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
------- |