 4 50% 0 0 o J 3 0 0 100% 0 0 1 0 0 0 Median (% sample) 12.5 12.5 25 25 25 25 % within 1 cone, of median 90.9 90.9 84.6 91.7 85.7 85.7 % beyond 1 cone, of median 9.09 9.09 15.4 8.33 14.3 14.3 145 ------- 9.12 Champia Chronic Preliminary Testing Results No interlaboratory test data were obtained for the Champia chronic test method (see Section 2.1); however, preliminary testing conducted by the referee laboratory (EnviroSystems, Inc.) provided limited single- laboratory data on the precision of the method. The results of preliminary testing are detailed in Appendix D and summarized here. In total, 17 Champia chronic tests were conducted by the referee laboratory during preliminary testing. Three tests were conducted on reference toxicant samples (filtered, sterilized natural seawater spiked with CuSO4), four tests were conducted on spiked receiving water samples (unfiltered and unsterilized natural seawater spiked with CuSO4), seven tests were conducted on unspiked effluent samples (municipal wastewater treatment plant effluent), and three tests were conducted on unspiked receiving water samples (unfiltered and unsterilized natural seawater). For the three reference toxicant tests conducted between 7/27/99 and 5/16/00, the IC25 averaged 0.228 jug Cu/L with a CV of 27.6%. For the four tests conducted on spiked receiving water, a CV of 49.7% was achieved. For the unspiked effluent and unspiked receiving water sample types, method precision could only be assessed in the testing of duplicate samples. Results from effluent and receiving water samples collected on different dates would reflect effluent and receiving water variability as well as method variability. Testing of duplicate effluent samples collected on 5/9/00 yielded a CV of 50.0%. Testing of duplicate receiving water samples collected on 5/23/00 could not be adequately compared due to an inconclusive test result from one of the duplicate samples (see Appendix D). 9.13 Holmesimysis Acute Preliminary Testing Results No interlaboratory test data were obtained for the Holmesimysis acute test method (see Section 2.1); however, preliminary testing conducted by the referee laboratory (MEC Analytical, Inc.) provided limited single- laboratory data for this method. The results of preliminary testing are detailed in Appendix D and summarized here. In total, five Holmesimysis acute tests were conducted by the referee laboratory during preliminary testing. Of these five tests, three exhibited control survival of 80% and failed to meet test acceptability criteria. It is believed that poor control survival in these tests was due to the use of a synthetic seawater rather than a natural seawater for organism holding and test dilution. The two remaining tests met test acceptability criteria for control survival, but test organisms were field-collected and their ages could not be confirmed as meeting the required test conditions (1-5 days; 24-h range in age). Additional tests were not conducted by the referee laboratory due to difficulties in obtaining field-collected gravid females (used to produce juveniles for test use) during the winter and spring months of the WET Variability Study (see Appendix D). 9.14 Results Summary This section summarizes the results of interlaboratory testing conducted during the WET Variability Study. In this study, EPA assessed the performance of the WET test methods by calculating successful test completion rates, false positive rates, and precision (i.e., CVs for point estimates and ranges for NOEC values) for each test method evaluated in interlaboratory testing. 146 ------- 9.14.1 Successful Test Completion Rate The successful test completion rate for each WET test method was calculated as the percentage of initiated and properly terminated tests that met the test acceptability criteria as specified in the WET method manuals (see Section 9.1.2). Table 9.63 summarizes the successful test completion rates achieved in the WET Variability Study. Successful test completion rates were 100% for 4 of the 10 test methods, indicating that no invalid tests were conducted for those methods. The successful test completion rate was above 94% for 8 of the 10 methods. Only the Ceriodaphnia chronic and Selenastrum chronic test methods achieved successful test completion rates below 90%. Table 9.63. Successful test completion rates for test methods evaluated in the WET Variability Study. Test method Ceriodaphnia acute Ceriodaphnia chronic Fathead acute Fathead chronic Selenastrum chronic (with EDTA) Selenastrum chronic (without EDTA) Mysidopsis chronic Sheepshead acute Sheepshead chronic Silverside acute Silverside chronic N 104 122 107 101 44 44 44 28 28 36 40 No. of invalid tests 5 22 0 2 16 15 1 0 0 2 0 Successful test completion rate (%) 95.2 82.0 100 98.0 63.6 65.9 97.7 100 100 94.4 100 The successful test completion rate for the Ceriodaphnia chronic test method (82.0%) was suppressed by poor performance in a subset of laboratories (see Section 9.3.1). Only 10 of the 34 participant laboratories performed invalid tests, but 8 of these laboratories performed invalid tests on 50% or more of the samples tested. The low rate of successful test completion in these 8 laboratories may have been influenced by the study's strict testing schedule, which required each test to be conducted on a given day and all tests to be conducted within a 15-day time period (see Section 2.2.5 and Table 2.4). When invalid tests conducted in a given laboratory were due to marginal or poor health of the test organism cultures, then it was logical that the laboratory would fail a high percentage of tests during this study because culture health was unlikely to fully recover within 15 days. Successful test completion rates for the Selenastrum chronic test method were 63.6% and 65.9% for tests conducted with and without EDTA, respectively. The lower successful test completion rates for this method appeared to be partially attributable to laboratory inexperience in using both the with and without EDTA techniques (see Section 9.6.1). Two laboratories that cultured organisms without EDTA and generally 147 ------- conducted tests without EDTA showed poor successful test completion (failing eight of eight tests) when EDTA was used. These laboratories failed all tests (8) conducted with EDTA and passed all but one test (7) without EDTA. Another laboratory that cultured organisms with EDTA and generally conducted tests with EDTA showed poor successful test completion (failing four of four tests) when the without EDTA procedure was used. This laboratory failed all tests (4 of 4) without EDTA and passed all tests (4 of 4) with EDTA. 9.14.2 False Positive Rate False positive rates were calculated as the percentage of valid tests showing toxicity in blank samples (see Section 9.1.3). Table 9.64 summarizes the false positive rates observed for test methods evaluated in the WET Variability Study. For the survival endpoint, false positive rates were 0.00% for all test methods evaluated. False positive rates for the reproduction endpoint were 3.70% for the Ceriodaphnia chronic test method and 0.00% for the Mysldopsls chronic test method. Table 9.64. False positive rates for test methods evaluated in the WET Variability Study. Test method Ceriodaphnia acute Ceriodaphnia chronic Fathead acute Fathead chronic Selenastmm chronic (with EDTA) Selenastrum chronic (without EDTA) Mysidopsis chronic Sheepshead acute Sheepshead chronic Silverside acute Silverside chronic N 33 27 27 24 5 6 7 7 7 6 7 False positive rate (%) Survival endpoint NOEC - 0.00 - 0.00 - - 0.00 - 0.00 - 0.00 LC50 0.00 0.00 0.00 0.00 - - 0.00 0.00 0.00 0.00 0.00 Growth endpoint NOEC - - - 4.35a 0.00 20.0b 0.00 - 0.00 - 0.00 IC25 - - - 4.17 0.00 33.3 0.00 - 0.00 - 0.00 Reproduction endpoint NOEC - 3.70 - - - - o.ooc - - - - IC25 - 3.70 - - - - o.ooc - - - - a N for the growth NOEC was 23. b N for the growth NOEC was 5. c N for the fecundity endpoint was 4. For the growth endpoint, false positive rates were 0.00% for all test methods except the fathead chronic test method and the Selenastrum chronic test method conducted without EDTA. For the fathead chronic test method, false positive rates were 4.17% for the growth IC25 and 4.35% for the growth NOEC. For the Selenastrum chronic test method conducted without EDTA, false positive rates were 33.3% for the growth IC25 and 20.0% for the growth NOEC. These relatively high false positive rates for the Selenastrum chronic test method may be due in part to a small sample size. The false positive rate of 33.3% reflects only 2 false 148 ------- positives out of 6 valid test results, and the 20.0% false positive rate reflects only 1 false positive out of 5 valid tests. When the growth IC50 was calculated for the same 6 tests, no false positives were observed. Also, no false positives were observed for the Selenastrum chronic test performed with the addition of EDTA. In summary, false positives were observed for only 3 of the 10 test methods (Ceriodaphnia chronic, fathead chronic, and Selenastrum chronic performed without EDTA), and the rate of false positives was below the theoretical false positive rate of 5% (based on the recommended 0.05 alpha level for hypothesis testing) for all test methods except for the Selenastrum chronic test method performed without EDTA. 9.14.3 Precision The precision of test methods evaluated in the WET Variability Study was estimated by calculating a CV for point estimates (i.e., LCSOs and IC25s). For test methods that included within-laboratory replication of sample types, CVs were calculated based on within-laboratory, between-laboratory, and total variance (see Section 9.1.4). These CVs were calculated independently for each sample type and averaged to obtain final estimates of within-laboratory, between-laboratory, and total variability. Table 9.65 shows the CVs calculated for each test method based on within-laboratory, between-laboratory, and total variability. As expected, the within- laboratory variability observed for most test methods was much lower than between-laboratory variability. Within-laboratory CVs ranged from 6.57% to 12.1% for LC50 values and from 6.98% to 25.2% for IC25 values. Within-laboratory variability was highest for the Selenastrum chronic test method, with CVs of 25.2% and 23.3% for growth IC25 results from tests conducted with and without EDTA, respectively. The within- laboratory variability for this method was much lower when calculated for the growth IC50 values. CVs for the Selenastrum growth IC50 were 5.82% with EDTA and 14.5% without EDTA. Within-laboratory CVs calculated in the WET Variability Study represent the variability of results between replicate samples tested simultaneously in a given laboratory. Therefore, these within-laboratory CVs are expected to be lower than previously reported within-laboratory CVs based on reference toxicants tested over time in a given laboratory (as in USEPA, 2000d). Between-laboratory variability was higher than within-laboratory variability for all test methods with the exception of the Selenastrum chronic test method performed with EDTA. Between-laboratory variability observed in the WET Variability Study ranged from 11.3% to 49.7% for LC50 values and from 14.6% to 72.0% for IC25 values. Similarly to within-laboratory CVs, the growth IC50 for the Selenastrum chronic test method was less variable than the growth IC25. The between-laboratory CV for the growth IC50 was 13.2% (versus 14.6% for the IC25) when the Selenastrum chronic test was conducted with EDTA and 43.9% (versus 72.0% for the IC25) when the Selenastrum chronic test was conducted without EDTA. The CVs calculated based on total variance were used to summarize the precision of test methods evaluated in the WET Variability Study. These precision estimates are averaged across sample types and incorporate both within-laboratory and between-laboratory components of method variability. Summarized precision estimates based on total variance are presented in Table 9.66. CVs for acute WET test methods ranged from 20.0% to 38.5%. CVs for chronic WET test methods ranged from 8.73% to 40.6% for LC50 values and from 10.5% to 58.5% for IC25 values. These CVs are well within the range of CVs previously reported for WET test methods. USEPA (1988) reported multilaboratory precision (CVs) of 22-167% (with a weighted mean of 50%) for acute WET methods testing reference toxicants. USEPA (1991) reported interlaboratory precision (CVs) of acute 149 ------- methods as 34.6% to 50.1% for the sheepshead acute method and 22.3% to 59.5% for the Mysldopsls acute method. Interlaboratory precision (CVs) of chronic methods was previously reported as 20.5%to 41.1% forthe Ceriodaphnia chronic test method, 31% for the fathead chronic method, and 44.2% for the sheepshead chronic method (USEPA, 1991). EPA intends to further comment on the results of the WET Variability Study and the significance of these results in subsequent rulemaking. Table 9.65. Within-laboratory, between-laboratory, and total variability observed for test methods evaluated in the WET Variability Study. Test method Ceriodaphnia acute Ceriodaphnia chronic Fathead acute Fathead chronic Selenastrum chronic (withEDTA) Selenastrum chronic (without EDTA) Mysidopsis chronic0 Sheepshead acuted Sheepshead chronicd Silverside acute Silverside chronic CV (%)a LC50 Within- laboratory 12.1 7.09 8.96 7.87 - - 6.57 - - 9.91 10.7 Between- laboratory 24.0 21.8 19.4 11.3 - - 27.3 - - 49.7 39.5 Total" 29.0 21.5 20.0 13.4 - - 31.2 26.0 8.73 38.5 40.6 IC25 Within- laboratory - 17.4 - 14.6 25.2 23.3 6.98 - - - 14.6 Between- laboratory - 27.6 - 15.0 14.6 72.0 38.3 - - - 42.3 Total" - 35.0 - 20.9 34.3 58.5 41.3 - 10.5 - 43.8 a Within-laboratory, between-laboratory, and total CVs presented are averaged across sample types. No within-laboratory replication was provided for the receiving water sample type, so CVs based on within and between-laboratory variance are averaged across only the reference toxicant and effluent sample types; CVs based on total variance are averaged across the reference toxicant, effluent, and receiving water sample types. See Sections 9.2 - 9.11 for precision estimates calculated independently for each sample type and variance component. b CVs based on total variance may not necessarily be greater than CVs based on within and between-laboratory variance because the CVs presented are averaged across sample types. c For the Mysidopsis chronic test method, CVs for the IC25 represent results for the growth endpoint. d Within and between-laboratory components of variability were not estimated for the sheepshead acute and chronic test methods because no within-laboratory replication was provided for these methods. 150 ------- Table 9.66. Summarized precision estimates (CVs) for test methods evaluated in the WET Variability Study. Test method Ceriodaphnia acute Ceriodaphnia chronic Fathead acute Fathead chronic Selenastrum chronic (with EDTA) Selenastrum chronic (without EDTA) Mysidopsis chronicb Sheepshead acute Sheepshead chronic Silverside acute Silverside chronic CV (%)a LC50 29.0 21.5 20.0 13.4 - - 31.2 26.0 8.73 38.5 40.6 IC25 - 35.0 - 20.9 34.3 58.5 41.3 - 10.5 - 43.8 a CVs presented are based on total variance and averaged across sample types. bFor the Mysidopsis chronic test method, CVs for the IC25 represent results for the growth endpoint. 151 ------- 10.0 REFERENCES American Society for Testing and Materials. 1997. Standard practice for conducting an interlaboratory study to determine the precision of a test method. E 691-92. In ASTM Standards onPrecision and Bias for Various Applications, 5th Ed. West Conshohocken, PA, pp.309-328. DynCorp. 1996. Directory of Environmental Testing Laboratories. S.J. Kolb, D.J. Lesko, B.W. Schorer, B.P. Moser, B.E. Lane, R.K. Holder, and K. Lee, eds., DynCorp, Alexandria, VA. Edison Electric Institute et al. v. EPA. Settlement Agreement, July 24, 1998. U.S. Court of Appeals, D.C. Circuit, No. 96-1062. SAS Institute. 2000. SAS Procedure Guide. Version 8, Gary, NC. Tidepool Scientific Software. 1996. ToxCalc Toxicity Data Analysis Software, User's Guide. Version 5.0, McKinleyville, CA. U.S. Environmental Protection Agency. 1988. Availability, Adequacy, and Comparability of Testing Procedures for the Analysis of Pollutants Established Under Section 304(h) of the Federal Water Pollution Control Act. EPA/600/9-87/030. U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory (currently, National Exposure Research Laboratory), Cincinnati, OH. U.S. Environmental Protection Agency. 1991. Technical Support Document for Water Quality-Based Toxics Control. EPA/505/2-90/001. U.S. Environmental Protection Agency, Office of Water Enforcement and Permits, and Office of Water Regulations and Standards, Washington, DC. U.S. Environmental Protection Agency. 1993. Methods for Measuring the Acute Toxicity of Effluents and Receiving Waters to Freshwater and Marine Organisms, 4th ed., EPA/600/4-90/027F. U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory (currently, National Exposure Research Laboratory), Cincinnati, OH. U.S. Environmental Protection Agency. 1994a. Short-term Methods for Estimating the Chronic Toxicity of Effluents and Receiving Waters to Freshwater Organisms, 3rd ed., EPA/600/4-91/002. U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory (currently, National Exposure Research Laboratory), Cincinnati, OH. U.S. Environmental Protection Agency. 1994b. Short-term Methods for Estimating the Chronic Toxicity of Effluents and Receiving Waters to Marine and Estuarine Organisms, 2nd ed., EPA/600/4-91/003. U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory (currently, National Exposure Research Laboratory), Cincinnati, OH. U.S. Environmental Protection Agency. 1996. Clarifications Regarding Flexibility in 40 CFR Part 136 Whole Effluent Toxicity (WET) Test Methods, April 10, 1996, memorandum from Tudor Davies, U.S. Environmental Protection Agency, Office of Science and Technology, Washington D.C. 152 ------- U.S. Environmental Protection Agency. 1999. Errata for Effluent and Receiving Water Toxicity Test Manuals: Acute Toxicity of Effluents and Receiving Waters to Freshwater and Marine Organisms; Short-term Methods for Estimating the Chronic Toxicity of Effluents and Receiving Waters to Freshwater Organisms; and Short-term Methods for Estimating the Chronic Toxicity of Effluents and Receiving Waters to Marine andEstuarine Organisms. January 1999. EPA/600/R-98/182. U.S. Environmental Protection Agency, Office of Research and Development, Duluth, MN. U.S. Environmental Protection Agency. 2000a. Method Guidance and Recommendations for Whole Effluent Toxicity (WET) Testing (40 CFRPart 136). EPA/821/B-00/004. U.S. Environmental Protection Agency, Office of Water, Washington, B.C. U.S. Environmental Protection Agency. 2000b. Preliminary Report: Interlaboratory Variability Study of EPA Short-term Chronic and Acute Whole Effluent Toxicity Test Methods, Vol. 1. EPA/821/R-00/028A. U.S. Environmental Protection Agency, Office of Water, Washington, B.C. U.S. Environmental Protection Agency. 2000c. Preliminary Report: Interlaboratory Variability Study of EPA Short-term Chronic and Acute Whole Effluent Toxicity Test Methods, Vol. 2: Appendix. EPA/821/R- 00/028B. U.S. Environmental Protection Agency, Office of Water, Washington, B.C. U.S. Environmental Protection Agency. 2000d. Understanding and Accounting for Method Variability in Whole Effluent Toxicity Applications Under the National Pollutant Discharge Elimination System Program. EPA/833/R-00/003. U.S. Environmental Protection Agency, Office of Wastewater Management, Washington, B.C. 153 ------- |