v>EPA
United Stales
Environmental Protection
Agency
Office of Research and
Development
Washington, DC 20460
EPA/600/4-91/021
February 1992
Short-Term
Methods for
Estimating the
Chronic Toxicity of
Effluents and
Receiving Waters to
Marine and Estuarine
Organisms
Final Draft
-------
EPA/600/4-91/021
February 1992
Final Draft
SHORT-TERM METHODS FOR ESTIMATING THE CHRONIC TOXICITY OF EFFLUENTS
AND RECEIVING WATERS TO MARINE AND ESTUARINE ORGANISMS
(Second Edition)
Edited by
Donald J. Klemrn1 and George E. Morrison"
Environmental Monitoring Systems Laboratory, Cincinnati, Ohio
Environmental Research Laboratory, Narragansett, Rhode Island
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY - CINCINNATI
OFFICE OF RESEARCH AND DEVELOPMENT
U. S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
Printed on Recycled Paper
-------
DISCLAIMER
This document has been reviewed by the Environmental Monitoring Systems
Laboratory - Cincinnati (EMSL-Cincinnati), U. S. Environmental Protection
Agency (USEPA), and approved for publication. The mention of trade names or
commercial products does not constitute endorsement or recommendation for use
11
-------
FOREWORD
Environmental measurements are required to determine the quality of
ambient waters and the character of waste effluent. The Environmental
Monitoring Systems Laboratory - Cincinnati (EMSL-Cincinnati) conducts research
to:
o Develop and evaluate analytical methods to identify and measure the
concentration of chemical pollutants in drinking waters, surface
waters, groundwaters, wastewaters, sediments, sludges, and solid
wastes.
o Investigate methods for the identification and measurement of viruses,
bacteria and other microbiological organisms in aqueous samples and to
determine the responses of aquatic organisms to water quality.
o Develop and operate a quality assurance program to support the
achievement of data quality objectives in measurements of pollutants
in drinking water, surface water, groundwater, wastewater, sediment
and sol id waste.
o Develop methods and models to detect and quantify responses in aquatic
and terrestrial organisms exposed to environmental stressors and to
correlate the exposure with effects on chemical and biological
indicators.
The Federal Water Pollution Control Act Amendments of 1972 (PL 92-500),
the Clean Water Act (CWA) of 1977 (PL 95-217) and the Water Quality Act of
1987 (PL 100-4) explicitly state that it is the national policy that the
discharge of toxic substances in toxic amounts be prohibited. Thus, the
detection of chronically toxic effluents plays an important role in
identifying and controlling toxic discharges to surface waters. This manual
is the second edition of the marine and estuarine chronic toxicity test manual
for effluents, first published (EPA/600/4-88/028) by EMSL-Cincinnati in May,
1988. It provides updated and standardized methods for estimating the chronic
toxicity of effluents and receiving waters to estuarine and marine organisms
for use by the U. S. Environmental Protection Agency (USEPA) regional
programs, the state programs, and the National Pollutant Discharge Elimination
System (NPDES) permittees.
Thomas A. Clark, Director
Environmental Monitoring Systems
Laboratory - Cincinnati
m
-------
PREFACE
This manual represents the second edition of the Agency's methods
manual for estimating the chronic toxicity of effluents and receiving waters
to marine and estuarine organisms initially published by EMSL-Cincinnati in
May, 1988. This edition reflects changes recommended by the Toxicity
Assessment Subcommittee of the EMSL-Cincinnati Biological Advisory Committee,
USEPA headquarters and regional staff, other Federal agencies, state and
interstate water pollution control programs, environmental protection groups,
trade associations, major industries, consulting firms, academic institutions
engaged in aquatic toxicology research, and other interested parties in the
private sector.
The membership of the Toxicity Assessment Subcommittee, EMSL -
Cincinnati Biological Advisory Committee is as follows:
James M. Lazorchak, Chairman
Environmental Monitoring Systems Laboratory - Cincinnati
William Peltier, Subcommittee Chairman,
Environmental Services Division, Region 4
Peter Nolan, Environmental Services Division, Region 1
Steve Ward, Environmental Services Division, Region 2
Ronald Preston, Environmental Services Division, Region 3
Charles Steiner, Environmental Services Division, Region 5
Evan Hornig, Environmental Services Division, Region 6
Terry Hollister, Environmental Services Division, Region 6
Michael Tucker, Environmental Services Division, Region 7
Loys Parrish, Environmental Services Division, Region 8
Peter Husby, Environmental Services Division, Region 9
Joseph Cummins, Environmental Services Division, Region 10
Gretchen Hayslip, Water Monitoring and Analysis Section, Region 10
Bruce Binkley, National Enforcement Investigations Center, Denver
Wesley Kinney, Environmental Monitoring Systems Laboratory - Las Vegas
George Morrison, Environmental Research Laboratory - Narragansett
Douglas Middaugh, Environmental Research Laboratory - Gulf Breeze
Teresa Norberg-King, Environmental Research Laboratory - Duluth
Donald Klemm, Environmental Monitoring Systems Laboratory Cincinnati
Philip Lewis, Environmental Monitoring Systems Laboratory - Cincinnati
Cornelius I. Weber, Environmental Monitoring Systems Laboratory -
Cincinnati
Margarete Heber, Human Health and Ecological Criteria Division, Office of
Science & Technology, Office of Water
Richard Swartz, Environmental Research Laboratory - Newport
Bruce Newton, Assessment and Watershed Protection Division, Office
of Wetlands, Oceans, and Watersheds, Office of Water
Christopher Zarba, Health and Ecological Criteria Division, Office of
Science & Technology, Office of Water
Daniel Rieder, Hazard Evaluation Division, Office of Pesticide Programs
Jerry Smrchek, Health and Environmental Review Division, Office of
Toxic Substances
IV
-------
PREFACE (CONTINUED)
Gail Hansen, Office of Solid Waste
Royal Nadeau, Emergency Response Team, Edison, New Jersey
James M. Lazorchak, Ph.D.
Chief, Bioassessment and Ecotoxicology Branch
Ecological Monitoring Research Division
Environmental Monitoring Systems Laboratory
P i nr» -i nr\ a+ i
Cincinnati
-------
ABSTRACT
This manual describes six short-term (one hour to nine days) estuarine and
marine methods for measuring the chronic toxicity of effluents and receiving
waters to five species: the sheepshead minnow, Cyprinodon variegatus; the
inland silverside, Mem'dia beryllina-, the mysid, Mysidopsis bahia; the sea
urchin, Arbacia punctulata; and the red macroalga, Champia parvula. The
methods include single and multiple concentration static renewal and static
nonrenewal toxicity tests for effluents and receiving waters. Also included
are guidelines on laboratory safety; quality assurance; facilities and
equipment; dilution water; effluent and receiving water sample collection,
preservation, shipping, and holding; test conditions; toxicity test data
analysis; report preparation; and organism culturing, holding, and handling.
Examples of computer input and output for Dunnett's Procedure, Probit
Analysis, and the Linear Interpolation Method are provided in the Appendices.
Source code listings of computer programs for Dunnett's Procedure and Probit
Analysis are also provided in the Appendices.
-------
CONTENTS
Foreword iii
Preface iv
Abstract vi
Figures x
Tables xv
Acknowledgments xxvi
Section Number Page
1. Introduction 1
2. Short-Term Methods for Estimating Chronic Toxicity 4
Introduction 4
Types of Tests 7
Static Tests 8
Flow-through Tests 8
Advantages and Disadvantages of Toxicity Test Types .... 8
3. Health and Safety 10
General Precautions 10
Safety Equipment 10
General Laboratory and Field Operations 10
Disease Prevention 11
Safety Manuals 11
Waste Disposal 11
4. Quality Assurance 12
Introduction 12
Facilities, Equipment, and Test Chambers 12
Test Organisms 13
Laboratory Water Used for Culturing and
Test Dilution Water 13
Effluent and Receiving Water Sampling
and Handling 13
Test Conditions 13
Quality of Test Organisms 14
Food Quality 14
Acceptability of Chronic Toxicity Test Results 15
Analytical Methods 15
Calibration and Standardization 15
Replication and Test Sensitivity 16
Variability in Toxicity Test Results 16
Test Precision 16
Demonstrating Acceptable Laboratory Performance 18
Documenting Ongoing Laboratory Performance 18
Reference Toxicants 19
Record Keeping 19
Video Tapes of USEPA Culture and Toxicity
Test Methods 20
Supplemental Reports for Training Video tapes 20
5. Facilities, Equipment, and Supplies 22
General Requirements 22
vii
-------
CONTENTS (CONTINUED)
Test Chambers 23
Cleaning Test Chambers and Laboratory Apparatus 23
Apparatus and Equipment for Culturing and
Toxicity Tests 24
Reagents and Consumable Materials 24
Test Organisms 24
Supplies 24
6. Test Organisms 26
Test Species 26
Sources of Test Organisms 27
Life Stage 28
Laboratory Culturing 28
Holding and Handling of Test Organisms 28
Transportation to the Test Site 29
Test Organism Disposal 30
7. Dilution Water 31
Types of Dilution Water 31
Standard, Synthetic Dilution Water 31
Use of Receiving Water as Dilution Water 32
Use of Tap Water as Dilution Water 35
Dilution Water Holding 36
8. Effluent and Receiving Water Sampling, Sample Handling,
and Sample Preparation for Toxicity Tests 37
Effluent Sampling 37
Effluent Sample Types 37
Effluent Sampling Recommendations 38
Receiving Water Sampling 39
Effluent and Receiving Water Sample Handling,
Preservation, and Shipping 40
Sample Receiving 41
Persistence of Effluent Toxicity during Sample
Shipment and Holding 41
Preparation of Effluent and Receiving Water Samples
for Toxicity Tests 41
Preliminary Toxicity Range-finding Tests 45
Multi-concentration (Definitive) Effluent
Toxicity Tests 45
Receiving Water Tests 46
9. Chronic Toxicity Test Endpoints and Data Analysis 47
Endpoints 47
Relationship between Endpoints determined by
Hypothesis Testing and Point Estimation Techniques. ... 48
Precision 50
Data Analysis 50
Choice of Analysis 52
Hypothesis Tests 53
Point Estimation Techniques 55
10. Report Preparation 59
Introduction 59
viii
-------
CONTENTS (CONTINUED)
Plant Operations 59
Sources of Effluent, Receiving Water, and Dilution Water . . 59
Test Methods 60
Test Organisms 60
Quality Assurance 60
Results 60
Conclusions and Recommendations 61
11. Sheepshead Minnow, Cyprinodon variegatus, Larval Survival
and Growth Test 62
12. Sheepshead Minnow, Cyprinodon variegatus, Embryo-larval
Survival and Teratogenicity Test 123
13. Inland Silverside, Mem'dia beryllina, Larval Survival
and Growth Test 159
14. Mysid, Mysidopsis bahia, Survival, Growth, and
Fecundity Test 216
15. Sea Urchin, Arbacia punctulata, Fertilization Test 296
16. Red Macroalga, Champia parvula, Reproduction Test 336
Selected References 375
Appendices 395
A. Independence, Randomization, and Outliers 397
B. Validating Normality and Homogeneity of Variance
Assumptions 404
C. Dunnett's Procedure 416
D. T-test with Bonferroni's Adjustment 463
E. Steel's Many-one Rank Test 468
F. Wilcoxon Rank Sum Test 473
G. Probit Analysis 479
H. Single Concentration Toxicity Test - Comparison
of Control with 100% Effluent or Receiving Water 497
I. Linear Interpolation Method 501
-------
FIGURES
SECTION 1-10
Number Page
1. Control (cusum) charts 21
2. Flowchart for statistical analysis of test data 58
SECTION 11
Number Page
1. Embryonic development of sheepshead minnow, Cyprinodon
van'egatus 76
2. Flowchart for statistical analysis of sheepshead
minnow, Cyprinodon variegatus, larval survival data 87
3. Plot of mean survival proportion data in Table 5 89
4. Plot of adjusted Probits and predicted regression line
from USEPA Probit Program 96
5. Flowchart for statistical analysis of sheepshead
minnow, Cyprinodon van'egatus, larval growth data 98
6. Plot of mean weight data from sheepshead minnow, Cyprinodon
van'egatus, larval survival and growth test 99
7. Plot of raw data, observed means and smoothed means,
for the sheepshead minnow, Cyprinodon van'egatus,
growth data from Tables 4 and 21 108
8. BOOTSTRP program output for the IC25 109
9. BOOTSTRP program output for the IC50 110
10. Data forms for sheepshead minnow, Cyprinodon van'egatus,
larval survival and growth test. Daily record of larval
survival and test conditions 119
11. Data forms for sheepshead minnow, Cyprinodon van'egatus,
larval survival and growth test. Dry weights of larvae 121
12. Data forms for sheepshead minnow, Cyprinodon van'egatus,
larval survival and growth test. Summary of test results 122
-------
FIGURES (CONTINUED)
SECTION 12
Number Page
1. Embryonic development of sheepshead minnow, Cyprinodon
variegatus 139
2. Flowchart for statistical analysis of sheepshead
minnow, Cyprinodon variegatus, embryo-larval data 146
3. Plot of sheepshead minnow, Cyprinodon variegatus, total
mortality data from the embryo-larval test 147
4. Plot of adjusted Probits and predicted regression
line 154
5. Data form for sheepshead minnow, Cyprinodon variegatus,
embryo-larval survival/teratogenicity test. Daily record of
embryo-larval survival/terata and test conditions 157
SECTION 13
Number Page
1. Glass chamber with sump area 162
2. Inland silverside, Mem'dia beryllina 172
3. Flowchart for statistical analysis of inland silverside,
Mem'da beryllina, survival data 182
4. Plot of mean survival proportion of inland silverside,
Mem'dia beryllina, larvae 184
5. Plot of adjusted Probits and predicted regression
line 193
6. Flowchart for statistical analysis of inland silverside,
Mem'dia beryllina, growth data 195
7. Plot of mean weights of inland silverside, Mem'dia beryllina,
larvae at each treatment 196
8. Plot of the raw data, the observed means and the smoothed
means, from Tables 13 and 20 205
9. BOOTSTRP program output for the IC25 206
10. BOOTSTRP program output for the IC50 207
xi
-------
FIGURES (CONTINUED)
SECTION 13 (CONTINUED)
Number
11. Data forms for inland silverside, Mem'dia beryllina,
larval survival and growth test. Daily record of larval
survival and test conditions 212
12. Data forms for inland silverside, Mem'dia beryllina,
larval survival and growth test. Dry weights of larvae 214
13. Data forms for inland silverside, Mem'dia beryllina, larval
survival and growth test. Summary of test results 215
SECTION 14
Number Page
1. Apparatus (brood chamber) for collection of
juvenile mysids, Mysidopsis bahia 230
2. Mature female mysid, Mysidopsis bahia, with eggs in oviducts . . . 234
3. Mature female mysid, Mysidopsis bahia, with eggs in oviducts
and developing embryos in the brood sac 235
4. Mature male mysid, Mysidopsis bahia 236
5. Immature mysid, Mysidopsis bahia, (A) lateral view,
(B) dorsal view 237
6. Flowchart for analysis of mysid, Mysidopsis bahia,
survival data 244
7. Plot of mean survival of mysids, Mysidopsis bahia,
at each treatment level 245
8. Plot of adjusted Probits and predicted regression
line 254
9. Flowchart for statistical analysis of mysid, Mysidopsis bahia,
growth data 255
10. Plot of mean values for mysid, Mysidopsis bahia, growth 257
11. Plot of raw data, observed means and smoothed means,
for the mysid, Mysidopsis bahia, growth data from
Tables 14 and 21 267
xii
-------
FIGURES (CONTINUED)
SECTION 14 (CONTINUED)
Number Page
12. BOOTSTRP program output for the IC25 268
13. BOOTSTRP program output for the IC50 269
14. Flowchart for statistical analysis of mysid, Mysidopsis
bahia, fecundity test 271
15. Proportion of female mysids, Mysidopsis bahia, with eggs 273
16. A plot of the mean proportion of females mysids,
Mysidopsis bahia, with eggs 283
17. BOOTSTRP program output for the IC25 284
18. BOOTSTRP program output for the IC50 285
19. Data sheet for water quality measurements 291
20. Data sheet for survival and fecundity data 292
21. Data sheet for dry weight measurements 294
SECTION 15
Number Page
1. Flowchart for statistical analysis of sea urchin,
Arbacia punctulata, data 313
2. Plot of mean percent of fertilized sea urchin,
Arbacia punctulata, eggs 314
3. BOOTSTRP program output for the IC25 324
4. BOOTSTRP program output for the IC50 325
5. Data sheet for (1) fertilization test using sea urchin,
Arbacia punctulata 333
6. Data sheet (2) for fertilization test using sea urchin,
Arbacia punctulata 334
7. Data sheet (3) for fertilization test using sea urchin,
Arbacia punctulata 335
xm
-------
FIGURES (CONTINUED)
SECTION 16
Number Page
1. Lite history of the red macroalga, Champia parvula 340
2. Apex of branch of female plant, showing sterile
hairs and reproductive hairs (trichogynes) 342
3. A portion of the male thallus showing spermatial sori 342
4. A magnified portion of a spermatial sorus 343
5. Apex of a branch on a mature female plant that was
exposed to spermatia from a male plant 343
6. A mature cystocarp 350
7. Comparison of a very young branch and an immature
cystocarp 350
8. An aborted cystocarp 350
9. Flowchart for statistical analysis of the red macroalga,
Champia parvula, data 356
10. Plot of mean number of cystocarps per plant 357
11. BOOTSTRP program output for the IC25 366
12. BOOTSTRP program output for the IC50 367
13. Data sheet for the red macroalga, Champia parvula,
sexual reproduction test. Receiving water summary sheet 373
14. Data sheet for the red macroalga, Champia parvula,
sexual reproduction test. Cystocarp data sheet 374
xiv
-------
TABLES
SECTION 1-10
Number Page
1. National Interlaboratory study of chronic toxicity test
precision, 1991: Summary of Responses Using Two Reference
Toxicants 17
2. Commercial sources of brine shrimp, Artemia, cysts 25
3. Preparation of GP2 artificial seawater using reagent
grade chemicals 33
4. Oxygen solubility (mg/L) in water at equilibrium with
air at 760 mm Hg 43
5. Percent un-ionized NH3 in aqueous ammonia solution:
temperature 15-26°C and pH 6.0-8.9 44
SECTION 11
Number Page
1. Reagent grade chemicals used in the preparation of GP2
artificial seawater for the sheepshead minnow, Cyprinodon
variegatus, toxicity test 68
2. Preparation of test solutions at a salinity of 20 °/oo,
using 20 °/oo salinity dilution water prepared from
natural seawater, hypersaline brine, or artificial
sea salts 82
3. Summary of test conditions and test acceptability criteria
for sheepshead minnow, Cyprinodon variegatus, larval survival
and growth test with effluent and receiving waters 84
4. Summary of survival and growth data for sheepshead minnow,
Cyprinodon variegatus, larvae exposed to an effluent for
seven days 86
5. Sheepshead minnow, Cyprinodon variegatus, survival data 88
6. Centered observations for Shapiro-Milk's example 88
7. Ordered centered observations for the Shapiro-Milk's
example 90
8. Coefficients and differences for Shapiro-Milk's example 91
xv
-------
TABLES (CONTINUED)
SECTION 11 (CONTINUED)
Number
9. Assigning ranks to the control and 6.25% effluent
concentration for Steel's Many-one rank test 92
10. Table of ranks 92
11. Rank sums 93
12. Data for Probit Analysis 94
13. Output for USEPA Probit Analysis program, version 1.4 95
14. Sheepshead minnow, Cypn'nodon variegatus, growth data 97
15. Centered observations for Shapiro-Milk's example 97
16. Ordered centered observations for Shapiro-Milk's example 100
17. Coefficients and differences for Shapiro-Milk's example 101
18. ANOVA table 103
19. ANOVA table for Dunnett's Procedure example 104
20. Calculated T-values 105
21. Sheepshead minnow, Cypn'nodon variegatus, mean growth
response after smoothing 106
22. Single-laboratory precision of the Sheepshead minnow,
Cypn'nodon variegatus, larval survival and growth test
performed in FORTY FATHOMS" artificial seawater, using
larvae from fish maintained and spawned in FORTY FATHOMS"
artificial seawater, and copper (CU) sulfate as a reference
toxicant 112
23. Single-laboratory precision of the Sheepshead minnow,
Cypn'nodon variegatus, larval survival and growth test
performed in FORTY FATHOMS" artificial seawater, using
larvae from fish maintained and spawned in FORTY FATHOMS"
artificial seawater, and sodium dodecyl sulfate (SDS) as
a reference toxicant 113
24. Single-laboratory precision of the sheepshead minnow,
Cypn'nodon variegatus, larval survival and growth test
performed in natural seawater, using larvae from fish
xvi
-------
TABLES (CONTINUED)
SECTION 11 (CONTINUED)
Number Page
maintained and spawned in natural seawater, and copper (CU)
sulfate as a reference toxicant 114
25. Single-laboratory precision of the sheepshead minnow,
Cyprinodon variegatus, larval survival and growth test
performed in natural seawater, using larvae from fish
maintained and spawned in natural seawater, and sodium
dodecyl sulfate (SDS) as a reference toxicant 115
26. Single-laboratory precision of the sheepshead minnow,
Cyprinodon variegatus, larval survival and growth test
performed in FORTY FATHOMS" artificial seawater, using
larvae from fish maintained and spawned in FORTY FATHOMSR
artificial seawater, and hexavalent chromium as a reference
toxicant 116
27. Comparison of Larval survival (LC50) and growth (IC50)
values for the sheepshead minnow, Cyprinodon variegatus,
exposed to sodium dodecyl sulfate (SDS) and copper (CU)
sulfate in GP2 artificial seawater medium or natural seawater. . . Ii7
28. Data from an interlaborataory study of the sheepshead minnow,
Cyprinodon variegatus, larval survival and growth test, using
an industrial effluent as a reference toxicant 118
SECTION 12
Number Page
1. Preparation of test solutions at a salinity of 20 °/oo,
using 20 °/oo natural or artificial seawater, hypersaline
brine, or artificial sea salts 129
2. Summary of test conditions and test acceptability criteria
for the sheepshead minnow, Cyprinodon variegatus, embryo-larval
survival and teratogenicity test with effluent and receiving
waters 142
3. Sheepshead minnow, Cyprinodon variegatus, embryo-larval
total mortality data 145
4. ANOVA table 145
5. ANOVA table for Dunnett's Procedure example 149
xvii
-------
TABLES (CONTINUED)
SECTION 12 (CONTINUED)
Number Page
6. Calculated T-values 150
7. Data for Probit Analysis 151
8. Output for USEPA Probit Analysis program, version 1.4 153
9. Single-laboratory precision of the sheepshead minnow,
Cyprinodon variegatus, embryo-larval survival and
teratogenicity test performed in HW MARINEMIX* artificial
seawater, using embryos from fish maintained and spawned
in HW MARINEMIXR artificial seawater using copper (CU)
sulfate as a reference toxicant 155
10. Single-laboratory precision of the sheepshead minnow,
Cyprinodon van'egatus, embryo-larval survival and
teratogenicity test performed in HW MARINEMIXR artificial
seawater, using embryos from fish maintained and spawned
in HW MARINEMIXR artificial seawater using sodium dodecyl
sulfate (SDS) as a reference toxicant 156
SECTION 13
Number Page
1. Reagent grade chemicals used in the preparation of GP2
artificial seawater for the inland silverside, Mem'dia
beryllina, toxicity test 165
2. Preparation of 3 L saline water from deionized water and a
hypersaline brine of 100 °/oo needed for test solutions at
20 °/oo salinity 168
3. Summary of test conditions and test acceptability criteria for
the inland silverside, Mem'dia beryllina, larval survival and
growth test with effluent and receiving waters 180
4. Inland silverside, Mem'dia beryllina, larval survival data .... 183
5. Centered observations for Shapiro-Wilk's example 183
6. Ordered centered observations for Sharpiro-Wilk's example 185
7. Coefficients and differences for Shapiro-Wilk's example 186
8. ANOVA table 188
xviii
-------
TABLES (CONTINUED)
SECTION 13 (CONTINUED)
Number Page
9. ANOVA table for Dunnett's Procedure example 189
10. Calculated T-values 190
11. Data for Probit Analysis 191
12. Output for USEPA probit analysis program, version 1.4 192
13. Inland silverside, Henidia beryllina, growth data 194
14. Centered observations for Shapiro-Milk's example 197
15. Ordered centered observations for Shapiro-Wilk's example 198
16. Coefficients and differences for Shapiro-Wilk's example 198
17. ANOVA table 200
18. ANOVA table for Dunnett's Procedure example 201
19. Calculated T-values 202
20. Inland silverside, Mem'dia beryllina, mean growth response
after smoothing 203
21. Single-laboratory precision of the inland silverside,
Mem'dia beryllina, survival and growth test performed
in natural seawater, using larvae from fish maintained
and spawned in natural seawater, and copper (CU) as a
reference toxicant 209
22. Single-laboratory precision of the inland silverside,
Mem'dia beryllina, survival and growth test performed
in natural seawater, using larvae from fish maintained
and spawned in natural seawater, and sodium dodecyl sulfate
as a reference toxicant 210
23. Comparison of the single-laboratory precision of the inland
silverside, Mem'dia beryllina, survival and growth IC50 values
exposed to the reference toxicants, sodium dodecyl sulfate
(SDS) or copper (CU) sulfate, in GP2 artificial
seawater medium or natural seawater 211
xix
-------
TABLES (CONTINUED)
SECTION 14
Number Page
1. Reagent grade chemicals used in the preparation of GP2
artificial seawater for the mysid, Mysidopsis bahia, toxicity
test ............................... 221
2. Quantities of effluent, deionized water, and hypersaline
brine (100 °/oo) needed to prepare 1800 ml volumes of test
solution with a salinity of 20 °/°° ...............
3. Summary of test conditions and test acceptability criteria
for the mysid, Mysidopsis bahia, seven day survival, growth,
and fecundity test with effluent and receiving waters ....... 239
4. Data for Mysidopsis bahia 7-day survival, growth, and
fecundity test ............... . .......... 241
5. Mysid, Mysidopsis bahia, survival data .............. 243
6. Centered observations for Shapiro-Milk's example ......... 246
7. Ordered centered observations for Shapiro-Milk's example ..... 247
8. Coefficients and differences for Shapiro-Wilk's example ...... 248
9. Assigning ranks to the control and 50% concentration level
for Steel"' s Many-one Rank Test .................. 250
10. Table of ranks .......................... 251
11. Rank sums ............................. 251
12. Data for Probit Analysis ..................... 252
13. Output for USEPA Probit Analysis program, version 1.4 ....... 253
14. Mysid, Mysidopsis bahia, growth data ............... 256
15. Centered observations for Shapiro-Wilk's example ......... 258
16. Ordered centered observations for Shapiro-Wilk's example ..... 259
17. Coefficients and differences for Shapiro-Wilk's example ...... 260
18. ANOVA table ............................ 262
19. ANOVA table for Dunnett's Procedure example ............ 263
xx
-------
TABLES (CONTINUED)
SECTION 14 (CONTINUED)
Number Page
20. Calculated T-values 264
21. Mysid, Mysidopsis bahia, mean growth response after smoothing. . . 265
22. Mysid, Mysidopsis bahia, fecundity data: Percent females
with eggs 272
23. Centered observations for Shapiro-Wilk's example 274
24. Ordered centered observations for Shapiro-Wilk's example 275
25. Coefficients and differences for Shapiro-Wilk's example 276
26. ANOVA table 278
27. ANOVA table for Dunnett's Procedure example 279
28. Calculated T-values 280
29. Mysid, Mysidopsis bahia, mean proportion of females with eggs. . . 281
30. Single-laboratory precision of the mysid, Mysidopsis bahia,
survival, growth, and fecundity test performed in natural
seawater, using juveniles from mysids cultured and spawned
in natural seawater and copper (CU) sulfate as a reference
toxicant 287
31. Single-laboratory precision of the mysid, Mysidopsis bahia,
survival, growth, and fecundity test performed in natural
seawater, using juveniles from mysids cultured and spawned
in natural seawater and sodium dodecyl sulfate (SDS) as a
reference toxicant 288
32. Comparison of survival (LC50), growth and fecundity (IC50)
results from 7-day tests with the mysid, Mysidopsis bahia,
using natural seawater and artificial seawater (GP2) as
dilution water and sodium dodecyl sulfate (SDS) as a
reference toxicant 289
33. Comparison of survival (LC50), growth and fecundity (IC50)
results from 7-day tests with the mysid, Mysidopsis bahia,
using, natural seawater and artificial seawater (GP2) as
dilution water and copper (CU) sulfate as a reference
toxicant 289
xxi
-------
TABLES (CONTINUED)
SECTION 14 (CONTINUED)
Number Pa9e
34. Control results from 7-day survival, growth, and fecundity
tests with the mysid, Mysidopsis bahia, using natural
seawater and artificial seawater (GP2) as a dilution
water 290
SECTION 15
Number Page
1. Preparation of test solutions at a salinity of 30 °/oo using
natural seawater, hypersaline brine, or artificial sea salts . . . 302
2. Reagent grade chemicals used in the preparation of GP2
artificial seawater for the sea urchin, Arbacia punctulata
toxicity test 303
3. Summary of test conditions and test acceptability criteria
for sea urchin, Arbacia punctulata, fertilization test with
effluent and receiving waters 310
4. Data from sea urchin, Arbacia punctulata, fertilization test . . . 312
5. Sea urchin, Arbacia punctulata, fertilization data 312
6. Centered observations for Shapiro-Wilk's example 315
7- Ordered centered observations for Shapiro-Wilk's example 315
8. Coefficients and differences for Shapiro-Wilk's example 316
9. ANOVA table 318
10. ANOVA table for Dunnett's Procedure example 319
11. Calculated T-values 320
12. Sea urchin, Arbacia punctulata, mean proportion of
fertilized eggs 322
13. Single-laboratory precision of the sea urchin,
Arbacia punctulata, fertilization test performed
in FORTY FATHOMSR artificial seawater, using gametes
from adults maintained in FORTY FATHOMS" artificial
seawater, or obtained directly form natural sources,
xxi i
-------
TABLES (CONTINUED)
SECTION 15 (CONTINUED)
Number Page
and copper (CU) sulfate and sodium dodecyl sulfate (SDS)
as reference toxicants 327
14. Single-laboratory precision of the sea urchin,
Arbacia punctulata, fertilization test performed
in FORTY FATHOMS" artificial seawater, using gametes
from adults maintained in FORTY FATHOMS" artificial
seawater, or obtained directly form natural sources,
and sodium dodecyl sulfate (SDS) as a reference toxicant 328
15. Single-laboratory precision of the sea urchin,
Arbacia punctulata, fertilization test performed
in natural seawater, using gametes from adults
maintained in natural seawater and copper (CU)
sulfate as a reference toxicant 329
16. Single-laboratory precision of the sea urchin,
Arbacia punctulata, fertilization test performed
in natural seawater, using gametes from adults
maintained in natural seawater and sodium dodecyl
sulfate (SDS) as a reference toxicant 330
17. Single-laboratory precision of the sea urchin,
Arbacia punctulata, fertilization test performed
in GP2, using gametes from adults maintained in
GP2 artificial seawater and copper (CU) sulfate and
sodium dodecyl sulfate (SDS) as reference toxicants 331
18. Single-laboratory precision of the sea urchin,
Arbacia punctulata, fertilization test performed
in natural seawater, using gametes from adults
maintained in natural seawater and copper (CU)
sulfate and sodium dodecyl sulfate (SDS) as
reference toxicants 332
SECTION 16
Number Page
1. Nutrients to be added to natural seawater and to
artificial seawater (GP2) described in Table 2 346
2. Reagent grade chemicals used in the preparation of GP2
artificial seawater for use in conjunction with natural
seawater for the red macroalga, Champia parvula, culturing
and toxicity testing 347
xxiii
-------
TABLES (CONTINUED)
SECTION 16 (CONTINUED)
Number Pa9e
3. Summary of test conditions and test acceptability criteria
for the red macroalga, Champia parvula, sexual reproduction
test with effluent and receiving waters 351
4. Data from the red macroalga, Champia parvula, effluent
toxicity test. Cystocarp counts for individual plants and
mean count per test chamber for each effluent concentration. . . . 354
5. The red macroalga, Champia parvula, sexual reproduction data . . . 355
6. Centered observations for Shapiro-Milk's example 355
7. Ordered centered observations for Shapiro-Wilk's example 358
8. Coefficients and differences for Shapiro-Wilk's example 359
9. ANOVA table 361
10. ANOVA table for Dunnett's Procedure example 362
11. Calculated T-values 363
12. Red macroalga, Champia parvula, mean number of cystocarps 364
13. Single-laboratory precision of the red macroalga,
Champia parvula, reproduction test performed in a
50/50 mixture of natural seawater and GP2 artificial
seawater, using gametes from adults cultured in natural
seawater. The reference toxicant used was copper (CU) sulfate . . 369
14. Single-laboratory precision of the red macroalga,
Champia parvula, reproduction test performed in a
50/50 mixture of natural seawater and GP2 artificial
seawater, using gametes from adults cultured in natural
seawater. The reference toxicant used was sodium dodecyl
sulfate (SDS) 370
15. Single-laboratory precision of the red macroalga,
Champia parvula, reproduction test in natural seawater
(30 /oo salinity). The reference toxicant used was copper
(CU) sulfate 371
XXIV
-------
TABLES (CONTINUED)
SECTION 16 (CONTINUED)
Number Page
16. Single-laboratory precision of the red macroalga,
Champia parvula, reproduction test in natural seawater
(30 /oo salinity). The reference toxicant used was
sodium dodecyl sulfate (SDS) 372
xxv
-------
ACKNOWLEDGMENTS
The principal authors of this document are Donald J. Klemm, Cornelius
I. Weber, Philip A. Lewis, James M. Lazorchak, Florence Fulk, and Timothy W.
Neiheisel, Environmental Monitoring Systems Laboratory Cincinnati; George E.
Morrison* Environmental Research Laboratory, Narragansett, Rhode Island;
Dennis M. McMullen, Technology Application Incorporated, Cincinnati, and Cathy
Poore, Computer Sciences Corporation, Cincinnati. Contributors to specific
sections of this manual are listed below.
1. Sections 1-10; General Guidelines
Margarete Heber, OST, Office of Water
Donald J. Klemm, EMSL - Cincinnati
James M. Lazorchak, EMSL - Cincinnati
George Morrison, ERL Narragansett
William H. Peltier, ESD, Region 4
2. Sections 11-16; Toxicity Test Methods
Margarete Heber, OST, Office of Water
Donald J. Klemm, EMSL Cincinnati
George Morrison, ERL Narragansett
William H. Peltier, ESD, Region 4
Quentin H. Pickering, EMSL Cincinnati
3. Data Analysis (Section 9, 11-16, and Appendices)
Florence Fulk, EMSL Cincinnati
Laura Gast, Computer Sciences Corporation (CSC)
Cathy Poore, Computer Sciences Corporation (CSC)
Review comments from the following persons are gratefully
acknowledged:
Pam Comeleo, Science Application International Corporation, Environmental
Research Laboratory, U.S. Environmental Protection Agency, Narragansett,
Rhode Island
Randy Comeleo, Science Application International Corporation, Environmental
Research Laboratory, U.S. Environmental Protection Agency, Narragansett,
Rhode Island
Philip A. Crocker, Water Quality Management Branch, U.S. Environmental
Protection Agency, Dallas, Texas
Dan Fisher, John Hopkins University, Queenstown, Maryland
Terry A. Hollister, Environmental Services Division, Houston Branch, U.S.
Environmental Protection Agency, Houston, Texas
Glen Modica, Science Application International Corporation, Environmental
Research Laboratory, U.S. Environmental Protection Agency, Narragansett,
Rhode Island
Michael G. Morton, Permits Branch, U.S. Environmental Protection Agency,
Dallas, Texas
xxvi
-------
ACKNOWLEDGMENTS (CONTINUED)
Peter Nolan, Environmental Services Division, U.S. Environmental Protection
Agency, Lexington, Massachusetts
Mark Tagliabue, Science Application International Corporation, Environmental
Research Laboratory, U.S. Environmental Protection Agency, Narragansett,
Rhode Island
Glenn Thursby, Environmental Research Laboratory, U.S. Environmental
Protection Agency, Narragansett, Rhode Island
Jerry Smrchek, Office of Pesticides and Toxic Substances, Environmental
Effects Branch, Health and Environmental Review Division, U.S.
Environmental Protection Agency, Washington, D.C.
Amy Wagner, Laboratory Support Section, U.S. Environmental Protection
Agency, San Francisco, California
Audrey Weber, Virginia Water Control Board, Glen Allen, Virginia
Very useful public comments on the first edition of the marine and
estuarine toxicity test methods (EPA/600/4-87/028) were received in response
to the proposed rule, published in the Federal Register, December 4, 1989
[FR 54(231):50216-50224], regarding the USEPA's intent to include short-term
chronic toxicity tests in Table IA, 40 CFR Part 136. These comments were
carefully considered in preparation of the second edition of the marine and
estuarine toxicity test methods manual (EPA/600/4-91/021), and are included in
the Public Docket for the rulemaking, located at room 2904, USEPA
Headquarters, Washington, D.C.
Materials in this manual were taken in part from the following sources:
Methods for Acute Toxicity Tests with Fish, Macroinvertebrates, and
Amphibians, Environmental Research Laboratory, U.S. Environmental Protection
Agency, Duluth, Minnesota, EPA-660/3-75-009 (USEPA, 1975); Handbook for
Analytical Quality Control in Water and Wastewater Laboratories, Environmental
Monitoring and Support Laboratory Cincinnati, U.S. Environmental Protection
Agency, Cincinnati, Ohio, EPA-600/4-79/019 (USEPA, 1979a); Interim NPDES
Compliance Biomonitoring Inspection Manual, Enforcement Division, Office of
Water Enforcement, U.S. Environmental- Protection Agency, Washington, D.C.
(USEPA, 1979c); Methods for Measuring the Acute Toxicity of Effluents to
Freshwater and Marine Organisms, Environmental Monitoring and Support
Laboratory - Cincinnati, U.S. Environmental Protection Agency, Cincinnati,
Ohio, EPA-600/4-85/013 (USEPA, 1985d); Short-term Methods for Estimating the
Chronic Toxicity of Effluents and Receiving Waters to Freshwater Organisms,
Environmental Monitoring and Support Laboratory - Cincinnati, U.S.
Environmental Protection Agency, Cincinnati, Ohio, EPA-600/4-85/014 (USEPA,
1985e); Schimmel, S.C., ed., Users Guide to the Conduct and Interpretation of
Complex Effluent Toxicity Tests at Estuarine/Marine Sites, Environmental
Research Laboratory - Narragansett (ERL-N), U.S. Environmental Protection
Agency, Narragansett, Rhode Island (USEPA, 1987d); NPDES compliance inspection
manual, Office of Water Enforcement and Permits (EN-338), U.S. Environmental
Protection Agency, Washington, D.C. (USEPA, 1988a); Short-term Methods for
Estimating the Chronic Toxicity of Effluents and Receiving Waters to Marine
xxvi i
-------
ACKNOWLEDGMENTS (CONTINUED)
and Estuarine Organisms, Environmental Monitoring and Support Laboratory
Cincinnati, U.S. Environmental Protection Agency, Cincinnati, Ohio, EPA-600/4-
87/028 (USEPA, 1988a); Short-term Methods for Estimating the Chronic Toxicity
of Effluents and Receiving Waters to Freshwater Organisms, Environmental
Monitoring Systems Laboratory Cincinnati, U.S. Environmental Protection
Agency, Cincinnati, Ohio, EPA/600/4-89/001 (USEPA, 1989b); Methods for
Measuring the Acute Toxicity of Effluents to Freshwater and Marine Organisms,
Environmental Monitoring Systems Laboratory Cincinnati, U.S. Environmental
Protection Agency, Cincinnati, Ohio, EPA/600/4-90/027 (USEPA, 1991c); Short-
term Methods for Estimating the Chronic Toxicity of Effluents and Receiving
Waters to Freshwater Organisms, Environmental Monitoring Systems Laboratory -
Cincinnati, U.S. Environmental Protection Agency, Cincinnati, Ohio, EPA/600/4-
91/022 (USEPA, 1992); and the Technical support document for water quality-
based toxic control, Office of Water Enforcement and Permits and Office of
Water Regulations and Standards, U.S. Environmental Protection Agency,
Washington, D.C., EPA/505/2-90-001 (USEPA, 1991a).
Five of the six methods in the manual were adapted from methods developed
at the Environmental Research Laboratory Narragansett. Individuals
responsible for specific methods are as follows: Melissa Hughes, Margarete
Heber, and Walter Berry, Science Applications International Corporation
(SAIC), and Steven Schimmel, USEPA, developed the sheepshead minnow larval
survival and growth test; Margarete Heber and Melissa Hughes, SAIC, Steven
Schimmel, USEPA, and David Bengtson, University of Rhode Island, developed the
inland silverside, Henidia beryllina, larval survival and growth test; Suzanne
Lussier, USEPA, Anne Kuhn and John Sewall, SAIC, developed the mysid,
Hysidopsia bahia, survival, growth, and fecundity test; Diane Nacci and
Raymond Walsh, SAIC, and Eugene Jackim, USEPA, developed the sea urchin,
Arbacia punctulata, fertilization test; and Glen Thursby, University of Rhode
Island, and Richard Steele, USEPA, developed the red macroalga, Champia
parvula, reproduction test.
Terry Hoi lister, U.S. Environmental Protection Agency, Region 6, Houston,
Texas, adapted the sheepshead minnow, Cyprinodon van'egatus, embryo-larval
survival and teratogenicity test from the fathead minnow, Pimephales promelas,
embryo-larval test in EPA/660/4-85/014 (USEPA, 1985e).
Debbie Hall, Bioassessment and Ecotoxicology Branch, and Mary Sullivan,
Quality Assurance Research Division, provided valuable secretarial assistance,
and Betty Thomas, Technical Information Manager, EMSL-Cincinnati, provided an
editorial review.
xxvm
-------
SECTION 1
INTRODUCTION
1.1 This manual describes chronic toxicity tests for use in the National
Pollutant Discharge Elimination System (NPDES) Permits Program to identify
effluents and receiving waters containing toxic materials in chronically toxic
concentrations. The methods included in this manual are referenced in Table
IA, 40 CFR Part 136 regulations and, therefore, constitute approved methods
for chronic toxicity tests. They are also suitable for determining the
toxicity of specific compounds contained in discharges. The tests may be
conducted in a central laboratory or on-site, by the regulatory agency or the
permittee.
1.2 The data are used for NPDES permits development and to determine
compliance with permit toxicity limits. Data can also be used to predict
potential acute and chronic toxicity in the receiving water, based on the
LC50, NOEC, IC25, or IC50 (see Section 9, Chronic Toxicity Test Endpoints And
Data Analysis) and appropriate dilution, application, and persistence factors.
The tests are performed as a part of self-monitoring permit requirements,
compliance biomonitoring inspections, toxics sampling inspections, and special
investigations. Data from chronic toxicity tests performed as part of permit
requirements are evaluated during compliance evaluation inspections and
performance audit inspections.
1.3 Modifications of these tests are also used in toxicity reduction
evaluations and toxicity identification evaluations to identify the toxic
components of an effluent, to aid in the development and implementation of
toxicity reduction plans, and to compare and control the effectiveness of
various treatment technologies for a given type of industry, irrespective of
the receiving water (USEPA, 1988c; USEPA, 1988d; USEPA, 1989c; USEPA, 1989d;
USEPA, 1991a; USEPA, 1991d; USEPA, 1991e).
1.4 This methods manual serves as a companion to the acute toxicity test
methods for freshwater and marine organisms (USEPA, 1991c), the short-term
chronic toxicity test methods for freshwater organisms (USEPA, 1992), and the
manual for evaluation of laboratories performing aquatic toxicity tests
(1991b).
1.5 Guidance for the implementation of toxicity tests in the NPDES program is
provided in the Technical Support Document for Water Quality-Based Toxics
Control (USEPA, 1991a).
1.6 These marine and estuarine short-term toxicity tests are similiar to
those developed for the freshwater organisms to evaluate the toxicity of
effluents discharged to estuarine and coastal marine waters under the NPDES
permit program. Methods are presented in this manual for five species from
four phylogenetic groups. Five of the six methods were developed and
extensively field tested by Environmental Research Laboratory - Narragansett
(ERL-N). The methods vary in duration from one hour and 20 minutes to nine
days.
1
-------
1.7 The five species for which toxicity test methods are provided are: the
sheepshead minnow, Cyprinodon variegatus; the inland silverside, Menidia
beryllina; the mysid, Mysidopsis bahia; the sea urchin, Arbacia punctulata;
and the red macroalga, Champia parvula.
1.7.1 The tests included in this document are based on the following methods:
1. "Guidance manual for conducting complex effluent and receiving
water larval fish growth/survival studies with the sheepshead
minnow, Cyprinodon variegatus," by Melissa M. Hughes, Margarete A.
Heber, Steven C. Schimmel and Walter J. Berry, 1987, Contribution
No. X104, Environmental Research Laboratory, U.S. Environmental
Protection Agency, Narragansett, Rhode Island (USEPA, 1987b).
2. "Guidance manual for rapid chronic toxicity test on effluents and
receiving waters with larval inland silversides, Menidia
beryllina," by Margarete A. Heber, Melissa M. Hughes, Steven C.
Schimmel, and David Bengtson, 1987, Contribution No. 792,
Environmental Research Laboratory, U.S. Environmental Protection
Agency, Narragansett, Rhode Island (USEPA, 1987c).
3. "Guidance manual for conducting seven-day, mysid survival/growth/
reproduction study using the estuarine mysid, Mysidopsis bahia,"
by Suzanne M. Lussier, Anne Kuhn, and John Sewall, 1987,
Contribution No. X106, Environmental Research Laboratory, U.S.
Environmental Protection Agency, Narragansett, Rhode Island
(USEPA, 1987f).
4. "Guidance manual for conducting sperm cell tests with the sea
urchin, Arbacia punctulata, for use in testing complex effluents,"
by Diane E. Nacci, Raymond Walsh, and Eugene Jackim, 1987,
Contribution No. X105, Environmental Research Laboratory, U.S.
Environmental Protection Agency, Narragansett, Rhode Island
(USEPA, 1987g).
5. "Guidance manual for conducting sexual reproduction tests with the
marine macroalga, Champia parvula, for use in testing complex
effluents," by Glenn B. Thursby and Richard L. Steele, 1987,
Contribution No. X103, Environmental Research Laboratory, U.S.
Environmental Protection Agency, Narragansett, Rhode Island
(USEPA, 1987d).
6. A nine-day, sheepshead minnow, Cyprinodon variegatus,
static-renewal, embryo-larval survival and teratogenicity test,
developed by Terry Hoi lister, USEPA, Region 6, Houston, Texas.
1.7.2 Four of the methods incorporate the chronic endpoints of growth or
reproduction (or both) in addition to lethality. The sheepshead minnow 9-day
embryo-larval survival and teratogenicity test incorporates teratogenic
effects in addition to lethality. The sea urchin sperm cell test uses
fertilization as an endpoint and has the advantage of an extremely short
exposure period (1 h and 20 min).
-------
1.8 The validity of the marine/estuarine methods in predicting adverse
ecological impacts of toxic discharges was demonstrated in field studies
(USEPA, 1986d).
1.9 These methods are restricted to use by or under the supervision of
analysts experienced in the use or conduct of aquatic toxicity testing and the
interpretation of data from aquatic toxicity testing. Each analyst must
demonstrate the ability to generate acceptable test results with these methods
using the procedures described in this methods manual.
1.10 The manual was prepared in the established EMSL-Cincinnati format
(USEPA, 1983).
-------
SECTION 2
SHORT-TERH METHODS FOR ESTIMATING CHRONIC TOXICITY
2.1 INTRODUCTION
2.1.1 The objective of aquatic toxicity tests with effluents or pure
compounds is to estimate the "safe" or "no-effect" concentration of these
substances, which is defined as the concentration which will permit normal
propagation of fish and other aquatic life in the receiving waters. The
endpoints that have been considered in tests to determine the adverse effects
of toxicants include death and survival, decreased reproduction and growth,
locomotor activity, gill ventilation rate, heart rate, blood chemistry,
histopathology, enzyme activity, olfactory function, and terata. Since it is
not feasible to detect and/or measure all of these (and other possible)
effects of toxic substances on a routine basis, observations in toxicity tests
generally have been limited to only a few effects, such as mortality, growth,
and reproduction.
2.1.2 Acute lethality is an obvious and easily observed effect which accounts
for its wide use in the early period of evaluation of the toxicity of pure
compounds and complex effluents. The results of these tests were usually
expressed as the concentration lethal to 50% of the test organisms (LC50) over
relatively short exposure periods (one-to-four days).
2.1.3 As exposure periods of acute tests were lengthened, the LC50 and lethal
threshold concentration were observed to decline for many compounds. By
lengthening the tests to include one or more complete life cycles and
observing the more subtle effects of the toxicants, such as a reduction in
growth and reproduction, more accurate, direct, estimates of the threshold or
safe concentration of the toxicant could be obtained. However, laboratory
life cycle tests may not accurately estimate the "safe" concentration of
toxicants because they are conducted with a limited number of species under
highly controlled, steady state conditions, and the results do not include the
effects of the stresses to which the organisms would ordinarily be exposed in
the natural environment.
2.1.4 An early published account of a full life cycle, fish toxicity test was
that of Mount and Stephan (1967). In this study, fathead minnows, Pimephales
promelas, were exposed to a graded series of pesticide concentrations
throughout their life cycle, and the effects of the toxicant on survival,
growth, and reproduction were measured and evaluated. This work was soon
followed by full life cycle tests using other toxicants and fish species.
2.1.5 McKim (1977) evaluated the data from 56 full life cycle tests, 32 of
which used the fathead minnow, Pimephales promelas, and concluded that the
embryo-larval and early juvenile life stages were the most sensitive stages.
He proposed the use of partial life cycle toxicity tests with the early life
stages (ELS) of fish to establish water quality criteria.
2.1.6 Macek and Sleight (1977) found that exposure of critical life stages of
-------
fish to toxicants provides estimates of chronically safe concentrations
remarkably similar to those derived from full life cycle toxicity tests. They
reported that "for a great majority of toxicants, the concentration which will
not be acutely toxic to the most sensitive life stages is the chronically safe
concentration for fish, and that the most sensitive life stages are the
embryos and fry." Critical life stage exposure was considered to be exposure
of the embryos during most, preferably all, of the embryogenic (incubation)
period, and exposure of the fry for 30 days post hatch for warm water fish
with embryogenic periods ranging from one-to-fourteen days, and for 60 days
post hatch for fish with longer embryogenic periods. They concluded that in
the majority of cases, the maximum acceptable toxicant concentration (MATC)
could be estimated from the results of exposure of the embryos during
incubation, and the larvae for 30 days post hatch.
2.1.7 Because of the high cost of full life cycle fish toxicity tests and the
emerging concensus that the ELS test data usually would be adequate for
estimating chronically safe concentrations, there was a rapid shift by aquatic
toxicologists to 30- to 90-day ELS toxicity tests for estimating chronically
safe concentrations in the late 1970s. In 1980, USEPA adopted the policy that
ELS test data could be used in establishing water quality criteria if data
from full life cycle tests were not available (USEPA, 1980a).
2.1.8 Published reports of the results of ELS tests indicate that the
relative sensitivity of growth and survival as endpoints may be species
dependent, toxicant dependent, or both. Ward and Parrish (1980) examined the
literature on ELS tests that used embryos and juveniles of the sheepshead
minnow, Cyprinodon variegatus, and found that growth was not a statistically
sensitive indicator of toxicity in 16 of 18 tests. They suggested that the
ELS tests be shortened to 14 days posthatch and that growth be eliminated as
an indicator of toxic effects.
2.1.9 In a review of the literature on 173 fish full life cycle and ELS tests
performed to determine the chronically safe concentrations of a wide variety
of toxicants, such as metals, pesticides, organics, inorganics, detergents,
and complex effluents, Woltering (1984) found that at the lowest effect
concentration, significant reductions were observed in fry survival in 57%,
fry growth in 36%, and egg hatchability in 19% of the tests. He also found
that fry survival and growth were very often equally sensitive, and concluded
that the growth response could be deleted from routine application of the ELS
tests. The net result would be a significant reduction in the duration and
cost of screening tests with no appreciable impact on estimating MATCs for
chemical hazard assessments. Benoit et al. (1982), however, found larval
growth to be the most significant measure of effect and survival to be equally
or less sensitive than growth in early life stage tests with four organic
chemicals.
2.1.10 Efforts to further reduce the length of partial life cycle toxicity
tests for fish without compromising their predictive value have resulted in
the development of an eight-day, embryo-larval survival and teratogenicity
test for fish and other aquatic vertebrates (USEPA, 1981e; Birge et al.,
1985), and a seven-day larval survival and growth test (Norberg and Mount,
1985).
-------
2.1.11 The similarity of estimates of chronically safe concentrations of
toxicants derived from short-term, embryo-larval survival and teratogenicity
tests to those derived from full life cycle tests has been demonstrated by
Birge et al. (1981), Birge and Cassidy (1983), and Birge et al. (1985).
2.1.12 Use of a seven-day, fathead minnow, Pimephales promelas, larval
survival and growth test was first proposed by Norberg and Mount at the 1983
annual meeting of the Society for Environmental Toxicology and Chemistry
(Norberg and Mount, 1983). This test was subsequently used by Mount and
associates in field demonstrations at Lima, Ohio (USEPA, 1984b), and at many
other locations. Growth was frequently found to be more sensitive than
survival in determining the effects of complex effluents.
2.1.13 Norberg and Mount (1985) performed three single toxicant fathead
minnow larval growth tests with zinc, copper, and DURSBANR, using dilution
water from Lake Superior. The results were comparable to, and had confidence
intervals that overlapped with, chronic values reported in the literature for
both ELS and full life cycle tests.
2.1.14 USEPA (1987b) and USEPA (1987c) adapted the fathead minnow larval
growth and survival test for use with the sheepshead minnow and the inland
silverside, respectively. When daily renewal 7-day sheepshead minnow larval
growth and survival tests and 28-day ELS tests were performed with industrial
and municipal effluents, growth was more sensitive than survival in seven out
of 12 larval growth and survival tests, equally sensitive in four tests, and
less sensitive in only one test. In four cases, the ELS test may have been
three to 10 times more sensitive to effluents than the larval growth and
survival test. In tests using copper, the No Observable Effect Concentrations
(NOECs) were the same for both types of test, and growth was the most
sensitive endpoint for both. In a four laboratory comparison, six of seven
tests produced identical NOECs for survival and growth (USEPA, 1987d). Data
indicate that the inland silverside is at least equally sensitive or more
sensitive to effluents and single compounds than the sheepshead minnow, and
can be tested over a wider salinity range, 5-30 °/oo (USEPA, 1987d).
2.1.15 Lussier et al. (1985) and USEPA (1987f) determined that survival and
growth are often as sensitive as reproduction in 28-day life cycle tests with
the mysid, Mysidopsis ba'nia.
2.1.16 Nacci and Jackim (1985) and USEPA (1987g) compared the results from
the sea urchin fertilization test, using organic compounds, with results from
acute toxicity tests using the freshwater organisms, fathead minnows,
Pimphales promelas, and Daphnia magna. The test was also compared to acute
toxicity tests using Atlantic silverside, Mem'dia mem'dia, and the mysid,
Mysidopsis bahia, and five metals. For six of the eight organic compounds,
the results of the fertilization test and the acute toxicity test correlated
well (r = 0.85). However, the results of the fertilization test with the
five metals did not correlate well with the results from the acute tests.
2.1.17 USEPA (1987e) evaluated two industrial effluents containing heavy
metals, five industrial effluents containing organic chemicals (including dyes
and pesticides), and 15 domestic wastewaters using the two-day red macroalga,
-------
Champia parvula, sexual reproduction test. Nine single compounds were used to
compare the effects on sexual reproduction using a two-week exposure and a
two-day exposure. For six of the nine compounds tested, the chronic values
were the same for both tests.
2.1.18 The use of short-term toxicity tests in the NPDES Program is
especially attractive because they provide a more direct estimate of the safe
concentrations of effluents in receiving waters than was provided by acute
toxicity tests, at an only slightly increased level of effort, compared to the
fish full life cycle chronic and 28-day ELS tests and the 28-day mysid life
cycle test.
2.2 TYPES OF TESTS
2.2.1 The selection of the test type will depend on the NPDES permit
requirements, the objectives of the test, the available resources, the
requirements of the test organisms, and effluent characteristics such as
fluctuations in effluent toxicity.
2.2.2 Effluent chronic toxicity is generally measured using a multi-
concentration, or definitive test, consisting of a control and a minimum of
five effluent concentrations. The tests are designed to provide dose-response
information, expressed as the percent effluent concentration that affects the
hatchability, gross morphological abnormalities, survival, growth, and/or
reproduction within the prescribed period of time (one hour and 20 minutes to
nine days). The results of the tests are expressed in terms of either the
highest concentration that has no statistically significant observed effect on
those responses when compared to the controls or the estimated concentration
that causes a specified percent reduction in responses verses the controls.
2.2.3 Use of pass/fail tests consisting of a single effluent concentration
(e.g., the receiving water concentration or RWC) and a control is not
recommended. If the NPDES permit has a whole effluent toxicity limit for
acute toxicity at the RWC, it is prudent to use that permit limit as the
midpoint of a series of five effluent concentrations. This will ensure that
there is sufficient information on the dose-response relationship. For
example, the effluent concentrations utilized in a test may be:
(1) 100% effluent, (2) (RWC + 100)/2, (3) RWC, (4) RWC/2, and (5) RWC/4. More
specifically, if the RWC = 50%, the effluent concentrations used in the
toxicity test would be 100%, 75%, 50%, 25%, and 12.5%.
2.2.4 Receiving (ambient) water toxicity tests commonly employ two
treatments, a control and the undiluted receiving water, but may also consist
of a series of receiving water dilutions.
2.2.5 A negative result from a chronic toxicity test does not preclude the
presence of toxicity. Also, because of the potential temporal variability in
the toxicity of effluents, a negative test result with a particular sample
does not preclude the possibility that samples collected at some other time
might exhibit chronic toxicity.
2.2.6 The frequency with which chronic toxicity tests are conducted under a
7
-------
given NPDES permit is determined by the regulatory agency on the basis of
factors such as the variability and degree of toxicity of the waste,
production schedules, and process changes.
2.2.7 Tests may be static (static nonrenewal or static renewal), or flow-
through if method is applicable.
2.3 STATIC TESTS
2.3.1 Static nonrenewal tests The test organisms are exposed to the same
test solution for the duration of the test.
2.3.2 Static renewal tests - The test organisms are exposed to a fresh
solution of the same concentration of sample every 24 h or other prescribed
interval, either by transferring the test organisms from one test chamber to
another, or by replacing all or a portion of solution in the test chambers.
2.4 FLOW-THROUGH TESTS
2.4.1 Two types of flow-through tests are in common use: (1) sample is pumped
continuously from the sampling point directly to the dilutor system; and (2)
grab or composite samples are collected periodically, placed in a tank
adjacent to the test laboratory, and pumped continuously from the tank to the
dilutor system. Because of the large volume (often 400 L/day) of effluent
normally required for flow-through tests, it is generally considered too
costly and impractical to conduct these tests off site at a central
laboratory.
2.5 ADVANTAGES AND DISADVANTAGES OF TOXICITY TEST TYPES
2.5.1 Static nonrenewal, short-term toxicity tests:
Advantages:
1. Simple and inexpensive
2. Very cost effective in determining compliance with permit
conditions.
3. Limited resources (space, manpower, equipment) required; would
permit staff to perform many more tests in the same amount of
time.
4. Smaller volume of effluent required than for static renewal or
flow-through tests.
Disadvantages:
1. Dissolved oxygen (DO) depletion may result from high chemical
oxygen demand (COD), biological oxygen demand (BOD), or metabolic
wastes.
2. Possible loss of toxicants through volatilization and/or
adsorption to the exposure vessels.
3. Generally less sensitive than static renewal or flow-through
tests, because the toxic substances may degrade or be adsorbed,
8
-------
thereby reducing the apparent toxicity. Also, there is less
chance of detecting slugs of toxic wastes, or other temporal
variations in waste properties.
2.5.2 Static renewal, short-term toxicity tests:
Advantages:
1. Reduced possibility of DO depletion from high COD and/or
BOD, or ill effects from metabolic wastes from organisms in the
test solutions.
2. Reduced possibility of loss of toxicants through volatilization
and/or adsorption to the exposure vessels.
3. Test organisms that rapidly deplete energy reserves are fed when
the test solutions are renewed, and are maintained in a healthier
state.
Disadvantages:
1. Require greater volume of effluent than nonrenewal tests.
2. Generally less sensitive than flow-through tests, because the
toxic substances may degrade or be adsorbed, thereby reducing the
apparent toxicity. Also, there is less chance of detecting slugs
of toxic wastes, or other temporal variations in waste properties.
2.5.3 Flow-through tests (if method applicable):
Advantages:
1. Provide a more representative evaluation of the acute toxicity of
the source, especially if sample is pumped continuously directly
from the source and its toxicity varies with time.
2. DO concentrations are more easily maintained in the test chambers.
3. A higher loading factor (biomass) may be used.
4. The possibility of loss of toxicant due to volatilization,
adsorption, degradation, and uptake is reduced.
Disadvantages:
1. Large volumes of sample and dilution water are required.
2. Test equipment is more complex and expensive, and requires more
maintenance and attention.
3. More space is required to conduct tests.
4. Because of the resources required, it would be very difficult to
perform multiple or overlapping sequential tests.
-------
SECTION 3
HEALTH AND SAFETY1
3.1 GENERAL PRECAUTIONS
3.1.1 Each laboratory should develop and maintain an effective health and
safety program, requiring an ongoing commitment by the laboratory management
and includes: (1) a safety officer with the responsibility and authority to
develop and maintain a safety program; (2) the preparation of a formal,
written, health and safety plan, which is provided to the laboratory staff;
(3) an ongoing training program on laboratory safety; and (4) regularly
scheduled, documented, safety inspections.
3.1.2 Collection and use of effluents in toxicity tests may involve
significant risks to personal safety and health. Personnel collecting
effluent samples and conducting toxicity tests should take all safety
precautions necessary for the prevention of bodily injury and illness which
might result from ingestion or invasion of infectious agents, inhalation or
absorption of corrosive or toxic substances through skin contact, and
asphyxiation due to a lack of oxygen or the presence of noxious gases.
3.1.3 Prior to sample collection and laboratory work, personnel should
determine that all necessary safety equipment and materials have been obtained
and are in good condition.
3.2 SAFETY EQUIPMENT
3.2.1 Personal Safety Gear
3.2.1.1 Personnel must use safety equipment, as required, such as rubber
aprons, laboratory coats, respirators, gloves, safety glasses, hard hats, and
safety shoes. Plastic netting on glass beakers, flasks and other glassware
minimizes breakage and subsequent shattering of the glass.
3.2.2 Laboratory Safety Equipment
3.2.2.1 Each laboratory (including mobile laboratories) should be provided
with safety equipment such as first aid kits, fire extinguishers, fire
blankets, emergency showers, chemical spill clean-up kits, and eye fountains.
3.2.2.2 Mobile laboratories should be equipped with a telephone to enable
personnel to summon help in case of emergency.
3.3 GENERAL LABORATORY AND FIELD OPERATIONS
3.3.1 Work with effluents should be performed in compliance with accepted
1Adapted from USEPA (1989b) and USEPA (1991c).
10
-------
rules pertaining to the handling of hazardous materials (see safety manuals
listed in Subsection 3.5). It is recommended that personnel collecting
samples and performing toxicity tests should not work alone.
3.3.2 Because the chemical composition of effluents is usually only poorly
known, they should be considered as potential health hazards, and exposure to
them should be minimized. Fume and canopy hoods over the toxicity test areas
must be used whenever possible.
3.3.3 It is advisable to cleanse exposed parts of the body immediately after
collecting effluent samples.
3.3.4 All containers should be adequately labeled to indicate their contents.
3.3.5 Staff should be familiar with safety guidelines on Material Safety Data
Sheets for reagents and other chemicals purchased from suppliers.
Incompatible materials should not be stored together. Good housekeeping
contributes to safety and reliable results.
3.3.6 Strong acids and volatile organic solvents employed in glassware
cleaning must be used in a fume hood or under an exhaust canopy over the work
area.
3.3.7 Electrical equipment or extension cords not bearing the approval of
Underwriter Laboratories must not be used. Ground-fault interrupters must be
installed in all "wet" laboratories where electrical equipment is used.
3.3.8 Mobile laboratories should be properly grounded to protect against
electrical shock.
3.4 DISEASE PREVENTION
3.4.1 Personnel handling samples which are known or suspected to contain
human wastes should be immunized against tetanus, typhoid fever, polio, and
hepatitis B.
3.5 SAFETY MANUALS
3.5.1 For further guidance on safe practices when collecting effluent samples
and conducting toxicity tests, check with the permittee and consult general
safety manuals, including USEPA (1986a), and Walters and Jameson (1984).
3.6 WASTE DISPOSAL
3.6.1 Wastes generated during toxicity testing must be properly handled and
disposed of in an appropriate manner. Each testing facility will have its own
waste disposal requirements based on local, state and Federal rules and
regulations. It is extremely important that these rules and regulations be
known, understood, and complied with by all persons responsible for, or
otherwise involved in, performing toxicity testing activities. Local fire
officials should be notified of any potentially hazardous conditions.
11
-------
SECTION 4
QUALITY ASSURANCE1
4.1 INTRODUCTION
4.1.1 Development and maintenance of a toxicity test laboratory quality
assurance (QA) program (USEPA, 1991b) requires an ongoing commitment by
laboratory management. Each toxicity test laboratory should (1) appoint a
quality assurance officer with the responsibility and authority to develop and
maintain a QA program, (2) prepare a quality assurance plan with stated data
quality objectives (DQOs), (3) prepare written descriptions of laboratory
standard operating procedures (SOPs) for culturing, toxicity testing,
instrument calibration, sample chain-of-custody procedures, laboratory sample
tracking system, glassware cleaning, etc., and (4) provide an adequate,
qualified technical staff for culturing and toxicity testing the organisms,
and suitable space and equipment to assure reliable data.
4.1.2 QA practices for toxicity testing laboratories must address all
activities that affect the quality of the final effluent toxicity data, such
as: (1) effluent sampling and handling; (2) the source and condition of the
test organisms; (3) condition of equipment; (4) test conditions; (5)
instrument calibration; (6) replication; (7) use of reference toxicants; (8)
record keeping; and (9) data evaluation.
4.1.3 Quality control practices, on the other hand, consist of the more
focused, routine, day-to-day activities carried out within the scope of the
overall QA program. For more detailed discussion of quality assurance and
general guidance on good laboratory practices and laboratory evaluation
related to toxicity testing, see FDA (1978), USEPA (1979d, 1980b, 1980c, and
1991b), DeWoskin (1984), and Taylor (1987).
4.1.4 Guidelines for the evaluation of laboratory performing toxicity tests
and laboratory evaluation criteria are found in USEPA (1991b).
4.2 FACILITIES, EQUIPMENT, AND TEST CHAMBERS
4.2.1 Separate test organism culturing and toxicity testing areas should be
provided to avoid possible loss of cultures due to cross-contamination.
Ventilation systems should be designed and operated to prevent recirculation
or leakage of air from chemical analysis laboratories or sample storage and
preparation areas into test organism culturing or toxicity testing areas, and
from toxicity test laboratories and sample preparation areas into culture
rooms.
4.2.2 Laboratory and toxicity test temperature control equipment must be
adequate to maintain recommended test water temperatures. Recommended
materials must be used in the fabrication of the test equipment which comes in
Adapted in part from USEPA (1989b), USEPA (1991b), and USEPA (1991c).
12
-------
contact with the effluent (see Section 5, Facilities, Equipment, and Supplies;
and specific toxicity test method).
4.3 TEST ORGANISMS
4.3.1 The test organisms used in the procedures described in this manual are
the sheepshead minnow, Cyprinodon variegatus; the inland silverside, Menidia
beryllina; the mysid, Mysidopsis bahia; the sea urchin, Arbacia punctulata-,
and the red macroalga, Champia parvula. The organisms used should be disease-
free and appear healthy, behave normally, feed well, and have low mortality in
cultures, during holding, and in test control. Test organisms should be
positively identified to species. Also, see Section 6, Test Organisms.
4.4 LABORATORY WATER USED FOR CULTURING AND TEST DILUTION WATER
4.4.1 The quality of water used for test organism culturing and for dilution
water used in toxicity tests is extremely important. Water for these two uses
should come from the same source. The dilution water used in effluent
toxicity tests will depend on the objectives of the study and logistical
constraints, as discussed in Section 7, Dilution Water. The dilution water
used in the toxicity tests may be natural seawater, hypersaline brine (100
°/oo) prepared from natural seawater, or artificial seawater prepared from
commercial sea salts, such as FORTY FATHOMS" or HW MARINEMIX", if recommended
in the method. GP2 synthetic seawater, made from reagent grade chemical salts
(30 °/o°) in conjunction with natural seawater, may also be used if
recommended. Hypersaline brine and artificial seawater can be used with
Champia parvula only if they are accompanied by at least 50% natural seawater.
Types of water are discussed in Section 5, Facilities, Equipment, and
Supplies. Water supplies used for culturing and test dilution water should be
analyzed quarterly, as a minimum, for toxic metals and organics. The
concentration of the metals, Al, As, Cr, Co, Cu, Fe, Pb, Ni, Zn, expressed as
total metal, should not exceed 1 jig/L each, and Cd, Hg, and Ag, expressed as
total metal, should not exceed 100 ng/L each. Total organochlorine pesticides
plus PCBs should be less than 50 ng/L (APHA, 1989). Individual pesticide
concentrations should not exceed the limits set in the USEPA National Water
Quality Guidelines.
4.5 EFFLUENT AND RECEIVING WATER SAMPLING AND HANDLING
4.5.1 Sample holding times and temperatures of effluent samples collected for
on-site and off-site testing must conform to conditions described in Section
8, Effluent and Receiving Water Sampling and Sample Handling.
4.6 TEST CONDITIONS
4.6.1 Water temperature and salinity must be maintained within the limits
specified for each test. The temperature of test solutions must be measured
by placing the thermometer or probe directly into the test solutions, or by
placing the thermometer in equivalent volumes of water in surrogate vessels
positioned at appropriate locations among the test vessels. Temperature
should be recorded continuously in at least one vessel during the duration of
each test. Test solution temperatures must be maintained within the limits
13
-------
specified for each test. DO and pH should be checked at the beginning of the
test and daily throughout the test period.
4.7 QUALITY OF TEST ORGANISMS
4.7.1 If the laboratory performs short-term chronic toxicity tests routinely
but does not have an ongoing test organism culturing program and must obtain
the test organisms from an outside source, the sensitivity of a batch of test
organisms must be determined with a reference toxicant in a short-term chronic
toxicity test performed monthly (see Subsections 4.14, 4.15, 4.16, and 4.17).
However, if the laboratory performs short-term chronic toxicity tests only
monthly or less frequently, a reference toxicant test must be performed
concurrently with each short-term chronic effluent and/or receiving water
toxicity test. The supplier should provide data with the shipment describing
the history of the sensitivity of organisms from the same source culture,
determined in monthly tests using a suitable reference toxicant. See Section
6, Test Organisms.
4.7.2 The supplier should also certify the species identification of the test
organisms, and provide the taxonomic reference (citation and page) or name(s)
of the taxonomic expert(s) consulted.
4.7.3 If the laboratory maintains breeding cultures, the sensitivity of the
offspring should be determined in a short-term chronic toxicity test performed
with a reference toxicant at least once each month (see Subsection 4.14, 4.15,
4.16, and 4.17). If preferred, this reference toxicant test may be performed
concurrently with an effluent toxicity test. However, if a given species of
test organism produced by inhouse cultures is used only monthly, or less
frequently in toxicity tests, a reference toxicant test must be performed
concurrently with each short-term chronic effluent and/or receiving water
toxicity test.
4.8 FOOD QUALITY
4.8.1 The quality of the food for fish and invertebrates is an important
factor in toxicity tests. This is especially true for the unsaturated fatty
acid content of brine shrimp nauplii, Artemia, used in mysid culturing.
Problems with the nutritional suitability of the food will be reflected in the
survival, growth, and reproduction of the test organisms in cultures and
toxicity tests. Artemia and other foods must be obtained as described in
Section 5, Facilities, Equipment, and Supplies.
4.8.2 The suitability of each new supply of food must be determined in a
side-by-side test, using two treatments with four replicates per treatment.
In this test, the response of control test organisms fed with the new food is
compared with the response of organisms fed a reference food or a previously
used, satisfactory food. For list of commercial sources of Artemia cysts, see
Table 2, Section 5, Facilities, Equipment, and Supplies.
4.8.3 New batches of food used in culturing and testing should be analyzed
for toxic organics and metals. If the concentration of total organochlorine
exceeds 0.15 jig/g wet weight, or the total concentration of total
14
-------
organochlorine pesticides plus PCBs exceeds 0.30 ng/g wet weight, or toxic
metals exceed 20 yg/g wet weight, the food should not be used (for analytical
methods, see USEPA, 1979b, 1982).
4.9 ACCEPTABILITY OF CHRONIC TOXICITY TEST RESULTS
4.9.1 The results of the sheepshead minnow, Cypn'nodon variegatus, inland
silverside, Mem'dia beryllina, or mysid, Mysidopsis bahia, tests are
acceptable if survival in the controls is 80% or greater. The sea urchin,
Arbacia punctulata, test requires control egg fertilization equal to or
exceeding 50%. However, greater than 90% fertilization may result in masking
toxic responses. The red macroalga, Champia parvula, test is acceptable if
survival is 100%, and the mean number of cystocarps per plant should equal or
exceed 10. If the sheepshead minnow, Cypn'ndon variegatus, larval survival
and growth test is begun with less-than-24-h old larvae, the mean dry weight
of the larvae in the control chambers at the end of the test should equal or
exceed 0.60 mg, if the weights are determined immediately, or 0.50 mg if the
larvae are preserved in a 4% formalin solution. If the inland silverside,
Mem'dia beryl Una, larval survival and growth test is begun with larvae seven
days old, the mean dry weight of the larvae in the control chambers at the end
of the test should equal or exceed 0.50 mg, if the weights are determined
immediately, or 0.43 mg if the larvae are preserved in a 4% formalin solution.
The mean mysid dry weight should be at least 0.20 mg. Automatic or hourly
feeding will generally provide control mysids with a dry weight of 0.30 mg.
At least 50% of the females should bear eggs at the end of the test, but mysid
fecundity is not a factor in test acceptability. However, fecundity must
equal or exceed 50% to be used as an endpoint in the test.
4.9.2 An individual test may be conditionally acceptable if temperature, DO,
and other specified conditions fall outside specifications, depending on the
degree of the departure and the objectives of the tests (see test conditions
and test acceptability criteria summaries). The acceptability of the test
will depend on the experience and professional judgment of the laboratory
investigator and the reviewing staff of the regulatory authority. Any
deviation from test specifications must be noted when reporting data from a
test.
4.10 ANALYTICAL METHODS
4.10.1 Routine chemical and physical analyses for culture and dilution water,
food, and test solutions must include established quality assurance practices
outlined in USEPA methods manuals (USEPA, 1979a; 1979b).
4.10.2 Reagent containers should be dated and catalogued when received from
the supplier, and the shelf life should not be exceeded. Also, working
solutions should be dated when prepared, and the recommended shelf life should
be observed.
4.11 CALIBRATION AND STANDARDIZATION
4.11.1 Instruments used for routine measurements of chemical and physical
parameters, such as pH, DO, temperature, conductivity, and salinity, must be
15
-------
calibrated and standardized according to instrument manufacturers procedures
as indicated in the general section on quality assurance (see USEPA Methods
150.1, 360.1, 170.1, and 120.1 in USEPA, 1979b). Calibration data are
recorded in a permanent log book.
4.11.2 Wet chemical methods used to measure hardness, alkalinity, and total
residual chlorine, must be standardized prior to use each day according to the
procedures for those specific USEPA methods (see USEPA Methods 130.2 and 310.1
in USEPA 1979b).
4.12 REPLICATION AND TEST SENSITIVITY
4.12.1 The sensitivity of the tests will depend in part on the number of
replicates per concentration, the probability level selected, and the type of
statistical analysis. If the variability remains constant, the sensitivity of
the test will increase as the number of replicates is increased. The minimum
recommended number of replicates varies with the objectives of the test and
the statistical method used for analysis of the data.
4.13 VARIABILITY IN TOXICITY TEST RESULTS
4.13.1 Factors which can affect test success and precision include: (1) the
experience and skill of the laboratory analyst; (2) test organism age,
condition, and sensitivity; (3) dilution water quality; (4) temperature
control; (5) and the quality and quantity of food provided. The results will
depend upon the species used and the strain or source of the test organisms,
and test conditions, such as temperature, DO, food, and water quality. The
repeatability or precision of toxicity tests is also a function of the number
of test organisms used at each toxicant concentration. Jensen (1972)
discussed the relationship between sample size (number of fish) and the
standard error of the test, and considered 20 fish per concentration as
optimum for Probit Analysis.
4.14 TEST PRECISION
4.14.1 The ability of the laboratory personnel to obtain consistent, precise
results must be demonstrated with reference toxicants before they attempt to
measure effluent toxicity. The single-laboratory precision of each type of
test to be used in a laboratory should be determined by performing at least
five or more tests with a reference toxicant.
4.14.2 Test precision can be estimated by using the same strain of organisms
under the same test conditions, and employing a known toxicant, such as a
reference toxicant.
4.14.3 Interlaboratory precision of chronic toxicity tests using two
reference toxicants with the mysid, Mysidopsis bahia, and the inland
silverside, Mem'dia beryllina, is listed in Table 1. Additional precision
data for each of the tests described in this manual are presented in the
sections describing the individual test methods.
4.14.4 Additional information on toxicity test precision is provided in the
16
-------
Technical Support Document for Water Quality-Based Toxic Control (see pp. 3-4,
and 11-15 in USEPA, 1991a).
4.14.5 In cases where the test data are used in the Probit Analysis (see
Section 9, Chronic Toxicity Test Endpoints and Data Analysis), precision can
be described by the mean, standard deviation, and relative standard deviation
(percent coefficient of variation, or CV) of the calculated endpoints from the
replicated tests. In cases where the test data are used in the Linear
Interpolation Method, precision can be estimated by empirical confidence
intervals derived by using the Bootstrap Method (see Section 9, Chronic
Toxicity Test Endpoints and Data Analysis). However, in cases where the
results are reported in terms of the No-Observed-Effect-Concentration (NOEC)
and Lowest-Observed-Effect-Concentration (LOEC) (see Section 9, Chronic
Toxicity Test Endpoints and Data Analysis), precision can only be described by
listing the NOEC-LOEC interval for each test. It is not possible to express
precision in terms of a commonly used statistic. However, when all tests of
the same toxicant yield the same NOEC-LOEC interval, maximum precision has
been attained. The "true" no effect concentration could fall anywhere within
the interval, NOEC ± (LOEC minus NOEC).
TABLE 1. NATIONAL INTERLABORATORY STUDY OF CHRONIC TOXICITY TEST
PRECISION. 1991: SUMMARY OF RESPONSES USING TWO REFERENCE
TOXICANTS*'2
Organism
Endpoint
No. Labs KCl(mg/L)*
SD
My si daps is
bahia
Survival, NOEC
Growth, IC25
Growth, IC50
Growth, NOEC
Fecundity, NOEC
34
26
22
32
25
NA
480
656
NA
NA
NA
3.47
3.17
NA
NA
NA
28.9
19.3
NA
NA
Organism
Endpoint
No. Labs CU(mg/L)4
SD
Me nidi a
beryl Una
Survival, NOEC
Growth, IC25
Growth, IC50
Growth, NOEC
19
13
12
17
NA
0.144
0.180
NA
NA
1.56
1.87
NA
NA
43.5
41.6
NA
1From a national study of interlaboratory precision of toxicity test data
performed in 1991 by the Environmental Monitoring Systems Laboratory -
Cincinnati, U.S. Environmental Protection Agency, Cincinnati, Ohio 45268.
Participants included federal, state, and private laboratories engaged in
NPDES permit compliance monitoring.
2Static renewal test, using 25 700 modified GP2 artificial seawater.
'Percent coefficient of variation = (standard deviation X 100)/mean.
^Expressed as mean.
17
-------
4.14.6 It should be noted here that the dilution factor selected for a test
determines the width of the NOEC-LOEC interval and the inherent maximum
precision of the test. As the absolute value of the dilution factor
decreases, the width of the NOEC-LOEC interval increases, and the inherent
maximum precision of the test decreases. When a dilution factor of 0.3 is
used, the NOEC could be considered to have a relative uncertainty as high as +
300%. With a dilution factor of 0.5, the NOEC could be considered to have a
relative variability of + 100%. As a result of the variability of different
dilution factors, USEPA recommends the use of a 0.5 dilution factor or
greater. Other factors which can affect test precision include: test
organism age, condition, and sensitivity; temperature control; and feeding.
4.15 DEMONSTRATING ACCEPTABLE LABORATORY PERFORMANCE
4.15.1 It is a laboratory's responsibility to demonstrate its ability to
obtain consistent, precise results with reference toxicants before it performs
toxicity tests with effluents for permit compliance purposes. To meet this
requirement, the intralaboratory precision, expressed as percent coefficient
of variation (CV%), of each type of test to be used in a laboratory should be
determined by performing five or more tests with different batches of test
organisms, using the same reference toxicant, at the same concentrations, with
the same test conditions (i.e., the same test duration, type of dilution
water, age of test organisms, feeding, etc.), and same data analysis methods.
A reference toxicant concentration series (0.5 or higher) should be selected
that will consistently provide partial mortalities at two or more
concentrations.
4.16 DOCUMENTING ONGOING LABORATORY PERFORMANCE
4.16.1 Satisfactory laboratory performance is demonstrated by performing at
least one acceptable test per month with a reference toxicant for each of the
toxicity test methods commonly used in the laboratory. For a given test
method, successive tests must be performed with the same reference toxicant,
at the same concentrations, in the same dilution water, using the same data
analysis methods.
4.16.2 A control chart should be prepared, using data from a minimum of five
toxicity tests, for each reference-toxicant-organism combination, and
successive toxicity endpoints (NOEC, IC25s, IC50s, LC50s) should be plotted
and examined to determine if the results are within prescribed limits (Figure
1). In this technique, a running plot is maintained for the toxicity values
(Xf) from successive tests with a given reference toxicant. The types of
control charts illustrated (USEPA, 1979a) are used to evaluate the cumulative
trend of results from a series of samples. At least five for Probit Analysis
results (such as LC50's), the mean (X) and upper and lower control limits
(+2S) are re-calculated with each successive point, until the statistics
stabilize. Precision may vary with the test species, reference toxicant, and
type of test.
4.16.3 Outliers, which are values falling outside the upper and lower control
limits, and trends of increasing or decreasing sensitivity, are readily
identified. At the P0 05 probability level, one in 20 tests would be expected
18
-------
to fall outside of the control limits by chance alone. For hypothesis testing
results, assuming that the same concentrations of reference toxicants are used
for each successive toxicity test, the NOEC from each test is entered on the
control chart. The values should fall within one concentration interval above
or below the central tendency.
4.16.4 If the toxicity value from a given test with the reference toxicant
falls well outside the expected range for the test organisms when using the
standard dilution water, the sensitivity of the organisms and the overall
credibility of the test system are suspect. In this case, the test procedure
should be examined for defects and should be repeated with a different batch
of test organisms.
4.16.5 Performance should improve with experience, and the control limits
should gradually narrow, as the statistics stabilize. However, control limits
of ± 2S, by definition, will be exceeded 5% of the time, regardless of how
well a laboratory performs. For this reason, good laboratories which develop
very narrow control limits may be penalized if a test result which falls just
outside the control limits is rejected de facto. The width of the control
limits should be considered in determining if data which exceed control limits
should be rejected.
4.17 REFERENCE TOXICANTS
4.17.1 Reference toxicants such as sodium chloride (NaCl), potassium chloride
(KC1), cadmium (CdCl2), copper (CuSOJ, sodium dodecyl sulfate (SDS), and
potassium dichromate (K2Cr207), are suitable for use in the NPDES Program and
other Agency programs requiring aquatic toxicity tests. EMSL-Cincinnati plans
to release USEPA-certified solutions of cadmium and copper for use as
reference toxicants in FY-92, through cooperative research and development
agreements with commercial suppliers, and will continue to develop additional
reference toxicants for future release. Interested parties can determine the
availability of "EPA Certified" reference toxicants by checking the EMSL-
Cincinnati electronic bulletin board, using a modem to access the following
telephone numbers: FTS 684-7610, or Commercial 513-569-7610. Standard
reference materials also can be obtained from commercial supply houses, or can
be prepared inhouse using reagent grade chemicals. The regulatory agency
should be consulted before reference toxicant(s) are selected and used.
4.18 RECORD KEEPING
4.18.1 Proper record keeping is important. A complete file must be
maintained for each individual toxicity test or group of tests on closely
related samples. This file must contain a record of the sample chain-of-
custody; a copy of the sample log sheet; the original bench sheets for the
test organism responses during the toxicity test(s); chemical analysis data on
the sample(s); detailed records of the test organisms used in the test(s),
such as species, source, age, date of receipt, and other pertinent information
relating to their history and health; information on the calibration of
equipment and instruments; test conditions employed; and results of reference
toxicant tests. Laboratory data should be recorded on a real-time basis to
prevent the loss of information or inadvertent introduction of errors into the
19
-------
record. Original data sheets should be signed and dated by the laboratory
personnel performing the tests.
4.18.2 The regulatory authority should retain records pertaining to discharge
permits. Permittees are required to retain records pertaining to permit
applications and compliance for a minimum of 3 years [40 CFR 122.41(j)(2)].
4.19 VIDEO TAPES OF USEPA CULTURE AND TOXICITY TEST METHODS
4.19.1 Ordering Information: National Audiovisual Center, National Archives
and Records Admin., Customer Services Section, 8700 Edgeworth Dr., Capitol
Heights, MD 20743-3701. Charge to VISA or MasterCard by calling toll-free
(800) 638-1300 or FAX your purchase order by calling (301) 763-6025. For
other information call (301) 763-1896.
4.19.2 Title of Videos: USEPA Culture and Toxicity Test Methods for Marine
and EstuariPvi Effluents.
4.19.2.1 Toxicity test methods for the red macroalga, Champia parvula; the
sheepshead minnow, Cypn'nodon variegatus; the inland silverside, Mem'dia
beryllina; and the sea urchin, Arbacia punctulata (Order Number A18545,
$85.00).
4.19.2.2 Mysids, Mysidopsis bahia, culture and toxicity test (Order Number
A18657, $75.00).
4.20 SUPPLEMENTAL REPORTS FOR TRAINING VIDEO TAPES
4.20.1 Ordering information: USEPA, Office of Research and Development,
Washington, D.C. 20460.
4.20.1.1 Sheepshead minnow, Cypn'nodon variegatus, and inland silverside,
Menidia beryllina, larval survival and growth toxicity tests (EPA/600/3-
90/075), 1990.
4.20.1.2 Red algae, Champia parvula, sexual reproduction (EPA/600/3-90/076),
1990.
4.20.1.3 Sperm cell test using the sea urchin, Arbacia punctulata,
(EPA/600-3-90/077), 1990.
4.21.2 Ordering information: USEPA, Office of Water (EN-336), Washington,
D.C. 20460.
4.21.2.2 Mysids, Mysidopsis bahia, survival, growth, and fecundity test
(EPA/505/8-90-006a), 1990.
20
-------
o
Lkl
O
UPPER CONTR.OL LIMIT
CENTRALTENDENCY
LOWER CONTROL LIMIT
10
20
i.
o
UJ
o
UPPER CONTROL LIMIT(X+2S)
CENTRALTENDENCY
LOWER CONTROL LIMIT (X - 2S)
I I I
05 10 15 20
TOXICITY TEST WITH REFERENCE TOXICANTS
n- 1
Where: X,- = Successive toxicity values from toxicity tests
n = Number of tests.
X = Mean toxicity value.
S = Standard deviation.
Figure 1. Control (cusum) charts: (A) hypothesis testing
results; (B) point estimates (LC, EC, or 1C).
21
-------
SECTION 5
FACILITIES, EQUIPMENT, AND SUPPLIES1
5.1 GENERAL REQUIREMENTS
5.1.1 Effluent toxicity tests may be performed in a fixed or mobile
laboratory. Facilities must include equipment for rearing and/or holding
organisms. Culturing facilities for test organisms may be desirable in fixed
laboratories which perform large numbers of tests. Temperature control can be
achieved using circulating water baths, heat exchangers, or environmental
chambers. Water used for rearing, holding, acclimating, and testing organisms
may be natural sea water or water made up from hypersaline brine derived from
natural sea water, or water made up from reagent grade chemicals (GP2) or
commercial (FORTY FATHOMS" or HW MARINEMIX") artificial sea salts when
specifically recommended in the method. Air used for aeration must be free of
oil and toxic vapors. Oil-free air pumps should be used where possible.
Particulates can be removed from the air using BALSTONR Grade BX or equivalent
filters (Balston, Inc., Lexington, Massachusetts), and oil and other organic
vapors can be removed using activated carbon filters (BALSTONR, C-l filter, or
equivalent).
5.1.2 The facilities must be well ventilated and free of fumes. Laboratory
ventilation systems should be checked to ensure that return air from chemistry
laboratories and/or sample handling areas is not circulated to test organism
culture rooms or toxicity test rooms, or that air from toxicity test rooms
does not contaminate culture areas. Sample preparation, culturing, and
toxicity testing areas should be separated to avoid cross-contamination of
cultures or toxicity test solutions with toxic fumes. Air pressure
differentials between such rooms should not result in a net flow of
potentially contaminated air to sensitive areas through open or loosely-
fitting doors. Organisms should be shielded from external disturbances.
5.1.3 Materials used for exposure chambers, tubing, etc., which come in
contact with the effluent and dilution water, should be chosen carefully.
Tempered glass and perfluorocarbon plastics (TEFLONR) should be used whenever
possible to minimize sorption and leaching of toxic substances. These
materials may be reused following decontamination. Containers made of
plastics, such as polyethylene, polypropylene, polyvinyl chloride, TYGONR,
etc., may be used as test chambers or to ship, store, and transfer effluents
and receiving waters, but they should not be reused unless absolutely
necessary, because they might carry over adsorbed toxicants from one test to
another, if reused. However, these containers may be repeatedly reused for
storing uncontaminated waters such as deionized or laboratory-prepared
dilution waters and receiving waters. Glass or disposable polystyrene
containers can be used as test chambers. The use of large (> 20 L) glass
carboys is discouraged for safety reasons.
Adapted from USEPA (1985d) and USEPA (1991c).
22
-------
5.1.4 New plastic products of a type not previously used should be tested for
toxicity before initial use by exposing the test organisms in the test system
where the material is used. Equipment (pumps, valves, etc.) which cannot be
discarded after each use because of cost, must be decontaminated according to
the cleaning procedures listed below (Subsection 5.3.2). Fiberglass, in
addition to the previously mentioned materials, can be used for holding,
acclimating, and dilution water storage tanks, and in the water delivery
system, but once contaminated with pollutants the fiberglass should not be
reused. All material should be flushed or rinsed thoroughly with the test
media before using in the test.
5.1.5 Copper, galvanized material, rubber, brass, and lead must not come in
contact with culturing, holding, acclimation, or dilution water, or with
effluent samples and test solutions. Some materials, such as several types of
neoprene rubber (commonly used for stoppers) may be toxic and should be tested
before use.
5.1.6 Silicone adhesive used to construct glass test chambers absorbs some
organochlorine and organophosphorus pesticides, which are difficult to remove.
Therefore, as little of the adhesive as possible should be in contact with
water. Extra beads of adhesive inside the containers should be removed.
5.2 TEST CHAMBERS
5.2.1 Test chamber size and shape are varied according to size of the test
organism. Requirements are specified in each toxicity test method.
5.3 CLEANING TEST CHAMBERS AND LABORATORY APPARATUS
5.3.1 New plasticware used for sample collection or organism exposure vessels
generally does not require rigorous cleaning, and it is usually sufficient to
soak the new containers overnight in seawater before use. New, disposable,
plastic test chambers may have to be rinsed with dilution water before use.
New glassware should be soaked overnight in 10% acid (see below) and also
should be soaked overnight in sea water.
5.3.2 All nondisposable sample containers, test vessels, pumps, tanks, and
other equipment that has come in contact with effluent must be washed after
use to remove surface contaminants, as described below.
1. Soak 15 minutes in tap water and scrub with detergent, or clean in
an automatic dishwasher.
2. Rinse twice with tap water.
3. Carefully rinse once with fresh dilute (10% V:V) nitric acid or
hydrochloric acid to remove scale, metals and bases. To prepare a
10% solution of acid, slowly add 10 ml of concentrated acid to 90
ml of deionized water.
4. Rinse twice with deionized water.
5. Rinse once with full-strength, pesticide-grade acetone to remove
organic compounds (use a fume hood or canopy).
6. Rinse three times with deionized water.
23
-------
5.3.3 All test chambers and equipment must be thoroughly rinsed with the
dilution water immediately prior to use in each test.
5.4 APPARATUS AND EQUIPMENT FOR CULTURING AND TOXICITY TESTS
5.4.1 Apparatus and equipment requirements for culturing and toxicity tests
are specified in each toxicity test method. Also, see USEPA (1991c).
5.4.2 Water Purification System
5.4.2.1 A good quality deionized water, providing 18 mega-ohm, laboratory
grade water, should be available in the laboratory and with sufficient
capacity for laboratory needs. Deionized water may be obtained from MILLIPORE
MILLI-Q , or equivalent system. If large quantities of high quality deionized
water are needed, it may be advisable to supply the laboratory grade water
deionizer with preconditioned water from a Culligen", Continental", or
equivalent, mixed-bed water treatment system. The recommended order of the
cartridges in a four-cartridge MILLI-Q" System is: (1) ion exchange, (2) ion
exchange, (3) carbon, and (4) organic cleanup (ORGANEX-QR), followed by a
final bacteria filter.
5.5 REAGENTS AND CONSUMABLE MATERIALS
5.5.1 Sources of Food for Culture and Toxicity Tests
5.5.1.1 All food should be tested for nutritional suitability and chemically
analyzed for organochlorine pesticides, PCBs, and toxic metals (see Section 4,
Quality Assurance).
1. Brine Shrimp, Artemia, Cysts -- A list of commercial sources is
provided in Table 2.
2. Frozen Adult Brine Shrimp, Artemia -- Available from San Francisco
Bay Brand, 8239 Enterprise Dr., Newark, California, 94560 (Phone:
415-792-7200) or available at most pet supply shops.
3. Flake Food -- TETRAMIN" and BIORILR are available at most pet
supply shops.
4. Feeding requirements and other specific foods are indicated in the
specific toxicity test method.
5.5.2 Reagents and consumable materials are specified in each toxicity test
method. Also, see Section 4, Quality Assurance.
5.6 TEST ORGANISMS
5.6.1 Test organisms are obtained from inhouse cultures or commercial
suppliers (see specific toxicity test method and Sections 4, Quality Assurance
and 6, Test Organisms).
5.7 SUPPLIES
5.7.1 See toxicity test methods (Sections 11-16) for specific supplies.
24
-------
TABLE 2. COMMERCIAL SOURCES OF BRINE SHRIMP, ARTEMIA, CYSTS1
Aquafauna Biomarine
P.O. Box 5
Hawthorne, CA 90250
Tel. (213) 973-5275
Fax. (213) 676-9387
(Great Salt Lake North Arm,
San Francisco Bay)
Argent Chemical
8702 152nd Ave. NE
Redmond, WA 98052
Tel. (800) 426-6258
Tel. (206) 855-3777
Fax. (206) 885-2112
(Platinum Label San Francisco Bay;
Gold Label - San Francisco Bay,
Brazil; Silver Label - Great
Salt Lake, Australia; Bronze
Label - China, Canada)
Bonneville Artemia International, Inc.
P.O. Box 511113
Salt Lake City, UT 84151-1113
Tel. (801) 972-4704
Fax. (801) 972-4795
Ocean Star International
P.O. Box 643
Snowville, UT 84336
Tel. (801) 872-8217
Fax (801) 872-8272
(Great Salt Lake)
Sanders Brine Shrimp Co.
3850 South 540 West
Ogden, UT 84405
Tel. (801) 393-5027
(Great Salt Lake)
Sea Critters Inc.
P.O. Box 1508
Tavernier, FL 33070
Tel. (305) 367-2672
Aquarium Products
180L Penrod Court
Glen Burnie, MD 21061
Tel. (800) 368-2507
Tel. (301) 761-2100
(Colombia)
Artemia Systems
Wiedauwkaai 79
B-9000 Ghent, Belgium
Tel. 011-32-91-534142
Fax. 011-32-91-536893
(For marine species AF grade)
[small nauplii], UL grade [large
nauplii], for freshwater species -
HI grade [small nauplii], EG grade
[large nauplii]
Golden West Artemia
411 East 100 South
Salt Lake City, UT 84111
Tel. (801) 532-1400
Fax. (801) 531-8160
Pennsylvania Pet Products
Box 191
Spring City, PA 19475
Tel. Not listed.
(Great Salt Lake)
San Francisco Bay Brand
8239 Enterprise Drive
Newark, CA 94560
Tel. (415) 792-7200
(Great Salt Lake, San Francisco
Bay)
Western Brine Shrimp
957 West South Temple
Salt Lake City, UT 84104
Tel. (801) 364-3642
Fax. (801) 534-0211
(Great Salt Lake)
1List provided by David A. Bengtson, University of Rhode Island, Narragansett
25
-------
SECTION 6
TEST ORGANISMS
6.1 TEST SPECIES
6.1.1 The species used in characterizing the chronic toxicity of effluents
and/or receiving waters will depend on the requirements of the regulatory
authority and the objectives of the test. It is essential that good quality
test organisms be readily available throughout the year from inhouse or
commercial sources to meet NPDES monitoring requirements. The organisms used
in toxicity tests must be identified to species. If there is any doubt as to
the identity of the test organisms, they should be sent to a taxonomic expert
for examination.
6.1.2 Toxicity test conditions and culture methods for the species listed in
Subsection 6.1.3 are provided in this manual (also, see USEPA, 1991c).
6.1.3 The organisms used in the short-term tests described in this manual are
the sheepshead minnow, Cyprinodon variegatus; the inland silverside, Mem'dia
beryllina; the mysid, Mysidopsis bahia; the sea urchin, Arbacia punctulata;
and the red macroalga, Champia parvula.
6.1.4 Some states have developed culturing and testing methods for indigenous
species that may be as sensitive or more sensitive than the species
recommended in Subsection 6.1.3. However, USEPA allows the use of indigenous
species only where state regulations require their use or prohibit importation
of the species in Subsection 6.1.3. Where state regulations prohibit
importation of non-native fishes or use of the recommended test species,
permission must be requested from the appropriate state agency prior to their
use. Required clearances should be obtained from state fisheries agencies
before arrangements are made for the interstate shipment of sheepshead minnows
or silversides.
6.1.5 Where states have developed culturing and testing methods for
indigenous species other than those recommended in this manual, data comparing
the sensitivity of the substitute species and one or more of the recommended
species must be obtained in side-by-side toxicity tests with reference
toxicants and/or effluents, to ensure that the species selected are at least
as sensitive as the recommended species. These data must be submitted to the
permitting authority (State or Region) if required. USEPA acknowledges that
reference toxicants prepared from pure chemicals may not always be
representative of effluents. However, because of the observed and/or
potential variability in the quality and toxicity of effluents, it is not
possible to specify a representative effluent.
6.1.6 Guidance for the selection of test organisms where the salinity of the
effluent and/or receiving water requires special consideration is provided in
the Technical Support Document for Water Quality-Based Toxics Control (USEPA,
1991a).
26
-------
1. Where the salinity of the receiving water is < l°/oo, freshwater
organisms are used regardless of the salinity of the effluent.
2. Where the salinity of the receiving water is > l°/oo, the choice of
organisms depends on state water quality standards and/or permit
requirements.
6.2 SOURCES OF TEST ORGANISMS
6.2.1 The test organisms in this manual can be cultured in the laboratory.
Culturing and handling procedures for each organism are described in the
respective test method sections. Also, see USEPA (1991c).
6.2.2 Inhouse cultures should be established wherever it is cost effective.
If inhouse cultures cannot be maintained, test organisms should be purchased
from experienced commercial suppliers (see USEPA, 1991c).
6.2.3. Starter cultures of the red macroalga, Champia parvula, are available
from the U.S. Environmental Protection Agency, Environmental Research
Laboratory - Narragansett, 27 Tarzwell Drive, Narragansett, RI 02882.
6.2.4 Sheepshead minnows, inland silversides, mysids, and sea urchins may be
purchased from commercial suppliers. However, some of these organisms (e.g.,
adult sheepshead minnows or adult inland silversides) may not always be
available from commercial suppliers and may have to be collected in the field
and brought back to the laboratory for spawning to obtain eggs and larvae.
6.2.5 If, because of their source, there is any uncertainty concerning the
identity of the organisms, it is advisable to have them examined by a
taxonomic specialist to confirm their identification. For detailed guidance
on identification, see the individual toxicity test methods.
6.2.6 Feral (Natural Occurring, Wild Caught) Organisms
6.2.6.1 The use of test organisms taken from the receiving water has strong
appeal, and would seem to be the logical approach. However, it is impractical
for the following reasons:
1. Sensitive organisms may not be present in the receiving water because
of previous exposure to the effluent or other pollutants.
2. It is often difficult to collect organisms of the required age and
quality from the receiving water.
3. Most states require collection permits, which may be difficult to
obtain. Therefore, it is usually more cost effective to culture the
organisms in the laboratory or obtain them from private, state, or
Federal sources. Fish such as sheepshead minnows and silversides, and
invertebrates such as mysids, are easily reared in the laboratory or
purchased.
4. The required QA/QC records, such as the single-laboratory precision
data, would not be available.
5. Since it is mandatory that the identity of test organisms is known to
the species level, it would be necessary to examine each organism
27
-------
caught in the wild to confirm its identity, which would usually be
impractical or, at the least, very stressful to the organisms.
6. Test organisms obtained from the wild must be observed in the
laboratory for a minimum of one week prior to use, to assure that they
are free of signs of parasitic or bacterial infections and other
adverse effects. Fish captured by electroshocking must not be used in
toxicity testing.
6.2.6.2 Guidelines for collection of natural occurring organisms are provided
in USEPA (1973, 1990a).
6.2.7 Regardless of their source, test organisms should be carefully observed
to ensure that they are free of signs of stress and disease, and in good
physical condition. Some species of test organisms, such as trout, can be
obtained from stocks certified as "disease-free."
6.3 LIFE STAGE
6.3.1 Young organisms are often more sensitive to toxicants than are adults.
For this reason, the use of early life stages, such as juvenile mysids and
larval fish, is required for all tests. There may be special cases, however,
where the limited availability of organisms will require some deviation from
the recommended life stage. In a given test, all organisms should be
approximately the same age and should be taken from the same source. Since
age may affect the results of the tests, it would enhance the value and
comparability of the data if the same species in the same life stages were
used throughout a monitoring program at a given facility.
6.4 LABORATORY CULTURING
6.4.1 Instructions for culturing and/or holding the recommended test
organisms are included in specified test methods (also, see USEPA, 1991c).
6.5 HOLDING AND HANDLING TEST ORGANISMS
6.5.1 Test organisms should not be subjected to changes of more than 3°C in
water temperature or 3 °/oo in salinity in any 12 h period.
6.5.2 Organisms should be handled as little as possible. When handling is
necessary, it should be done as gently, carefully, and quickly as possible to
minimize stress. Organisms that are dropped or touch dry surfaces or are
injured during handling must be discarded. Dipnets are best for handling
larger organisms. These nets are commercially available or can be made from
small-mesh nylon netting, silk bolting cloth, plankton netting, or similar
material. Wide-bore, smooth glass tubes (4 to 8 mm inside diameter) with
rubber bulbs or pipettors (such as a PROPIPETTE" or other pipetter) should be
used for transferring smaller organisms such as daphnids, mysids, and larval
fish.
6.5.3 Holding tanks for fish are supplied with a good quality water (see
Section 5, Facilities, Equipment, and Supplies) with a flow-through rate of at
least two tank-volumes per day. Otherwise, use a recirculation system where
28
-------
the water flows through an activated carbon or undergravel filter to remove
dissolved metabolites. Culture water can also be piped through high intensity
ultraviolet light sources for disinfection, and to photodegrade dissolved
organics.
6.5.4 Crowding should be avoided. The DO must be maintained at a minimum of
4.0 mg/L. The solubility of oxygen depends on temperature, salinity, and
altitude. Aerate if necessary.
6.5.5 Fish should be fed as much as they will eat at least once a day with
live brine shrimp nauplii, Artemia, or frozen adult brine shrimp or dry food
(frozen food should be completely thawed before use). Adult brine shrimp can
be supplemented with commercially prepared food such as TETRAMIN" or BIORIL*
flake food, or equivalent. Excess food and fecal material should be removed
from the bottom of the tanks at least twice a week by siphoning.
6.5.6 Fish should be observed carefully each day for signs of disease,
stress, physical damage, and mortality. Dead and abnormal specimens should be
removed as soon as observed. It is not uncommon to have some fish (5-10%)
mortality during the first 48 h in a holding tank because of individuals that
refuse to feed on artificial food and die of starvation.
6.5.7 A daily record of feeding, behavioral observations, and mortality
should be maintained.
6.6 TRANSPORTATION TO THE TEST SITE
6.6.1 Organisms are transported from the base or supply laboratory to a
remote test site in culture water or standard dilution water in plastic bags
or large-mouth screw-cap (500 ml) plastic bottles in styrofoam coolers.
Adequate DO is maintained by replacing the air above the water in the bags
with oxygen from a compressed gas cylinder, and sealing the bags. Another
method commonly used to maintain sufficient DO during shipment is to aerate
with an airstone which is supplied from a portable pump. The DO concentration
must not fall below 4.0 mg/L.
6.6.2 Upon arrival at the test site, organisms are transferred to receiving
water if receiving water is to be used as the test dilution water. All but a
small volume of the holding water (approximately 5%) is removed by siphoning,
and replaced slowly over a 10 to 15 min period with dilution water. If
receiving water is used as dilution water, caution must be exercised in
exposing the test organisms to it, because of the possibility that it might be
toxic. For this reason, it is recommended that only approximately 10% of the
test organisms be exposed initially to the dilution water. If this group does
not show excessive mortality or obvious signs of stress in a few hours, the
remainder of the test organisms are transferred to the dilution water.
6.6.3 A group of organisms must not be used for a test if they appear to be
unhealthy, discolored, or otherwise stressed, or if mortality appears to
exceed 10% preceding the test. If the organisms fail to meet these criteria,
the entire group must be discarded and a new group obtained. The mortality
may be due to the presence of toxicity, if receiving water is used as dilution
29
-------
water, rather than a diseased condition of the test organisms. If the
acclimation process is repeated with a new group of test organisms and
excessive mortality occurs, it is recommended that an alternative source of
dilution water be used.
6.6.4 The marine organisms can be used at all concentrations of effluent by
adjusting the salinity of the effluent to salinities specified for the
appropriate species test condition or to the salinity approximating that of
the receiving water, by adding sufficient dry ocean salts, such as FORTY
FATHOMS", or equivalent, GP2, or hypersaline brine.
6.6.5 Saline dilution water can be prepared with deionized water or a
freshwater such as well water or a suitable surface water. If dry ocean salts
are used, care must be taken to ensure that the added salts are completely
dissolved and the solution is aerated 24 h before the test organisms are
placed in the solutions. The test organisms should be acclimated in synthetic
saline water prepared with the dry salts. Caution: addition of dry ocean
salts to dilution water may result in an increase in pH. (The pH of estuarine
and coastal saline waters is normally 7.5 - 8.3).
6.6.6 All effluent concentrations and the control(s) used in a test should
have the same salinity. The change in salinity upon acclimation at the
desired test dilution should not exceed 6 °/oo. The required salinities for
culturing and toxicity tests with estuarine and marine species are listed in
the test method sections.
I
6.7 TEST ORGANISM DISPOSAL
6.7.1 When the toxicity test(s) is concluded, all test organisms (including
controls) should be humanely destroyed and disposed of in an appropriate
manner.
30
-------
SECTION 7
DILUTION WATER
7.1 TYPES OF DILUTION WATER
7.1.1 The type of dilution water used in effluent toxicity tests will depend
largely on the objectives of the study.
7.1.1.1 If the objective of the test is to estimate the chronic toxicity of
the effluent, which is a primary objective of NPDES permit-related toxicity
testing, a synthetic (standard) dilution water is used. If the test organisms
have been cultured in water which is different from the test dilution water, a
second set of controls, using culture water, should be included in the test.
7.1.1.2 If the objective of the test is to estimate the chronic toxicity of
the effluent in uncontaminated receiving water, the test may be conducted
using dilution water consisting of a single grab sample of receiving water (if
non-toxic), collected outside the influence of the outfall, or with other
uncontaminated receiving water or standard dilution water having approximately
the same salinity as the receiving water. Seasonal variations in the quality
of receiving waters may affect effluent toxicity. Therefore, the salinity of
saline receiving water samples should be determined before each use. If the
test organisms have been cultured in water which is different from the test
dilution water, a second set of controls, using culture water, should be
included in the test.
7.1.1.3 If the objective of the test is to determine the additive or
mitigating effects of the discharge on already contaminated receiving water,
the test is performed using dilution water consisting of receiving water
collected outside the influence of the outfall. A second set of controls,
using culture water, should be included in the test.
7.2 STANDARD, SYNTHETIC DILUTION WATER
7.2.1 Standard, synthetic, dilution water is prepared with deionized water
and reagent grade chemicals (GP2) or commercial sea salts (FORTY FATHOMS", HW
MARINEMIXR). The source water for the deionizer can be ground water,
receiving water, or tap water.
7.2.2 Deionized Water used to Prepare Standard, Synthetic, Dilution Water
7.2.2.1 Deionized water is obtained from a MILLIPORE MILLI-Q* or equivalent
system. It is advisable to provide a preconditioned (deionized) feed
water by using a Culligan, Continental, or equivalent system in front of the
MILLI-Q* System to extend the life of the MILLI-Q" cartridges (see Section 5,
Facilities, Equipment, and Supplies).
7.2.2.2 The recommended order of the cartridges in a four-cartridge deionizer
(i.e., MILLI-QR System or equivalent) is: (1) ion exchange, (2) ion exchange,
31
-------
(3) carbon, and (4) organic cleanup (such as ORGANEX-QR, or equivalent),
followed by a final bacteria filter.
7.2.3 Standard, Synthetic Seawater
7.2.3.1 To prepare 20 L of a standard, synthetic, reconstituted seawater
(modified GP2), using reagent grade chemicals (Table 3), with a salinity of
31°/oo, follow the instructions below. Other salinities can be prepared by
making the appropriate dilutions. Larger or smaller volumes of modified GP2
can be prepared by using proportionately larger or smaller amounts of salts
and dilution water.
1. Place 20 L of MILLI-QR or equivalent deionized water in a
properly cleaned plastic carboy.
2. Weigh reagent grade salts listed in Table 3 and add, one at a
time, to the deionized water. Stir well after adding each salt.
3. Aerate the final solution at a rate of 1 L/h for 24 h.
4. Check the pH and salinity.
7.2.3.2 Synthetic seawater can also be prepared by adding commercial sea
salts, such as FORTY FATHOMS", HW MARINEMIX", or equivalent, to deionized
water. For example, thirty-one parts per thousand (31°/oo) FORTY FATHOMS" can
be prepared by dissolving 31 g of sea salts per liter of deionized water. The
salinity of the resulting solutions should be checked with a refractometer.
7.2.4 Artificial seawater is to be used only if specified in the method.
EMSL-Cincinnati has found FORTY FATHOMS" artifical sea salts (Marine
Enterprises, Inc., 8755 Mylander Lane, Baltimore, Maryland 21204; phone:
301-321-1189) suitable for maintaining and spawning the sheepshead minnow,
Cyprinodon variegatus, and for its use in the sheepshead minnow larval
survival and growth test, suitable for maintaining and spawning the inland
silverside, Mem'dia beryllina, and for its use in the inland silverside larval
survival and growth test, suitable for culturing and maintaining mysid shrimp,
Hysidopsis bahia, and its use in the mysid shrimp survival, growth, and
fecundity test, and suitable for maintaining sea urchins, Arbacia punctulata,
and for its use in the sea urchin fertilization test. The USEPA Region 6
Houston Laboratory has successfully used HW MARINEMIX* (Hawaiian Marine
Imports Inc., P.O. Box 218687, Houston, Texas 77218, phone 713-492-7864) sea
salts to maintain and spawn sheepshead minnows, and perform the larval
survival and growth test and the embryo-larval survival and teratogenicity
test. Also, HW MARINEMIXR sea salts has been used successfully to culture and
maintain the mysid brood stock and perform the mysid survival, growth,
fecundity test. An artificial seawater formulation, GP2 (Spotte et al.,
1984), Table 3, has been used by the Environmental Research Laboratory-
Narragansett, RI for all but the embryo-larval survival and teratogenicity
test. The suitability of GP2 as a medium for culturing organisms has not been
determined.
7.3 USE OF RECEIVING WATER AS DILUTION WATER
7.3.1 If the objectives of the test require the use of uncontaminated
receiving water as dilution water, and the receiving water is uncontaminated,
32
-------
TABLE 3. PREPARATION OF GP2 ARTIFICIAL SEAWATER USING REAGENT GRADE
rUFMTTAI C1.2-3
CHEMICALS
Compound
1.
2.
3.
4.
5.
6.
7.
8.
9.
NaCl
Na2S04
KC1
KBr
Na2B407 . 10 H20
MgCl2 . 6 H20
CaCl2 . 2 H20
SrCl2 . 6 H20
NaHC03
Concentration
(9/L)
21.03
3.52
0.61
0.088
0.034
9.50
1.32
0.02
0.17
Amount (g)
Required for
20 L
420.6
70.4
12.2
1.76
0.68
190.0
26.4
0.400
3.40
Modified GP2 from Spotte et al. (1984)
2The constituent salts and concentrations were taken from
USEPA (1990b). The salinity is 30.89 g/L.
3GP2 can be diluted with deionized (DI) water to the desired test salinity.
33
-------
it may be possible to collect a sample of the receiving water close to the
outfall, but should be away from or beyond the influence of the effluent.
However, if the receiving water is contaminated, it may be necessary to
collect the sample in an area "remote" from the discharge site, matching as
closely as possible the physical and chemical characteristics of the receiving
water near the outfall.
7.3.2 The sample should be collected immediately prior to the test, but never
more than 96 h before the test begins. Except where it is used within 24 h,
or in the case where large volumes are required for flow through tests, the
sample should be chilled to 4°C during or immediately following collection,
and maintained at that temperature prior to use in the test.
7.3.3 The investigator should collect uncontaminated water having a salinity
as near as possible to the salinity of the receiving water at the discharge
site. Water should be collected at slack high tide, or within one hour after
high tide. If there is reason to suspect contamination of the water in the
estuary, it is advisable to collect uncontaminated water from an adjacent
estuary. At times it may be necessary to collect water at a location closer
to the open sea, where the salinity is relatively high. In such cases,
deionized water or uncontaminated freshwater is added to the saline water to
dilute it to the required test salinity. Where necessary, the salinity of a
surface water can be increased by the addition of artificial sea salts, such
as FORTY FATHOMS", HW MARINEMIX*, or equivalent, GP2, a natural seawater of
higher salinity, or hypersaline brine. Instructions for the preparation of
hypersaline brine by concentrating natural seawater are provided below.
7.3.4 Receiving water containing debris or indigenous organisms, that may be
confused with or attack the test organisms, should be filtered through a sieve
having 60 ^m mesh openings prior to use.
7.3.5 Hypersaline Brine
7.3.5.1 Hypersaline brine (HSB) has several advantages that make it desirable
for use in toxicity testing. It can be made from any high quality, filtered
seawater by evaporation, and can be added to deionized water to prepare
dilution water, or to effluents or surface waters to increase their salinity.
7.3.5.2 The ideal container for making HSB from natural seawater is one
that (1) has a high surface to volume ratio, (2) is made of a noncorrosive
material, and (3) is easily cleaned (fiberglass containers are ideal).
Special care should be used to prevent any toxic materials from coming in
contact with the seawater being used to generate the brine. If a heater is
immersed directly into the seawater, ensure that the heater material's do not
corrode or leach any substances that would contaminate the brine. One
successful method used is a thermostatically controlled heat exchanger made
from fiberglass. If aeration is used, use only oil-free air compressors to
prevent contamination.
7.3.5.3 Before adding seawater to the brine generator, thoroughly clean
the generator, aeration supply tube, heater, and any other materials that
34
-------
will be in direct contact with the brine. A good quality biodegradable
detergent should be used, followed by several thorough deionized water
rinses. High quality (and preferably high salinity) seawater should be
filtered to at least 10 ^m before placing into the brine generator. Water
should be collected on an incoming tide to minimize the possibility of
contamination.
7.3.5.4 The temperature of the seawater is increased slowly to 40°C.
The water should be aerated to prevent temperature stratification and to
increase water evaporation. The brine should be checked daily (depending on
the volume being generated) to ensure that the salinity does not exceed
100 /oo and that the temperature does not exceed 40°C. Additional
seawater may be added to the brine to obtain the volume of brine required.
7.3.5.5 After the required salinity is attained, the MSB should be
filtered a second time through a 1-^m filter and poured directly into
portable containers (20-L CUBITAINERS" or polycarbonate water cooler jugs are
suitable). The containers should be capped and labelled with the date the
brine was generated and its salinity. Containers of MSB should be stored in
the dark and maintained under room temperature until used.
7.3.5.6 If a source of MSB is available, test solutions can be made by
following the directions below. Thoroughly mix together the deionized water
and brine before mixing in the effluent.
7.3.5.7 Divide the salinity of the MSB by the expected test salinity to
determine the proportion of deionized water to brine. For example, if the
salinity of the brine is 100 °/°o and the test is to be conducted at
25 °/oo, 100 °/oo divided by 25 °/oo = 4.0. The proportion of brine is
1 part in 4 (one part brine to three parts deionized water).
7.3.5.8 To make 1 L of seawater at 25 °/°° salinity from a hypersaline
brine of 100 °/oo, 250 Ml of brine and 750 Ml of deionized water are
required.
7.4 USE OF TAP WATER AS DILUTION WATER
7.4.1 The use of tap water in the reconstituting of synthetic (artificial)
seawater as dilution water is discouraged unless it is dechlorinated and fully
treated. Tap water can be dechlorinated by deionization, carbon filtration,
or the use of sodium thiosulfate. Use of 3.6 mg/L (anhydrous) sodium
thiosulfate will reduce 1.0 mg chlorine/L (APHA, 1989, pp. 9-32). Following
dechlorination, total residual chlorine should not exceed 0.01 mg/L. Because
of the possible toxicity of thiosulfate to test organisms, a control lacking
thiosulfate should be included in toxicity tests utilizing thiosulfate-
dechlorinated water.
7.4.2 To be adequate for general laboratory use following dechlorination, the
tap water is passed through a deionizer and carbon filter to remove toxic
metals and organics, and to control hardness and alkalinity.
35
-------
7.5 DILUTION WATER HOLDING
7.5.1 A given batch of dilution water should not be used for more than 14
days following preparation because of the possible build up of bacterial,
fungal, or algal slime growth and the problems associated with it. The
container should be kept covered and the contents should be protected from
light.
36
-------
SECTION 8
EFFLUENT AND RECEIVING WATER SAMPLING, SAMPLE HANDLING,
AND SAMPLE PREPARATION FOR TOXICITY TESTS
8.1 EFFLUENT SAMPLING
8.1.1 The effluent sampling point should be the same as that specified in the
NPDES discharge permit (USEPA, 1988a). Conditions for exception would be:
(1) better access to a sampling point between the final treatment and the
discharge outfall; (2) if the processed waste is chlorinated prior to
discharge, it may also be desirable to take samples prior to contact with the
chlorine to determine toxicity of the unchlorinated effluent; or (3) in the
event there is a desire to evaluate the toxicity of the influent to municipal
waste treatment plants or separate wastewater streams in industrial facilities
prior to their being combined with other wastewater streams or non-contact
cooling water, additional sampling points may be chosen.
8.1.2 The decision on whether to collect grab or composite samples is based
on the objectives of the test and an understanding of the short and long-term
operations and schedules of the discharger. If the effluent quality varies
considerably with time, which can occur where holding times are short, grab
samples may seem preferable because of the ease of collection and the
potential of observing peaks (spikes) in toxicity. However, the sampling
duration of a grab sample is so short that full characterization of an
effluent over a 24-h period would require a prohibitively large number of
separate samples and tests. Collection of a 24-h composite sample, however,
may dilute toxicity spikes, and average the quality of the effluent over the
sampling period. Sampling recommendations are provided below (also see USEPA,
1991a).
8.1.3 Aeration during collection and transfer of effluents should be
minimized to reduce the loss of volatile chemicals.
8.1.4 Details of date, time, location, duration, and procedures used for
effluent sample and dilution water collection should be recorded.
8.2 EFFLUENT SAMPLE TYPES
8.2.1 The advantages and disadvantages of effluent grab and composite samples
are listed below:
8.2.1.1 Grab Samples
Advantages:
1. Easy to collect; require a minimum of equipment and on-site time.
2. Provide a measure of instantaneous toxicity. Toxicity spikes are
not marked by dilution.
37
-------
Disadvantages:
1. Samples are collected over a very short period of time and on a
relatively infrequent basis. The chances of detecting a spike in
toxicity would depend on the frequency of sampling, and the
probability of missing spikes is high.
8.2.1.2 Composite Samples:
Advantages:
1. A single effluent sample is collected over a 24-h period.
2. The sample is collected over a much longer period of time than grab
samples and contains all toxicity spikes.
Disadvantages:
1. Sampling equipment is more sophisticated and expensive, and must be
placed on-site for at least 24 h.
2. Toxicity spikes may not be detected because they are masked by
dilution with less toxic wastes.
8.3 EFFLUENT SAMPLING RECOMMENDATIONS
8.3.1 When tests are conducted on-site, test solutions can be renewed daily
with freshly collected samples.
8.3.2 When tests are conducted off-site, a minimum of three samples are
collected. If these samples are collected on Test Days 1, 3, and 5, the first
sample would be used for test initiation, and for test solution renewal on Day
2. The second sample would be used for test solution renewal on Days 3 and 4.
The third sample would be used for test solution renewal on Days 5, 6, and 7.
8.3.3 Sufficient sample must be collected to perform the required toxicity
and chemical tests. A 4-L (1-gal) CUBITAINERR will provide sufficient sample
volume for most tests.
8.3.4 The following effluent sampling methods are recommended:
8.3.4.1 Continuous Discharges
1. If the facility discharge is continuous, but the calculated retention
time of the continuously discharged effluent is less than 14 days and
the variability of the effluent toxicity is unknown, at a minimum,
four grab samples or four composite samples are collected over a 24-h
period. For example, a grab sample is taken every 6 h (total of four
samples) and each sample is used for a separate toxicity test, or four
successive 6-h composite samples are taken and each is used in a
separate test.
2. If the calculated retention time of a continuously discharged effluent
is greater than 14 days, or if it can be demonstrated that the
38
-------
wastewater does not vary more than 10% in toxicity over a 24-h period,
regardless of retention time, a single grab sample is collected for a
single toxicity test.
3. The retention time of the effluent in the wastewater treatment
facility may be estimated from calculations based on the volume of the
retention basin and rate of wastewater inflow. However, the
calculated retention time may be much greater than the actual time
because of short-circuiting in the holding basin. Where
short-circuiting is suspected, or sedimentation may have reduced
holding basin capacity, a more accurate estimate of the retention time
can be obtained by carrying out a dye study.
8.3.4.2 Intermittent Discharges
8.3.4.2.1 If the facility discharge is intermittent, a grab sample is
collected midway during each discharge period. Examples of intermittent
discharges are:
1. When the effluent is continuously discharged during a single 8-h work
shift (one sample is collected), or two successive 8-h work shifts
(two samples are collected).
2. When the facility retains the wastewater during an 8-h work shift, and
then treats and releases the wastewater as a batch discharge (one
sample is collected).
3. When the facility discharges wastewater to an estuary only during an
outgoing tide, usually during the 4 h following slack high tide (one
sample is collected).
8.3.4.3 At the end of a shift, clean up activities may result in the
discharge of a slug of toxic waste, which may require sampling and testing.
8.4 RECEIVING WATER SAMPLING
8.4.1 Logistical problems and difficulty in securing sampling equipment
generally preclude the collection of composite receiving water samples for
toxicity tests. Therefore, it is common practice to collect a single grab
sample and use it throughout the test.
8.4.2 The sampling point is determined by the objectives of the test. At
estuarine and marine sites, samples should be collected at mid-depth.
8.4.3 To determine the extent of the zone of toxicity in the receiving water
at estuarine and marine effluent sites, receiving water samples are collected
at several distances away from the discharge. The time required for the
effluent-receiving-water mixture to travel to sampling points away from the
effluent, and the rate and degree of mixing, may be difficult to ascertain.
Therefore, it may not be possible to correlate receiving water toxicity with
effluent toxicity at the discharge point unless a dye study is performed. The
toxicity of receiving water samples from five stations in the discharge plume
39
-------
can be evaluated using the same number of test vessels and test organisms as
used in one effluent toxicity test with five effluent dilutions.
8.5 EFFLUENT AND RECEIVING WATER SAMPLE HANDLING, PRESERVATION, AND SHIPPING
8.5.1 Unless the samples are used in an on-site toxicity test the day of
collection, it is recommended that they be held at 4°C until used to inhibit
microbial degradation, chemical transformations, and loss of highly volatile
toxic substances.
8.5.2 Composite samples should be chilled as they are collected. Grab
samples should be chilled immediately following collection.
8.S.3 If the effluent has been chlorinated, total residual chlorine must be
measured immediately following sample collection.
8.5.4 Sample "holding time," as defined here, begins when the last grab
sample of a series is collected, or when the composite sampling period is
completed. If the data from the samples are to be acceptable for use in the
NPDES Program, the lapsed time from collection to first use of the sample in
test initiation or test solution renewal should not exceed 36 h. The results
of tests using samples held more than 36 h may not reflect the true toxicity
of the effluent at the time of collection. The sampling schedule should be
adjusted so that the lapsed time from sample collection to shipment is held to
a minimum. However, in the event logistical constraints preclude a 36-h
holding time, permission to use samples held longer than 36 h must be obtained
from the NPDES permitting authority (see Subsection 8.7.1).
8.5.5 To minimize the loss of toxicity due to volatilization of toxic
constituents, all sample containers should be "completely" filled, leaving no
air space between the contents and the lid.
8.5.6 Samples Used in On-Site Tests
8.5.6.1 Samples collected for on-site tests should be used within 24 h.
8.5.7 Samples Shipped to Off Site Facilities
8.5.7.1 Samples collected for off site toxicity testing are to be chilled to
4°C during or immediately after collection, and shipped iced to the performing
laboratory. Sufficient ice should be placed with the sample in the shipping
container to ensure that ice will still be present when the sample arrives at
the laboratory and is unpacked.
8.5.7.2 Samples may be shipped in one or more 4-L (1-gal) CUBITAINERSR or new
plastic "milk" jugs. All sample containers should be rinsed with source water
before being filled with sample. After use with receiving water or effluents,
CUBITAINERS and plastic jugs are punctured to prevent reuse.
8.5.7.3 Several sample shipping options are available, including Express
Mail, air express, bus, and courier service. Express Mail is delivered seven
40
-------
days a week. Saturday and Sunday shipping and receiving schedules of private
carriers vary with the carrier.
8.6 SAMPLE RECEIVING
8.6.1 Upon arrival at the laboratory, samples are logged in and the
temperature is measured and recorded. If the samples are not
immediately prepared for testing, they are stored at 4°C until used.
8.6.2 Every effort must be made to initiate the test with an effluent sample
on the day of arrival in the laboratory, and the sample holding time should
not exceed 36 h unless prior arrangements have been made with the NPDES
permitting authority.
8.7 PERSISTENCE OF EFFLUENT TOXICITY DURING SAMPLE SHIPMENT AND HOLDING
8.7.1 The persistence of the toxicity of an effluent prior to its use in a
toxicity test is of interest in assessing the validity of toxicity test data,
and in determining the possible effects of allowing an extension of the
holding time. Where an extension in holding time is requested by a permittee
(see Subsection 8.5.4), information on the effects of the extension on the
toxicity of samples must be obtained by performing a multi-concentration
chronic toxicity test on effluent samples held 36 h, and comparing the results
with those obtained in a toxicity test with the same samples after they have
been held for the requested, longer period. The portion of the sample set
aside for the second test must be held under the same conditions as during
shipment and holding.
8.8 PREPARATION OF EFFLUENT AND RECEIVING MATER SAMPLES FOR TOXICITY TESTS
8.8.1 Adjust the sample salinity to the level appropriate for objectives of
the study using hypersaline brine or artificial sea salts.
8.8.2 When aliquots are removed from the sample container, the head space
above the remaining sample should be held to a minimum. Air which enters a
container upon removal of sample should be expelled by compressing the
container before reclosing, if possible (i.e., where a CUBITAINER used), or
by using an appropriate discharge valve (spigot).
8.8.3 It may be necessary to first coarse-filter samples through a NYLONR
sieve having 2- to 4-mm mesh openings to remove debris and/or break up large
floating or suspended solids. If samples contain indigenous organisms that
may attack or be confused with the test organisms, the samples must be
filtered through a sieve with 60 ^m mesh openings. Since filtering may
increase the dissolved oxygen (DO) in an effluent, the DO should be determined
prior to filtering. Low dissolved oxygen concentrations will indicate a
potential problem in performing the test. Caution: filtration may remove some
toxicity.
8.8.4 If the samples must be warmed to bring them to the prescribed test
temperature, supersaturation of the dissolved oxygen and nitrogen may become a
problem. To avoid this problem, the effluent and dilution water are checked
41
-------
with a DO probe after reaching test temperature and, if the DO is greater than
100% saturation or lower than 4.0 mg/L, based on temperature and salinity, the
solutions are aerated moderately (approximately 500 mL/min) for a few minutes,
using an airstone, until the DO is lowered to 100% saturation (Table 4) or
until the DO is within the prescribed range (>4.0 mg/L). Caution: avoid
excessive aeration.
8.8.4.1 Aeration during the test may alter the results and should be used
only as a last resort to maintain the required DO. Aeration can reduce the
apparent toxicity of the test solutions by stripping them of highly volatile
toxic substances, or increase their toxicity by altering the pH. However, the
DO in the test solution must not be permitted to fall below 4.0 mg/L.
8.8.4.2 In static tests (non-renewal or renewal) low DOs may commonly occur
in the higher concentrations of wastewater. Aeration is accomplished by
bubbling air through a pipet at the rate of 100 bubbles/min. If aeration is
necessary, all test solutions must be aerated. It is advisable to monitor the
DO closely during the first few hours of the test. Samples with a potential
DO problem generally show a downward trend in DO within 4 to 8 h after the
test is started. Unless aeration is initiated during the first 8 h of the
test, the DO may be exhausted during an unattended period, thereby
invalidating the test.
8.8.5 At a minimum, pH, conductivity or salinity, and total residual
chlorine are measured in the undiluted effluent or receiving water, and pH and
conductivity are measured in the dilution water.
8.8.5.1 It is recommended that total alkalinity and total hardness also be
measured in the undiluted effluent test water and the dilution water.
8.8.6 Total ammonia is measured in effluent and receiving water samples where
toxicity may be contributed by un-ionized ammonia (i.e., where total ammonia
>5 mg/L). The concentration (mg/L) of un-ionized (free) ammonia in a sample
is a function of temperature and pH, and is calculated using the percentage
value obtained from Table 5, under the appropriate pH and temperature, and
multiplying it by the concentration (mg/L) of total ammonia in the sample.
8.8.7 Effluents and receiving waters can be dechlorinated using 6.7 mg/L
anhydrous sodium thiosulfate to reduce 1 mg/L chlorine (APHA, Standard
Methods, 17th Edition, 1989; p. 9-32; note that the amount of thiosulfate
required to dechlorinate effluents is greater than the amount needed to
dechlorinate tap water). Since thiosulfate may contribute to sample toxicity,
a thiosulfate control should be used in the test in addition to the normal
dilution water control.
8.8.8 The DO concentration in the samples should be near saturation prior to
use. Aeration will bring the DO and other gases into equilibrium with air,
minimize oxygen demand, and stabilize the pH. However, aeration during
collection, transfer, and preparation of samples should be minimized to reduce
the loss of volatile chemicals.
42
-------
Table 4. OXYGEN SOLUBILITY (MG/L) IN WATER AT EQUILIBRIUM WITH AIR AT
760 MM HG (AFTER RICHARDS AND CORWIN, 1956)
TEMP
0
1
2
3
4
5
6
8
10
12
14
16
18
20
22
24
26
28
30
32
SALINITY (°/oo)
0
14.
13.
13.
13.
12.
12.
12.
11.
10.
10.
10.
9.
9.
8.
8.
8.
8.
7.
7.
7.
2
8
4
1
7
4
1
5
9
5
0
6
2
9
6
3
1
8
6
3
5
13.8
13.4
13.0
12.7
12.3
12.0
11.7
11.2
10.7
10.2
9.7
9.3
9.0
8.6
8.4
8.1
7.8
7.6
7.4
7.1
10
13.4
13.0
12.6
12.3
12.0
11.7
11.4
10.8
10.3
9.9
9.5
9.1
8.7
8.4
8.1
7.8
7.6
7.4
7.1
6.9
15
12.9
12.6
12.2
11.9
11.6
11.3
11.0
10.5
10.0
9.6
9.2
8.8
8.5
8.1
7.9
7.6
7.4
7.2
6.9
6.7
20
12.5
12.2
11.9
11.6
11.3
11.0
10.7
10.2
9.7
9.3
8.9
8.5
8.2
7.9
7.6
7.4
7.2
7.0
6.7
6.5
25
12.1
11.8
11.5
11.2
10.9
10.6
10.3
9.8
9.4
9.0
8.6
8.3
8.0
7.7
7.4
7.2
7.0
6.8
6.5
6.3
30
11.7
11.4
11.1
10.8
10.5
10.2
10.0
9.5
9.1
8.7
8.3
8.0
7.7
7.4
7.2
6.9
6.7
6.5
6.3
6.1
35
11.2
11.0
10.7
10.4
10.1
9.8
9.6
9.2
8.8
8.4
8.1
7.7
7.5
7.2
6.9
6.7
6.5
6.3
6.1
5.9
40
10.8
10.6
10.3
10.0
9.8
9.5
9.3
8.9
8.5
8.1
7.8
7.5
7.2
6.9
6.7
6.5
6.3
6.1
5.9
5.7
43
10.6
10.3
10.0
9.8
9.5
9.3
9.1
8.7
8.3
7.9
7.6
7.3
7.1
6.8
6.6
6.4
6.1
6.0
5.8
5.6
43
-------
TABLE 5. PERCENT UN-IONIZED NH, IN AQUEOUS AMMONIA SOLUTIONS:
TEMPERATURE 15-26°C AND pH 6.0-8.91
pH TEMPERATURE (°C)
6.0
6.1
6.2
6.3
6.4
6.5
6.6
6.7
6.8
6.9
7.0
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
7.9
8.0
8.1
8.2
8.3
8.4
8.5
8.6
8.7
8.8
8.9
15
0.0274
0.0345
0.0434
0.0546
0.0687
0.0865
0.109
0.137
0.172
0.217
0.273
0.343
0.432
0.543
0.683
0.858
1.08
1.35
1.70
2.13
2.66
3.33
4.16
5.18
6.43
7.97
9.83
12.07
14.7
17.9
16
0.0295
0.0372
0.0468
0.0589
0.0741
0.0933
0.117
0.148
0.186
0.234
0.294
0.370
0.466
0.586
0.736
0.925
1.16
1.46
1.83
2.29
2.87
3.58
4.47
5.56
6.90
8.54
10.5
12.9
15.7
19.0
17
0.0318
0.0400
0.0504
0.0634
0.0799
0.1005
0.127
0.159
0.200
0.252
0.317
0.399
0.502
0.631
0.793
0.996
1.25
1.57
1.97
2.46
3.08
3.85
4.80
5.97
7.40
9.14
11.2
13.8
16.7
20.2
18
0.0343
0.0431
0.0543
0.0683
0.0860
0.1083
0.136
0.171
0.216
0.271
0.342
0.430
0.540
0.679
0.854
1.07
1.35
1.69
2.12
2.65
3.31
4.14
5.15
6.40
7.93
9.78
12.0
14.7
17.8
21.4
19
0.0369
0.0464
0.0584
0.0736
0.0926
0.1166
0.147
0.185
0.232
0.292
0.368
0.462
0.581
0.731
0.918
1.15
1.45
1.82
2.28
2.85
3.56
4.44
5.52
6.86
8.48
10.45
12.8
15.6
18.9
22.7
20
0.0397
0.0500
0.0629
0.0792
0.0996
0.1254
0.158
0.199
0.250
0.314
0.396
0.497
0.625
0.786
0.988
1.24
1.56
1.95
2.44
3.06
3.82
4.76
5.92
7.34
9.07
11.16
13.6
16.6
20.0
24.0
21 22
0.0427 0.0459
0.0537 0.0578
0.0676 0.0727
0.0851 0.0915
0.107 0.115
0.135 0.145
0.170 0.182
0.214 0.230
0.269 0.289
0.338 0.363
0.425 0.457
0.535 0.575
0.672 0.722
0.845 0.908
1.061 1.140
1.33 1.43
1.67 1.80
2.10 2.25
2.62 2.82
3.28 3.52
4.10 4.39
5.10 5.46
6.34 6.78
7.85 8.39
9.69 10.3
11.90 12.7
14.5 15.5
17.6 18.7
21.2 22.5
25.3 26.7
23
0.0493
0.0621
0.0781
0.0983
0.124
0.156
0.196
0.247
0.310
0.390
0.491
0.617
0.776
0.975
1.224
1.54
1.93
2.41
3.02
3.77
4.70
5.85
7.25
8.96
11.0
13.5
16.4
19.8
23.7
28.2
24
0.0530
0.0667
0.0901
0.1134
0.133
0.167
0.210
0.265
0.333
0.419
0.527
0.663
0.833
1.05
1.31
1.65
2.07
2.59
3.24
4.04
5.03
6.25
7.75
9.56
11.7
14.4
17.4
21.0
25.1
29.6
25
0.0568
0.0716
0.0901
0.1134
0.143
0.180
0.226
0.284
0.358
0.450
0.566
0.711
0.893
1.12
1.41
1.77
2.21
2.77
3.46
4.32
5.38
6.68
8.27
10.2
12.5
15.2
18.5
22.2
26.4
31.1
26
0.0610
0.0768
0.0966
0.1216
0.153
0.193
0.242
0.305
0.384
0.482
0.607
0.762
0.958
1.20
1.51
1.89
2.37
2.97
3.71
4.62
5.75
7.14
8.82
10.9
13.3
16.2
19.5
23.4
27.8
32.6
Table provided by Teresa Norberg-King, Environmental Research Laboratory,
Duluth, Minnesota. Also see Emerson et al. (1975), Thurston et al. (1974),
and USEPA (1985d).
44
-------
8.8.9 Mortality due to pH alone may occur if the pH of the sample falls
outside the range of 6.0 - 9.0. Thus, the presence of other forms of toxicity
(metals and organics) in the sample may be masked by the toxic effects of low
or high pH. The question about the presence of other toxicants can be
answered only by performing two parallel tests, one with an adjusted pH, and
one without an adjusted pH. Freshwater samples are adjusted to pH 7.0, and
marine samples are adjusted to pH 8.0, by adding IN NaOH or IN HC1 dropwise,
as required, being careful to avoid overadjustment.
8.9 PRELIMINARY TOXICITY RANGE-FINDING TESTS
8.9.1 USEPA Regional and State personnel generally have observed that it is
not necessary to conduct a toxicity range-finding test prior to initiating a
static, chronic, definitive toxicity test. However, when preparing to perform
a static test with a sample of completely unknown quality, or before
initiating a flow-through test, it is advisable to conduct a preliminary
toxicity range-finding test.
8.9.2 A toxicity range-finding test ordinarily consists of a down-scaled,
abbreviated static acute test in which groups of five organisms are exposed to
several widely-spaced sample dilutions in a logarithmic series, such as 100%,
10.0%, 1.00%, and 0.100%, and a control, for 8-24 h. Caution: if the sample
must also be used for the full-scale definitive test, the 36-h limit on
holding time (Subsection 8.5.4) must not be exceeded before the definitive
test is initiated.
8.9.3 It should be noted that the toxicity (LC50) of a sample observed in a
range-finding test may be significantly different from the toxicity observed
in the follow-up, chronic, definitive test because: (1) the definitive test is
longer; and (2) the test may be performed with a sample collected at a
different time, and possibly differing significantly in the level of toxicity.
8.10 MULTICONCENTRATION (DEFINITIVE) EFFLUENT TOXICITY TESTS
8.10.1 The tests recommended for use in determining discharge permit
compliance in the NPDES program are multiconcentration, or definitive, tests
which provide (1) a point estimate of effluent toxicity in terms of an IC25 or
IC50, or (2) a no-observed-adverse-effect concentration (NOEC) defined in
terms of mortality, and obtained by hypothesis testing. The tests may be
static renewal or static non-renewal.
8.10.2 The tests consist of a control and a minimum of five effluent
concentrations commonly selected to approximate a geometric series, such as
100%, 50%, 25%, 12.5%, and 6.25%, by using a dilution factor of 0.5.
8.10.3 These tests are also to be used in determining compliance with permit
limits on the mortality of the receiving water concentration (RWC) of
effluents by bracketing the RWC with effluent concentrations in the following
manner: (1) 100% effluent, (2) [RWC + 100J/2, (3) RWC, (4) RWC/2, and (5)
RWC/4. For example, where the RWC = 50%, the effluent concentrations used in
the test would be 100%, 75%, 50%, 25%, and 12.5%.
45
-------
8.10.4 If acute/chronic ratios are to be determined by simultaneous acute and
short-term chronic tests with a single species, using the same sample, both
types of tests must use the same test conditions, i.e., pH, temperature, water
hardness, salinity, etc.
8.11 RECEIVING WATER TESTS
8.11.1 Receiving water toxicity tests generally consist of 100% receiving
water and a control. The total salinity of the control should be comparable
to the receiving water.
8.11.2 The data from the two treatments are analyzed by hypothesis testing to
determine if test organism survival in the receiving water differs
significantly from the control. Four replicates and 10 organisms per
replicate are required for each treatment (see Summary of Test Conditions and
Test Acceptability Criteria in the specific test method).
8.11.3 In cases where the objective of the test is to estimate the degree of
toxicity of the receiving water, a definitive, multiconcentration test is
performed by preparing dilutions of the receiving water, using a > 0.5
dilution series, with a suitable control water.
46
-------
SECTION 9
CHRONIC TOXICITY TEST ENDPOINTS AND DATA ANALYSIS
9.1 ENDPOINTS
9.1.1 The objective of chronic aquatic toxicity tests with effluents and pure
compounds is to estimate the highest "safe" or "no-effect concentration" of
these substances. For practical reasons, the responses observed in these
tests are usually limited to hatchability, gross morphological abnormalities,
survival, growth, and reproduction, and the results of the tests are usually
expressed in terms of the highest toxicant concentration that has no
statistically significant observed effect on these responses, when compared to
the controls. The terms currently used to define the endpoints employed in
the rapid, chronic and sub-chronic toxicity tests have been derived from the
terms previously used for full life cycle tests. As shorter chronic tests
were developed, it became common practice to apply the same terminology to the
endpoints. The terms used in this manual are as follows:
9.1.1.1 Safe Concentration The highest concentration of toxicant that will
permit normal propagation of fish and other aquatic life in receiving waters.
The concept of a "safe concentration" is a biological concept, whereas the
"no-observed-effect concentration" (below) is a statistically defined
concentration.
9.1.1.2 No-Observed-Effect-Concentration (NOEC) - The highest concentration
of toxicant to which organisms are exposed in a full life cycle or partial
life cycle (short-term) test, that causes no observable adverse effects on the
test organisms (i.e., the highest concentration of toxicant in which the
values for the observed responses are not statistically significantly
different from the controls). This value is used, along with other factors,
to determine toxicity limits in permits.
9.1.1.3 Lowest-Observed-Effect-Concentration (LOEC) The lowest
concentration of toxicant to which organisms are exposed in a life cycle or
partial life cycle (short-term) test, which causes adverse effects on the test
organisms (i.e., where the values for the observed responses are statistically
significantly different from the controls).
9.1.1.4 Effective Concentration (EC) A point estimate of the toxicant
concentration that would cause an observable adverse affect on a quanta!, "all
or nothing," response (such as death, immobilization, or serious
incapacitation) in a given percent of the test organisms, calculated by point
estimation techniques. If the observable effect is death or immobility, the
term, Lethal Concentration (LC), should be used (see Subsection 9.1.1.5). A
certain EC or LC value might be judged from a biological standpoint to
represent a threshold concentration, or lowest concentration that would cause
an adverse effect on the observed response.
9.1.1.5 Lethal Concentration (LC) - The toxicant concentration that would
cause death in a given percent of the test population. Identical to EC when
47
-------
the observable adverse effect is death. For example, the LC50 is the
concentration of toxicant that would cause death in 50% of the test
population.
9.1.1.6 Inhibition Concentration (1C) The toxicant concentration that would
cause a given percent reduction in a nonquantal biological measurement for the
test population. For example, the IC25 is the concentration of toxicant that
would cause a 25% reduction in mean young per female or in growth for the test
population, and the IC50 is the concentration of toxicant that would cause a
50% reduction.
9.2 RELATIONSHIP BETWEEN ENDPOINTS DETERMINED BY HYPOTHESIS TESTING AND POINT
ESTIMATION TECHNIQUES
9.2.1 If the objective of chronic aquatic toxicity tests with effluents and
pure compounds is to estimate the highest "safe or no-effect concentration" of
these substances, it is imperative to understand how the statistical endpoints
of these tests are related to the "safe" or "no-effect" concentration. NOECs
and LOECs are determined by hypothesis testing (Dunnett's Test, a t-test with
the Bonferroni adjustment, Steel's Many-One Rank Test, or the Wilcoxon Rank
Sum Test), whereas, LCs, ICs, and ECs are determined by point estimation
techniques (Probit Analysis or Linear Interpolation Method). There are
inherent differences between the use of a NOEC or LOEC derived from hypothesis
testing to estimate a "safe" concentration, and the use of a LC, 1C, EC, or
other point estimates derived from curve fitting, interpolation, etc.
9.2.2 Most point estimates, such as the LC, 1C, or EC are derived from a
mathematical model that assumes a continuous dose-response relationship. By
definition, any LC, 1C, or EC value is an estimate of some amount of adverse
effect. Thus the assessment of a "safe" concentration must be made from a
biological standpoint rather than with a statistical test. In this instance,
the biologist must determine some amount of adverse effect that is deemed to
be "safe," in the sense that from a practical biological viewpoint it will not
affect the normal propagation of fish and other aquatic life in receiving
waters. Thus, to use a point estimate such as an LC, 1C, or EC to determine a
"safe" concentration would require the specification by biologists or
toxicologists of what level of adverse effect would be deemed acceptable or
"safe".
9.2.3 The use of NOECs and LOECs, on the other hand, assumes either (1) a
continuous dose-response relationship, or (2) a noncontinuous (threshold)
model of the dose-response relationship.
9.2.3.1 In the case of a continuous dose-response relationship, it is also
assumed that adverse effects that are not "statistically observable" are also
not important from a biological standpoint, since they are not pronounced
enough to test statistically significant against some measure of the natural
variability of responses.
9.2.3.2 In the case of noncontinuous dose-response relationships, it is
assumed that there exists a true threshold, or concentration below which there
is no adverse effect on aquatic life, and above which there is an adverse
48
-------
effect. The purpose of the statistical analysis in this case is to estimate
as closely as possible where that threshold lies.
9.2.3.3 In either case, it is important to realize that the amount of the
adverse effect that is statistically observable (LOEC) or not observable
(NOEC) is highly dependent on all aspects of the experimental design, such as
the number of concentrations of toxicant, number of replicates per
concentration, number of organisms per replicate, and use of randomization.
Other factors that affect the sensitivity of the test include the choice of
statistical analysis, the choice of an alpha level, and the amount of
variability between responses at a given concentration.
9.2.3.4 Where the assumption of a continuous dose-response relationship is
made, by definition some amount of adverse effect might be present at the
NOEC, but is not great enough to be detected by hypothesis testing.
9.2.3.5 Where the assumption of a noncontinuous dose-response relationship
is made, the NOEC would indeed be an estimate of a "safe" or "no-effect"
concentration if the amount of adverse effect that appears at the threshold is
great enough to test as statistically significantly different from the
controls in the face of all aspects of the experimental design mentioned
above. If, however, the amount of adverse effect at the threshold were not
great enough to test as statistically different, some amount of adverse effect
might be present at the NOEC. In any case, the estimate of the NOEC with
hypothesis testing is always dependent on the aspects of the experimental
design mentioned above. For this reason, the reporting and examination of
some measure of the sensitivity of the test (either the minimum significant
difference or the percent change from the control that this minimum difference
represents) is extremely important.
9.2.4 In summary, the assessment of a "safe" or "no-effect" concentration
cannot be made from the results of statistical analysis alone, unless (1) the
assumptions of a strict threshold model are accepted, and (2) it is assumed
that the amount of adverse effect present at the threshold is statistically
detectable by hypothesis testing. In this case, estimates obtained from a
statistical analysis are indeed estimates of a "no-effect" concentration. If
the assumptions are not deemed tenable, then estimates from a statistical
analysis can only be used in conjunction with an assessment from a biological
standpoint of what magnitude of adverse effect constitutes a "safe"
concentration. In this instance, a "safe" concentration is not necessarily a
truly "no-effect" concentration, but rather a concentration at which the
effects are judged to be of no biological significance.
9.2.5 A better understanding of the relationship between endpoints derived by
hypothesis testing (NOECs) and point estimation techniques (LCs, ICs, and ECs)
would be very helpful in choosing methods of data analysis. Norberg-King
(1991) reported that the IC25s were comparable to the NOECs for 23 effluent
and reference toxicant data sets analyzed. The data sets included short-term
chronic toxicity tests for the sea urchin, Arbacia punctulata, the sheepshead
minnow, Cypn'nodon variegatus, and the red macroalga, Champia parvula. Birge
et al. (1985) reported that LCls derived from Probit Analyses of data from
short-term embryo-larval tests with reference toxicants were comparable to
49
-------
NOECs for several organisms. Similarly, USEPA (1988e) reported that the IC25s
were comparable to the NOECs for a set of Cen'odaphm'a dubia chronic tests
with a single reference toxicant. However, the scope of these comparisons was
very limited, and sufficient information is not yet available to establish an
overall relationship between these two types of endpoints, especially when
derived from effluent toxicity test data.
9.3 PRECISION
9.3.1 Hypothesis Tests
9.3.1.1 When hypothesis tests are used to analyze toxicity test data, it is
not possible to express precision in terms of a commonly used statistic. The
results of the test are given in terms of two endpoints, the No-Observed-
Effect Concentration (NOEC) and the Lowest-Observed-Effect Concentration
(LOEC). The NOEC and LOEC are limited to the concentrations selected for the
test. The width of the NOEC-LOEC interval is a function of the dilution
series, and differs greatly depending on whether a dilution factor of 0.3 or
0.5 is used in the test design. Therefore, USEPA recommends the use of the
0.5 dilution factor (see Section 4, Quality Assurance; Subsection 4.14.6, Test
Precision). It is not possible to place confidence limits on the NOEC and
LOEC derived from a given test, and it is difficult to quantify the precision
of the NOEC-LOEC endpoints between tests. If the data from a series of tests
performed with the same toxicant, toxicant concentrations, and test species,
were analyzed with hypothesis tests, precision could only be assessed by a
qualitative comparison of the NOEC-LOEC intervals, with the understanding that
maximum precision would be attained if all tests yielded the same NOEC-LOEC
interval. In practice, the precision of results of repetitive chronic tests
is considered acceptable if the NOECs vary by no more than one concentration
interval above or below a central tendency. Using these guidelines, the
"normal" range of NOECs from toxicity tests using a 0.5 dilution factor
(two-fold difference between adjacent concentrations), would be four-fold.
9.3.2 Point Estimation Techniques
9.3.2.1 Point estimation techniques have the advantage of providing a point
estimate of the toxicant concentration causing a given amount of adverse
(inhibiting) effect, the precision of which can be quantitatively assessed (1)
within tests by calculation of 95% confidence limits, and (2) across tests by
calculating a standard deviation and coefficient of variation.
9.4 DATA ANALYSIS
9.4.1 Role of the Statistician
9.4.1.1 The use of the statistical methods described in this manual for
routine data analysis does not require the assistance of a statistician.
However, the interpretation of the results of the analysis of the data from
any of the toxicity tests described in this manual can become problematic
because of the inherent variability and sometimes unavoidable anomalies in
biological data. If the data appear unusual in any way, or fail to meet the
necessary assumptions, a statistician should be consulted. Analysts who are
50
-------
not proficient in statistics are strongly advised to seek the assistance of a
statistician before selecting the method of analysis and using any of the
results.
9.4.1.2 The statistical methods recommended in this manual are not the only
possible methods of statistical analysis. Many other methods have been
proposed and considered. Certainly there are other reasonable and defensible
methods of statistical analysis of this kind of toxicity data. Among
alternative hypothesis tests some, like Williams' Test, require additional
assumptions, while others, like the bootstrap methods, require computer-
intensive computations. Alternative point estimations approaches most
probably would require the services of a statistician to determine the
appropriateness of the model (goodness of fit), higher order linear or
nonlinear models, confidence intervals for estimates generated by inverse
regression, etc. In addition, point estimation or regression approaches would
require the specification by biologists or toxicologists of some low level of
adverse effect that would be deemed acceptable or safe. The statistical
methods contained in this manual have been chosen because they are (1)
applicable to most different toxicity test data sets for which they are
recommended, (2) powerful statistical tests, (3) hopefully "easily" understood
by nonstatisticians, and (4) amenable to use without a computer, if necessary.
9.4.2 Plotting of the Data
9.4.2.1 The data should be plotted, both as a preliminary step to help detect
problems and unsuspected trends or patterns in the responses, and as an aid in
interpretation of the results. Further discussion and plotted sets of data
are included in the methods and the Appendices.
9.4.3 Data Transformations
9.4.3.1 Transformations of the data, (e.g., arc sine square root and logs),
are used where necessary to meet assumptions of the proposed analyses,
such as the requirement for normally distributed data.
9.4.4 Independence, Randomization, and Outliers
9.4.4.1 Statistical independence among observations is a critical assumption
in all statistical analyses of toxicity data. One of the best ways to ensure
independence is to properly follow rigorous randomization procedures.
Randomization techniques should be employed at the start of the test,
including the randomization of the placement of test organisms in the test
chambers and randomization of the test chamber location within the array of
chambers. Discussions of statistical independence, outliers and
randomization, and a sample randomization scheme, are included in Appendix A.
9.4.5 Replication and Sensitivity
9.4.5.1 The number of replicates employed for each toxicant concentration is
an important factor in determining the sensitivity of chronic toxicity tests.
51
-------
Test sensitivity generally increases as the number of replicates is increased,
but the point of diminishing returns in sensitivity may be reached rather
quickly. The level of sensitivity required by a hypothesis test or the
confidence interval for a point estimate will determine the number of
replicates, and should be based on the objectives for obtaining the toxicity
data.
9.4.5.2 In a statistical analysis of toxicity data, the choice of a
particular analysis and the ability to detect departures from the assumptions
of the analysis, such as the normal distribution of the data and homogeneity
of variance, is also dependent on the number of replicates. More than the
minimum number of replicates may be required in situations where it is
imperative to obtain optimal statistical results, such as with tests used in
enforcement cases or when it is not possible to repeat the tests. For
example, when the data are analyzed by hypothesis testing, the nonparametric
alternatives cannot be used unless there are at least four replicates at each
toxicant concentration.
9.4.6 Recommended Alpha Levels
9.4.6.1 The data analysis examples included in the manual specify an alpha
level of 0.01 for testing the assumptions of hypothesis tests and an alpha
level of 0.05 for the hypothesis tests themselves. These levels are common
and well accepted levels for this type of analysis and are presented as a
recommended minimum significance level for toxicity data analysis.
9.5 CHOICE OF ANALYSIS
9.5.1 The recommended statistical analysis of most data from chronic toxicity
tests with aquatic organisms follows a decision process illustrated in the
flowchart in Figure 2. An initial decision is made to use point estimation
techniques (Probit Analysis or Linear Interpolation Method) and/or to use
hypothesis testing (Dunnett's Test, the t-test with the Bonferroni adjustment,
Steel's Many-one Rank Test, or Wilcoxon Rank Sum Test). If hypothesis testing
is chosen, subsequent decisions are made on the appropriate procedure for a
given set of data, depending on the results of tests of assumptions, as
illustrated in the flowchart. A specific flow chart is included in the
analysis section for each test.
9.5.2 Since a single chronic toxicity test might yield information on more
than one parameter (such as survival, growth, and reproduction), the lowest
estimate of a "no-observed-effect concentration" for any of the
responses would be used as the "no-observed-effect concentration" for
each test. It follows logically that in the statistical analysis of the data,
concentrations that had a significant toxic effect on one of the observed
responses would not be subsequently tested for an effect on some other
response. This is one reason for excluding concentrations that have shown a
statistically significant reduction in survival from a subsequent hypothesis
test for effects on another parameter such as reproduction. A second
reason is that the exclusion of such concentrations usually results in a more
powerful and appropriate statistical analysis. In performing the point
estimation techniques recommended in this manual, an all-data approach is
52
-------
used. For example, data from concentrations above the NOEC for survival are
included in determining ICp estimates using the Linear Interpolation Method.
9.5.3 Analysis of Growth and Reproduction Data
9.5.3.1 Growth data from the sheepshead minnow, Cyprinodon variegatus, and
inland silverside, Mem'dia beryllina, larval survival and growth tests, and
the mysid, Mysidopsis bahia, survival, growth, and fecundity test, are
analyzed using hypothesis testing according to the flowchart in Figure 2. The
above mentioned growth data may also be analyzed by generating a point
estimate with the Linear Interpolation Method.
9.5.3.2 Fecundity data from the mysid, Mysidopsis bahia, test may be analyzed
using hypothesis testing after an arc sine transformation according to the
flowchart in Figure 2. The fecundity data from the mysid test may also be
analyzed by generating a point estimate with the Linear Interpolation Method.
9.5.3.3 Reproduction data from the red macroalga, Champia parvula, test are
analyzed using hypothesis testing as illustrated in Figure 2. The
reproduction data from the red macroalga test may also be analyzed by
generating a point estimate with the Linear Interpolation Method.
9.5.4 Analysis of the Sea Urchin, Arbacia punctuTata, Fertilization Data
9.5.4.1 Data from the sea urchin, Arbacia punctulata, fertilization test may
be analyzed by hypothesis testing after an arc sine transformation according
to the flowchart in Figure 2. The fertilization data from the sea urchin test
may also be analyzed by generating a point estimate with the Linear
Interpolation Method.
9.5.5 Analysis of Mortality Data
9.5.5.1 Mortality data from the sheepshead minnow, Cyprinodon variegatus, and
inland silverside, Mem'dia beryllina, larval survival and growth tests, the
sheepshead minnow embryo-larval survival and teratogenicity test, and the
mysid survival, growth, and reproduction test are analyzed by Probit Analysis,
if appropriate (see discussion below). The mortality data can also be
analyzed by hypothesis testing, after an arc sine transformation (see
Appendices B-F), according to the flowchart in Figure 2.
9.6 HYPOTHESIS TESTS
9.6.1 Dunnett's Procedure
9.6.1.1 Dunnett's Procedure is used to determine the NOEC. The procedure
consists of an analysis of variance (ANOVA) to determine the error term, which
is then used in a multiple comparison method for comparing each of the
treatment means with the control mean, in a series of paired tests (see
Appendix C). Use of Dunnett's Procedure requires at least three replicates
per treatment to check the assumptions of the test. In cases where the
numbers of data points (replicates) for each concentration are not equal, a
53
-------
t-test may be performed with Bonferroni's adjustment for multiple comparisons
(see Appendix D), instead of using Dunnett's Procedure.
9.6.1.2 The assumptions upon which the use of Dunnett's Procedure is
contingent are that the observations within treatments are normally
distributed, with homogeneity of variance. Before analyzing the data, these
assumptions must be tested using the procedures provided in Appendix B.
9.6.1.3 If, after suitable transformations have been carried out, the
normality assumptions have not been met, Steel's Many-One Rank Test should be
used if there are four or more data points (replicates) per toxicant
concentration. If the numbers of data points for each toxicant concentration
are not equal, the Wilcoxon Rank Sum Test with Bonferroni's adjustment should
be used (see Appendix F).
9.6.1.4 Some indication of the sensitivity of the analysis should be provided
by calculating (1) the minimum difference between means that can be detected
as statistically significant, and (2) the percent change from the control mean
that this minimum difference represents for a given test.
9.6.1.5 A step-by-step example of the use of Dunnett's Procedure is provided
in Appendix C.
9.6.2 T-Test with the Bonferroni Adjustment
9.6.2.1 The t-test with the Bonferroni adjustment is used as an alternative
to Dunnett's Procedure when the number of replicates is not the same for all
concentrations. This test sets an upper bound of alpha on the overall error
rate, in contrast to Dunnett's Procedure, for which the overall error rate is
fixed at alpha. Thus, Dunnett's Procedure is a more powerful test.
9.6.2.2 The assumptions upon which the use of the t-test with the Bonferroni
adjustment is contingent are that the observations within treatments are
normally distributed, with homogeneity of variance. These assumptions must be
tested using the procedures provided in Appendix B.
9.6.2.3 The estimate of the safe concentration derived from this test is
reported in terms of the NOEC. A step-by-step example of the use of a t-test
with the Bonferroni adjustment is provided in Appendix D.
9.6.3 Steel's Many-one Rank Test
9.6.3.1 Steel's Many-one Rank Test is a multiple comparison method for
comparing several treatments with a control. This method is similar to
Dunnett's Procedure, except that it is not necessary to meet the assumption
of normality. The data are ranked, and the analysis is performed on the
ranks rather than on the data themselves. If the data are normally or nearly
normally distributed, Dunnett's Procedure would be more sensitive (would
detect smaller differences between the treatments and control). For data that
are not normally distributed, Steel's Many-one Rank Test can be much more
efficient (Hodges and Lehmann, 1956).
54
-------
9.6.3.2 It is necessary to have at least four replicates per toxicant
concentration to use Steel's test. Unlike Dunnett's Test, the sensitivity of
this test cannot be stated in terms of the minimum difference between
treatment means and the control mean.
9.6.3.3 The estimate of the safe concentration is reported as the NOEC. A
step-by-step example of the use of Steel's Many-One Rank Test is provided in
Appendix E.
9.6.4 Wilcoxon Rank Sum Test
9.6.4.1 The Wilcoxon Rank Sum Test is a nonparametric test for comparing a
treatment with a control. The data are ranked and the analysis proceeds
exactly as in Steel's Test except that Bonferroni's adjustment for multiple
comparisons is used instead of Steel's tables. When Steel's test can be used
(i.e., when there are equal numbers of data points per toxicant concentra-
tion), it will be more powerful (able to detect smaller differences as
statistically significant) than the Wilcoxon Rank Sum Test with Bonferroni's
adjustment.
9.6.4.2 The estimate of the safe concentration is reported as the NOEC. A
step-by-step example of the use of the Wilcoxon Rank Sum Test is provided in
Appendix F.
9.6.5 A Caution in the Use of Hypothesis Testing
9.6.5.1 If in the calculation of an NOEC by hypothesis testing, two tested
concentrations cause statistically significant adverse effects, but an
intermediate concentration did not cause statistically significant effects,
the results should be used with extreme caution.
9.7 POINT ESTIMATION TECHNIQUES
9.7.1 Probit Analysis
9.7.1.1 Probit Analysis is used to estimate an LCI, LC50, EC1, or EC50 and
the associated 95% confidence interval. The analysis consists of adjusting
the data for mortality in the control, and then using a maximum likelihood
technique to estimate the parameters of the underlying log tolerance
distribution, which is assumed to have a particular shape.
9.7.1.2 The assumption upon which the use of Probit Analysis is contingent
is a normal distribution of log tolerances. If the normality assumption is
not met, and at least two partial mortalities are not obtained, Probit
Analysis should not be used. In cases where Probit Analysis is not
appropriate, the LC50 and confidence interval may be estimated by the
Spearman-Karber method, the trimmed Spearman-Karber method, or the Graphical
method (USEPA, 1991c). If a test results in 100% survival and 100% mortality
in adjacent treatments (all or nothing effect), a LC50 may be estimated using
the Graphical Method.
55
-------
9.7.1.3 It is important to check the results of Probit Analysis to determine
if the analysis is appropriate. The chi-square test for heterogeneity
provides one good test of appropriateness of the analysis. In cases where
there is a significant chi-square statistic, where there appears to be
systematic deviation from the model, or where there are few data in the
neighborhood of the point to be estimated, Probit results should be used with
extreme caution.
9.7.1.4 The natural rate of occurrence of a measured response, such as
mortality in the test organisms (referred to as the natural spontaneous
response), may be used to adjust the results of the Probit Analysis if such a
rate is judged to be different from zero. If a reliable, consistent estimate
of the natural spontaneous response can be determined from historical data,
the historical occurrence rate may be used to make the adjustment. In cases
where historical data are lacking, the spontaneous occurrence rate should
optimally be estimated from all the data as part of the maximum likelihood
procedure. However, this can require sophisticated computer software. An
acceptable alternative is to estimate the natural occurrence rate from the
occurrence rate in the controls. In this instance, greater than normal
replication in the controls would be beneficial.
9.7.1.5 A discussion of Probit Analysis, along with a computer program for
performing the Probit Analysis, are included in Appendix G. The current
version of the Probit program makes no distinction between EC and LC
endpoints, and labels all results as EC values. When the response in question
is mortality, the EC values output by the program should be treated as the
corresponding LC values.
9.7.2 Linear Interpolation Method
9.7.2.1 The Linear Interpolation Method (see Appendix I, Linear Interpolation
Method) is a procedure to calculate a point estimate of the effluent or other
toxicant concentration [Inhibition Concentration, (1C)] that causes a given
percent reduction (e.g., 25%, 50%, etc.) in the reproduction or growth of the
test organisms. The procedure was designed for general applicability in the
analysis of data from short-term chronic toxicity tests.
9.7.2.2 Use of the Linear Interpolation Method is based on the assumptions
that the responses (1) are monotonically non-increasing (the mean response for
each higher concentration is less than or equal to the mean response for the
previous concentration), (2) follow a piece-wise linear response function, and
(3) are from a random, independent, and representative sample of test data.
The assumption for piece-wise linear response cannot be tested statistically,
and no defined statistical procedure is provided to test the assumption for
monotonicity. Where the observed means are not strictly monotonic by
examination, they are adjusted by smoothing. In cases where the responses at
the low toxicant concentrations are much higher than in the controls, the
smoothing process may result in a large upward adjustment in the control mean.
9.7.2.3 The inability to test the monotonicity and piece-wise linear
assumptions for this method makes it difficult to assess when the method is,
or is not, producing reliable results. Therefore, the method should be used
56
-------
with caution when the results of a toxicity test approach an "all or nothing"
response from one concentration to the next in the concentration series, and
when it appears that there is a large deviation from monotonicity. See
Appendix H for a more detailed discussion of the use of this method and a
computer program available for the calculations.
57
-------
DATA
I
POINT
ESTIMATION
\ -
ENDPOINT ESTIMATE
LC, EC, 1C
NORMAL
HOMOGENEOUS VARIANCE
N0 PQMAI
REP
YES
T-TESTWITH
BONFERRONI ^
ADJUSTMENT
(SURVIVAL, GROWTH, REPRODUCTION, ETC.
ir
HYPOTHESIS TESTING
1
TRANSFORMATION?
I
SHAPIRO-WILK
DISTRIBUTION
MOM NOR
S TEST
HETEROGE
UADIAK
BARTLtl 1 S TEST
NO STATIST!*
RECOM
1
NUMBER OF
LICATES?
NO
:AL ANALYSIS , _
WENDED
YE
t
EQUAL NUMBER OF
REPLICATES?
YES
If \ '
NETT'S STEEL'S MANY-ONE WILCC
EST RANK TEST QnK|CCQ
I
ENDPOINT ESTIMATES
NOEC.LOEC
MAL DISTRIBUTION
NEOUS
CE
^^.
v
4 OR MORE
REPLICATES?
S
NO
^^
1
)XON RANK SUM
TEST WITH
RONI ADJUSTMENT
Figure 2. Flowchart for statistical analysis of test data.
58
-------
SECTION 10
REPORT PREPARATION1
The toxicity data are reported, together with other appropriate data
The following general format and content are recommended for the report:
10.1 INTRODUCTION
1. Permit number
2. Toxicity testing requirements of permit
3. Plant location
4. Name of receiving water body
5. Contract Laboratory (if the test was performed under contract)
a. Name of firm
b. Phone number
c. Address
10.2 PLANT OPERATIONS
1. Product(s)
2. Raw materials
3. Operating schedule
4. Description of waste treatment
5. Schematic of waste treatment
6. Retention time (if applicable)
7. Volume of waste flow (MGD, CFS, GPM)
8. Design flow of treatment facility at time of sampling
10.3 SOURCE OF EFFLUENT, RECEIVING WATER, AND DILUTION WATER
1. Effluent Samples
a. Sampling point
b. Sample collection method
c. Collection dates and times
d. Mean daily discharge on sample collection date
e. Lapsed time from sample collection to delivery
f. Sample temperature when received at the laboratory
g. Physical and chemical data
2. Receiving Water Samples
a. Sampling point
b. Collection dates and times
c. Sample collection method
d. Physical and chemical data
e. Tide stages
f. Sample temperature when received at the laboratory
'Adapted in part from USEPA (1985d) and USEPA (1991c).
59
-------
g. Lapsed time from sample collection to delivery
3. Dilution Water Samples
a. Source
b. Collection date and time
c. Pretreatment
d. Physical and chemical characteristics
10.4 TEST METHODS
1. Toxicity test method used (title, number, source)
2. End point(s) of test
3. Deviation(s) from reference method, if any, and the reason(s)
4. Date and time test started
5. Date and time test terminated
6. Type of volume and test chambers
7. Volume of solution used per chamber
8. Number of organisms used per test chamber
9. Number of replicate test chambers per treatment
10. Acclimation of test organisms (temperature and salinity mean and
range)
11. Test temperature (mean and range)
12. Specify if aeration was needed
13. Feeding frequency, and amount and type of food
14. Test salinity (mean and range)
10.5 TEST ORGANISMS
1. Scientific name and how determined
2. Age
3. Life stage
4. Mean length and weight (where applicable)
5. Source
6. Diseases and treatment (where applicable)
7. Taxonomic key used for species identification
10.6 QUALITY ASSURANCE
1. Reference toxicant used routinely; source
2. Date and time of most recent reference toxicant test;
test results and current control (cusum) chart
3. Dilution water used in reference toxicant test
4. Results (NOEC or, where applicable, LOEC, LC50, EC50,
IC25 and/or IC50)
5. Physical and chemical methods used
10.7 RESULTS
1. Provide raw toxicity data in tabular form, including daily records
of affected organisms in each concentration (including controls),
and plots of toxicity data
2. Provide table of LC50s, NOECs, IC25, IC50, etc.
60
-------
3. Indicate statistical methods to calculate endpoints
4. Provide summary table of physical and chemical data
5. Tabulate QA data
10.8 CONCLUSIONS AND RECOMMENDATIONS
1. Relationship between test endpoints and permit limits,
2. Action to be taken.
61
-------
SECTION 11
TEST METHOD1'2
SHEEPSHEAD MINNOW, CYPRINODON VARIEGATUS,
LARVAL SURVIVAL AND GROWTH TEST
METHOD 1004
1. SCOPE AND APPLICATION
1.1 This method estimates the chronic toxicity of effluents and receiving
waters to the sheepshead minnow, Cypn'nodon van'egatus, using newly hatched
larvae in a seven-day, static-renewal test. The effects include the
synergistic, antagonistic, and additive effects of all the chemical, physical,
and biological components which adversely affect the physiological and
biochemical functions of the test species.
1.2 Daily observations on mortality make it possible to also calculate acute
toxicity for desired exposure periods (i.e., 24-h, 48-h, 96-h LCBOs).
1.3 Detection limits of the toxicity of an effluent or chemical are organism
dependent.
1.4 Single or multiple excursions in toxicity may not be detected using 24-h
composite samples. Also, because of the long sample collection period
involved in composite sampling, and because the test chambers are not sealed,
highly volatile and highly degradable toxicants in the source may not be
detected in the test.
1.5 This method is commonly used in one of two forms: (1) a definitive test,
consisting of a minimum of five effluent concentrations and a control, and (2)
a receiving water test(s), consisting of one or more receiving water
concentrations and a control.
2. SUMMARY OF METHOD
2.1 Larvae (preferably less than 24-h old) are exposed in a static renewal
system for seven days to different concentrations of effluent or to receiving
water. Test results are based on the survival and weight of the larvae in
test solutions, compared to controls.
3. INTERFERENCES
3.1 Toxic substances may be introduced by contaminants in dilution water,
glassware, sample hardware, and testing equipment (see Section 5, Facilities
and Equipment).
The format used for this method was taken from USEPA (1983).
This method was adapted in part from USEPA (1985e) and was based
on USEPA (1987b).
62
-------
3.2 Adverse effects of low dissolved oxygen concentrations (DO), high
concentrations of suspended and/or dissolved solids, and extremes of pH, may
mask the effects of toxic substances.
3.3 Improper effluent sampling and handling may adversely affect test
results (see Section 8, Effluent and Receiving Water Sampling, Sample
Handling, and Sample preparation for Toxicity Tests).
3.4 Pathogenic and/or predatory organisms in the dilution water and
effluent may affect test organism survival, and confound test results.
3.5 Food added during the test may sequester metals and other toxic
substances and reduce the apparent toxicity of the test substance. However,
in a growth test the nutritional needs of the organisms must be satisfied,
even if feeding has the potential to confound test results.
4. SAFETY
4.1 See Section 3, Health and Safety.
5. APPARATUS AND EQUIPMENT
5.1 Facilities for holding and acclimating test organisms.
5.2 Brine shrimp, Artemia, culture unit -- see Subsection 6.14 below and
Section 4, Quality Assurance.
5.3 Sheepshead minnow culture unit -- see Subsection 6.15 below. The
maximum number of larvae required per test will range from a maximum of 360,
if 15 larvae are used in each of four replicates, to a minimum of 180 per
test, if 10 larvae are used in each of three replicates. It is preferable
to obtain the test organisms from an inhouse culture unit. If it is not
feasible to culture fish inhouse, embryos or newly hatched larvae can be
obtained from other sources if shipped in well oxygenated saline water in
insulated containers.
5.4 Samplers -- automatic sampler, preferably with sample cooling
capability, that can collect a 24-h composite sample of 5 L.
5.5 Environmental chamber or equivalent facility with temperature control
(25 ± 1°C).
5.6 Water purification system -- Millipore Milli-QR, deionized water (DI) or
equivalent (see Section 5, Facilities, Equipment, and Supplies).
5.7 Balance -- Analytical, capable of accurately weighing to 0.0001 g.
5.8 Reference weights, Class S -- for checking performance of balance.
Weights should bracket the expected weights of the weighing boats and the
expected weights of the weighing boats plus fish.
5.9 Drying oven -- 105°C, for drying larvae.
63
-------
5.10 Air pump -- for oil-free air supply.
5.11 Air lines, and air stones -- for aerating water containing embryos or
larvae, or for supplying air to test solutions with low DO.
5.12 pH and DO meters -- for routine physical and chemical measurements.
Unless the test is being conducted to specifically measure the effect of one
of the above parameters, a portable, field-grade instrument is acceptable.
5.13 Standard or micro-Winkler apparatus -- for determining DO (optional).
5.14 Dissecting microscope -- for checking embryo viability.
5.15 Desiccator -- for holding dried larvae.
5.16 Light box -- for counting and observing larvae.
5.17 Refractometer -- for determining salinity.
5.18 Thermometers, glass or electronic, laboratory grade -- for measuring
water temperatures.
5.19 Thermometers, bulb-thermograph or electronic-chart type -- for
continuously recording temperature.
5.20 Thermometer, National Bureau of Standards Certified (see USEPA METHOD
170.1, USEPA, 1979b) -- to calibrate laboratory thermometers.
5.21 Test chambers -- four (minimum of three) for each concentration and
control. Borosilicate glass 1000 mL beakers or modified Norberg and Mount
(1985) glass chambers used in the short-term inland silverside test may be
used. It is recommended that each chamber contain a minimum of 50 mL/larvae
and allow adequate depth of test solution (5.0 cm). To avoid potential
contamination from the air and evaporation of water from the test solutions,
the chambers should be covered during the test.
5.22 Beakers -- six Class A, borosilicate glass or non-toxic plasticware,
1000 ml for making test solutions.
5.23 Wash bottles -- for deionized water, for washing embryos from
substrates and containers, and for rinsing small glassware and instrument
electrodes and probes.
5.24 Crystallization dishes, beakers, culture dishes (1 L), or equivalent -
for incubating embryos.
5.25 Volumetric flasks and graduated cylinders -- Class A, borosilicate
glass or non-toxic plastic labware, 10-1000 ml for making test solutions.
5.26 Separatory funnels, 2-L -- two to four for culturing Artemia nauplii.
5.27 Pipets, volumetric -- Class A, 1-100 mL.
64
-------
5.28 Pipets, automatic -- adjustable, 1-100 ml_.
5.29 Pipets, serological -- 1-10 mL, graduated.
5.30 Pipet bulbs and fillers -- PROPIPET", or equivalent.
5.31 Droppers, and glass tubing with fire polished edges, 4 mm ID -- for
transferring larvae.
5.32 Siphon with bulb and clamp -- for cleaning test chambers.
5.33 Forceps -- for transferring dead larvae to weighing boats.
5.34 NITEXR or stainless steel mesh sieves, < 150 [im, 500 urn, 3-5 mm --
for collecting Artemia nauplii and fish embryos, and for spawning baskets,
respectively. (Available from Sterling Marine Products, 18 Label Street,
Montclair, NJ 07042; phone 201-783-9800).
6. REAGENTS AND CONSUMABLE MATERIALS
6.1 Sample containers -- for sample shipment and storage (see Section 8,
Effluent and Receiving Water Sampling and Sample Handling).
6.2 Data sheets (one set per test) -- for data recording.
6.3 Vials, marked -- 18-24 per test, containing 4% formalin or 70% ethanol,
to preserve larvae. (Optional).
6.4 Weighing boats, aluminum -- 18-24 per test.
6.5 Tape, colored -- for labelling test chambers.
6.6 Markers, water-proof -- for marking containers, etc.
6.7 Buffer, pH 7, (or as per instructions of instrument manufacturer) --
for standards and calibration check (see USEPA Method 150.1, USEPA, 1979b).
6.8 Membranes and filling solutions for dissolved oxygen probe (see USEPA
Method 360.1, USEPA, 1979b), or reagents -- for modified Winkler analysis.
6.9 Laboratory quality control samples and standards -- for calibration of
the above methods.
6.10 Reference toxicant solutions (see Section 4, Quality Assurance,
Subsections 4.7, 4.14, 4.15, 4.16, and 4.17).
6.11 Formalin (4%) or 70% ethanol -- for use as a preservative for the fish
larvae.
6.12 Reagent water -- defined as distilled or deionized water that does not
contain substances which are toxic to the test organisms (see subsection 5.6
above).
65
-------
6.13 Effluent, receiving water, and dilution water -- see Section 7, Dilution
Water, and Section 8, Effluent and Receiving Water Sampling, Sample Handling,
and Sample Preparation for Toxicity Tests.
6.13.1 Saline test and dilution water -- The salinity of the test water
must be in the range of 20 to 32 °/oo. The salinity should vary by no
more than + 2 °/oo among the chambers on a given day. If effluent and
receiving water tests are conducted concurrently, the salinities of these
tests should be similar. This test is not recommended for salinities less
than 20 °/oo (see Subsection 10.3).
6.13.2 The overwhelming majority of industrial and sewage treatment
effluents entering marine and estuarine systems contain little or no
measurable salts. Exposure of sheepshead minnow larvae to these effluents
will require adjustments in the salinity of the test solutions. It is
important to maintain a constant salinity across all treatments. In
addition, it may be desirable to match the test salinity with that of the
receiving water. Two methods are available to adjust salinities -- a
hypersaline brine derived from natural seawater or artificial sea salts.
6.13.3 Hypersaline brine (100 °/oo salinity): Hypersaline brine (HSB)
has several advantages that make it desirable for use in toxicity testing.
It can be made from any high quality, filtered seawater by evaporation, and
can be added to the effluent or to deionized water to increase the
salinity. HSB derived from natural seawater contains the necessary trace
metals, biogenic colloids, and some of the microbial components necessary
for adequate growth, survival, and/or reproduction of marine and estuarine
organisms, and may be stored for prolonged periods without any apparent
degradation. However, if 100 °/oo salinity HSB is used as a diluent, the
maximum concentration of effluent that can be tested will be 80% at
20 °/oo salinity and 70% at 30 °/oo salinity.
6.13.3.1 The ideal container for making brine from natural seawater is one
that (1) has a high surface to volume ratio, (2) is made of a non-corrosive
material, and (3) is easily cleaned (fiberglass containers are ideal).
Special care should be used to prevent any toxic materials from coming in
contact with the seawater being used to -generate the brine. If a heater is
immersed directly into the seawater, ensure that the heater materials do not
corrode or leach any substances that would contaminate the brine. One
successful method used is a thermostatically controlled heat exchanger made
from fiberglass. If aeration is used, use only oil-free air compressors to
prevent contamination.
6.13.3.2 Before adding seawater to the brine generator, thoroughly clean
the generator, aeration supply tube, heater, and any other materials that
will be in direct contact with the brine. A good quality biodegradable
detergent should be used, followed by several (at least three) thorough
deionized water rinses.
6.13.3.3 High quality (and preferably high salinity) seawater should be
filtered to at least 10 urn before placing into the brine generator. Water
66
-------
should be collected on an incoming tide to minimize the possibility of
contamination.
6.13.3.4 The temperature of the seawater is increased slowly to 40°C.
The water should be aerated to prevent temperature stratification and to
increase water evaporation. The brine should be checked daily (depending on
volume being generated) to ensure that the salinity does not exceed
100 °/oo and that the temperature does not exceed 40°C. Additional
seawater may be added to the brine to obtain the volume of brine required.
6.13.3.5 After the required salinity is attained, the brine should be
filtered a second time through a 1 urn filter and poured directly into
portable containers, such as 20-L (5 gal) cubitainers or polycarbonate water
cooler jugs. The containers should be capped and labelled with the date the
brine was generated and its salinity. Containers of brine should be stored
in the dark and maintained at room temperature until used.
6.13.3.6 If a source of hypersaline brine is available, test solutions can
be made by following the directions below. Thoroughly mix together the
deionized water and brine before adding the effluent.
6.13.3.7 Divide the salinity of the hypersaline brine by the expected test
salinity to determine the proportion of deionized water to brine. For
example, if the salinity of the brine is 100 °/oo and the test is to be
conducted at 20 °/oo, IQO °/oo divided by 20 °/oo = 5.0. The proportion of
brine is 1 part in 5 (one part brine to four parts deionized water).
6.13.3.8 To make 1 L of sea water at 20 °/oo salinity from a hypersaline
brine of 100 °/oo, divide 1 L (1000 ml) by 5.0. The result, 200 ml, is
the quantity of brine needed to make 1 L of seawater. The difference, 800
ml, is the quantity of deionized water required.
6.13.4 Artificial sea salts: FORTY FATHOMS" brand sea salts (Marine
Enterprises, Inc., 8755 Mylander Lane, Baltimore, Maryland 21204; phone
301-321-1189) have been used successfully at the USEPA EMSL-Cincinnati
Newtown Facility to maintain and spawn sheephead minnows and perform the
larval survival and growth test (see Section 7, Dilution Water). In addition,
a slightly modified version of the GP2 medium (Spotte et al., 1984) has been
successfully used to perform the sheepshead minnow survival and growth test
(Table 1). The use of GP2 for holding and culturing of adults is not
recommended at this time.
6.13.4.1 The FORTY FATHOMS" brand synthetic sea salts are packaged in plastic
bags and mixed with deionized water or equivalent. The instructions on the
package of sea salts should be followed carefully, and the salts should be
mixed in a separate container -- not in the culture tank. The deionized water
used in hydration should be in the temperature range of 21-26°C. Seawater
made from artificial sea salts is conditioned (see Spotte 1973 et al., 1974;
Bower, 1983) before it is used for culturing or testing. After adding the
water, place an air stone in the container, cover, and aerate the solution
mildly for 24 hours before use.
67
-------
6.13.4.2 The GP2 reagent grade chemicals (Table 1) should be mixed with
deionized (DI) water or its equivalent in a container other than the culture
or testing tanks. The deionized water used for hydration should be between
21-26°C. The artificial seawater must be conditioned (aerated) for 24 hours
before use as the testing medium. If the solution is to be autoclaved, sodium
bicarbonate is added after the solution has cooled. A stock solution of
sodium bicarbonate is made up by dissolving 33.6 gm NaHC03 in 500 ml of
deionized water. Add 2.5 ml of this stock solution for each liter of the GP2
artificial seawater.
TABLE 1. REAGENT GRADE CHEMICALS USED IN THE PREPARATION OF GP2 ARTIFICIAL
SEAWATER FOR THE SHEEPSHEAD MINNOW, CYPRINODON VARIEGATUS,
TOXICITY TEST1'2'3
Compound Concentration Amount (g)
(g/L) Required for
20 L
1.
2.
3.
4.
5.
6.
7.
8.
9.
NaCl
Na2S04
KC1
KBr
Na2B407 . 10 H20
MgCl2 . 6 H20
CaCl2 . 2 H20
SrCl2 . 6 H20
NaHCO,
21.03
3.52
0.61
0.088
0.034
9.50
1.32
0.02
0.17
420.6
70.4
12.2
1.76
0.68
190.0
26.4
0.400
3.40
'Modified GP2 from Spotte et al. (1984).
The constituent salts and concentrations were taken from
USEPA (1990b). The salinity is 30.89 g/L.
GP2 can be diluted with deionized (DI) water to the desired test salinity.
68
-------
6.14 BRINE SHRIMP, ARTEMIA, NAUPLII (see USEPA et al., 1991c).
6.14.1 Newly-hatched Artemia nauplii are used as food for sheepshead minnow
larvae in toxicity tests and in the maintenance of continuous stock cultures.
Although there are many commercial sources of brine shrimp cysts, the
Brazilian or Colombian strains are currently preferred because the supplies
examined have had low concentrations of chemical residues and produce nauplii
of suitably small size. For commercial sources of brine shrimp, Artemia,
cysts, see Table 2 (Section 5, Facilities, Equipment, and Supplies) and
Section 4, Quality Assurance, Subsection 4.8 Food Quality.
6.14.2 Each new batch of Artemia cysts must be evaluated for size (Vanhaecke
and Sorgeloos, 1980, and Vanhaecke et al., 1980) and nutritional suitability
(see Leger et al., 1985, 1986) against known suitable reference cysts by
performing a side-by-side larval growth test using the "new" and "reference"
cysts. The "reference" cysts used in the suitability test may be a
previously tested and acceptable batch of cysts, or may be obtained from the
Quality Assurance Research Division, Environmental Monitoring Systems
Laboratory, Cincinnati, Ohio 45268. A sample of newly-hatched Artemia nauplii
from each new batch of cysts should be chemically analyzed. The Artemia cysts
should not be used if the concentration of total organochlorine pesticides
exceeds 0.15 \ig/g wet weight or the total concentration of organochlorine
pesticides plus PCBs exceeds 0.30 [ig/g wet weight. (For analytical methods
see USEPA, 1982).
6.14.3 Artemia nauplii are obtained as follows:
1. Add 1 L of seawater, or a solution prepared by adding 35.0 g
uniodized salt (NaCl) or artificial sea salts to 1 L deionized water,
to a 2-L separatory funnel, or equivalent.
2. Add 10 mL Artemia cysts to the separatory funnel and aerate for 24-h
at room temperature. (Hatching time varies with incubation
temperature and the geographic strain of Artemia used). (See USEPA e
al., 1991c and ASTM designation E1203, 1987, for details on Artemia
culture and quality control).
3. After 24 h, cut off the air supply in the separatory funnel. Artemia
nauplii are phototactic, and will concentrate at the bottom if a dark
cloth or paper towel is placed over the top of the separatory funnel
for 5-10 min. To prevent mortality, do not leave the concentrated
nauplii at the bottom of the funnel more than 10 min without aeration.
4. Drain the nauplii into a cup or funnel fitted with a < 150 [im Nitex
or stainless steel screen, and rinse with seawater or equivalent
before used.
6.14.4 Testing Artemia nauplii as food for toxicity test organisms.
6.14.4.1 The primary criterion for acceptability of each new supply of brine
shrimp cysts is the ability of the nauplii to support good survival and
growth of the sheepshead minnow larvae (see Subsection 12. ACCEPTABILITY OF
69
-------
TEST RESULTS). The larvae used to evaluate the suitability of the brine
shrimp nauplii must be of the same geographical origin, species, and stage
of development as those used routinely in the toxicity tests. Sufficient
data to detect differences in survival and growth should be obtained by
using three replicate test vessels, each containing a minimum of 15 larvae,
for each type of food.
6.14.4.2 The feeding rate and frequency, test vessels, volume of control
water, duration of the test, and age of the nauplii at the start of the
test, should be the same as used for the routine toxicity tests.
6.14.4.3 Results of the brine shrimp nutrition assay, where there are only
two treatments, can be evaluated statistically by use of a t-test. The
"new" food is acceptable if there are no statistically significant
differences in the survival and growth of the larvae fed the two sources of
nauplii.
6.15 SHEEPSHEAD MINNOWS
6.15.1 Brood Stock
6.15.1.1 Adult sheepshead minnows for use as brood stock may be obtained by
seine in Gulf of Mexico and Atlantic coast estuaries, from commercial
sources, or from young fish raised to maturity in the laboratory. Feral
brood stocks and first generation laboratory fish are preferred, to minimize
inbreeding.
6.15.1.2 To detect disease and to allow time for acute mortality due to the
stress of capture, field-caught adults are observed in the laboratory a
minimum of two weeks before using as a source of gametes. Injured or
diseased fish are discarded.
6.15.1.3 Sheepshead minnows can be continuously cultured in the laboratory
from eggs to adults. The larvae, juvenile, and adult fish should be kept in
appropriate size rearing tanks, maintained at ambient laboratory
temperature. The larvae should be fed sufficient newly-hatched Artemia
nauplii daily to assure that live nauplii are always present. Juveniles are
fed frozen adult brine shrimp and a commercial flake food, such as TETRA
SM-80", available from Tetra Sales (U.S.A), 201 Tabor Rd. Morris Plains,
New Jersey 07950, phone 800-526-0650, or MARDEL AQUARIAN* Tropical Fish
Flakes, available from Mardel Laboratories, Inc., 1958 Brandon Court,
Glendale Heights, Illinois 60139, phone 312-351-0606, or equivalent. Adult
fish (age one month) are fed flake food three or four times daily,
supplemented with frozen adult brine shrimp.
6.15.1.3.1 Sheepshead minnows reach sexual maturity in three-to-five months
after hatch, and have an average standard length of approximately 27 mm for
females and 34 mm for males. At this time, the males begin to exhibit
sexual dimorphism and initiate territorial behavior. When the fish reach
sexual maturity and are to be used for natural spawning, the temperature
should be controlled at 18-20°C.
70
-------
6.15.1.4 Adults can be maintained in natural or artificial seawater in a
flow-through or recirculating, aerated system consisting of an all-glass
aquarium, or a "Living Stream" (Figid Unit, Inc., 3214 Sylvania Ave, Toledo,
Ohio 43613, phone 419-474-6971), or equivalent.
6.15.1.5 The system is equipped with an undergravel or outside biological
filter of shells (see Spotte (1973) or Bower (1983) for conditioning the
biological filter), or a cartridge filter, such as a MAGNUMR Filter,
available from Carolina Biological Supply Co., Burlington, North Carolina
27215, phone 800-334-5551, or an EHEIM* Filter, available from Hawaiian
Marine Imports Inc., P.O. Box 218687, Houston, Texas 77218, phone
713-492-7864, or equivalent, at a salinity of 20-30 °/oo and a photoperiod
of 14 h light/10 h dark.
6.15.2 Obtaining Embryos for Toxicity Tests1
6.15.2.1 Embryos can be shipped to the laboratory from an outside source or
obtained from adults held in the laboratory. Ripe eggs can be obtained
either by natural spawning or by intraperitoneal injection of the females
with human chorionic gonadotrophin (HCG) hormone, available from United
States Biochemical Corporation, Cleveland, Ohio 44128, phone 216-765-5000.
If the culturing system for adults is temperature controlled, natural
spawning can be induced. Natural spawning is preferred because repeated
spawnings can be obtained from the same brood stock, whereas with hormone
injection, the brood stock is sacrificed in obtaining gametes.
6.15.2.2 It should be emphasized that the injection and hatching schedules
given below are to be used only as guidelines. Response to the hormone
varies from stock to stock and with temperature. Time to hatch and percent
viable hatch also vary among stocks and among batches of embryos obtained
from the same stock, and are dependent on temperature, DO, and salinity.
The coordination of spawning and hatching is further complicated by the fact
that, even under the most ideal conditions, embryos spawned over a 24-h
period may hatch over a 72-h period. Therefore, it is advisable (especially
if natural spawning is used) to obtain fertilized eggs over several days to
ensure that a sufficient number of newly hatched larvae (less than 24 h old)
will be available to initiate a test.
6.15.2.3 Forced Spawning
6.15.2.3.1 HCG is reconstituted with sterile saline or Ringer's solution
immediately before use. The standard HCG vial contains 1,000 IU to be
reconstituted in 10 ml of saline. Freeze-dried HCG which comes with
premeasured and sterilized saline is the easiest to use. Use of a 50 IU
dose requires injection of 0.05 ml of reconstituted hormone solution.
Reconstituted HCG may be used for several weeks if kept in the refrigerator.
6.15.2.3.2 Each female is injected intraperitoneally with 50 IU HCG on two
consecutive days, starting at least 10 days prior to the beginning of a
Adapted from USEPA (1978b).
71
-------
test. Two days following the second injection, eggs are stripped from the
females and mixed with sperm derived from excised macerated testes. At
least ten females and five males are used per test to ensure that there is a
sufficient number (400) of viable embryos.
6.15.2.3.3 HCG is injected into the peritoneal cavity, just below the skin,
using as small a needle as possible. A 50 IU dose is recommended for
females approximately 27 mm in standard length. A larger or smaller dose
may be used for fish which are significantly larger or smaller than 27 mm.
With injections made on days one and two, females which are held at 25°C
should be ready for stripping on days 4, 5, and 6. Ripe females should show
pronounced abdominal swelling, and release at least a few eggs in response
to a gentle squeeze. Injected females should be isolated from males. It
may be helpful if fish that are to be injected are maintained at 20°C
before injection, and the temperature raised to 25°C on the day of the
first injection.
6.15.2.3.4 Prepare the testes immediately before stripping the eggs from
the females. Remove the testes from three-to-five males. The testes are
paired, dark grey organs along the dorsal midline of the abdominal cavity.
If the head of the male is cut off and pulled away from the rest of the
fish, most of the internal organs can be pulled out of the body cavity,
leaving the testes behind. The testes are placed in a few ml of seawater
until the eggs are ready.
6.15.2.3.5 Strip the eggs from the females, into a dish containing 50-100
ml of seawater, by firmly squeezing the abdomen. Sacrifice the females and
remove the ovaries if all the ripe eggs do not flow out freely. Break up
any clumps of ripe eggs and remove clumps of ovarian tissue and underripe
eggs. Ripe eggs are spherical, approximately 1 mm in diameter, and almost
clear.
6.15.2.3.6 While being held over the dish containing the eggs, the testes
are macerated in a fold of NITEXR screen (250-500 urn mesh) dampened with
seawater. The testes are then rinsed with seawater to remove the sperm from
tissue, and the remaining sperm and testes are washed into the dish. Let
the eggs and milt stand together for 10-15 min, swirling occasionally.
6.15.2.3.7 Pour the contents of the dish into a beaker, and insert an
airstone. Aerate gently, such that the water moves slowly over the eggs,
and incubate at 25 C for 60-90 min. After incubation, wash the eggs on a
Nitex screen and resuspend them in clean seawater. Examine the eggs
periodically under a dissecting microscope until they are in the 2-8 cell
stage. (The stage at which it is easiest to tell the developing embryos
from the abnormal embryos and unfertilized eggs; see Figure 1). The eggs
can then be gently rolled on a Nitex screen and culled (Subsection 6.15.2.5)
6.15.2.4 Natural Spawning
6.15.2.4.1 Cultures of adult fish to be used for spawning are maintained at
18-20°C until embryos are required. When embryos are required, raise the
temperature to 25°C in the morning, seven or eight days before the beginning
72
-------
of a test. That afternoon, transfer the adult fish (generally, at least
females and three males) to a spawning chamber (approximately, 20X35X22 cm
high; USEPA, 1978b), which is a basket constructed of 3-5 mm nylon mesh, made
to fit a 57-L (15 gal) aquarium. Eggs will fall through the bottom of the
basket and onto a collecting screen (250-500 \M mesh) below the basket. Allow
the embryos to collect for 24 h. Embryos are washed from the screen, checked
for viability, and placed in incubation dishes. Replace the screens until a
sufficient number of embryos have been collected. One-to-three spawning
aquaria can be used to collect the required number of embryos to run a
toxicity test. To help keep the embryos clean, the adults are fed while the
screens are removed.
6.15.2.5 Incubation
6.15.2.5.1 Four hours post-fertilization, the embryos obtained by natural
or forced spawning are rolled gently with a finger on a 250-500 \im nylon
screen to remove excess fibers and tissue. The embryos have adhesive threads
and tend to adhere to each other. Gentle rolling on the screen facilitates
the culling process described below. To reduce fungal contamination of the
newly spawned embryos after they have been manipulated, they should be placed
in a 250 urn sieve and briskly sprayed with seawater from a squeeze bottle.
6.15.2.5.2 Under a dissecting microscope, separate and discard abnormal
embryos and unfertilized eggs. While they are checked, the embryos are
maintained in sea water at 25°C. The embryos should be in Stages C-G,
Figure 1.
6.15.2.5.3 If the test is prepared with four replicates of 15 larvae at
each of six treatments (five effluent concentrations and a control), and the
combined mortality of eggs and larvae prior to the start of the test is less
than 20%, approximately 400 viable embryos are required at this stage.
6.15.2.5.4 Embryos are demersal. They should be aerated and incubated at
25°C, at a salinity of 20-30 °/oo and a 14-h photoperiod. The embryos
can be cultured in either a flow-through or static system, using aquaria or
crystallization dishes. However, if the embryos are cultured in dishes, it
is essential that aeration and daily water changes be provided, and the
dishes be covered to reduce evaporation that may cause increased salinity.
One-half to three-quarters of the sea water from the culture vessels can be
poured off and the incubating embryos retained. Embryos cultured in this
manner should hatch in six or seven days.
6.15.2.5.5 At 48 h post-fertilization, embryos are examined under a
microscope to determine development and survival. Embryos should be in
Stages I and J, Figure 1. Discard dead embryos. Approximately 360 viable
embryos are required at this stage.
7. EFFLUENT AND RECEIVING WATER COLLECTION, PRESERVATION, AND STORAGE
7.1 See Section 8, Effluent and Receiving Water Sampling, Sample Handling,
and Sample Preparation for Toxicity Tests.
73
-------
8. CALIBRATION AND STANDARDIZATION
8.1 See Section 4, Quality Assurance.
9. QUALITY CONTROL
9.1 See Section 4, Quality Assurance.
10. TEST PROCEDURES
10.1 TEST SOLUTIONS
10.1.1 Receiving Waters
10.1.1.1 The sampling point is determined by the objectives of the test. At
estuarine and marine sites, samples are collected at mid-depth. Receiving
water toxicity is determined with samples used directly as collected or with
samples passed through a 60 \im NITEXR filter and compared without dilution,
against a control. Using four replicate chambers per test, each containing
500-750 ml, and 400 mL for chemical analysis, would require approximately 3400
mL or more of sample per test.
10.1.2 Effluents
10.1.2.1 The selection of the effluent test concentrations should be based
on the objectives of the study. A dilution factor of 0.5 is commonly used. A
dilution factor of 0.5 provides precision of + 100%, and allows for testing of
concentrations between 6.25% and 100% effluent using only five effluent
concentrations (6.25%, 12.5%, 25.0%, 50.0%, and 100%). Test precision shows
little improvement as dilution factors are increased beyond 0.5 and declines
rapidly if smaller dilution factors are used. Therefore, USEPA recommends a
dilution factor of 0.5. If 100 °/oo salinity MSB is used as a diluent, the
maximum concentration of effluent that can be tested will be 80% at 20 °/oo
salinity and 70% at 30 °/oo salinity.
10.1.2.2 If the effluent is known or suspected to be highly toxic, a lower
range of effluent concentrations should be used (such as 25%, 12.5%, 6.25%,
3.12%, and 1.56%). If a high rate of mortality is observed during the first
l-to-2 h of the test, additional dilutions should be added at the lower range
of effluent concentrations.
10.1.2.3 The volume of effluent required for daily renewal of four replicates
per concentration, each containing 750 mL of test solution, is approximately 5
L. Prepare enough test solution (approximately 3400 mL) at each effluent
concentration to provide 400 mL additional volume for chemical analyses (Table
2).
10.1.2.4 The salinity of effluent and receiving water tests for sheepshead
minnows should be between 20 /oo and 30 °/oo. If concurrent effluent
and receiving water testing occurs, the effluent test salinity should
closely approximate that of the receiving water test. If an effluent is
tested alone, select a salinity between 20 °/oo and 30 °/oo, whichever
74
-------
comes closest to the salinity of the receiving waters. Table 2 illustrates
the quantities of effluent, artificial sea salts, hypersaline brine, or
seawater needed to prepare 3 L of test solution at each treatment level for
tests performed at 20 /oo salinity.
10.1.2.5 Just prior to test initiation (approximately one hour), the
temperature of sufficient quantity of the sample to make the test solutions
should be adjusted to the test temperature (25 + 1°C) and maintained at that
temperature during the addition of dilution water.
10.1.2.6 Higher effluent concentrations (i.e., 25%, 50%, and 100%) may
require aeration to maintain adequate dissolved oxygen concentrations.
However, if one solution is aerated, all concentrations must be aerated.
Aerate effluent as it warms and continue to gently aerate test solutions in
the test chambers for the duration of the test.
10.1.2.7 Tests should begin as soon as possible, preferably within 24 h
after sample collection. The maximum holding time following retrieval of the
sample from the sampling device should not exceed 36 h for off site toxicity
tests unless permission is granted by the permitting authority. In no
case should the sample be used in a test more than 72 h after sample
collection.
10.1.3 Dilution Water
10.1.3.1 Dilution water may be natural seawater (receiving water),
hypersaline brine prepared from natural seawater, or artificial seawater
prepared from FORTY FATHOMSR or GP2 sea salts (see Table 1 and Section 7,
Quality Assurance). Other artificial sea salts may be used for culturing
sheepshead minnows and for the larval survival and growth test if the control
criteria for acceptability of test data are satisfied.
10.2 START OF THE TEST
10.2.1 If the embryos have been incubating at 25°C, 30 °/oo salinity,
and a 14-h photoperiod, for five-to-six days with aeration and daily water
renewals, approximately 24 h prior to hatching, the salinity of the sea
water in the incubation chamber may be reduced from 30 °/oo to the test
salinity, if lower than 30 °/oo. in addition to maintaining good water
quality, reducing the salinity and/or changing the water may also help to
initiate hatching over the next 24 h. A few larvae may hatch 24 h ahead of
the majority. Remove these larvae and reserve them in a separate dish,
maintaining the same culture conditions. It is preferable to use only the
larvae that hatch in the 24 h prior to starting the test. However, if
sufficient numbers of larvae do not hatch within the 24-h period, the larvae
that hatch prior to 24 h are added to the test organisms. The test organisms
are then randomly selected for the test. When eggs or larvae must be shipped
to the test site from a remote location, it may be necessary to use larvae
older than 24-h because of the difficulty in coordinating test organism
75
-------
Figure 1. Embryonic development of sheepshead minnow, Cyprinodon
variegatus: A. Mature unfertilized egg, showing attachment filaments and
micropyle, X33; B. Blastodisc fully developed; C,D. Blastodisc, 8 cells;
E. Blastoderm, 16 cells; F. Blastoderm, late cleavage stage; G. Blastoderm
with germ ring formed, embryonic shield developing; H. Blastoderm covers
over 3/4 of yolk, yolk noticeably constricted; I. Early embryo. From
Kuntz (1916).
76
-------
Figure 1. (Continued). Embryonic development of sheepshead minnow,
Cyprinodon variegatus: J. Embryo 48 h after fertilization, now segmented
throughout, pigment on yolk sac and body, otoliths formed; K. Posterior
portion of embryo free from yolk and moves freely within egg membrane, 72
h after fertilization; L. Newly hatched fish, actual length 4 mm; M.
Larval fish 5 days after hatching, actual length 5 mm; N. Young fish 9 mm
in length; 0. Young fish 12 mm in length. From Kuntz (1916).
77
-------
shipments with field operations. However, in the latter case, the larvae
should not be more than 48-h old at the start of the test and should all be
within 24-h of the same age.
10.2.2 Label the test chambers with a marking pen and use color coded tape to
identify each treatment and replicate. A minimum of five effluent
concentrations and a control should be used for each study. Each treatment
(including controls) should have four (minimum of three) replicates. For
exposure chambers, use 1000 ml beakers, non-toxic disposable plasticware, or
glass chambers with a sump area as illustrated in the inland silverside test
method.
10.2.3 Distribute the test solutions to the test chambers.
10.2.4 The test is started by randomly placing larvae from the common pool
into each test chamber until each chamber contains 15 (minimum of 10) larvae,
for a total of 60 larvae for each treatment (minimum of three replicates).
See Appendix A for an example of randomization. The amount of water added to
the chambers when transferring the larvae should be kept to a minimum to avoid
unnecessary dilution of the test concentrations.
10.2.5 The chambers may be placed on a light table to facilitate counting the
larvae.
10.2.6 Randomize the position of the test chambers at the beginning of the
test (see Appendix A). Maintain the chambers in this configuration throughout
the test. Preparation of a position chart may be helpful.
10.3 LIGHT, PHOTOPERIOD, SALINITY, AND TEMPERATURE
10.3.1 The light quality and intensity should be at ambient laboratory
levels, which is approximately 10-20 \iE/m2/s, or 50 to 100 foot candles
(ft-c), with a photoperiod of 16 h light and 8 h darkness. The test salinity
should be in the range of 20 to 30 °/oo to accommodate receiving waters that
may fall within this range. Conduct of this test at salinities less than 20
°/oo may cause an unacceptably low growth response and thereby invalidate the
test. The salinity should vary by no more than + 2 °/oo among the chambers on
a given day. If effluent and receiving water tests are conducted
concurrently, the salinities of these tests should be similar. The water
temperature in the test chambers should be maintained at 25 + 1°C.
10.4 DISSOLVED OXYGEN (DO) CONCENTRATION
10.4.1 Aeration may affect the toxicity of effluents and should be used only
as a last resort to maintain a satisfactory DO. The DO concentrations should
be measured on new solutions at the start of the test (Day 0) and before daily
renewal of test solutions on subsequent days. The DO should not fall below
4.0 mg/L (see Section 8, Effluent and Receiving Water Sampling, Sample
Handling, and Sample Preparation for Toxicity Tests). If it is necessary to
aerate, all treatments and the control should be aerated. The aeration rate
should not exceed 100 bubbles/min, using a pipet with an orifice of
approximately 1.5 mm, such as a 1-mL Kimax serological pipet, No. 37033, or
78
-------
equivalent. Care should be taken to ensure that turbulence resulting from
aeration does not cause undue stress on the fish.
10.5 FEEDING
10.5.1 Sheepshead minnow larvae are fed newly-hatched (less than 24-h old)
Artemia nauplii once a day from hatch Day 0 through Day 6. Feed 0.10 g
nauplii per test chamber on Days 0-2, and 0.15 g nauplii per test chamber on
Days 3-6. The larvae are not fed on Day 7. Equal amounts of nauplii must
be added to each replicate to reduce the variability in larval weight.
Sufficient numbers of nauplii should be fed to assure that some remain alive
overnight in the test chambers. An adequate but not excessive amount should
be provided to each replicate on a daily basis. Feeding excessive amounts
of nauplii will result in a depletion in DO to a lower than acceptable level
(below 4.0 mg/L).
10.5.2 On Days 0-2, weigh 4 g wet weight or pipette 4 ml of concentrated,
rinsed Artemia nauplii for a test with five treatments and a control.
Resuspend the 4 g Artemia in 80 mL of natural or artificial sea water in a 100
ml beaker. Aerate or swirl Artemia to maintain a thoroughly mixed suspension
of nauplii. Dispense 2 ml Artemia suspension by pipette or adjustable
syringe to each test chamber. Collect only enough Artemia in the pipette or
syringe for one test chamber or settling of Artemia may occur, resulting in
unequal amounts of Artemia being distributed to the replicate test
chambers. On Days 3-6, weigh 6 g wet weight or pipette 6 ml Artemia
suspension for a test with five treatments and a control. Resuspend the 6 g
Artemia in 80 mL of natural or artificial sea water in a 100 ml beaker.
Aerate or swirl as 2 ml is dispensed to each test chamber. If the survival
rate in any test chamber on any day falls below 50%, reduce the volume of
Artemia added to that test chamber by one-half (i.e., from 2 mL to 1 mL) and
continue feeding one-half the volume through Day 6. Record the time of
feeding (Figure 7).
10.6 DAILY CLEANING OF TEST CHAMBERS
10.6.1 Before the daily renewal of test solutions, uneaten and dead Artemia,
dead fish larvae, and other debris are removed from the bottom of the test
chambers with a siphon hose. As much of the uneaten Artemia as possible
should be siphoned from each chamber to ensure that the larvae principally eat
newly hatched nauplii. Alternately, a large pipet (50 mL), fitted with a
safety pipet filler or rubber bulb, can be used. Because of their small size
during the first few days of the tests, larvae are easily drawn into the
siphon tube when cleaning the test chambers. By placing the test chambers on
a light box, inadvertent removal of live larvae can be greatly reduced because
they can be more easily seen. If the water siphoned from the test chambers is
collected in a white plastic tray, the live larvae caught in the siphon can be
retrieved and returned to the chambers. Any incidence of removal of live
larvae from the test chambers by the siphon during cleaning, and subsequent
return to the chambers, should be noted in the test records.
79
-------
10.7 OBSERVATIONS DURING THE TEST
10.7.1 Routine Chemical and Physical Observations
10.7.1.1 At a minimum, the following measurements are made and recorded
(see Figure 2).
10.7.1.2 DO is measured at the beginning and end of each 24-h exposure
period in one test chamber at each test concentration and in the control.
10.7.1.3 Temperature, pH, and salinity are measured at the end of each 24-h
exposure period in one test chamber at each test concentration and in the
control. The pH is measured in the effluent sample each day.
10.7.2 Routine Biological Observations
10.7.2.1 The number of live larvae in each test chamber are recorded daily
(Figure 7), and the dead larvae are discarded.
10.7.2.2 Protect the larvae from unnecessary disturbance during the test by
carrying out the daily test observations, solution renewals, and removal of
dead larvae, carefully. Make sure the larvae remain immersed during the
performance of the above operations.
10.8 TEST SOLUTION RENEWAL
10.8.1 The test solutions are renewed daily using freshly prepared
solution, immediately after cleaning the test chambers. For on-site
toxicity studies, fresh effluent and receiving water samples used in
toxicity tests should be collected daily, and no more than 24 h should
elapse between collection of the sample and use in the test (see
Section 8, Effluent and Receiving Water Sampling, Sample Handling, and Sample
Preparation for Toxicity Tests). For off-site tests, a minimum of three
samples must be collected, preferably on days one, three, and five. Maintain
the samples at 4°C until used.
10.8.2 Approximately one hour before the test, warm the sample(s) (if
stored at 4°C) to 25 ± 1°C.
10.8.3 For test solution renewal, the water level in each chamber is
lowered to a depth of 7 to 10 mm, which leaves 15 to 20% of the test
solution. New test solution (750 mL) should be added slowly by pouring down
the side of the test chamber to avoid excessive turbulence and possible
injury to the larvae.
10.8.4 The higher effluent concentrations (i.e., 25, 50, and 100%) may
require aeration to maintain adequate DOs. However, if one solution is
aerated, then all treatments must be aerated. Gently aerate test solutions
in the test chambers so that the larvae are not disturbed.
80
-------
10.9 TERMINATION OF THE TEST
10.9.1 The test is terminated after seven days of exposure. At
termination, dead larvae are removed and surviving larvae in each test chamber
are counted and immediately prepared for drying and weighing, or are preserved
as a group in 4% formalin or 70% ethanol and dried and weighed at a later
date.
10.9.2 For immediate drying and weighing, siphon or pour live larvae onto
a 500 urn mesh screen in a large beaker to retain the larvae and allow
Artemia and debris to be rinsed away. Rinse the larvae with deionized water
to wash away salts that might contribute to the dry weight, and sacrifice
the larvae by placing them in an ice bath of deionized water to prepare for
drying.
10.9.3 Small aluminum weighing boats can be used to dry and weigh the
larvae. Mark for identification an appropriate number of small aluminum
weighing boats (one per replicate). Weigh to the nearest 0.01 mg, and
record the weights (Figure 11).
10.9.4 Immediately prior to the drying, rinse the preserved larvae in
distilled water. The group of rinsed larvae from each test chamber is
transferred to a tared weighing boat and dried at 60°C for 24 h or at
105°C for a minimum of 6 h. Immediately upon removal from the drying
oven, the weighing boats are placed in a desiccator until weighed, to
prevent the absorption of moisture from the air. Weigh to the nearest
0.01 mg all weighing boats containing dried larvae and subtract the tare
weight to determine the dry weight of larvae in each replicate. Record the
weights (Figure 11). For each test chamber, divide the final dry weight by
the number of larvae surviving in the test chamber to determine the average
individual dry weight, and record (Figure 11). Complete the summary data
sheet (Figure 12) after calculating the average measurements and
statistically analyzing the dry weights and percent survival.
11. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA
11.1 A summary of test conditions and test acceptability criteria is listed
in Table 3.
12. ACCEPTABILITY OF TEST RESULTS
12.1 The tests are acceptable if (1) the average survival of control larvae
equals or exceeds 80%, and (2) the average dry weight of unpreserved control
larvae is equal to or greater than 0.60 mg, or (3) the average dry weight of
preserved control larvae is equal to or greater than 0.50 mg. The above
minimum weights presume that the age of the larvae at the start of the test is
less than or equal to 24 h.
13. DATA ANALYSIS
13.1 GENERAL
81
-------
TABLE 2. PREPARATION OF TEST SOLUTIONS AT A SALINITY OF 20 °/oo, USING 20
°/oo SALINITY DILUTION WATER PREPARED FROM NATURAL SEAWATER,
HYPERSALINE BRINE, OR ARTIFICIAL SEA SALTS
Solutions To Be Combined
Effluent
Solution
1
2
3
4
5
Control
Effluent
Cone. (%)
1001'2
50
25
12.5
6.25
0.0
Volume of
Effluent Solution
6800 mL
3400 mL Solution 1
3400 mL Solution 2
3400 mL Solution 3
3400 mL Solution 4
Volume of Diluent
Seawater (20 °/oo)
+ 3400 mL
+ 3400 mL
+ 3400 mL
+ 3400 mL
3400 mL
Total
17000 mL
1This illustration assumes: (1) the use of 750 mL of test solution in each of
four replicates and 400 mL for chemical analysis (total of 3400 mL) for the
control and each of five concentrations of effluent (2) an effluent dilution
factor of 0.5, and (3) the effluent lacks appreciable salinity. A sufficient
initial volume (6800 mL) of effluent is prepared by adjusting the salinity to
the desired level. In this example, the salinity is adjusted by adding
artificial sea salts to the 100% effluent, and preparing a serial dilution
using 20 °/oo seawater (natural seawater, hypersaline brine, or artificial
seawater). Following addition of salts, the effluent is stirred for 1 h to
assure that the salts have dissolved. The salinity of the initial 6800 mL of
100% effluent is adjusted to 20 °/oo by adding 136 g of dry artificial sea
salts (FORTY FATHOMS"). Test concentrations are then made by mixing
appropriate volumes of salinity-adjusted effluent and 20 %o salinity
dilution water to provide 6800 mL of solution for each concentration. If
hypersaline brine alone (100 °/oo) is used to adjust the salinity of the
effluent, the highest concentration of effluent that could be achieved would
be 80% at 20 /oo salinity. When dry sea salts are used to adjust the
salinity of the effluent, it may be desirable to use a salinity control
prepared under the same conditions and used to determine survival and growth.
2The same procedures would be followed in preparing test concentrations at
other salinities between 20 °/oo and 30 °/oo: (1) the salinity of the bulk
(initial) effluent sample would be adjusted to the appropriate salinity using
artificial sea salts or hypersaline brine, and (2) the remaining effluent
concentrations would be prepared by serial dilution, using a large batch
(17000 mL) of seawater for dilution water, which had been prepared at the
same salinity as the effluent, using natural seawater, or hypersaline or
artificial sea salts and deionized water.
82
-------
13.1.1 Tabulate and summarize the data. A sample set of survival and
growth response data is listed in Table 4.
13.1.2 The endpoints of toxicity tests using the sheepshead minnow larvae
are based on the adverse effects on survival and growth. The LC50, the IC25,
and the IC50 are calculated using point estimation techniques (see Section 9,
Chronic Toxicity Test Endpoints and Data Analysis). LOEC and NOEC values, for
survival and growth, are obtained using a hypothesis test approach such as
Dunnett's Procedure (Dunnett, 1955) or Steel's Many-one Rank Test (Steel,
1959; Miller, 1981). See the Appendices for examples of the manual
computations, program listings, and examples of data input and program output.
13.1.3 The statistical tests described here must be used with a knowledge
of the assumptions upon which the tests are contingent. Tests for normality
and homogeneity of variance are included in Appendix B. The assistance of
a statistician is recommended for analysts who are not proficient in
statistics.
13.2 EXAMPLE OF ANALYSIS OF SHEEPHEAD MINNOW SURVIVAL DATA
13.2.1 Formal statistical analysis of the survival data is outlined in
Figure 2. The response used in the analysis is the proportion of animals
surviving in each test or control chamber. Separate analyses are performed
for the estimation of the NOEC and LOEC endpoints and for the estimation of
the LC50 endpoint. Concentrations at which there is no survival in any of the
test chambers are excluded from statistical analysis of the NOEC and LOEC, but
included in the estimation of the LC50 endpoint.
13.2.2 For the case of equal numbers of replicates across all
concentrations and the control, the evaluation of the NOEC and LOEC endpoints
is made via a parametric test, Dunnett's Procedure, or a nonparametric test,
Steel's Many-one Rank Test, on the arc sine transformed data. Underlying
assumptions of Dunnett's Procedure, normality and homogeneity of variance, are
formally tested. The test for normality is the Shapiro-Wilk's Test, and
Bartlett's Test is used to determine the homogeneity of variance. If either
of these tests fail, the nonparametric test, Steel's Many-one Rank Test, is
used to determine the NOEC and LOEC endpoints. If the assumptions of
Dunnett's Procedure are met, the endpoints are estimated by the parametric
procedure.
13.2.3 If unequal numbers of replicates occur among the concentration
levels tested, there are parametric and nonparametric alternative analyses.
The parametric analysis is a t-test with the Bonferroni adjustment. The
Wilcoxon Rank Sum Test with the Bonferroni adjustment is the nonparametric
alternative. For detailed information on the Bonferroni adjustment, see
Appendix D.
83
-------
TABLE 3 SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA FOR THE
SHEEPSHEAD MINNOW, CYPRINODON VARIEGATUS, LARVAL SURVIVAL AND GROWTH
TEST WITH EFFLUENTS AND RECEIVING WATERS
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
concentrations:
10. Age of test organisms:
11. Larvae/test chamber
12. Replicate
chambers/concentrati on:
13. Source of food:
14. Feeding regime:
15. Cleaning:
16. Aeration:
Static renewal
20 °/oo to 32 °/oo ± 2 °/oo
25 ± 1°C
Ambient laboratory illumination
10-20 \it/m2/s (50-100 ft-c) (ambient
lab levels)
16 h light, 8 h darkness
600 mL 1 L beakers or equivalent
500 - 750 mL/replicate (loading and
DO restrictions must be met)
Daily
Newly hatched larvae (less than
24 h old); 24-h range in age
15 larvae/chamber, (minimum of 10)
4 (minimum of 3)
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
None, unless DO falls below
4.0 mg/L, then aerate all
chambers. Rate should be
less than 100 bubbles/min
84
-------
TABLE 3. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA FOR THE
SHEEPSHEAD MINNOW, CYPRINODON VARIEGATUS, LARVAL SURVIVAL AND GROWTH
TEST WITH EFFLUENTS AND RECEIVING WATERS (CONTINUED)
17. Dilution water:
18. Test concentrations:
19. Dilution factor:
20. Test duration:'
21. Effects measured:
22. Test acceptability
criteria:
23. Sampling requirements:
24. Estimated maximum sample
volume required:
Uncontaminated source of natural
seawater, artificial seawater (GP2,
FORTY FATHOMS", or equivalent) or
hypersaline brine mixed with deionized
water
Effluents: Minimum of five effluent
concentrations and a control
Receiving waters: 100% receiving water and
a control
Effluents: > 0.5 series
Receiving waters: None, or > 0.5 series
7 days
Survival and growth (weight)
80% or greater survival in controls,
0.60 mg or greater average dry weight of
unpreserved control larvae, or 0.50 mg
or greater average dry weight of preserved
(4% formalin) control larvae
For on-site tests, samples are collected
daily, and used within 24 h of the time
they are are removed from the sampling
device. For off-site tests, a minimum of
three samples are collected on days one,
three, and five with a maximum holding time
of 36 h before first used (See Section 11,
Subsection 10.8.1).
6 L per day
85
-------
TABLE 4. SUMMARY OF SURVIVAL AND GROWTH DATA FOR SHEEPSHEAD MINNOW,
CYPRINODON VARIEGATUS, LARVAE EXPOSED TO AN EFFLUENT FOR SEVEN
DAYS1
Effl.
Cone.
(%)
0.0
6.25
12.5
25.0
50.0
100.0
Proportion of
Survival in Repl
Chambers
A
1.0
1.0
1.0
1.0
0.8
0.0
B
1.0
1.0
1.0
1.0
0.8
0.0
C
1.0
0.9
1.0
1.0
0.7
0.0
icate
D
1.0
1.0
1.0
0.8
0.6
0.0
Mean
Prop.
Surv
1
0
1
0
0
0
.00
.98
.00
.95
.73
.00
Avg Dry Wgt (mg) In Mean
Replicate Chambers Dry Wgt
A B C D (mg)
1.29
1.27
1.32
1.29
0.78
--
1.32
1.00
1.37
1.33
0.70
--
1.59
1.08
1.35
1.20
0.66
--
1.27
0.97
1.34
1.17
0.77
--
1.368
1.080
1.345
1.248
0.728
--
1Four replicates of 10 larvae each.
13.2.4 Probit Analysis (Finney, 1971) is used to estimate the LC50. In
this analysis, the total mortality data from all test replicates at a given
concentration are combined. If the data do not fit the Probit model, the
Spearman-Karber method, the trimmed Spearman-Karber method, or the Graphical
method may be used (USEPA, 1991c).
13.2.5 Example of Analysis of Survival Data
13.2.5.1 This example uses the survival data from the Sheepshead Minnow
Larval Survival and Growth Test. The proportion surviving in each replicate
must first be transformed by the arc sine square root transformation
procedure described in Appendix B. The raw and transformed data, means and
standard deviations of the transformed observations at each effluent
concentration and control are listed in Table 5. A plot of the survival
proportions is provided in Figure 3. Since there was 100% mortality in all
four replicates for the 100% concentration, it was not included in the
statistical analysis and was considered a qualitative mortality effect.
86
-------
STATISTICAL ANALYSIS OF SHEEPSHEAD MINNOW LARVAL
SURVIVAL AND GROWTH TEST
SURVIVAL
SURVIVAL DATA
PROPORTION SURVIVING
ENDPOINT ESTIMATE
LC50
ARC SINE
TRANSFORMATION
\
:
SHAPIRO-WILK'S TEST
NORMAL DISTRIBUTION
NON-NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
BARTLETT'S TEST
HETEROGENEOUS
VARIANCE
NO
1
EQUAL NUMBER OF
REPLICATES?
v
YES
EQUAL NUMBER OF
REPLICATES?
NO
YES
T-TEST WITH
BONFERRONI
ADJUSTMENT
DUNNETT'S
TEST
STEEL'S MANY-ONE
RANK TEST
WILCOXON RANK SUM
TEST WITH
BONFERRONI ADJUSTMENT
ENDPOINT ESTIMATES
NOEC.LOEC
Figure 2. Flowchart for statistical analysis of sheepshead
minnow, Cypn'nodon variegatus, larval survival data,
87
-------
TABLE 5. SHEEPSHEAD MINNOW, CYPRINODON VARIEGATUS, SURVIVAL DATA
Replicate Control
Effluent Concentration^)
6.25 12.5 25.0 50.0
RAW
ARC SINE
TRANSFORMED
MEAN(Y1)
Q
i
A
B
C
D
A
B
C
D
1.0
1.0
1.0
1.0
1.412
1.412
1.412
1.412
1.412
0.0
1
1.0
1.0
0.9
1.0
1.412
1.412
1.249
1.412
1.371
0.007
2
1.0
1.0
1.0
1.0
1.412
1.412
1.412
1.412
1.412
0.0
3
1.0
1.0
1.0
0.8
1.412
1.412
1.412
1.107
1.336
0.023
4
0.8
0.8
0.7
0.6
1.107
1.107
0.991
0.886
1.023
0.011
5
13.2.6 Test for Normality
13.2.6.1 The first step of the test for normality is to center the
observations by subtracting the mean of all observations within a
concentration from each observation in that concentration. The centered
observations are summarized in Table 6.
TABLE 6. CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S EXAMPLE
Effluent Concentration (%)
Replicate
Control
6.25
12.5
25.0
50.0
A
B
C
D
0.0
0.0
0.0
0.0
0.041
0.041
-0.122
0.041
0.0
0.0
0.0
0.0
0.076
0.076
0.076
-0.229
0.084
0.084
-0.032
-0.137
88
-------
CONNECTS THE MEAN VALUE FOR EACH CONCENTRATION
CO
UD
§
§
1.0
0.9
0.8 H
0.7
0.6 -
0.5 -
0.4
0.3
0.2
0.1 -
0.0 :
0
.00
6.25
12.50
EFFLUI:NT CONCENTRATION
25.00
50.00
Figure 3. Plot of mean survival proportion data in Table 5.
-------
13.2.6.2 Calculate the denominator, D, of the statistic:
D = I (X,- X)2
Where: X- = the ith centered observation
X = the overall mean of the centered observations
n = the total number of centered observations
13.2.6.3 For this set of data: n = 20
X = J_ (-0.001) = 0.000
20
D = 0.1236
13.2.6.4 Order the centered observations from smallest to largest
«(1) _ «(2) _ ^(n)
where X(O denotes the ith ordered observation. The ordered
observations for this example are listed in Table 7.
13.2.6.5 From Table 4, Appendix B, for the number of observations, n,
obtain the coefficients a^ a2, ... ak where k is n/2 if n is even and (n-l)/2
if n is odd. For the data in this example, n = 20 and k = 10. The a, values
are listed in Table 8.
TABLE 7. ORDERED CENTERED OBSERVATIONS FOR THE SHAPIRO-MILK'S EXAMPLE
i
1
2
3
4
5
6
7
8
9
10
xci>
-0.229
-0.137
-0.122
-0.032
0.0
0.0
0.0
0.0
0.0
0.0
i
11
12
13
14
15
16
17
18
19
20
X<»
0.0
0.0
0.041
0.041
0.041
0.076
0.076
0.076
0.084
0.084
90
-------
13.2.6.6 Compute the test statistic, W, as follows:
k
W = 1 [ S a, (X(n-'"+1) - X(i)) ]2
D i=l
The differences x(n"i+1) - X(1) are listed in Table 7. For the data
in this example,
W = _1 (0.3178)2 = 0.8171
0.1236
TABLE 8. COEFFICIENTS AND DIFFERENCES FOR SHAPIRO-WILK'S EXAMPLE
a,-
1
2
3
4
5
6
7
8
9
10
0.4734
0.3211
0.2565
0.2085
0.1686
0.1334
0.1013
0.0711
0.0422
0.0140
0.313
0.221
0.198
0.108
0.076
0.041
0.041
0.041
0.0
0.0
X(20)
X(19)
X(18)
X(17)
X06)
X(15)
X(H,
X(13)
X(12,
X(11)
- x<1>
- x(2)
- x<3)
- x(4)
X(5)
X(6)
- x(7>
- x(8)
- x(9)
- x<10)
13.2.6.7 The decision rule for this test is to compare W as calculated
in 14.2.6.6 to a critical value found in Table 6, Appendix B. If the
computed W is less than the critical value, conclude that the data are
not normally distributed. For the data in this example, the critical
value at a significance level of 0.01 and n = 20 observations is 0.868.
Since W = 0.817 is less than the critical value, conclude that the data
are not normally distributed.
13.2.6.8 Since the data do not meet the assumption of normality, Steel's
Many-one Rank Test will be used to analyze the survival data.
13.2.7 Steel's Many-one Rank Test
13.2.7.1 For each control and concentration combination, combine the
data and arrange the observations in order of size from smallest to
largest. Assign the ranks (1, 2, ..., 8) to the ordered observations
with a rank of 1 assigned to the smallest observation, rank of 2 assigned
91
-------
to the next larger observation, etc. If ties occur when ranking, assign
the average rank to each tied observation.
13.2.7.2 An example of assigning ranks to the combined data for the control
and 6.25% effluent concentration is given in Table 9. This ranking procedure
is repeated for each control/concentration combination. The complete set of
rankings is summarized in Table 10. The ranks are next summed for each
effluent concentration, as shown in Table 11.
TABLE 9. ASSIGNING RANKS TO THE CONTROL AND 6.25% EFFLUENT CONCENTRATION
FOR STEEL'S MANY-ONE RANK TEST
Rank
1
5
5
5
5
5
5
5
Transformed
Proportion
Surviving
1.249
1.412
1.412
1.412
1.412
1.412
1.412
1.412
Effluent
Concentration
(%)
6.25
6.25
6.25
6.25
Control
Control
Control
Control
TABLE 10. TABLE OF RANKS
Repli
cate
A
B
C
D
Control
1.412 (5,4.5,5,6.5)
1.412 (5,4.5,5,6.5)
1.412 (5,4.5,5,6.5)
1.412 (5,4.5,5,6.5)
Effluent Concentration (%)
6.25 12.5 25.0
1.412 (5)
1.412 (5)
1.249 (1)
1.412 (5)
1.412 (4.5)
1.412 (4.5)
1.412 (4.5)
1.412 (4.5)
1.412 (5)
1.412 (5)
1.412 (5)
1.107 (1)
50.0
1.107 (3.5)
1.107 (3.5)
0.991 (2)
0.886 (1)
92
-------
TABLE 11. RANK SUMS
Effluent Concentration (%) Rank Sum
6.25
12.5
25.0
50.0
16
18
16
10
13.2.7.3 For this example, we want to determine if the survival in any
of the effluent concentrations is significantly lower than the survival
in the control. If this occurs, the rank sum at that concentration would
be significantly lower than the rank sum of the control. Thus we are only
concerned with comparing the rank sums for the survival at each of the various
effluent concentrations with some "minimum" or critical rank sum, at or below
which the survival would be considered significantly lower than the control.
At a significance level of 0.05, the minimum rank sum in a test with four
concentrations (excluding the control) and four replicates is 10 (See Table 5,
Appendix E).
13.2.7.4 Since the rank sum for the 50% effluent concentration is equal to
the critical value, the proportion surviving in the 50% concentration is
considered significantly less than that in the control. Since no other rank
sums are less than or equal to the critical value, no other concentrations
have significantly lower proportion surviving than the control. Hence, the
NOEC and the LOEC are assumed to be the 25% and 50% concentrations,
respectively.
13.2.8 Probit Analysis
13.2.8.1 The data used for the Probit Analysis is summarized in Tables 12.
For the Probit Analysis, the 100% effluent concentration with 100% mortality
was included. To perform the Probit Analysis, run the USEPA Probit Analysis
Program. An example of the program input and output is supplied in Appendix
G. The current version of the Probit program makes no distinction between EC
and LC endpoints, and labels all results as EC values. Because the response
here is mortality, the EC values output by the program should be treated as
the corresponding LC values.
13.2.8.2 The computer output for the example is found in Table 13 and Figure
4. Since there is 100% survival in the controls, there is no need to adjust
for control mortality. The test for heterogeneity was significant, thus
confidence limits for the LC values could not be calculated. Probit Analysis
does not appear appropriate in this case.
93
-------
TABLE 12. DATA FOR PROBIT ANALYSIS
Effluent Concentration (%)
Control 6.25 12.5 25.0 50.0 100.0
Number Dead
Number Exposed
0
40
1
40
0
40
2
40
11
40
40
40
13.3 EXAMPLE OF ANALYSIS OF SHEEPSHEAD MINNOW GROWTH DATA
13.3.1 Formal statistical analysis of the growth data is outlined in
Figure 5. The response used in the statistical analysis is mean weight
per replicate. The IC25 and IC50 can be calculated for the growth data via a
point estimation technique (see Section 9, Chronic Toxicity Test Endpoints and
Data Analysis). Hypothesis testing can be used to obtain an NOEC for growth.
Concentrations above the NOEC for survival are excluded from the hypothesis
test for growth effects.
13.3.2 The statistical analysis using hypothesis testing consists of a
parametric test, Dunnett's Procedure, and a nonparametric test, Steel's
Many-one Rank Test. The underlying assumptions of the Dunnett's Procedure,
normality and homogeneity of variance, are formally tested. The test for
normality is the Shapiro-Wilk's Test and Bartlett's Test is used to test for
homogeneity of variance. If either of these tests fail, the nonparametric
test, Steels' Many-one Rank Test, is used to determine the NOEC and LOEC
endpoints. If the assumptions of Dunnett's Procedure are met, the endpoints
are determined by the parametric test.
13.3.3 Additionally, if unequal numbers of replicates occur among the
concentration levels tested there are parametric and nonparametric
alternative analyses. The parametric analysis is a t-test with the Bonferroni
adjustment. The Wilcoxon Rank Sum Test with the Bonferroni adjustment is the
nonparametric alternative. For detailed information on the Bonferroni
adjustment, see Appendix D.
94
-------
TABLE 13. OUTPUT FOR USEPA PROBIT ANALYSIS PROGRAM,
VERSION 1.4
EPA PROBIT ANALYSIS PROGRAM
USED FOR CALCULATING EC VALUES
Version 1.4
Probit Analysis of Sheepshead Minnow Larval Survival Data
Cone.
6.2500
12.5000
25.0000
50.0000
100.0000
Number
Exposed
40
40
40
40
40
Number
Resp.
1
0
2
11
40
Observed
Proportion
Responding
0.0250
0.0000
0.0500
0.2750
1.0000
Adjusted
Proportion
Responding
0.0250
0.0000
0.0500
0.2750
1.0000
Predicted
Proportion
Responding
0.0000
0.0035
0.0783
0.4462
0.8741
Chi - Square Heterogeneity » 734.142
WARNING
Significant heterogeneity exists. The results reported
for this data set may not be valid. The results should
be interpreted with appropriate caution.
NOTE
Slope not significantly different from zero.
EC fiducial limits cannot be computed.
Mu
Sigma
Parameter
1.730763
0.234925
Estimate
Std. Err.
95% Confidence Limits
Intercept
Slope
-2.367291
4.256671
15.651289
9.093670
( -52.169689,
( -24.679386,
Theoretical Spontaneous Response Rate = 0.0000
47.435108)
33.192726)
Estimated EC Values and Confidence Limits
Point
EC 1.00
EC 5.00
EC10.00
EC15.00
EC50.00
EC85.00
EC90.00
EC95.00
EC99.00
Cone.
15.2847
22.0971
26.8957
30.7105
53.7977
94.2410
107.6078
130.9761
189.3522
Lower Upper
95% Confidence Limits
95
-------
Probit Analysis of Sheepshead Minnow Larval Survival Data
PLOT OF ADJUSTED PROBITS AND PREDICTED REGRESSION LINE
Probiv
10+ °
8+
7+
6+
5+
.. o
3+0
0+0
-+ + + + + + +_
EC01 EC10 EC25 EC50 EC75 EC90 EC99
Figure 4. Plot of adjusted probits and predicted regression line
from USEPA Probit Program
96
-------
13.3.4 The data, mean and standard deviation of the observations at each
concentration including the control are listed in Table 14. A plot of
the mean weights for each treatment is provided in Figure 6. Since there
is no survival in the 100% concentration, it is not considered in the
growth analysis. Additionally, since there is significant mortality in
the 50% effluent concentration, its effect on growth is not considered.
TABLE 14. SHEEPSHEAD MINNOW, CYPRINODON VARIEGATUS, GROWTH DATA
Effluent Concentration (%)
Replicate Control
6.25
12.5
25.0 50.0 100.0
A
B
C
D
Mean(Y,-)
S?
1
1
1
1
1
1
0
1
.29
.32
.59
.27
.37
.0224
1.
0.
1.
0.
1.
0.
2
27
998
08
97
08
0183
1
1
1
1
1
0
3
.32
.37
.35
.34
.34
.0004
1
1
1
1
1
0
4
.29
.33
.20
.17
.25
.0056 -
5
-
-
-
.
-
6
13.3.5 Test for Normality
13.3.5.1 The first step of the test for normality is to center the
observations by subtracting the mean of all the observations within a
concentration from each observation in that concentration. The centered
observations are summarized in Table 15.
TABLE 15. CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S EXAMPLE
Replicate Control
Effluent Concentration (%)
6.25 12.5 25.0
A
B
C
D
-0.08
-0.05
0.22
-0.10
0.19
-0.08
0.00
-0.11
-0.02
0.03
0.01
-0.00
0.04
0.08
0.05
-0.08
97
-------
STATISTICAL ANALYSIS OF SHEEPSHEAD MINNOW LARVAL
SURVIVAL AND GROWTH TEST
GROWTH
GROWTH DATA
MEAN DRY WEIGHT
POINT ESTIMATION
HYPOTHESIS TESTING
(EXCLUDING CONCENTRATIONS
ABOVE NOEC FOR SURVIVAL)
ENDPOINT ESTIMATE
IC25, IC50
i
SHAPIRO-WILK'S TEST
NORMAL DISTRIBUTION
NON-NORMAL DISTRIBUTION
BARTLETT'S TEST
HOMOGENEOUS VARIANCE
HETEROGENEOUS
VARIANCE
NO
I
EQUAL NUMBER OF
REPLICATES?
YES
JL
EQUAL NUMBER OF
REPLICATES?
NO
YES
T-TESTWITH
BONFERRONI
ADJUSTMENT
DUNNETT'S
TEST
STEEL'S MANY-ONE
RANK TEST
WILCOXON RANK SUM
TEST WITH
BONFERRONI ADJUSTMENT
ENDPOINT ESTIMATES
NOEC.LOEC
Figure 5. Flowchart for statistical analysis of sheepshead minnow,
Cypn'nodon van'egatus, larval growth data.
98
-------
1.6 -
1.5 -
1.4 -_
1.3 :
1.2 :
_. 1.1
S j
& 1.0
§ 0.9
I: 0.8 H
fe n,
Q °'7
§ 0.6
0.5
0.4 H
0.3
0.2 H
0.1
0.0 H
0.00
CONNECTS MEAN VALUE FOR EACH CONCENTRATION
REPRESENTS THE CRITICAL VALUE FOR DUNNETTS TEST
(ANY PROPORTION BELOW THIS VALUE WOULD BE
SIGNIFICANTLY DIFFERENT FROM THE CONTROL)
6.25 12.50
EFFLUENT CONCENTRATION (%)
25.00
Figure 6. Plot of mean weight data from sheepshead minnow,
Cyprinodon van'egatus, larval survival and growth test.
-------
13.3.5.2 Calculate the denominator, D, of the test statistic:
D = I (X,. X)2
Where: Xs = the ith centered observation
X = the overall mean of the centered observations
n = the total number of centered observations.
For this set of data, n = 16
X = JL(-0.006) = 0.000
16
D = 0.1402
13.3.5.3 Order the centered observations from smallest to largest:
v(D V<2> Y
A ~ A ... A
Where X0) is the ith ordered observation. These ordered observations
are listed in Table 16.
TABLE 16. ORDERED CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S EXAMPLE
i
1
2
3
4
5
6
7
8
x<"
-0.11
-0.10
-0.08
-0.08
-0.08
-0.05
-0.05
-0.02
i
9
10
11
12
13
14
15
16
x«<>
-0.00
0.00
0.01
0.03
0.04
0.08
0.19
0.22
13.3.5.4 From Table 4, Appendix B, for the number of observations, n,
obtain the coefficients a,, a2, ..., ak where k is n/2 if n is even and (n-l)/2
if n is odd. For the data in this example, n = 16, k = 8. The a- values are
listed in Table 17. '
13.3.5.5 Compute the test statistic, W, as follows:
k
W = I [ S a, (X(n-'"+1) X(i)) ]2
D i = l
100
-------
The differences x(n"i+1) - X
X(15)
X(H)
x
X(12)
XC1D
xcio>
X(9)
- X<1)
- X(2)
- x(3)
- x(4)
- x(5)
- x(6)
x<7)
- x(8)
13.3.6 Test for Homogeneity of Variance
13.3.6.1 The test used to examine whether the variation in mean dry
weight is the same across all effluent concentrations including the
control, is Bartlett's Test (Snedecor and Cochran, 1980). The test
statistic is as follows:
P _ P ,
[ ( S V,) In S2 E V, In S,.2 ]
Where: V,- = degrees of freedom for each effluent concen-
tration and control, V,- = (n,- 1)
101
-------
p = number of levels of effluent concentration
including the control
-2 ( S V, S,-2)
S2 = 1-1
p
S Vf
1=1
C = 1 + ( 3(p-l))-1 [ E 1/V, ( S V,)'1 ]
i=l 1=1
Where:
In = loge
i = 1, 2, ..., p where p is the number of concentrations
including the control
n( = the number of replicates for concentration i.
13.3.6.2 For the data in this example, (See Table 14) all effluent
concentrations including the control have the same number of replicates
(n, = 4 for all i). Thus, V, - 3 for all i.
13.3.6.3 Bartlett's statistic is therefore:
B = [(12)ln(0.0117) 3 E ln(S?) 1/1.139
i = l
= [12(-4.4482) - 3(-20.809)]/1.12
= 9.049/1.139
= 7.945
13.3.6.4 B is approximately distributed as chi square with p 1 degrees of
freedom, when the variances are in fact the same. Therefore, the appropriate
critical value for this test, at a significance level of 0.01 with three
degrees of freedom, is 11.345. Since B = 7.945 is less than the critical
value of 11.345, conclude that the variances are not different.
13.3.7 Dunnett's Procedure
13.3.7.1 To obtain an estimate of the pooled variance for the Dunnett's
Procedure, construct an ANOVA table as described in Table 18.
102
-------
TABLE 18. ANOVA TABLE
Source
Between
Within
Total
Where: p
N
df
P - 1
N - p
N 1
= number
- total
Sum of Squares
(SS)
SSB
SSW
SST
effluent concentrations
number of observations n.
Mean Square(MS)
(SS/df)
S2 = SSB/(p-l)
Sy = SSW/(N-p)
including the control
+ n, ... +n,.
ni
= number of observations in concentration i
SSB
Z T,2/n, - G2/N
Between Sum of Squares
p ni
SST = Z Z Y2, - G2/N
-i < J
Total Sum of Squares
SSW = SST - SSB
Within Sum of Squares
G = the grand total of all sample observations, G = z T,-
i=l
T,- = the total of the replicate measurements for
concentration "i"
YJ- = the jth observation for concentration "i" (represents
the mean dry weight of the fish for effluent
concentration i in test chamber j)
13.3.7.2 For the data in this example:
N
T,
T
n, = n, = n, = 4
16
T, =
12
13
*
Y14 = 5.47
Y24 = 4.32
, 3 V* = 5.38
Y + Y" + Y« + Yl! = 4.99
*
"
'42
r3
'43
'44
T, + T2 + T3
20.16
SSB = Z T,2/n, - G2/N
103
-------
J_(102.43) - (20.158)2 = 0.21
4 16
SST = S S'Y?, - G2/N
p
SMfJ
i-1 j=l
= 25.74 (20.16)2 - 0.34
16
SSW = SST - SSB = 0.34 - 0.21 = 0.130
SB = SSB/(p-l) = 0.21/(4-l) = 0.070
Su = SSW/(N-p) = 0.13/(16-4) = 0.01
13.3.7.3 Summarize these calculations in the ANOVA table (Table 19)
TABLE 19. ANOVA TABLE FOR DUNNETT'S PROCEDURE EXAMPLE
Source
Between
Within
Total
df
3
12
15
Sum of Squares
(SS)
0.21
0.13
0.34
Mean Square(MS)
(SS/df)
0.07
0.01
13.3.7.4 To perform the individual comparisons, calculate the t
statistic for each concentration, and control combination as follows:
*i
( Y, - Y, )
Su
Where: Y, = mean dry weight for effluent concentration i
Y! = mean dry weight for the control
Su = square root of within mean square
n1 = number of replicates for control
n,- = number of replicates for concentration i.
13.3.7.5 Table 20 includes the calculated t values for each
104
-------
concentration and control combination. In this example, comparing the
6.25% concentration with the control, the calculation is as follows:
( 1.37 1.08 )
t2 = = 4.10
[ 0.10 / (1/4) + (1/4) ]
TABLE 20. CALCULATED T-VALUES
Effluent Concentration(%) i t1
6.25
12.5
25.0
2
3
4
4.10
0.42
1.70
13.3.7.6 Since the purpose of this test is to detect a significant reduction
in mean weight, a one-sided test is appropriate. The critical value for this
one-sided test is found in Table 5, Appendix C. For an overall alpha level of
0.05, 12 degrees of freedom for error and three concentrations (excluding the
control) the critical value is 2.29. The mean weight for concentration "i" is
considered significantly less than the mean weight for the control if t- is
greater than the critical value. Since t2 is greater than 2.29, the 6.25%
concentration has significantly lower growth than the control. However, the
12.5% and 25% concentrations do not exhibit this effect. Hence the NOEC and
the LOEC for growth cannot be calculated.
13.3.7.7 To quantify the sensitivity of the test, the minimum significant
difference (MSD) that can be statistically detected may be calculated.
MSD = d Su / (1/n,) + (1/n)
Where: d = the critical value for the Dunnett's procedure
Su = the square root of the within mean square
n = the common number of replicates at each concentration
(this assumes equal replication at each concentration)
n1 = the number of replicates in the control.
13.3.7.8 In this example:
MSD = 2.29 (0.10) / (1/4) + (1/4)
= 2.29 (0.10)(0.707)
= 0.16
13.3.7.9 Therefore, for this set of data, the minimum difference that can be
detected as statistically significant is 0.16 mg.
105
-------
13.3.7.10 This represents a 12% reduction in mean weight from the control.
13.3.8 Calculation of the 1C
13.3.8.1 The growth data in Table 4 are utilized in this example. As can be
seen, the observed means are not monotonically non-increasing with respect to
concentration. Therefore, it is necessary to smooth the means prior to
calculating the_IC. In the following discussion, the observed means are
represented by Y,- and the smoothed means by Mr
13.3.8.2 Starting with_the control mean, Y1 = 1.368 and Y2 = 1.080, we see
that YT > Y2. Set M, = Yr Comparing Y2 to Y3 = 1.345, Y2 < Y3.
14.3.8.3 Calculate the smoothed means:
M2 = M3 = (Y2 + Y3)/2 = 1.2125
13.3.8.4 Since Y4 = 1.247 is larger than M3, average Y4 with the previous
concentrations:
M2 = M3 = M4 = (M2 + M3 + YJ/3 = 1.224
13.3.8.5 Since Y5 = 0.728 < M4, set M. = 0.728. Table 21 contains the
smoothed means and Figure 7 gives a plot of the smoothed response curve.
TABLE 21. SHEEPSHEAD MINNOW, CYPRINODON VARIEGATUS
MEAN GROWTH RESPONSE AFTER SMOOTHING
Effluent
Cone.
(%)
Control
6.25
12.50
25.00
50.00
i
1
2
3
4
5
MI
(mg)
1.368
1.224
1.224
1.224
0.728
13.3.8.6 An IC25 and IC50 can be estimated using the Linear Interpolation
Method. A 25% reduction in mean growth, compared to the controls, would
result in a mean dry weight of 1.026 mg, where M^l-p/100) = 1.368(1-25/100).
A 50% reduction in mean growth, compared to the controls, would result in a
mean dry weight of 0.684 mg. Examining the means and their associated
concentrations (Table 4), the response, 1.026 mg, is bracketed by C, = 25.0%
effluent and C5 = 50.0% effluent. The response (0.728 mg) at the highest
106
-------
effluent concentration (50.0%) is greater than 50% of the control (0.684 mg)
Thus, the IC50 is specified as greater than 50% effluent.
13.3.8.7 Using Equation 1 from Appendix J, the estimate of the IC25 is
calculated as follows:
ICp = C, + [M,(l p/100) MJ
(MJ+1 Md)
IC25 = 25.0 + [1.368(1 - 25/100) 1.224] (50.00 25.00)
(0.728 1.224)
= 35.0%.
13.3.8.8 When the Bootstrap program (BOOTSTRP) was used to analyze this set
of data, requesting 80 resamples, the mean estimate of the IC25 was 34.9788%,
with a standard deviation of 2.7670% (coefficient of variation = 12.7%). The
empirical 95% confidence interval for the true mean was (29.5307%, 39.7882%).
The BOOTSTRP computer program output for the IC25 for this data set is shown
in Figure 8.
13.3.8.9 When the Bootstrap program (BOOTSTRP) was used to analyze this set
of data for the IC50, requesting 80 resamples, the output indicated that none
of the concentration response means were less than 50% of the control. Thus,
the IC50 could not be calculated. The BOOTSTRP computer program output is
shown in Figure 9.
107
-------
o
00
1.8
1.7
1.6
1.5
1.2 -
1.0 :
0.9 ^
0.8 ^
0.7 ^
0.6 :
INDIVIDUAL REPLICATE MEAN WEIGHT
CONNECTS THE OBSERVED MEAN VALUE
CONNECTS THE SMOOTHED MEAN VALUE
0.00
6.25 12.50 25.00
EFFLUENT CONCENTRATION (%)
50.00
Figure 7. Plot of raw data, observed means, and smoothed means for the sheepshead
minnow, Cyprinodon van'egatus, growth data from Tables 4 and 21.
-------
THE NUMBER OF RESAMPLES IS 80
*** LISTING OF GROUP CONCENTRATIONS (% EFF=) AND RESPONSE MEANS ***
CONC. (%EFF) RESPONSE MEAN MEAN AFTER POOLING
0.000 1.368 1.368
6.250 1.080 1.224
12.500 1.345 1.224
25.000 1.247 1.224
50.000 0.728 0.728
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
FROM THE INPUT SAMPLE IS 34.9937.
************************************************************
* BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY *
* OF THE ESTIMATED ICp *
************************************************************
THE MEAN OF THE BOOTSTRAP ESTIMATES IS 34.9788.
THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS 2.7670.
AN EMPIRICAL 95.0% CONFIDENCE INTERVAL FOR THE
BOOTSTRAP ESTIMATE IS ( 29.5307, 39.7882).
Figure 8. BOOTSTRP program output for the IC25.
109
-------
THE NUMBER OF RESAMPLES IS 80
*** LISTING OF GROUP CONCENTRATIONS (% EFF.) AND RESPONSE MEANS ***
CONC. (%EFF) RESPONSE MEAN MEAN AFTER POOLING
0.000 1.368 1.368
6.250 1.080 1.224
12.500 1.345 1.224
25.000 1.247 1.224
50.000 0.728 0.728
*** NO LINEAR INTERPOLATION ESTIMATE CAN BE CALCULATED FROM THE INPUT
DATA, SINCE NONE OF THE (POSSIBLY POOLED) GROUP RESPONSE MEANS
WERE LESS THAN 50.0% OF THE CONTROL RESPONSE MEAN.
************************************************************
* BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY *
* OF THE ESTIMATED ICp *
THE MEAN OF THE BOOTSTRAP ESTIMATES IS 49.0466.
THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS .6845.
AN EMPIRICAL ****% CONFIDENCE INTERVAL FOR THE
BOOTSTRAP ESTIMATE IS ( 47.9205, 49.9405).
*** NOTE: THE ABOVE BOOTSTRAP CALCULATIONS WERE BASED ON 10
INSTEAD OF 80 RESAMPLINGS. THOSE RESAMPLES NOT
USED HAD ESTIMATES ABOVE THE HIGHEST CONCENTRATION / % EFF.
Figure 9. BOOTSTRP program output for the IC50.
110
-------
14. PRECISION AND ACCURACY
14.1 PRECISION
14.1.1 Data on the single-laboratory precision of the sheepshead minnow
larval survival and growth test using FORTY FATHOMSR artificial seawater,
natural seawater, and GP2 with copper sulfate, sodium dodecyl sulfate, and
hexavalent chromium, as reference toxicants, are given in Tables 22-27. The
IC25, IC50, or LC50 data (coefficient of variation), indicating acceptable
precision for the reference toxicants (copper, sodium dodecyl sulfate, and
hexavalent chromium), are also listed in these Tables.
14.1.2 Data from a study of multilaboratory test precision, involving a total
of seven tests by four participating laboratories, are listed in Table 28. The
laboratories reported very similar results, indicating good interlaboratory
precision. The coefficient of variation (IC25) was 44.2% and (IC50) was
56.9%, indicating acceptable precision.
14.2 ACCURACY
14.2.1 The accuracy of toxicity tests cannot be determined.
Ill
-------
TABLE 22. SINGLE-LABORATORY PRECISION OF THE SHEEPSHEAD MINNOW, CYPRINODON
VARIEGATUS, LARVAL SURVIVAL AND GROWTH TEST PERFORMED IN FORTY
FATHOMS" ARTIFICIAL SEAWATER, USING LARVAE FROM FISH MAINTAINED
AND SPAWNED IN FORTY FATHOMS* ARTIFICIAL SEAWATER, USING COPPER
(CU) AS A REFERENCE TOXICANT1'2'3'4'5
Test NOEC
Number (n9/L)
1 50
2 <50*
3 <50*
4 50
5 <50*
6 50
7 50
8 50
n: 5
Mean: NA
CV(%): NA
IC25
(H9/L)
0.1133
0.0543
0.0418
0.0632
0.0577
0.0483
0.0796
0.1235
8
0.0727
41.82
IC50
(M/L)
0.1523
0.0975
0.0714
0.0908
0.0998
0.1325
0.1597
0.2364
8
0.1300
40.87
Most
Sensitive
Endpoint
S
G
G
S
S
G
G
G
1Data from USEPA (1989a) and USEPA (1991a).
2Tests performed by Donald J. Klemm, Bioassessment and Ecotoxicology Branch,
Newtown Facility, Environmental Monitoring Systems Laboratory - Cincinnati.
All tests were performed using Forty Fathoms synthetic seawater.
Three replicate exposure chambers, each with 15 larvae, were used for
the control and each copper concentration. Copper concentrations
used in Tests 1-6 were: 50, 100, 200, 400, and 800 \ig/L. Copper
concentrations in Tests 7-8 were: 25, 50, 100, 200 and 400 tig/L.
4Adults collected in the field.
5For a discussion of the precision of data from chronic toxicity
tests see Section 4, Quality Assurance.
*Lowest concentration tested was 0.05 mg/L (NOEC Range: >50* - 50 jig/L).
112
-------
TABLE 23. SINGLE-LABORATORY PRECISION OF THE SHEEPSHEAD MINNOW, CYPRINID'ON
VARIEGATUS, LARVAL SURVIVAL AND GROWTH TEST PERFORMED IN FORTY
FATHOMSR ARTIFICIAL SEAWATER. USING LARVAE FROM FISH MAINTAINED
AND SPAWNED IN FORTY FATHOMS* ARTIFICIAL SEAWATER, USING SODIUM
DODECYL SULFATE (SDS) AS A REFERENCE TOXICANT1'2'3'4'5'6
Test
Number
1
2
3
4
5
6
n:
Mean:
CV(%):
NOEC
(mg/L)
1.0
1.0
1.0
0.5
1.0
0.5
6
NA
NA
IC25
(mg/L)
1.2799
1.4087
2.3051
1.9855
1.1901
1.1041
6
1.5456
31.44
IC50
(mg/L)
1.5598
1.8835
2.8367
2.6237
1.4267
1.4264
6
1.9595
31.82
Most
Sensitive
Endpoint
S
S
S
G
S
G
from USEPA (1989a) and USEPA (1991a).
2Tests performed by Donald J. Klemm, Bioassessment nd Ecotoxicology Branch,
Newtown Facility, Environmental Monitoring Systems Laboratory - Cincinnati
3A11 tests were performed using Forty Fathoms synthetic seawater.
Three replicate exposure chambers, each with 15 larvae, were used for
the control and each SDS concentration. SDS concentrations in
Tests 1-2 were: 1.0, 1.9, 3.9, 7.7, and 15.5 mg/L. SDS concentrations
in Tests 3-6 were: 0.2, 0.5, 1.0, 1.9, and 3.9 mg/L.
4Adults collected in the field.
5For a discussion of the precision of data from chronic toxicity
tests see Section 4, Quality Assurance.
6NOEC Range: 0.5 -1.0 mg/L (this represents a difference of one exposure
concentration).
113
-------
TABLE 24. SINGLE-LABORATORY PRECISION OF THE SHEEPSHEAD MINNOW, CYPRINIDON
VARIEGATUS, LARVAL SURVIVAL AND GROWTH TEST PERFORMED IN NATURAL
SEAWATER, USING LARVAE FROM FISH MAINTAINED AND SPAWNED IN NATURAL
SEAWATER, USING COPPER (CU) SULFATE AS A REFERENCE TOXICANT1'2'3'4'5-*
Test
Number
1
2
3
4
5
n:
Mean:
CV(%):
NOEC
(H9/L)
125
31
125
125
125
5
NA
NA
IC25
(H9/L)
320.3
182.3
333.4
228.4
437.5
5
300.4
33.0
IC50
(ugA)
437.5
323.0
484.4
343.8
NC*
4
396.9
19.2
Most
Sensitive
Endpoint
S
G
S
S
S
1Data from USEPA (1989a) and USEPA (1991a).
Tests performed by George Morrison and Elise Torello, Environmental
Research Laboratory, U. S. Environmental Protection Agency,
Narragansett, Rhode Island.
3Three replicate exposure chambers, each with 10-15 larvae, were
used for the control and each copper concentration. Copper
concentrations were: 31, 63, 125, 250, and 500 ng/L.
4NOEC Range: 31 - 125 jig/L (this represents a difference of two exposure
concentrations).
5Adults collected in the field.
6For a discussion of the precision of data from chronic toxicity
tests see Section 4, Quality Assurance.
*No linear interpolation estimate could be calculated from the data, since
none of the group response means were less than 50 percent of the control
response mean.
114
-------
TABLE 25. SINGLE-LABORATORY PRECISION OF THE SHEEPSHEAD MINNOW, CYPRINIDON
VARIEGATUS, LARVAL SURVIVAL AND GROWTH TEST PERFORMED IN NATURAL
SEAWATER, USING LARVAE FROM FISH MAINTAINED AND SPAWNED IN NATURAL
SEAWATER. USING SODIUM DODECYL SULFATE (SDS) AS A REFERENCE
TOXICANT1'2'3'4-5'6
Test NOEC
Number (mg/L)
1 2.5
2 1.3
3 1.3
4 1.3
5 1.3
n: 5
Mean: NA
CV(%): NA
IC25
(mg/L)
2.9
NCI
1.9
2.4
1.5
4
2.2
27.6
IC50
(mg/L)
3.6
NC2
2.4
NC2
1.8
3
2.6
35.3
Most
Sensitive
Endpoint
S
G
S
G
S
1Data from USEPA (1989a) and USEPA (1991a).
2Tests performed by George Morrison and Elise Torello, Environmental
Research Laboratory, U. S. Environmental Protection Agency,
Narragansett, Rhode Island.
3Three replicate exposure chambers, each with 10-15 larvae, were
used for the control and each SDS concentration. SDS concentrations
were: 0.3, 0.6, 1.3, 2.5, and 5.0 mg/L.
4NOEC Range: 1.3 - 2.5 mg/L (this represents a difference of one exposure
concentration).
5Adults collected in the field.
6For a discussion of the precision of data from chronic toxicity
tests see Section 4, Quality Assurance.
NCI = No linear interpolation estimate could be calclulated from the data,
since none of the group response means were less than 75 percent of
the control response mean.
NC2 = No linear interpolation estimate could be calculated from the data,
since none of the group response means were less than 50 percent of
the control response mean.
115
-------
TABLE 26. SINGLE-LABORATORY PRECISION OF THE SHEEPSHEAD MINNOW,
CYPRINODON VARIEGATUS, LARVAL SURVIVAL AND GROWTH TEST
PERFORMED IN FORTY FATHOMS" ARTIFICIAL SEAWATER, USING
LARVAE FROM FISH MAINTAINED AND SPAWNED IN FORTY FATHOMS"
ARTIFICIAL SEAWATER, AND HEXAVALENT CHROMIUM AS THE
REFERENCE TOXICANT1'*'3'4'5
Test NOEC
Number (mg/L)
1 2.0
2 1.0
3 4.0
4 2.0
5 1.0
n: 5
Mean: NA
CV(%): NA
IC25
(mg/L)
5.8
2.9
6.9
2.4
3.1
5
4.2
47.6
IC50
(mg/L)
11.4
9.9
11.5
9.2
10.8
5
10.6
9.7
Most
Sensitive
Endpoint
G
G
G
G
G
1Tests performed by Donald J. Klemm, Bioassessment and Ecotoxioology Branch,
Newtown Facility, Environmental Monitoring Systems Laboratory - Cincinnati.
2A11 tests were performed using Forty Fathoms synthetic seawater.
Three replicate exposure chambers, each with 15 larvae, were used for
the control and each hexavalent chromium concentration. Hexavalent
chromium concentrations used in Tests 1-5 were: 1.0, 2.0, 4.0, 8.0, 16.0,
and 32.0 mg/L.
3NOEC Range: 1.0 - 4.0 mg/L (this represents a difference of four exposure
concentrations).
4Adults collected in the field.
5For a discussion of the precision of data from chronic toxicity
tests see Section 4, Quality Assurance.
116
-------
TABLE 27. COMPARISON OF LARVAL SURVIVAL (LC50) AND GROWTH (IC50) VALUES FOR
THE SHEEPSHEAD MINNOW, CYPRINODON VARIEGATUS, EXPOSED TO SODIUM
DODECYL SULFATE (SDS) AND COPPER (CU) SULFATE IN GP2 ARTIFICIAL
SEAWATER MEDIUM OR NATURAL SEAWATER1'^'3'4
Survival Growth
SDS (mg/L) GP2 NSW GP2 NSW
Mean
CV (%)
7.49
8.70
8.38
8.19
7.7
8.13
8.87
8.85
8.62
4.9
7.39
8.63
8.48
8.17
8.3
8.41
8.51
9.33
8.75
5.8
COPPER (ng/L) GP2 NSW GP2 NSW
Mean
CV (%)
455
467
390
437
9.4
412
485
528
475
12.3
341
496
467
435
18.9
333
529
776
546
40.7
1Tests performed by George Morrison and Glen Modica, Environmental
Research Laboratory, U. S. Environmental Protection Agency,
Narragansett, Rhode Island.
2Three replicate exposure chambers, each with 10-15 larvae, were
used for the control and each SDS concentration. SDS concentrations
were: 0.3, 0.6, 1.3, 2.5, and 5.0 mg/L.
3Adults collected in the field.
4For a discussion of the precision of data from chronic toxicity
tests see Section 4, Quality Assurance.
117
-------
TABLE 28. DATA FROM INTERLABORATORY STUDY OF THE SHEEPSHEAD MINNOW,
CYPRIHODON VARIEGATUS, LARVAL SURVIVAL AND GROWTH TEST USING
AN INDUSTRIAL EFFLUENT AS A REFERENCE TOXICANT1'2'3'4
Test
Number
Most Sensitive Endpoint
NOEC
IC25
IC50
Laboratory A
Laboratory B
Laboratory C
Laboratory D
1
2
1
2
3.2 (S,G)
3.2 (S,G)
3.2 (S,G)
3.2 (S,G)
1.0 (S)
7.4 (S)
7.6 (S)
5.7 (G)
5.7 (G)
4.7 (S)
1Data from USEPA (1987d), USEPA (1989a), and USEPA (1991a).
Affluent concentrations were: 0.32, 1.0, 3.2, 10.0, and 32.0%.
3NOEC Range: 1.0 - 3.2 percent (this represents a difference of
one exposure concentration).
4Endpoints: G = growth; S = survival.
7.4 (G)
14.3 (G)
9.7 (G)
8.8 (G)
7.2 (S)
n:
Mean:
CV(%):
1 3.2 (S,G)
2 1.0 (G)
7
NA
NA
7.4 (G)
5.2 (S)
7
5.5
44.2
24.7 (G)
7.2 (S)
7
11.3
56.9
118
-------
Figure 10. Data forms for sheepshead minnow, Cypn'nodon variegatus, larval survival and growth
test. Daily record of larval survival and test conditions.
Test Dates:
Type Effluent:
Species:
Field
Lab
Test
Effluent Tested:
CONCENTRATION:
REPLICATE:
DAYS
LIVE
LARVAE
TEMP
I"C|
SALINITY
(°-'.J
DO
(mg/l|
LARVAE'
DRY WT
0
1
2
3
4
5
MEAN WEIGHT/
LARVAE Imgl ISO
6
7
REPLICATE:
0
1
LARVAE
DRY WT
2
3
4
5
6
MEAN WEIGHT
LARVAE Imgl > S D
7
REPLICATE:
0
1
ULAHVAE
DRV WT
2
3
4
5
6
7
REPLICATE:
0
1
1LARVAE
DRV WT
2
3
4
5
6
MEAN WEIGHT
LARVAE (mg| i S D
7
CONCENTRATION:
LIVE
LARVAE
TEMP
<°CI
SALINITY
l°'~l
DO
|mg/l)
LARVAE/
our WT
MEAN WEIGHT/
LARVAE |mg| t S D
LARVAE
DRY WT
MEAN WEIGHT
LARVAE Imgl ' S D
LARVAE
DRV WT
LARVAE
DHV WT
MEAN WEIGHT
LARVAE Imgl I SO
CONCENTRATION:
LIVE
LARVAE
TEMP
PC)
SALINITY
(° -1
DO
Img/ll
LARVAE/
DRY WT
MEAN WEIGHT/
LARVAE Imgl t S O
LARVAE
DRV WT
MEAN WEIGHT
LARVAE Imgl I S D
LARVAE
DRV WT
LARVAE'
DRY WT
MEAN WEIGHT-
LARVAE Imgl 1 S D
COMMENTS:
TIME
FED
Adapted from USEPA (19875)
-------
Figure 10. Data forms for sheepshead minnow, Cyprinodon variegaius, larval survival and growth
test. Daily record of larval survival and test conditions.1 Continued.
Test Dates:
Type Effluent:
Species:
Field
Lab
Test
Effluent Tested.
CONCENTRATION:
REPLICATE:
DAYS
LIVE
LARVAE
TEMP
(°CI
SALINITY
("'-I
DO
-------
Test Dates:
Figure 11. Data forms for sheepshead minnow, Cyprinodon variegatus,^
larval survival and growth test. Dry weights of larvae.
Species:
Pan
Cone.
&
Rep.
Initial
Wt.
(mg)
Final
Wt.
(mg)
Diff.
(mg)
#
Larvae
Av. Wt./
Larvae
(mg)
'Adapted from USEPA (1987b).
121
-------
Figure 12. Data forms for shee'pshead minnow, Cyprinodon variegatus,
larval survival and growth test. Summary of test results.1
Test Dates:
Effluent Tested:
Species:
TREATMENT
# LIVE
LARVAE
SURVIVAL
(%)
MEAN DRY WT /
LARVAE (mg)
±S.D.
SIGNIF. DIFF.
FROM CONTROL
(0)
MEAN
TEMPERATURE
(oC)
±S.D.
MEAN SALINITY
0/00
±S.D.
AV. DISSOLVED
OXYGEN
(mg./L) ±S.D.
COMMENTS:
Adapted from USEPA (1987b).
122
-------
SECTION 12
TEST METHOD1'2
SHEEPSHEAD MINNOW, CYPRINODON VARIEGATUS,
EMBRYO-LARVAL SURVIVAL AND TERATOGENICITY TEST
METHOD 1005
1. SCOPE AND APPLICATION
1.1 This method estimates the chronic toxicity of effluents and receiving
waters to the sheepshead minnow, Cyprinodon variegatus, using embryos and
larvae in a nine-day, static renewal test. The effects include the
synergistic, antagonistic, and additive effects of all the chemical, physical,
and biological components which adversely affect the physiological and
biochemical functions of the test organisms. The test is useful in screening
for teratogens because organisms are exposed during embryonic development.
The test has several advantages over the larval growth test because feeding is
not required and the larvae are not dried and weighed.
1.2 Daily observations on mortality make it possible to also calculate acute
toxicity for desired exposure periods (i.e., 24-h, 48-h, 96-h LCSOs).
1.3 Detection limits of the toxicity of an effluent or chemical substance are
organism-dependent.
1.4 Single or multiple excursions in toxicity may not be detected using 24-h
composite samples. Also, because of the long sample collection period
involved in composite sampling, and because the test chambers are not sealed,
highly volatile and highly degradable toxicants in the source may not be
detected in the test.
1.5 This test is commonly used in one of two forms: (1) a definitive test,
consisting of a minimum of five effluent concentrations and a control, and (2)
a receiving water test(s), consisting of one or more receiving water
concentrations and a control.
2. SUMMARY OF METHOD
2.1 Sheepshead minnow embryos and larvae are exposed in a static renewal
system, from shortly after'fertilization of the eggs through four days
posthatch (total of nine days), to different concentrations of effluent or to
receiving water. Test results are based on the total frequency of both
mortality and gross morphological deformities (terata).
3. INTERFERENCES
1The format used for this method was taken from USEPA (1983).
2This method was adapted from materials provided by Terry Hoi lister, USEPA,
Region 6 Laboratory, Houston, Texas, and from USEPA (1981e), USEPA (1985e),
and USEPA (1987b).
123
-------
3.1 Toxic substances may be introduced by contaminants in dilution water,
glassware, sample hardware, and testing equipment (see Section 5, Facilities,
Equipment, Supplies).
3.2 Adverse effects of low dissolved oxygen concentrations (DO), high
concentrations of suspended and/or dissolved solids, and extremes of pH may
mask the effect of toxic substances.
3.3 Improper effluent sampling and handling may adversely affect test results
(see Section 8, Effluent and Receiving Water Sampling, Sample Handling, and
sample Preparation for Toxicity Tests).
3.4 Pathogenic and/or predatory organisms in the dilution water and effluent
may affect test organism survival, and confound test results.
4. SAFETY
4.1 See Section 3, Health and Safety.
5. APPARATUS AND EQUIPMENT
5.1 Facilities for holding and acclimating test organisms.
5.2 Sheepshead minnow culture unit -- see Subsection 6.13 below. To perform
toxicity tests on-site or in the laboratory, sufficient numbers of newly
fertilized eggs must be available, preferably from an inhouse sheepshead
minnow culture unit. If necessary, embryos can be obtained from outside
sources if shipped in well oxygenated water in insulated containers.
5.2.1 A test using 15 embryos per test vessel and four replicates per
concentration, will require 360 newly-fertilized embryos at the start of the
test (Table 5). A test with a minimum of 10 embryos per test vessel and three
replicates per concentration, and with five effluent concentrations and a
control, will require a minimum of 180 embryos at the start of the test.
5.3 Brine shrimp, Artemia, culture unit -- for feeding sheepshead minnow
larvae in the continuous culture unit (see Subsection 6.12 below).
5.4 Samplers -- automatic sampler, preferably with sample cooling capability,
that can collect a 24-h composite sample of 5 L, and maintain sample
temperature at 4°C.
5.5 Environmental chamber or equivalent facility with temperature control
(25 ± 1°C).
5.6 Water purification system -- Millipore Milli-QR, deionized water (DI) or
equivalent.
5.7 Balance -- analytical, capable of accurately weighing to 0.0001 g.
Note: An analytical balance is not needed for this test but is needed for
other specified toxicity test methods with growth endpoints.
124
-------
5.8 Reference weights, Class S -- for checking the performance of the
balance. The reference weights should bracket the expected weights of
reagents, and the expected weights of the weighing boats and the weights of
the weighing boats plus larvae, used in Artemia suitability studies.
5.9 Air pump -- for oil free air supply.
5.10 Air lines, and air stones -- for aerating water containing embryos,
larvae, or supplying air to test solution with low DO.
5.11 pH and DO meters -- for routine physical and chemical measurements.
Unless the test is being conducted to specifically measure the effect of one
of the above parameters, a portable, field-grade instrument is acceptable.
5.12 Standard or micro-Winkler apparatus -- for determining DO (optional).
5.13 Dissecting microscope -- for examining embryos and larvae.
5.14 Light box -- for counting and observing embryos and larvae.
5.15 Refractometer -- for determining salinity.
5.16 Thermometers, glass or electronic, laboratory grade -- for measuring
water temperatures.
5.17 Thermometers, bulb-thermograph or electronic-chart type -- for
continuously recording temperature.
5.18 Thermometer, National Bureau of Standards Certified (see USEPA METHOD
170.1, USEPA, 1979b) -- to calibrate laboratory thermometers.
5.19 Test chambers -- four (minimum of three), borosilicate glass or
non-toxic plastic labware per test concentration. The chambers should be
covered during the test to avoid potential contamination from the air. Care
must be taken to avoid inadvertently removing embryos or larvae when test
solutions are decanted from the chambers. The covers are removed only for
observation and removal of dead organisms.
5.20 Beakers -- six Class A, borosilicate glass or non-toxic plasticware,
1000 ml for making test solutions.
5.21 Wash bottles -- for deionized water, for washing embryos from substrates
and containers, and for rinsing small glassware and instrument electrodes and
probes.
5.22 Volumetric flasks and graduated cylinders -- Class A, borosilicate glass
or non-toxic plastic labware, 10-1000 ml for making test solutions.
5.23 Pipets, volumetric -- Class A, 1-100 ml.
5.24 Pipets, automatic -- adjustable, 1-100 ml.
125
-------
5.25 Pipets, serological -- 1-10 ml, graduated.
5.26 Pipet bulbs and fillers -- PROPIPETR, or equivalent.
5.27 Droppers and glass tubing with fire polished aperatures, 4 mm ID -- for
transferring embryos and larvae.
5.28 Siphon with bulb and clamp -- for cleaning test chambers.
5.29 NITEXR or stainless steel mesh sieves, -- < 150 \im, 500 urn, and 3-5 mm
-- for collecting Artemia and fish embryos, and for spawning baskets,
respectively (NITEXR is available from Sterling Marine Products, 18 Label
Street, Montclair, New Jersey 07042, phone 201-783-9800).
6. REAGENTS AND CONSUMABLE MATERIALS
6.1 Sample containers -- for sample shipment and storage (see Section 8,
Effluent and Receiving Water Sampling and Sample Handling).
6.2 Data sheets (one set per test) -- for data recording (see Fig. 5).
6.3 Tape, colored -- for labelling test chambers.
6.4 Markers, water-proof -- for marking containers, etc.
6.5 Buffers, pH 4, 7, and 10 (or as per instructions of instrument
manufacturer) for standards and calibration check (see USEPA Method 150.1,
USEPA, 1979b).
6.6 Membranes and filling solutions for dissolved oxygen probe (see USEPA
Method 360.1, USEPA, 1979b), or reagents for modified Winkler analysis.
6.7 Laboratory quality assurance samples and standards for the above
methods.
6.8 Reference toxicant solutions (see Section 4. Quality Assurance,
Subsections 4.7, 4.14, 4.15, 4.16, and 4.17).
6.9 Reagent water -- defined as distilled or deionized water that does not
contain substances which are toxic to the test organisms (see Subsection 5.6
above).
6.10 Effluent, receiving water, and dilution water -- see Section 7, Dilution
Water, and Section 8, Effluent and Receiving Water Sampling, Sample Handling,
and Sample Preparation for Toxicity Tests.
6.11 Saline test and dilution water -- The overwhelming majority of
industrial and sewage treatment effluents entering marine and estuarine
systems contain little or no measurable salts. Exposure of sheepshead minnow
embryos to these effluents will require adjustments in the salinity of the
test solutions. This test has been successfully performed over a range of
salinity of 6 /oo to 59 /oo salinity. It is important to maintain a
126
-------
constant salinity across all treatments. Also, the salinity should vary by no
more than + 2 /oo among the chambers on a given day.
6.11.1 If effluent and receiving water tests are conducted concurrently, the
salinities of these tests should be similar. In addition, it may be desirable
to match, the test salinity with that of the receiving water. Two methods are
available to adjust salinities -- a supersaline brine derived from natural
seawater or artificial sea salts.
6.11.2 Hypersaline brine (100 °/oo salinity): Hypersaline brine (MSB) has
several advantages that make it desirable for use in toxicity testing. It can
be made from any high quality, filtered seawater by evaporation, and can be
added to the effluent or to deionized water to increase the salinity. HSB
derived from natural seawater contains the necessary trace metals, biogenic
colloids, and some of the microbial components necessary for adequate growth,
survival, and/or reproduction of marine and estuarine organisms, and may be
stored for prolonged periods without any apparent degradation. However, the
concentration of effluent that can be tested using HSB is limited to 80% at 20
°/oo salinity, and 70% at 30 °/oo salinity.
6.11.2.1 The ideal container for making brine from natural seawater is one
that (1) has a high surface to volume ratio, (2) is made of a non-corrosive
material, and (3) is easily cleaned (fiberglass containers are ideal).
Special care should be used to prevent any toxic materials from coming in
contact with the seawater being used to generate the brine. If a heater is
immersed directly into the seawater, ensure that the heater materials do not
corrode or leach any substances that would contaminate the brine. One
successful method used is a thermostatically controlled heat exchanger made
from fiberglass. If aeration is used, use only oil-free air compressors to
prevent contamination.
6.11.2.2 Before adding seawater to the brine generator, thoroughly clean the
generator, aeration supply tube, heater, and any other materials that will be
in direct contact with the brine. A good quality biodegradable detergent
should be used, followed by several (at least three) thorough deionized water
rinses.
6.11.2.3 High quality (and preferably high salinity) seawater should be
filtered to at least 10 \im before placing into the brine generator. Water
should be collected on an incoming tide to minimize the possibility of
contamination.
6.11.2.4 The temperature of the seawater is increased slowly to 40°C. The
water should be aerated to prevent temperature stratification and to increase
water evaporation. The brine should be checked daily (depending on volume
being generated) to ensure that salinity does not exceed 100 °/oo and that the
temperature does not exceed 40°C. Additional seawater may be added to the
brine to obtain the volume of brine required.
6.11.2.5 After the required salinity is attained, the brine should be
filtered a second time through a \-\nm filter and poured directly into portable
containers, such as 20-L (5-gal) cubitainers or polycarbonate watercooler
127
-------
jugs. The containers should be capped and labelled with the date the brine
was generated and its salinity. Containers of brine should be stored in the
dark and maintained at room temperature until used.
6.11.2.6 If a source of hypersaline brine is available, test solutions can be
made by following the directions below. Thoroughly mix together the deionized
water and brine before mixing in the effluent.
6.11.2.7 Divide the salinity of the hypersaline brine by the expected test
salinity to determine the proportion of deionized water to brine. For
example, if the salinity of the brine is 100 /oo and the test is to be
conducted at 20 °/oo, 100 °/oo divided by 20 °/oo = 5.0. The proportion of
brine is 1 part in 5 (one part brine to four parts deionized water).
6.11.2.8 To make 1 L of sea water at 20 °/oo salinity from a hypersaline
brine of 100 °/oo, divide 1 L (1000 ml) by 5.0. The result, 200 ml, is the
quantity of brine needed to make 1 L of sea water. The difference, 800 ml, is
the quantity of deionized water required.
6.11.2.9 Table 1 illustrates the composition of test solutions at 20 °/oo if
they are prepared by serial dilution of effluent with 20 °/oo salinity
seawater.
6.11.3 Artificial sea salts: HW MARINEMIXR brand sea salts (Hawaiian Marine
Imports Inc., 10801 Kempwood, Suite 2, Houston, Texas 77043) have been used
successfully at the USEPA Houston laboratory to culture sheepshead minnows and
perform the embryo-larval survival and teratogenicity test. EMSL-Cincinnati
has found FORTY FATHOMS" artifical sea salts (Marine Enterprises, Inc., 8755
Mylander Lane, Baltimore, Maryland 21204; phone: 301-321-1189), to be
suitable for culturing sheepshead minnows and for performing the larval
survival and growth test and embryo-larval test. Artificial sea salts may be
used for culturing sheepshead minnows and for the embryo larval test if the
criteria for acceptability of test data are satisfied (see Subsection 11).
6.11.3.1 Synthetic sea salts are packaged in plastic bags and mixed with
deionized water or equivalent. The important thing is to follow the
instructions on the package of sea salts carefully and to mix the salts in a
separate container -- not the culture tank. The deionized water used in
hydration should be in the temperature range of 21-26°C. Seawater made from
artificial sea salts is conditioned (see Spotte, 1973; Bower, 1983) before it
is used for culturing or testing. After adding the water, place an airstone
in the container, cover, and aerate the solution mildly for at least 24 h
before use.
128
-------
TABLE 1. PREPARATION OF TEST SOLUTIONS AT A SALINITY OF 20 °/oo, USING 20
°/oo NATURAL OR ARTIFICIAL SEAWATER, HYPERSALINE BRINE, OR
ARTIFICIAL SEA SALTS
Effluent
Solution
Effluent
Cone.
Solutions To Be Combined
Volume of
Effluent
Solution
Volume of Diluent
Seawater (20 °/oo)
1
2
3
4
5
Control
1001'2
50
25
12.5
6.25
0.0
4000 mL
2000 mL Solution 1
2000 mL Solution 2
2000 mL Solution 3
2000 mL Solution 4
+ 2000 mL
+ 2000 mL
+ 2000 mL
+ 2000 mL
2000 mL
Total
10000 mL
1This illustration assumes: (1) the use of 400 mL of test solution in each of
four replicates and 400 mL for chemical analysis (total of 2000 mL) for the
control and five concentrations of effluent (2) an effluent dilution factor
of 0.5, and (3) the effluent lacks appreciable salinity. A sufficient
initial volume (4000 mL) of effluent is prepared by adjusting the salinity to
the desired level. In this example, the salinity is adjusted by adding
artificial sea salts to the 100% effluent, and preparing a serial dilution
using 20 °/oo seawater (natural seawater, hypersaline brine, or artificial
seawater). The salinity of the initial 4000 mL of 100% effluent is adjusted
to 20 °/oo by adding 80 g of dry artificial sea salts (HW MARINEMIX or FORTY
FATHOMSR), and mixing for 1 h. Test concentrations are then made by mixing
appropriate volumes of salinity adjusted effluent and 20 °/oo salinity
dilution water to provide 4000 mL of solution for each concentration. If
hypersaline brine alone (100 °/oo) is used to adjust the salinity of the
effluent, trie highest concentration of effluent that could be achieved would
be 80% at 20 °/oo salinity, and 70% at 30 °/oo salinity.
2The same procedures would be followed in preparing test concentrations at
other salinities between 20 °/°° and 30 °/00: U) Tne salinity of the bulk
(initial) effluent sample would be adjusted to the appropriate salinity using
artificial sea salts or hypersaline brine, and (2) the remaining effluent
concentrations would be prepared by serial dilution, using a large batch
(10 L) of seawater for dilution water, which had been prepared at the same
salinity as the effluent, using natural seawater, hypersaline and deionized
water.
129
-------
6.12 BRINE SHRIMP, ARTEMIA, CULTURE (see USEPA, 1991c).
6.12.1 If a sheepshead continuous culture unit is established, newly-hatched
Artemia nauplii will be needed for feeding the larvae, and a brine shrimp
culture unit should be prepared. Although there are many commercial sources
of brine shrimp cysts, the Brazilian or Colombian strains are currently
preferred because the supplies examined have had low concentrations of
chemical residues and produce nauplii of suitably small size. For commercial
sources of brine shrimp, Artemia, cysts, see Table 2 (Section 5, Facilities,
Equipment, and Supplies) and Section 4, Quality Assurance, Subsection 4.8 Food
Quality.
6.12.2 Each new batch of Artemia cysts must be evaluated for size (Vanhaecke
and Sorgeloos, 1980, and Vanhaecke et al., 1980) and nutritional suitability
(see Leger et al., 1985, 1986) against known suitable reference cysts by
performing a side by side larval growth test using the "new" and "reference"
cysts. The "reference" cysts used in the suitability test may be a previously
tested and acceptable batch of cysts, or may be obtained from the Quality
Assurance Research Division, Environmental Monitoring Systems Laboratory,
Cincinnati, Ohio 45268. A sample of newly hatched Artemia nauplii from each
new batch of cysts should be chemically analyzed. The Artemia cysts should
not be used if the concentration of total organic chlorine exceeds 0.15 jig/g
wet weight or the total concentration of organochlorine pesticides plus PCBs
exceeds 0.30 pg/g wet weight. For analytical methods see USEPA (1982).
6.12.3 Artemia nauplii are obtained as follows:
1. Add 1 L of seawater, or a solution prepared by adding 35.0 g
uniodized salt (NaCl) or artificial sea salts to 1 L of deionized
water, to a 2-L separatory funnel, or equivalent.
2. Add 10 mL Artemia cysts to the separatory funnel and aerate for
24 h at 27°C. (Hatching time varies with incubation temperature
and the geographic strain of Artemia used. See USEPA (1985d),
USEPA et al. (1991c), and ASTM designation E1203, 1987, for
details on Artemia culture and quality control).
3. After 24 h, cut off the air supply in the separatory funnel.
Artemia nauplii are phototactic, and will concentrate at the
bottom if a dark cloth or paper towel is placed over the top of
the separatory funnel for 5-10 min. To prevent mortality, do not
leave the concentrated nauplii at the bottom of the funnel more
than 10 min without aeration.
4. Drain the nauplii into a funnel fitted with a < 150 urn Nitex
screen, and rinse with seawater or equivalent before use.
6.12.4 Testing Artemia nauplii as food for sheepshead minnow culturing.
6.12.4.1 The primary criterion for acceptability of each new supply of brine
shrimp cysts is the ability of the nauplii to support good survival and growth
of the sheepshead minnow larvae. The larvae used to evaluate the suitability
of the brine shrimp nauplii must be of the same geographical origin, species,
and stage of development as those used routinely in the larval survival and
growth tests. Sufficient data to detect differences in survival and growth
130
-------
should be obtained by using three replicate test vessels, each containing a
minimum of 15 larvae, for each type of food.
6.12.4.2 The feeding rate and frequency, test vessels, volume of control
water, duration of the test, and age of the nauplii at the start of the test,
should be the same as used for the routine toxicity tests.
6.12.4.3 Results of the brine shrimp nutrition assay, where there are only
two treatments, can be evaluated statistically by use of a t-test. The "new"
food is acceptable if there are no statistically significant differences in
the survival and growth of the larvae fed the two sources of nauplii.
6.12.4.4 The average seven-day survival of larvae should be 80% or greater,
and (2) the average dry weight of larvae should be 0.60 mg or greater, if
dried and weighed immediately after the test, or (3) the average dry weight of
larvae should be 0.50 mg or greater, if the larvae are preserved in 4%
formalin before drying and weighing. The above minimum weights presume that
the age of the larvae at the start of the test is not greater than 24 h.
6.13 SHEEPSHEAD MINNOWS
6.13.1 Brood Stock
6.13.1.1 Adult sheepshead minnows for use as brood stock may be obtained Dy
seine in Gulf of Mexico and Atlantic coast estuaries, from commercial sources,
or from young fish raised to maturity in the laboratory. Feral brood stocks
and first generation laboratory fish are preferred, to minimize inbreeding.
6.13.1.2 To detect disease and to allow time for acute mortality due to the
stress of capture, field-caught adults are observed in the laboratory a
minimum of two weeks before using as a source of gametes. Injured or diseased
fish are discarded.
6.13.1.3 Sheepshead minnows can be continuously cultured in the laboratory
from eggs to adults. The larvae, juvenile, and adult fish should be kept in
appropriate size rearing tanks, maintained at ambient laboratory temperature.
The larvae should be fed sufficient newly hatched Artemia nauplii daily to
assure that live nauplii are always present. At the juvenile stage, they are
fed frozen adult brine shrimp and a commercial flake food, such as TETRA
SM-80R, available from Tetra Sales (U.S.A), 201 Tabor Road, Morris Plains, New
Jersey 07950, phone 800-526-0650, or MARDEL AQUARIAN" Tropical Fish Flakes,
available from Mardel Laboratories, Inc., 1958 Brandon Court, Glendale
Heights, Illinois 60139, phone 312-351-0606, or equivalent. Adult fish are
fed flake food three or four times daily, supplemented with frozen adult brine
shrimp.
6.13.1.3.1 Sheepshead minnows reach sexual maturity in three-to-five months
after hatch, and have an average standard length of approximately 27 mm for
females and 34 mm for males. At this time, the males begin to exhibit sexual
dimorphism and initiate territorial behavior. When the fish reach sexual
maturity and are to be used for natural spawning, the temperature should be
controlled at 18-20°C.
131
-------
6.13.1.4 Adults can be maintained in natural or artificial sea water in a
flow through or recirculating, aerated system consisting of an all-glass
aquarium, or a "Living Stream" (Figid Unit, Inc., 3214 Sylvania Avenue,
Toledo, Ohio 43613, phone 419-474-6971), or equivalent.
6.13.1.5 The system is equipped with an undergravel or outside biological
filter of shells (see Spotte, 1973 or Bower, 1983 for conditioning the
biological filter), or a cartridge filter, such as a MAGNUMR Filter, available
from Carolina Biological Supply Co., Burlington, North Carolina 27215, phone
800-334-5551, or an EHEIMR Filter, available from Hawaiian Marine Imports
Inc., P.O. Box 218687, Houston, Texas 77218, phone 713-492-7864, or
equivalent, at a salinityof 20-30 °/oo and a photoperiod of 14 h light/10 h
dark.
6.13.2 Obtaining Embryos for Toxicity Tests1
6.13.2.1 Embryos can be shipped to the laboratory from an outside source or
obtained from adults held in the laboratory. Ripe eggs can be obtained either
by natural spawning or by intraperitoneal injection of the females with human
chorionic gonadotrophin (HCG) hormone, available from United States
Biochemical Corporation, Cleveland, Ohio 44128, phone 216-765-5000. If the
culturing system for adults is temperature controlled, natural spawning can be
induced. Natural spawning is preferred because repeated spawnings can be
obtained from the same brood stock, whereas with hormone injection, the brood
stock is sacrificed in obtaining gametes.
6.13.2.2 It should be emphasized that the injection and hatching schedules
given below are to be used only as guidelines. Response to the hormone varies
from stock to stock and with temperature. Time to hatch and percent hatch
also vary among stocks and among batches of embryos obtained from the same
stock, and are dependent on temperature, DO, and salinity.
6.13.2.3 Forced Spawning
6.13.2.3.1 HCG is reconstituted with sterile saline or Ringer's solution
immediately before use. The standard HCG vial contains 1,000 IU to be
reconstituted in 10 ml of saline. Freeze-dried HCG which comes with
premeasured and sterilized saline is the easiest to use. Use of a 50 IU dose
requires injection of 0.05 ml of reconstituted hormone solution.
Reconstituted HCG may be used for several weeks if kept in the refrigerator.
6.13.2.3.2 Each female is injected intraperitoneally with 50 IU HCG on two
consecutive days, starting at least 4 days prior to the beginning of a test.
Two days following the second injection, eggs are stripped, from the females
and mixed with sperm derived from excised macerated testes. At least ten
females and five males are used per test to ensure that there is a sufficient
number of viable embryos.
Adapted from USEPA (1978b).
132
-------
6.13.2.3.3 HCG is injected into the peritoneal cavity, just below the skin,
using as small a needle as possible. A 50 IU dose is recommended for females
approximately 27 mm in standard length. A larger or smaller dose may be used
for fish which are significantly larger or smaller than 27 mm. With
injections made on days one and two, females which are held at 25°C should be
ready for stripping on Day 4. Ripe females should show pronounced abdominal
swelling, and release at least a few eggs in response to a gentle squeeze.
Injected females should be isolated from males. It may be helpful if fish
that are to be injected are maintained at 20°C before injection, and the
temperature raised to 25°C on the day of the first injection.
6.13.2.3.4 Prepare the testes immediately before stripping the eggs from the
females. Remove the testes from three-to-five males. The testes are paired,
dark grey organs along the dorsal midline of the abdominal cavity. If the
head of the male is cut off and pulled away from the rest of the fish, most of
the internal organs can be pulled out of the body cavity, leaving the testes
behind. The testes are placed in a few ml of seawater until the eggs are
ready.
6.13.2.3.5 Strip the eggs from the females, into a dish containing 50-100 ml
of seawater, by firmly squeezing the abdomen. Sacrifice the females and
remove the ovaries if all the ripe eggs do not flow out freely. Break up any
clumps of ripe eggs and remove clumps of ovarian tissue and underripe eggs.
Ripe eggs are spherical, approximately 1 mm in diameter, and almost clear.
6.13.2.3.6 While being held over the dish containing the eggs, the testes are
macerated in a fold of NITEXR screen (250-500 \im mesh) dampened with seawater.
The testes are then rinsed with seawater to remove the sperm from tissue, and
the remaining sperm and testes are washed into the dish. Let the eggs and
milt stand together for 10-15 min, swirling occasionally.
6.13.2.3.7 Pour the contents of the dish into a beaker, and insert an
airstone. Aerate gently, such that the water moves slowly over the eggs, and
incubate at 25°C for 60-90 min. After incubation, wash the eggs on a Nitex
screen and resuspend them in clean seawater.
6.13.2.4 Natural Spawning
6.13.2.4.1 Short-term (Demand) Embryo Production
6.13.2.4.1.1 Adult fish should be maintained at 18-20°C in a temperature
controlled system. To obtain embryos for a test, adult fish (generally, at
least eight-to-ten females and three males) are transferred to a spawning
chamber, with a photoperiod of 16 h light/8 h dark and a temperature of 25°C,
two days before the beginning of the test. The spawning chambers are
approximately 20 X 35 X 22 cm high (Hansen et al., 1978), and consist of a
basket of 3-5 mm NITEXR mesh, made to fit into a 57-L (15 gal) aquarium.
Spawning generally will begin within 24 h or less. The embryos will fall
through the bottom of the basket and onto a collecting screen (250-500 urn
mesh) or tray below the basket. The collecting tray should be checked for
embryos the next morning. The number of eggs produced is highly variable.
The number of spawning units required to provide the embryos needed to perform
133
-------
a toxicity test is determined by experience. If the trays do not contain
sufficient embryos after the first 24 h, discard the embryos, replace the
trays, and collect the embryos for another 24 h or less. To help keep the
embryos clean, the adults are fed while the screens are removed.
6.13.2.4.1.2 The embryos are collected in a tray placed on the bottom of the
tank. The collecting tray consists of ± 150 pm NITEXR screen attached to a
rigid plastic frame. The collecting trays with newly-spawned, embryos are
removed from the spawning tank, and the embryos are collected from the screens
by washing them with a wash bottle or removing them with a fine brush. The
embryos from several spawning units may be pooled in a single container to
provide a sufficient number to conduct the test(s). The embryos are
transferred into a petri dish or equivalent, filled with fresh culture water,
and are examined using a dissecting microscope or other suitable magnifying
device. Damaged and infertile eggs are discarded (see Fig. 1). It is
strongly recommended that the embryos be obtained from fish cultured inhouse,
rather than from outside sources, to eliminate the uncertainty of damage
caused by shipping and handling that may not be observable, but which might
affect the results of the test.
6.13.2.4.1.3 After sufficient embryos are collected for the test, the adult
fish are returned to the (18-20°C) culture tanks.
6.13.2.4.2 Sustained Natural Embryo Production
6.13.2.4.2.1 Sustained (long-term), daily, embryo production can be achieved
by maintaining mature fish in tanks, such as a (285-L or 75-gal) LIVING
STREAM" tank, at a temperature of 23-25°C. Embryos are produced daily, and
when needed, embryo "collectors" are placed on the bottom of the tank on the
afternoon preceding the start of the test. The next morning, the embryo
collectors are removed and the embryos are washed into a shallow glass culture
dish using artificial seawater.
6.13.2.4.2.2 Four embryo collectors, approximately 20 cm X 45 cm, will
approximately cover the bottom of the 285-L tank. The collectors are
fabricated from plastic fluorescent light fixture diffusors (grids), with
cells approximately 14 mm deep X 14 mm square. A screen consisting of 500 jim
mesh is attached to one side (bottom) of the grid with silicone adhesive. The
depth and small size of the grid protects the embryos from predation by the
adult fish.
6.13.2.4.2.3 The brood stock is replaced annually with feral stock.
6.13.2.5 Test Organisms
6.13.2.5.1 Embryos spawned over a less than 24-h period, are used for the
test. These embryos may be used immediately to start a test or may be placed
in a suitable container and transported for use at a remote location. When
overnight transportation is required, embryos should be obtained when they are
no more than 8-h old. This permits the tests at the remote site to be started
with less than 24-h old embryos. Embryos should be transported or shipped in
clean, insulated containers, in well aerated or oxygenated fresh sea water or
134
-------
aged artificial sea water of correct salinity, and should be protected from
extremes of temperature and any other stressful conditions during transport.
Instantaneous changes of water temperature when embryos are transferred from
culture unit water to test dilution water, or from transport container water
to on-site test dilution, should be less than 2°C. Instantaneous changes of
pH, dissolved ions, osmotic strength, and DO should also be kept to a minimum.
6.13.2.5.2 The number of embryos needed to start the test will depend on the
number of tests to be conducted and the objectives. If the test is conducted
with four replicate test chambers (minimum of three) at each toxicant
concentration and in the control, with 15 embryos (minimum of 10) in each test
chamber, and the combined mortality of embryos prior to the start of the test
is less than 20%, 400 viable embryos are required for the test.
7. SAMPLE COLLECTION, PRESERVATION AND HANDLING
7.1 See Section 8, Effluent and Receiving Water Sampling, Sample Handling,
and Sample Preparation for Toxicity Tests.
8. CALIBRATION AND STANDARDIZATION
8.1 See Section 4, Quality Assurance.
9. QUALITY CONTROL
9.1 See Section 4, Quality Assurance.
10. TEST PROCEDURES
10.1 TEST SOLUTIONS
10.1.1 Receiving Waters
10.1.1.1 The sampling point is determined by the objectives of the test. At
estuarine and marine sites, samples are usually collected at mid-depth.
Receiving water toxicity is determined with samples used directly as collected
or with samples passed through a 60 jim NITEXR filter and compared without
dilution, against a control. Using four replicate chambers per test, each
containing 400-500 mL, and 400 mL for chemical analysis, would require
approximately 2400 mL or more of sample per test.
10.1.2 Effluents
10.1.2.1 The selection of the effluent test concentration should be based on
the objectives of the study. A dilution factor of 0.5 is commonly used. A
dilution factor of 0.5 provides precision of ± 100%, and allows for testing of
concentrations between 6.25% and 100% effluent using only five effluent
concentrations (6.25%, 12.5%, 25%, 50%, and 100%). Test precision shows
little improvement as dilution factors are increased beyond 0.5 and declines
rapidly if smaller dilution factors are used. Therefore, USEPA recommends a
dilution factor of 0.5. If 100 °/oo salinity HSB is as a diluent, the maximum
135
-------
concentration of effluent that can be tested will be 80% at 20 °/oo and 70% at
30 °/oo salinity.
10.1.2.2 If the effluent is known or suspected to be highly toxic, a lower
range of effluent concentrations should be used (such as 25%, 12.5%, 6.25%,
3.12%, and 1.56%). If a high rate of mortality is observed during the first
l-to-2 h of the test, additional dilutions at the lower range of effluent
concentrations should be added.
10.1.2.3 The volume of effluent required to initiate the test and for daily
renewal of four replicates (minimum of three) per concentration for five
concentrations of effluent and a control, each containing 400 ml of test
solution, is approximately 4 L. Prepare enough test solution (approximately
3000 ml) at each effluent concentration to refill the test chambers and
provide at least 400 ml additional volume for chemical analyses.
10.1.2.4 Maintain the effluent at 4°C. Plastic containers such as 8-20 L
cubitainers have proven successful for effluent collection and storage.
10.1.2.5 Approximately one hour before use, warm a sufficient volume of
chilled effluent or receiving water sample(s) to the test temperature (25 +
1°C) and maintain it at that temperature until portions are added to
the dilution water.
11.1.2.6 The higher effluent concentrations (i.e., 25%, 50%, and 100%) may
require aeration to maintain adequate dissolved oxygen concentrations.
However, if one solution is aerated, all concentrations must be aerated.
Aerate effluent as it warms and continue to gently aerate test solutions in
the test chambers for the duration of the test.
10.2 START OF THE TEST
10.2.1 Tests should begin as soon as possible, preferably within 24 h after
sample collection. For on-site toxicity studies, no more than 24 h should
elapse between collection of the effluent and use in a embryo-larval study.
If the persistence of the sample toxicity is not known, the maximum holding
time following retrieval of the sample from the sampling device should not
exceed 36 h for off-site toxicity studies unless permission is granted by the
permitting authority. In no case should the test be started more than 72 h
after sample collection.
10.2.2 Label the test chambers with a marking pen and identify each treatment
and replicate with various colored-coded tape. A minimum of five effluent
concentrations and a control should be selected for each study. Each concen-
tration (including controls) is to have four replicates (minimum of three).
Use 500 ml beakers, crystallization dishes, nontoxic disposable plastic
labware, or equivalent for test chambers.
10.2.3 Prepare the test solutions (see Table 1) and add to the test chambers.
10.2.4 Gently agitate and mix the embryos to be used in the test in a large
container so that eggs from different spawns are evenly dispersed.
136
-------
10.2.5 The test is started by randomly placing embryos from the common pool,
using a small bore (2 mm), fire polished, glass tube calibrated to contain
approximately the desired number of embryos, into each of four replicate test
chamber, until each chamber contains 15 embryos (minimum of 10), for a total
of 60 embryos for each treatment (four replicates recommended, three minimum).
See Appendix A for an example of randomization. The amount of water added to
the chambers when transferring the embryos should be kept to a minimum to
avoid unnecessary dilution of the test concentrations.
10.2.6 After the embryos have been distributed to each test chamber, examine
and count them. Remove and discard damaged or infertile eggs and replace with
undamaged embryos. It may be more convenient and efficient to transfer
embryos to intermediate containers of dilution water for examination and
counting. After the embryos have been examined and counted in the
intermediate container, assign them to the appropriate test chamber and
transfer them with a minimum of dilution water.
10.2.7 Randomize the position of the test chambers at the beginning of the
test (see Appendix A). Maintain the chambers in this configuration throughout
the test. A position chart may be helpful.
10.3 LIGHT, PHOTOPERIOD, TEMPERATURE, AND SALINITY
10.3.1 The light quality and intensity should be at ambient laboratory
levels, approximately 10-20 nE/m2/s, or 50 to 100 foot candles (ft-c), with a
photoperiod of 16 h of light and 8 h of darkness. The test water temperature
should be maintained at 25 ± 1°C. The salinity should be 5 to 32 ± 2 °/oo to
accommodate receiving waters that may fall within this range.
10.4 DISSOLVED OXYGEN (DO) CONCENTRATION
10.4.1 Aeration may affect the toxicity of effluents and should be used only
as a last resort to maintain satisfactory DO. The DO should not fall below
4.0 mg/L. If it is necessary to aerate, all treatments and the control should
be aerated. The rate should not exceed 100 bubbles/min, using a pipet with a
1-2 mm orifice, such as a 1-mL Kimax Serological Pipet No. 37033, or
equivalent. Care should be taken to ensure that turbulence resulting from the
aeration does not cause undue physical stress to the fish.
10.5 FEEDING
10.5.1 Feeding is not required.
10.6 TEST SOLUTION RENEWAL
10.6.1 The test solutions are adjusted to the correct salinity and renewed
daily using freshly collected samples. During the daily renewal process, 7-10
mm of water is left in the chamber to ensure that the embryos and larvae
remain submerged during the renewal process. New test solution (400 mL)
should be added slowly by pouring down the side of the test chamber to avoid
exposing the embryos and larvae to excessive turbulence.
137
-------
10.6.2 Prepare test solutions daily, making a minimum of five concentrations
and a control. If concurrent effluent and receiving water testing occurs, the
effluent test salinity should closely approximate that of the receiving water
test. If an effluent is tested alone, select a salinity which approximately
matches the salinity of the receiving waters. Table 1 illustrates the
quantities of effluent, sea water, deionized water, and artificial sea salts
needed to prepare 3 L of test solution at each effluent concentration for
tests conducted at 20 °/oo salinity.
10.7 ROUTINE CHEMICAL AND PHYSICAL DETERMINATIONS
10.7.1 At a minimum, the following measurements are made and recorded (see
Figure 5).
10.7.1.1 DO is measured at the beginning and end of each 24-h exposure period
at all test concentrations and in the control.
10.7.1.2 Temperature, pH, and salinity are measured at the end of each 24-h
exposure period at all test concentrations and in the control.
10.8 OBSERVATIONS DURING THE TEST
10.8.1 At the end of the first 24 h of exposure, before renewing the test
solutions, examine and count the embryos. Remove the dead embryos (milky
colored and opaque) and record the number. If the rate of mortality or fungal
infection exceeds 20% in the control chambers, or if excessive
nonconcentration related mortality occurs, terminate the test and start a new
test with new embryos. If the above mortality conditions do not occur,
continue the test for the full nine days.
10.8.2 At 25°C, hatching begins on about the sixth day. After hatching
begins, count the number of dead and live embryos and the number of hatched,
dead, live, and deformed and/or debilitated larvae, daily (see Figure 1 for
illustrations of morphological development of embryo and larva). Deformed
larvae are those with gross morphological abnormalities such as curved
spines, lack of appendages, lack of fusiform shape (non-distinct mass), a
colored beating heart in an opaque mass, lack of mobility, abnormal swimming,
or other characteristics that preclude survival. Remove dead embryos and dead
and deformed larvae as previously discussed and record the numbers for all
test observations (see Figure 5).
138
-------
Figure 1. Embryonic development of sheepshead minnow, Cyprinodon variegatus:
A. Mature unfertilized egg, showing attachment filaments and micropyle, X33;
B. Blastodisc fully developed; C,D. Blastodisc, 8 cells; E. Blastoderm, 16
cells; F. Blastoderm, late cleavage stage; G. Blastoderm with germ ring
formed, embryonic shield developing; H. Blastoderm covers over 3/4 surface of
yolk, yolk noticeably constricted; I. Early embryo. From Kuntz (1916).
139
-------
Figure 1. (Continued). Embryonic development of sheepshead minnow,
Cyprinodon variegatus: J. Embryo 48 h after fertilization, now segmented
throughout, pigment on yolk sac and body, otoliths formed; K. Posterior
portion of embryo free from yolk and moves freely within egg membrane, 72 h
after fertilization; L. Newly hatched fish, actual length 4 mm; M. Larval fish
5 days after hatching, actual length 5 mm; N. Young fish 9 mm in length; 0.
Young fish 12 mm in length. From Kuntz (1916).
140
-------
10.8.3 Protect the embryos and larvae from unnecessary disturbance during the
test by carefully carrying out the daily test observations, solution renewals,
and removal of dead organisms. Make sure the test organisms remain immersed
during the performance of the above operations.
10.9 TERMINATION OF THE TEST
10.9.1 The test is terminated after nine days of exposure, or four days
post-hatch, whichever comes first. Count the number of surviving, dead, and
deformed and/or debilitated larvae, and record the numbers of each. The
deformed larvae are treated as dead. Keep a separate record of the total
number of deformed larvae for use in reporting the teratogenicity of the test
solution.
11. ACCEPTABILITY OF TEST RESULTS
11.1 For the test results to be acceptable, survival in the controls must be
at least 80% or better.
12. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA
12.1 A summary of test conditions and test acceptability criteria is listed
in Table 2.
13. DATA ANALYSIS
13.1 GENERAL
13.1.1 Tabulate and summarize the data.
13.1.2 The endpoints of this toxicity test are based on total mortality,
combined number of dead embryos, dead larvae, and deformed larvae. The EC
endpoints are calculated using Probit Analysis (Finney, 1971). LOEC and NOEC
values, for total mortality, are obtained using a hypothesis test approach
such as Dunnett's Procedure (Dunnett, 1955) or Steel's Many-one Rank Test
(Steel, 1959; Miller, 1981). See the Appendices for examples of the manual
computations, program listings, and examples of data input and program output.
13.1.3 The statistical tests described here must be used with a knowledge of
the assumptions upon which the tests are contingent. The assistance of a
statistician is recommended for analysts who are not proficient in statistics.
141
-------
TABLE 2. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA FOR THE
SHEEPSHEAD MINNOW, CYPRINODON VARIEGATUS, EMBRYO LARVAL SURVIVAL AND
TERATOGENICITY TEST WITH EFFLUENTS AND RECEIVING WATERS
1.
2.
3.
4.
5.
Test type:
Salinity:
Temperature:
Light quality:
Light intensity:
Static renewal
5 °/oo to 32 °/oo ± 2 °/oo
25 ± 1°C
Ambient laboratory light
10-20 jiE/m2/s, or 50-100 ft-c
6. Photoperiod:
7. Test chamber size:
8. Test solution volume:
9. Renewal of test concentration:
10. Age of test organisms:
11. No. of embryos/chamber:
12. Replicate test chambers/
concentration:
13. Embryos per concentration:
14. Feeding regime:
15. Aeration:
16. Dilution water:
17. Test concentrations:
(ambient laboratory levels)
16 h light, 8 h dark
400-500 mL
250-400 mL per replicate
Daily
less than 24 h old
15 (minimum of 10)
4 (minimum of 3)
60 (minimum of 30)
Feeding not required
None unless DO falls below 4.0 mg/L
Uncontaminated source of sea water;
deionized water mixed with
artificial sea salts (HW MARINEMIXR,
FORTY FATHOMS", or equivalent), or
hypersaline brine
Effluents: Minimum of five effluent
concentrations and a control
Receiving waters: 100% receiving water
and a control
142
-------
TABLE 2. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA FOR THE
SHEEPSHEAD MINNOW, CYPRINODON VARIEGATUS, EMBRYO LARVAL SURVIVAL AND
TERATOGENICITY TEST WITH EFFLUENTS AND RECEIVING WATERS (CONTINUED)
18. Dilution factor:
19. Test duration:
20. Effects measured:
21. Test acceptability
criteria:
22. Sampling requirements;
23. Estimated maximum sample
volume required:
Effluent: > 0.5 series
Receiving waters: None, or > 0.5
series
9 days
Percent hatch; percent larvae dead
or with debilitating morphological
and/or behavior abnormalities such
as: gross deformities; curved spine;
disoriented, abnormal swimming
behavior; surviving normal larvae
from original embryos
80% or greater survival in controls
For on-site tests, samples are
collected daily, and used within 24 h
of the time they are removed from the
sampling device. For off-site tests,
a minimum of three samples are
collected on days one, three, and five
with a maximum holding time of 36 h
before first use (see Section 12,
Subsection 10.2.1).
5 L per day
143
-------
13.2 EXAMPLE OF ANALYSIS OF SHEEPSHEAD MINNOW EMBRYO-LARVAL SURVIVAL AND
TERATOGENICITY DATA
13.2.1 Formal statistical analysis of the total mortality data is outlined in
Figure 2. The response used in the analysis is the total mortality proportion
in each test or control chamber. Separate analyses are performed for the
estimation of the NOEC and LOEC endpoints and for the estimation of the EC1,
ECS, EC10 and EC50 endpoints. Concentrations at which there is no survival in
any of the test chambers are excluded from the statistical analysis of the
NOEC and LOEC, but included in the estimation of the EC endpoints.
13.2.2 For the case of equal numbers of replicates across all concentrations
and the control, the evaluation of the NOEC and LOEC endpoints is made via a
parametric test, Dunnett's Procedure, or a nonparametric test, Steel's
Many-one Rank Test, on the arc sine transformed data. Underlying assumptions
of Dunnett's Procedure, normality and homogeneity of variance, are formally
tested. The test for normality is the Shapiro-Milk's Test, and Bartlett's
Test is used to determine the homogeneity of variance. If either of these
tests fail, the nonparametric test, Steel's Many-one Rank Test, is used to
determine the NOEC and LOEC endpoints. If the assumptions of Dunnett's
Procedure are met, the endpoints are estimated by the parametric procedure.
13.2.3 If unequal numbers of replicates occur among the concentration levels
tested, there are parametric and nonparametric alternative analyses. The
parametric analysis is a t-test with the Bonferroni adjustment. The Wilcoxon
Rank Sum Test with the Bonferroni adjustment is the nonparametric alternative.
For detailed information on the Bonferroni adjustment, see Appendix D.
13.2.4 Probit Analysis (Finney, 1971) is used to estimate the concentration
that causes a specified percent decrease in survival from the control. In
this analysis, the total mortality data from all test replicates at a given
concentration are combined.
13.2.5 In this example, total mortality, expressed as a proportion (combined
total number of dead embryos, dead larvae and deformed larvae divided by the
number of embryos at start of test), is the response of interest. The total
mortality proportion in each replicate must first be transformed by the arc
sine transformation procedure described in Appendix B. The raw and
transformed data, means and standard deviations of the transformed
observations at each SOS concentration and control are listed in Table 3. A
plot of the data is provided in Figure 3. Since there is 100% total mortality
in both replicates for the 8.0 mg/L concentration, it is not included in this
statistical analysis and is considered a qualitative mortality effect.
144
-------
TABLE 3. SHEEPSHEAD MINNOW, CYPRINODON VARIEGATUS, EMBRYO-LARVAL TOTAL
MORTALITY DATA
SDS Concentration (mq/L)
Repl
RAW
ARC SINE
TRANS-
FORMED
MEAN(Yf)
pfc
i
icate
A
B
A
B
Control
0.1
0.1
0.322
0.322
0.322
0.0
1
0.5
0.0
0.2
0.159
0.464
0.311
0.046
2
1.0
0.0
0.1
0.159
0.322
0.240
0.013
3
2.0
0.3
0.1
0.580
0.322
0.451
0.033
4
4.0
0.9
0.7
1.249
0.991
1.120
0.033
5
8.0
1.0
1.0
-
13.2.6 Test for Normality
13.2.6.1 Since only two replicates were run at each concentration level, the
test for normality is invalid. Additionally, a nonparametric alternative to
Dunnett's Procedure is not available with only duplicates. Thus the only
information that can be derived from the data is from Dunnett's Procedure.
However, the results from this test should be interpreted with caution since
the assumptions of the test are in question.
13.2.7 Dunnett's Procedure
13.2.7.1 To obtain an estimate of the pooled variance for the Dunnett's
Procedure, construct an ANOVA table as described in Table 4.
TABLE 4. ANOVA TABLE
Source
Total
df
Sum of Squares
(SS)
N - 1
SST
Mean Square(MS)
(SS/df)
Between
Within
p - 1
N - p
SSB
SSW
2
SB = SSB/(p-l)
2
Sw = SSW/(N-p)
145
-------
STATISTICAL ANALYSIS OF SHEEPSHEAD MINNOW EMBRYO-LARVAL
SURVIVAL AND TERATOGENECITY TEST
TOTAL MORTALITY:
TOTAL NUMBER OF DEAD EMBRYOS,
DEAD LARVAE, AND DEFORMED LARVAE
ENDPOINT ESTIMATE
ECs
ARC SINE
TRANSFORMATION
\
SHAPIRO-WILK'S TEST
NORMAL DISTRIBUTION
NON-NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
BARTLETTS TEST
HETEROGENEOUS
VARIANCE
EQUAL NUMBER OF
REPLICATES?
EQUAL NUMBER OF
REPLICATES?
\
YES
*
STEEL'S MANY-ONE
RANKTEST
T
WILCOXON RANK SUM
TEST WITH
BONFERRONI ADJUSTMENT
ENDPOINT ESTIMATES
NOEC.LOEC
Figure 2. Flowchart for statistical analysis of sheepshead
minnow, Cyprinodon variegatus, embryo-larval data.
146
-------
0.0
CONNECTS THE MEAN VALUE FOR EACH CONCENTRATION
REPRESENTS THE CRITICAL VALUE FOR DUNNETT'S TEST
(ANY PROPORTION ABOVE THIS VALUE WOULD BE
SIGNIFICANTLY DIFFERENT FROM THE CONTROL)
1.0 2.0
SDS CONCENTRATION (MG/L)
4.0
8.0
Figure 3. Plot of sheepshead minnow, Cyprinodon variegatus, total mortality
data from the embryo-larval test.
-------
Where: p = number of SDS concentration levels including the control
N = total number of observations n1 + n? . . . +np
n,- = number of observations in concentration i
p
SSB = X Tj2/n,. - G2/N Between Sum of Squares
> >
SST = S 2 Y2: G2/N Total Sum of Squares
i=l j-1
SSW = SST SSB Within Sum of Squares
P
G = the grand total of all sample observations, G = z T,.
i=l
T, = the total of the replicate measurements for
concentration "i"
Yy = the jth observation for concentration "i" (represents
the proportion of total mortality for SDS concentration
i in test chamber j)
13.2.7.2 For the data in this example:
n1 = n2 = n3 = n4 = n5 = 2
N = 1 0
T1 = YII + Y12 + Y13 = 0.644
T2 = Y21 + Y22 + Y23 = 0.623
T3 = Y31 + Y32 + Y33 = 0.481
T4 = Y41 + Y42 + Y43 = 0.902
T5 = Y,i + Y52 + Y53 = 2.240
G = T , + T2 + T3 + T4 = 4.890
P ? .
SSB = S Tj/n,. - G2/N
= 1 (6.865) - (4.890)2 = 1.041
2 10
P "i 2 ,
SST = 2 z Yf! G2/N
i * T J
= 3.559 - (4.890)2 = 1.168
10
148
-------
SSW = SST SSB = 1.168 - 1.041 = 0.127
Sj = SSB/(p-l) = 1.041/(5-l) = 0.260
$1 = SSW/(N-p) = 0.127/(10-5) = 0.025
13.2.7.3 Summarize these calculations in the ANOVA table (Table 5)
TABLE 5. ANOVA TABLE FOR DUNNETT'S PROCEDURE EXAMPLE
Source
Between
Within
Total
df
4
5
9
Sum of Squares
(SS)
1.041
0.127
1.168
Mean Square(MS)
(SS/df)
0.260
0.025
13.2.7.4 To perform the individual comparisons, calculate the t
statistic for each concentration, and control combination as follows:
( Y, Y, )
Sw / (1/n,) + (1/n,-)
Where: Y( = mean proportion of total mortality for SDS concentration i
Y, = mean proportion of total mortality for the control
Su = square root of within mean square
n, = number of replicates for the control
n,- = number of replicates for concentration i.
Since we are looking for an increased response in percent of total mortality
over control, the control mean is subtracted from the mean at a concentration.
13.2.7.5 Table 6 includes the calculated t values for each concentration and
control combination. In this example, comparing the 0.5 mg/L concentration
with the control the calculation is as follows:
( 0.311 - 0.322 )
t2 = = 0.570
[ 0.158 / (1/2) + (1/2) ]
149
-------
TABLE 6. CALCULATED T-VALUES
SDS Concentration (mg/L)
0.5
1.0
2.0
4.0
2
3
4
5
-0.070
-0.519
0.816
5.051
13.2.7.6 Since the purpose of this test is to detect a significant increase
in total mortality, a one-sided test is appropriate. The critical value for
this one-sided test is found in Table 5, Appendix C. For an overall alpha
level of 0.05, five degrees of freedom for error and four concentrations
(excluding the control) the critical value is 2.85. The mean proportion of
total mortality for concentration "i" is considered significantly less than
the mean proportion of total mortality for the control if t= is greater than
the critical value. Therefore, only the 4.0 mg/L concentration has a
significantly higher mean proportion of total mortality than the control.
Hence the NOEC is 2.0 mg/L and the LOEC is 4.0 mg/L.
13.2.7.7 To quantify the sensitivity of the test, the minimum significant
difference (MSD) that can be detected statistically may be calculated.
MSD = d Su /
(1/n,) + (1/n)
Where: d = the critical value for the Dunnett's procedure
Su = the square root of the within mean square
n = the common number of replicates at each concentration
(this assumes equal replication at each concentration)
n. = the number of replicates in the control.
13.2.7.8 In this example:
MSD = 2.85 (0.158) / (1/2) + (1/2)
= 2.85 (0.158)(1.0)
= 0.450
13.2.7.9 The MSD (0.450) is in transformed units. To determine the MSD in
terms of percent survival, carry out the following conversion.
1. Add the MSD to the transformed control mean.
0.322 + 0.450 = 0.772
150
-------
2. Obtain the untransformed values for the control mean and the sum
calculated in 1.
[Sine (0.322) ]* = 0.1
[Sine (0.772) ]2 = 0.487
3. The untransformed MSD (MSDJ is determined by subtracting the
untransformed values from step 2.
MSDU = 0.487 0.1 = 0.387
13.2.7.10 Therefore, for this set of data, the minimum difference in mean
proportion of total mortality between the control and any SDS concentration
that can be detected as statistically significant is 0.387.
13.2.7.11 This represents a 387% increase in mortality from the control.
13.2.8 Probit Analysis
13.2.8.1 The data used for the Probit Analysis is summarized in Table 7. For
the Probit Analysis, the SDS concentration with 100% total mortality in both
replicates is considered. To perform the Probit Analysis Program, run the
USEPA Probit Analysis Program. Examples of the program output are provided in
in Table 8 and Figure 4.
13.2.8.2 For this example, the chi-square test for heterogeneity was not
significant. Thus Probit Analysis appears to be appropriate for this set of
data.
TABLE 7. DATA FOR PROBIT ANALYSIS
SDS Concentration (mq/L)
Number Dead
Number Exposed
Control 0.5
2 2
20 20
1.0 2.0
1 4
20 20
4.0 8.0
16 20
20 20
14. PRECISION AND ACCURACY
14.1 PRECISION
14.1.1 Single-Laboratory Precision
14.1.1.1 Data on the single-laboratory precision of the sheepshead minnow
embryo-larval survival and teratogenicity test are available for eight tests
with copper sulfate and five tests with sodium dodecyl sulfate (USEPA, 1989a)
151
-------
The data for the first five tests show that the same NOEC AND LOEC, 240 ng
Cu/L and 270 [ig Cu/L, respectively, were obtained in all five tests, which is
the maximum level of precision that can be attained. Three additional tests
(6-8) were performed with narrower (20 yg) concentration intervals, to more
precisely identify the threshold concentration. The NOEC and LOEC for these
tests are 200 \ig and 220 jig Cu/L, respectively. For sodium dodecyl sulfate,
the NOEC'S and LOEC'S for all tests are 2.0 and 4.0 mg/L, respectively. The
precision, expressed as the coefficient of variation (CV%), is indicated in
Tables 9-10. For copper (CU), the coefficient of variation, depending on the
endpoint used, ranges from 2.5% to 6.1% which indicates excellent precision.
For sodium dodecyl sulfate (SDS), the coefficient of variation, depending on
the endpoint used, ranges from 11.7% to 51.2% indicating acceptable precision.
14.1.2 Multilaboratory Precision
14.1.2.1 Data on the multilaboratory precision of this test are not yet
available.
14.2 ACCURACY
14.2.1 The accuracy of toxicity tests cannot be determined.
152
-------
TABLE 8. OUTPUT FOR USEPA PROBIT ANALYSIS PROGRAM,
VERSION 1.4
EPA PROBIT ANALYSIS PROGRAM
USED FOR CALCULATING EC VALUES
Version 1.4
Probit Analysis of Sheepshead Minnow Embryo-Larval Survival and Teratogenicity Data
Cone.
Control
0.5000
1.0000
2.0000
4.0000
8.0000
Number
Exposed
20
20
20
20
20
20
Number
Resp.
2
2
1
4
16
20
Observed
Proportion
Responding
0.1000
0.1000
0.0500
0.2000
0.8000
1.0000
Adjusted
Proportion
Responding
0.0000
0.0174
-.0372
0.1265
0.7816
1.0000
Predicted
Proportion
Responding
0.0841
0.0000
0.0007
0.1179
0.7914
0.9975
Chi - Square Heterogeneity
Mu
Sigma
Parameter
Spontaneous
Response Rate
0.479736
0.150766
Estimate
0.441
Std. Err.
95* Confidence Limits
Intercept
Slope
1.818003
6.632814
0.976915
1.804695
( -0.096749,
( 3.095611,
3.732756)
10.170017)
0.084104
0.036007
0.013529,
0.154678)
Estimated EC Values and Confidence Limits
Point
.00
.00
EC 1.
EC 5.
EC10.00
EC15.00
EC50.00
EC85.00
EC90.00
EC95.00
EC99.00
Cone.
1.3459
1.7051
9343
1061
0181
3250
7093
5.3423
6.7680
Lower Upper
95% Confidence Limits
0.4533
0.7439
0.9654
1.1484
2.2676
3.5656
3.8443
4.2566
5.0712
1.9222
2.2689
2.4871
2.6523
3.6717
6.3827
7.5099
9.6486
15.6871
153
-------
Probit Analysis of Shcepshead Minnow Embryo-Larval Survival and Teratogenicity Data
PLOT OF ADJUSTED PROBITS AND PREDICTED REGRESSION LINE
Probit
10+ o
9+
8+
7+
6+
. .o.
4 +
.0
3+
2+
0+0
_+ + + + + + +_
EC01 EC10 EC25 EC50 EC75 EC90 EC99
Figure 4. Plot of adjusted probits and predicted regression line.
154
-------
TABLE 9. SINGLE-LABORATORY PRECISION OF THE SHEEPSHEAD MINNOW, CYPRINODON
VARIEGATUS, EMBRYO-LARVAL SURVIVAL AND TERATOGENICITY TEST
PERFORMED IN HW MARINEMIX* ARTIFICIAL SEAWATER, USING EMBRYOS FROM
FISH MAINTAINED AND SPAWNED IN HW MARINEMIXR ARTIFICIAL SEAWATER
USING COPPER (CU) AS REFERENCE TOXICANT1'2'3'4'5'6'7
Test
Number
1
2
3
4
5
6
7
8
n:
Mean:
CV(%):
EC1
(H9/L)
173
*
*
182
171
*
*
195
4
180
6.1
EC5
(MA)
189
*
*
197
187
*
*
203
4
194
3.8
EC10
(H9/L)
198
*
*
206
197
*
*
208
4
202
2.8
EC50
(MA)
234
*
*
240
234
*
*
226
4
233
2.5
NOEC
(MA)
240
240
240
240
240
<240
220
220
7
NA
NA
1Data from USEPA (1989a) and USEPA (1991a).
2Tests performed by Terry Hollister, Aquatic Biologist, Houston Facility,
Environmental Services Division, Region 6, U. S. Environmental Protection
Agency, Houston, Texas.
^Cyprinodon variegatus embryos used in the tests were less than 20 h
old when the tests began. Two replicate test chambers were used for the
control and each toxicant concentration. Ten embryos were randomly added to
each test chamber containing 250 mL of test or control water. Solutions
were renewed daily. The temperature and salinity of the test solutions were
24 + 1°C and 20°/oo, respectively.
4Copper test concentrations were prepared using copper sulfate. Copper
concentrations for Tests 1-5 were: 180, 210, 240, 270, and 300 pg/L.
Copper concentrations for Test 6 were: 220, 240, 260, 280, and 300 \ig/L.
Copper concentrations for Tests 7-8 were: 200, 220, 240, 260, and 280
Tests were conducted over a two-week period.
5Adults collected in the field.
6NOEC Range: 200 - 240 \iq/L (this represents a difference of two exposure
concentrations).
7For a discussion of the precision of data from chronic toxicity tests see
Section 4, Quality Assurance.
*Data do not fit the Probit model.
155
-------
TABLE 10. SINGLE-LABORATORY PRECISION OF THE SHEEPSHEAD MINNOW, CYPRINODON
VARIEGATUS, EMBRYO-LARVAL SURVIVAL AND TERATOGENICITY TEST
PERFORMED IN HW MARINEMIXR ARTIFICIAL SEAWATER, USING EMBRYOS FROM
FISH MAINTAINED AND SPAWNED IN HW MARINEMIXR ARTIFICIAL SEAWATER
USING SODIUM DODECYL SULFATE (SDS) AS REFERENCE TOXICANT1'2'3'4'5'6'7
Test
Number
1
2
3
4
5
n:
Mean.
CV(%):
EC1
(mg/L)
1.7
*
0.4
1.9
1.3
4
1.3
51.2
EC5
(mg/L)
2.0
*
0.7
2.2
1.7
4
1.6
41.6
EC10
(mg/L)
2.2
*
0.9
2.4
1.9
4
1.9
35.0
EC50
(mg/L)
3.1
*
2.5
3.3
3.0
4
2.9
11.7
NOEC
(mg/L)
2.0
4.0
2.0
2.0
2.0
5
NA
NA
1Data from USEPA (1989a) and USEPA (1991a).
2Tests performed by Terry Hollister, Aquatic Biologist, Houston Facility,
Environmental Services Division, Region 6, U. S. Environmental Protection
Agency, Houston, Texas.
3Cypn'nodon variegatus embryos used in the tests were less than 20 h
old when the tests began. Two replicate test chambers were used for the
control and each toxicant concentration. Ten embryos were randomly added to
each test chamber containing 250 mL of test or control water. Solutions
were renewed daily. The temperature and salinity of the test solutions were
24 + 1°C and 20 /oo, respectively.
SDS concentrations for all tests were: 0.5, 1.0, 2.0, 4.0, and 8.0 mg/L.
Tests were conducted over a three-week period.
5Adults collected in the field.
6NOEC Range: 2.0 - 4.0 mg/L (this represents a difference of two exposure
concentrations).
7For a discussion of the precision of data from chronic toxicity tests see
Section 4, Quality Assurance.
*Data do not fit the Probit model.
156
-------
Figure 5. Data form for Sheepshead minnow, Cyprinodon variegatus, embryo-larval
survival/teratogenicity test. Daily record of embryo-larval survival/terata and
test conditions.
Test Dates:
Species:
Type Effluent:
Field:
Lab:
Test:
Effluent Tested:
Original pH:
Salinity:
D.O.:
CONCENTRATION:
REPLICATE I:
DAYS
#Live/Dead
Embryo-Larvae
Terata
Temp. (°C)
Salinity (ppt)
D.O. (mg/L)
PH
0
1
2
3
4
CONCENTRATION:
REPLICATE II:
DAYS
#Live/Dead
Embryo-larvae
Terata
Temp. (°C)
Salinity (ppt)
D.O. (mg/L)
PH
0
1
2
3
4
5
6
5
b
7
8 j
9
7
8
9
Comments:
Note:Final endpoint for this test is total mortality (combined total number of
dead embryos, dead larvae, and deformed larvae). See Subsection 10.8 and 13.
157
-------
Figure 5 Data form for Sheepshead minnow, Cypnnodon vanegatus, embryo-larval
survival/teratogenicity test. Daily record of embryo-larval survival/terata and
test conditions (Continued).
CONCENTRATION:
REPLICATE III:
DAYS
#Live/Dead
Embryo-Larvae
Terata
Temp. (OC)
Salinity (ppt)
D.O. (mg/L)
PH
0
1
2
3
r 4
5
6
7
8
9
CONCENTRATION:
REPLICATE IV:
DAYS
#Live/Dead
Embryo-larvae
Terata
Temp. (°C)
Salinity (ppt)
D.O. (mg/L)
PH
0
1
2
3
4
5
6
7
8
9
Comments:
Note: Final endpoint for this test is total mortality (combined total number of
dead embryos, dead larvae, and deformed larvae). See Subsection 10.8 and 13.
158
-------
SECTION 13
TEST METHOD1'2
INLAND SILVERSIDE, HENIDIA BERYLLINA. LARVAL SURVIVAL AND GROWTH
METHOD 1006
1. SCOPE AND APPLICATION
1.1 This method estimates the chronic toxicity of effluents and receiving
waters to the inland silverside, Menidia beryllina, using seven-to-eleven day
old larvae in a seven day, static renewal test. The effects include the
synergistic, antagonistic, and additive effects of all the chemical, physical,
and biological components which adversely affect the physiological and
biochemical functions of the test species.
1.2 Daily observations on mortality make it possible to also calculate acute
toxicity for desired exposure periods (i.e., 24-h, 48-h, 96-h LCSOs).
1.3 Detection limits of the toxicity of an effluent or chemical substance are
organi sm-dependent.
1.4 Single or multiple excursions in toxicity may not be detected using 24-h
composite samples. Also, because of the long sample collection period
involved in composite sampling, and because the test chambers are not sealed,
highly volatile and highly degradable toxicants in the source may not be
detected in the test.
1.5 This test is commonly used in one of two forms: (1) a definitive test,
consisting of a minimum of five effluent concentrations and a control, and (2)
a receiving water test(s), consisting of one or more receiving water
concentrations and a control.
2. SUMMARY OF METHOD
2.1 Seven-to-eleven day old larvae are exposed in a static renewal system for
seven days to different concentrations of effluent or to receiving water.
Test results are based on the survival and growth (increase in weight) of the
larvae as compared to the control.
3. INTERFERENCES
3.1 Toxic substances may be introduced by contaminants in dilution water,
glassware, sample hardware, and testing equipment (see Section 5, Facilities,
Equipment, and Supplies).
3.2 Adverse effects of low dissolved oxygen (DO) concentrations, high
'The format used for this method was taken from USEPA (1983).
2This method was adapted from USEPA (1987c).
159
-------
concentrations of suspended and/or dissolved solids, and extremes of pH, may
mask or confound the effects of toxic substances.
3.3 Improper effluent sampling and handling may adversely affect test results
(see Section 8, Effluent and Receiving Water Sampling, Sample Handling, and
Sample Preparation for Toxicity Tests).
3.4 Pathogenic and/or predatory organisms in the dilution water and effluent
may affect test organism survival, and confound test results.
3.5 Food added during the test may sequester metals and other toxic
substances and confound test results.
4. SAFETY
4.1 See Section 3, Health and Safety.
5. APPARATUS AND EQUIPMENT
5.1 Facilities for holding and acclimating test organisms.
5.2 Brine shrimp, Artemia, culture unit -- see Subsection 6.16 below and
Section 7, Quality Assurance.
5.3 Mem'dia beryllina culture unit -- Subsection 6.17 below, Middaugh and
Hemmer (1984), Middaugh et al. (1986), USEPA (1987a) and USEPA (1991c) for
detailed culture methods. This test requires from 180 to 360 seven-to-eleven-
day-old larvae. It is preferable to obtain the test organisms from an inhouse
culture unit. If it is not feasible to culture fish inhouse, embryos or
larvae can be obtained from other sources by shipping them in well oxygenated
saline water in insulated containers.
5.4 Samplers -- automatic sampler, preferably with sample cooling capability,
that can collect a 24-h composite sample of 5 L.
5.5 Environmental chamber or equivalent facility with temperature control
(25 ± 1°C).
5.6 Water purification system -- Millipore Milli-QR, Deionized water (DI) or
equivalent.
5.7 Balance, analytical -- capable of accurately weighing to 0.0001 g.
5.8 Reference weights, Class S -- for checking performance of balance.
Weights should bracket the expected weights of the weighing boats and the
expected weights of the weighing boats plus fish.
5.9 Drying oven -- 105°C, for drying larvae.
5.10 Air pump -- for oil-free air supply.
160
-------
5.11 Air lines, plastic or pasteur pipettes, or air stones, -- for gently
aerating water containing the fragile larvae or for supplying air to test
solution with low DO.
5.12 pH and DO (non-stirring probe) meters -- for routine physical and
chemical measurements. Unless the test is being conducted to specifically
measure the effect of one of the above parameters, a portable, field-grade
instrument is acceptable.
5.13 Standard or micro-Winkler apparatus -- for calibrating DO (optional).
5.14 Desiccator -- for holding dried larvae.
5.15 Light box -- for counting and observing larvae.
5.16 Refractometer -- for determining salinity.
5.17 Thermometers, glass or electronic, laboratory grade -- for measuring
water temperatures.
5.18 Thermometers, bulb-thermograph or electronic chart type -- for
continuously recording temperature.
5.19 Thermometer, National Bureau of Standards Certified (see USEPA METHOD
170.1, USEPA, 1979b) -- to calibrate laboratory thermometers.
5.20 Test chambers -- four (minimum of three) chambers per concentration.
The chambers should be borosilicate glass or nontoxic disposable plastic
labware. To avoid potential contamination from the air, the chambers should
be covered during the test.
5.20.1 Each test chamber for the inland silverside should contain a minimum
of 750 ml of test solution. A modified Norberg-Mount (1985) chamber (Figure
1), constructed of glass and silicone cement, has been used successfully for
this test. This type of chamber holds an adequate column of test solution and
incorporates a sump area from which test solutions can be siphoned and renewed
without disturbing the fragile inland silverside larvae. Modifications for
the chamber are as follows: 1) 200 urn mesh nylon screen instead of stainless
steel screen; and 2) thin pieces of glass rods cemented with silicone to the
NYLONR screen to reinforce the bottom and sides to produce a sump area in one
end of the chamber. Avoid excessive use of silicone, while still ensuring
that the chambers do not leak and the larvae cannot get trapped or escape into
the sump area. Once constructed, check the chambers for leaks and repair if
necessary. Soak the chambers overnight in sea water (preferably in flowing
water) to cure the silicone cement before use. Other types of glass test
chambers, such as the 1000 ml beakers used in the short-term sheepshead minnow
larval survival and growth test, may be used. It is recommended that each
chamber contain a minimum of 50 ml per larvae and allow adequate depth of test
solution (5.0 cm).
5.21 Beakers -- six Class A, borosilicate glass or non-toxic plasticware,
1000 ml for making test solutions.
161
-------
9cm
GLASS
REINFORCEMENTS
9cm
SUMP
Figure 1. Glass chamber with sump area. Modified from Norberg and
Mount (1985). From USEPA (1987c).
162
-------
5.22 Mini-Winkler bottles -- for dissolved oxygen calibrations.
5.23 Wash bottles -- for deionized water, for washing embryos from substrates
and containers, and for rinsing small glassware and instrument electrodes and
probes.
5.24 Crystallization dishes, beakers, culture dishes, or equivalent -- for
incubating embryos.
5.25 Volumetric flasks and graduated cylinders -- Class A, borosilicate glass
or non-toxic plastic labware, 10-1000 ml for making test solutions.
5.26 Separatory funnels, 2-1 -- Two-four for culturing Artemia.
5.27 Pipets, volumetric -- Class A, 1-100 ml.
5.28 Pipets, automatic -- adjustable, 1-100 ml.
5.29 Pipets, serological -- 1-10 ml, graduated.
5.30 Pipet bulbs and fillers -- PROPIPETR, or equivalent.
5.31 Droppers, and glass tubing with fire polished edges, 4 mm ID -- for
transferring larvae.
5.32 Siphon with bulb and clamp -- for cleaning test chambers.
5.33 Forceps -- for transferring dead larvae to weighing boats.
5.34 NITEXR mesh sieves (< 150 urn and 500 \im) -- for collecting brine shrimp,
Artemia nauplii, and fish larvae.
6. REAGENTS AND CONSUMABLE MATERIALS
6.1 Sample containers -- for sample shipment and storage (see Section 8,
Effluent and Receiving Water Sampling, Sample Handling, and Sample Preparation
for Toxicity Tests).
6.2 Data sheets (one set per test) -- for data recording (Figures 7, 8, and
9).
6.3 Tape, colored -- for labelling test chambers
6.4 Markers, water-proof -- for marking containers, etc.
6.5 Vials, marked -- 24/test, containing 4% formalin or 70% ethanol, to
preserve larvae. (Optional).
6.6 Weighing boats, aluminum -- 26/test (2 extra).
6.7 pH buffers 4, 7, and 10 (or as per instructions of instrument
163
-------
manufacturer) for standards and calibration check (see USEPA Method 150.1,
USEPA, 1979b).
6.8 Membranes and solutions for dissolved oxygen probe (see USEPA Method
360.1, USEPA, 1979b), or reagents for modified Winkler analysis.
6.9 Laboratory quality assurance samples and standards for the above methods.
6.10 Reference toxicant solutions (see Section 4, Quality Assurance,
Subsections 4.7, 4.14, 4.15. 4.16, and 4.17).
6.11 Formalin (4%) or 70% ethanol for use as a preservative for the fish
larvae.
6.12 Reagent water -- defined as distilled or deionized water that does not
contain substances which are toxic to the test organisms (see paragraph 6.6
above).
6.13 Effluent, receiving water, and dilution water -- (see Section 7,
Dilution Water, and Section 8, Effluent and Surface Water Sampling, Sample
Handling, and Sample Preparation for Toxicity Tests).
6.13.1 Saline test and dilution water -- The salinity of the test water must
be in the range of 5 to 32 °/oo. The salinity should vary by no more than
+ 2 °/oo among the chambers on a given day. If effluent and receiving water
tests are conducted concurrently, the salinities of these tests should be
similar.
6.13.2 The overwhelming majority of industrial and sewage treatment effluents
entering marine and estuarine systems contain little or no measurable salts.
Exposure of Menidia beryl Una larvae to these effluents will require
adjustments in the salinity of the test solutions. It is important to
maintain a constant salinity across all treatments. In addition, it may be
desirable to match the test salinity with that of the receiving water.
Artificial sea salts or hypersaline brine (100 °/°o) derived from natural
seawater may be used to adjust the salinities. However, the use of
hypersaline brine will limit the concentration of effluent that be tested to
70% at 30 °/oo salinity and 80% at 20 °/oo salinity.
6.13.2.1 Artificial sea salts: A modified GP2 artificial seawater
formulation (Table 1) has been successfully used to perform the inland
silversides survival and growth tests. The use of GP2 for holding and
culturing of adults is not recommended at this time.
6.13.2.2 The GP2 artificial sea salts (Table 1) should be mixed with
deionized (DI) water or its equivalent in a container other than the culture
or testing tanks. The deionized water used for hydration should be between
21-26°C. The artificial seawater must be conditioned (aerated) for 24 hours
before use as the testing medium. If the solution is to be autoclaved, sodium
bicarbonate is added after the solution has cooled. A stock solution of
sodium bicarbonate is made up by dissolving 33.6 gm NaHC03 in 500 ml of
164
-------
TABLE 1. REAGENT GRADE CHEMICALS USED IN THE PREPARATION OF GP2 ARTIFICIAL
SEAWATER FOR THE INLAND SILVERSIDE, MENIDIA BERYLLINA, TOXICITY
TEST1'2'3
Compound Concentration Amount (g)
(g/L) Required for
20 L
1.
2.
3.
4.
5.
6.
7.
8.
9.
NaCl
Na2S04
KC1
KBr
Na2B407 . 10 H20
MgCl2 . 6 H20
CaCl2 . 2 H20
SrCl2 . 6 H20
NaHC03
21.03
3.52
0.61
0.088
0.034
9.50
1.32
0.02
0.17
420.6
70.4
12.2
1.76
0.68
190.0
26.4
0.400
3.40
Modified GP2 from Spotte et al. (1984)
2The constituent salts and concentrations were taken from
USEPA (1990b). The salinity is 30.89 g/L.
3GP2 can be diluted with deionized (DI) water to the desired test salinity.
165
-------
deionized water. Add 2.5 ml of this stock solution for each liter of the GP2
artificial seawater.
6.13.3 Hypersaline brine: Hypersaline brine (HSB) has several advantages
that make it desirable for use in toxicity testing. It can be made from any
high quality, filtered seawater by evaporation, and can be added to the
effluent or to deionized water to increase the salinity. HSB derived from
natural seawater contains the necessary trace metals, biogenic colloids, and
some of the microbial components necessary for adequate growth, survival,
and/or reproduction of marine and estuarine organisms, and may be stored for
prolonged periods without any apparent degradation.
6.13.3.1 The ideal container for making HSB from natural seawater is one that
(1) has a high surface to volume ratio, (2) is made of a noncorrosive
material, and (3) is easily cleaned (fiberglass is ideal). Special care
should be used to prevent any toxic materials from coming in contact with the
seawater being used to generate the brine. If a heater is immersed directly
into the seawater, ensure that the heater materials do not corrode or leach
any substances that would contaminate the brine. One successful method used
is a thermostatically controlled heat exchanger made from fiberglass. If
aeration is used, use only oil free air compressors to prevent contamination.
6.13.3.2 Before adding seawater to the brine generator, thoroughly clean the
generator, aeration supply tube, heater, and any other materials that will be
in direct contact with the brine. A good quality biodegradable detergent
should be used, followed by several (at least three) thorough deionized water
rinses. High quality (and preferably high salinity) seawater should be
filtered to at least 10 ^m before placing into the brine generator. Water
should be collected on,an incoming tide to minimize the possibility of
contamination.
6.13.3.3 The temperature of the seawater is increased slowly to 40°C. The
water should be aerated to prevent temperature stratification and to increase
water evaporation. The brine should be checked daily (depending on volume
being generated) to ensure that salinity does not exceed 100 °/oo and that the
temperature does not exceed 40°C. Additional seawater may be added to the
brine to obtain the volume of brine required.
6.13.3.4 After the required salinity is attained, the brine should be
filtered a second time through a 1 ^m filter and poured directly into portable
containers. Twenty-liter cubitainers or polycarbonate water cooler jugs are
suitable. The containers should be capped and labelled with the date the
brine was generated and its salinity. Containers of HSB should be stored in
the dark and maintained at room temperature until used.
6.13.3.5 If a source of HSB is available, test solutions can be made by
following the directions below. Thoroughly mix together the deionized water
and brine before mixing in the effluent.
6.13.3.6 Divide the salinity of the HSB by the expected test salinity to
determine the proportion of deionized water to brine. For example, if the
salinity of the HSB is 100 %o and the test is to be conducted at 20 °/oo,
166
-------
100 °/oo divided by 20 °/oo = 5.0. The proportion of brine is one part in
five (one part brine to four parts deionized water).
6.13.3.7 To make 1 L of seawater at 20 °/oo salinity from a MSB of 100 °/oo,
divide 1 L (1000 ml) by 5.0. The result, 200 ml, is the quantity of HSB
needed to make 1 L of seawater. The difference, 800 ml, is the quantity of
deionized water required.
6.13.3.8 Table 2 illustrates the composition of test solutions at 20 °/oo if
they are made by combining effluent (0 °/oo), deionized water and HSB at
100 °/oo salinity. The volume (ml) of brine required is determined by using
the amount calculated above. In this case, 200 ml of brine is required for 1
L; therefore, 600 ml would be required for 3 L of solution. The volumes of
HSB required are constant. The volumes of deionized water are determined by
subtracting the volumes of effluent and brine from the total volume of
solution: 3000 ml - ml effluent - mL brine = ml deionized water.
6.14 ROTIFER CULTURE
6.14.1 At hatching Mem'dia beryllina larvae are too small to ingest Artemia
nauplii and must be fed rotifers, Brachionus plicatilis. The rotifers can be
maintained in continous culture when fed algae (see Subsection 6.15 and USEPA,
1987a). Rotifers are cultured in 10-15 L Pyrex carboys (with a drain spigot
near the bottom) at 25-28°C and 25-35 °/oo salinity. Four 12-L culture
carboys should be maintained simultaneously to optimize production. Clean
carboys should be filled with autoclaved seawater. Alternatively, one can use
an immersion heater to heat saline water in the carboy to 70-80°C for 1 h.
6.14.2 When the water has cooled to 25-28°C, aerate and add a start-up sample
of rotifers (50 rotifers/mL) and food (about 1 L of a dense algal culture).
The carboys should be checked daily to ensure that adequate food is available
and that the rotifer density is adequate. If the water appears clear, drain 1
L of culture water and replace it with algae. Excess water can be removed
through the spigot drain and filtered through a < 60 \im mesh screen. Rotifers
collected on the screen should be returned to the culture. If a more precise
measure of the rotifer population is needed, rotifers collected from a known
volume of water can be resuspended in a smaller volume, killed with formalin
and counted in a Sedgwick-Rafter cell. If the density exceeds 50 rotifers/mL,
the amount of food per day should be increased to 2 L of algae suspension.
The optimum density of approximately 300-400 rotifers/mL may be reached in 7
10 days and is sustainable for 2-3 weeks. At these densities, the rotifers
should be cropped daily. Keeping the carboys away from light will reduce the
amount of attached algae on the carboy walls. When detritus accumulates,
populations of ciliates, nematodes, or harpacticoid copepods that may have
been inadvertently introduced can rapidly take over the culture. If this
occurs, discard the cultures.
6.15 ALGAL CULTURES
6.15.1 Tetraselmus suecica or Chlorella sp. (see USEPA, 1987a) can be
cultured in 20-L polycarbonate carboys that are normally used for bottled
drinking water. Filtered seawater is added to the carboys and then autoclaved
167
-------
(110°C for 30 min). After cooling to room temperataure, the carboys are
placed in a controlled temperature chamber at 18-20°C. One liter of T.
suecica or Chlorella sp. starter culture and 100 ml of nutrients are added to
each carboy.
TABLE 2. PREPARATION OF 3 L SALINE WATER FROM DEIONIZED WATER AND A
HYPERSALINE BRINE OF 100 %>o NEEDED FOR TEST SOLUTIONS AT
20 °/oo SALINITY.
Effluent
Concentration
(%)
80.0
40.0
20.0
10.0
5.00
Control
Total
Volume of
Effluent
(0 °/oo)
(mL)
2400.0
1200.0
600.0
300.0
150.0
0.0
4650.0
Volume of
De ionized
Water
(ml)
0.0
1200.0
1800.0
2100.0
2250.0
2400.0
9750.0
Volume of
Hypersaline
Brine
(mL)
600.0
600.0
600.0
600.0
600.0
600.0
3600.0
Total
Volume
(mL)
3000.0
3000.0
3000.0
3000.0
3000.0
3000.0
18000.0
6.15.2 Formula for algal culture nutrients
6.15.2.1 Add 180 g NaNO,, 12 g NaH2P04, and 6.16 g EDTA to 12 L of deionized
water. Mix with a magnetic stirrer until all salts are dissolved (at least
1 h).
6.15.2.2 Add 3.78 g FeCl3 ' 6 H20 and stir again. The solution should be
bright yellow.
6.15.2.3 The algal culture is vigorously aerated via a pipette inserted
through a foam stopper at the top of the carboy. A dense algal culture should
develop in 7 - 10 days and should be used by Day 14. Thus, start-up of
cultures should be made on a daily or every second day basis. Approximately 6
- 8 continuous cultures will meet the feeding requirements of four 12-L
168
-------
rotifer cultures. When emptied, carboys are washed with soap and water and
rinsed thoroughly with deionized water before reuse.
6.16 BRINE SHRIMP, ARTEMIA, NAUPLII (see USEPA, 1991c).
6.16.1 Although there are many commercial sources of brine shrimp cysts, the
Brazilian or Colombian strains are being used because the supplies examined
have had low concentrations of chemical residues and produce nauplii of
suitably small size. For commercial sources of brine shrimp, Artemia, cysts,
see Table 2 (Section 5, Facilities, Equipment, and Supplies) and Section 4,
Quality Assurance, Subsection 4.8 Food Quality. Each new batch of Artemia
cysts must be evaluated for size (Vanhaecke and Sorgeloos, 1980, and Vanhaecke
et al., 1980) and nutritional suitability (see Leger et al., 1985, 1986)
against known suitable reference cysts by performing a side by side larval
growth test using the "new" and "reference" cysts. The "reference" cysts used
in the suitability test may be a previously tested and acceptable batch of
cysts, or may be obtained from the Quality Assurance Research Division,
Environmental Monitoring Systems Laboratory, Cincinnati, Ohio, 45268, phone:
513-569-7325. A sample of newly-hatched Artemia nauplii from each new batch
of cysts should be chemically analyzed to establish that the concentration of
total organic chlorine does not exceed 0.15 \ig/g wet weight or that the total
concentration of organochlorine pesticides plus PCBs does not exceed 0.3 jig/g
wet weight. If those concentrations are exceeded, the Artemia should not be
used. For analytical methods, see USEPA (1982).
6.16.2 Artemia nauplii are obtained as follows:
1. Add 1-L of seawater, or a solution prepared by adding 35.0 g uniodized
salt (NaCl) or artificial sea salts to 1 L of deionized water, to a 2-L
separatory funnel or equivalent.
2. Add 10 ml Artemia cysts to the separatory funnel and aerate for 24 h at
27°C (Hatching time varies with incubation temperature and the
geographic strain of Artemia used. See USEPA (1985d), USEPA (1991c),
for details on Artemia culture and quality control.)
3. After 24 h, cut off the air supply in the separatory funnel. Artemia
nauplii are phototactic. Therefore, place a dark cloth or paper towel
over the top of the separatory funnel for 5-10 min, allowing the
nauplii to concentrate at the bottom. Leaving the nauplii concentrated
on the bottom much longer than 10 min without aeration will result in
mortality.
4. Drain the nauplii into a funnel fitted with a 150 \im NITEXR screen,
and rinse with sea water or equivalent before use.
6.16.3 Testing Artemia nauplii as food for bioassay organisms.
6.16.3.1 The primary criterion for acceptability of each new supply of brine
shrimp cysts is the ability of the nauplii to support good survival and growth
of the inland silverside larvae. The larvae used to evaluate the suitability
of the brine shrimp nauplii must be of the same geographical origin, species,
and stage of development as those used routinely in the toxicity tests. Three
replicate test vessels, each containing 15 larvae, used for each type of food,
will provide sufficient data to detect differences in survival and growth.
169
-------
6.16.3.2 The feeding rate and frequency, test vessels and volume of control
water, duration of the test, and age of the nauplii at the start of the test,
should be the same as used for the routine toxicity tests.
6.16.3.3 Results of the brine shrimp nutrition assay, where there are only
two treatments, can be evaluated statistically by use of a t-test. The "new"
food is acceptable if there are no statistically significant differences in
the survival and growth of the larvae fed the two sources of nauplii.
6.16.4 Use of Artemia nauplii as food for inland silverside, Menidia
beryl Una, larvae.
6.16.4.1 Menidia beryl Una larvae begin feeding on newly hatched Artemia
nauplii about five days after hatching, and are fed Artemia nauplii daily
throughout the 7-day larval survival and growth test. Survival of Menidia
beryl Una larvae 7-9 days old is improved by feeding newly hatched (< 24 hours
old) Artemia nauplii. Equal amounts of Artemia nauplii must be fed to each
replicate test chamber to minimize the variability of larval weight.
Sufficient numbers of nauplii should be fed to ensure that some remain alive
overnight in the test chambers. An adequate but not excessive amount should
be provided to each replicate on a daily basis. Feeding excessive amounts of
nauplii will result in a depletion in DO to below an acceptable level (below
4.0 mg/L). As much of the uneaten Artemia nauplii as possible should be
siphoned from each chamber prior to test solution renewal to ensure that the
larvae principally eat newly hatched nauplii.
6.17 INLAND SILVERSIDE, MENIDIA BERYLLINA, BROOD STOCK
6.17.1 The inland silverside, Menidia beryllina, is one of three species in
the atherinid family that are amenable to laboratory culture; and one of four
atherinid species used for chronic toxicity testing. Several atherinid
species have been utilized successfully for early life stage toxicity tests
using field collected (Goodman et al., 1985) and laboratory reared adults
(Middaugh and Takita, 1983; Middaugh and Hemmer, 1984; and USEPA, 1987a). The
inland silverside, Menidia beryllina, populates a variety of habitats from
Cape Cod, Massachusetts, to Florida and west to Vera Cruz, Mexico (Johnson,
1975). It can tolerate a wide range of temperature, 2.9 32.5°C (Tagatz and
Dudley, 1961; Smith, 1971) and salinity, of 0 58 °/oo (Simmons, 1957;
Renfro, 1960), having been reported from the freshwaters of the Mississippi
River drainage basin (Chernoff et al., 1981) to hypersaline lagoons (Simmons,
1957). Ecologically, Menidia spp. are important as major prey for many
prominent commercial species (e.g., bluefish (Pomatomus saltatrix), mackerel
(Scomber scombrus), and striped bass (Morone saxatilis) (Bigelow and
Schroeder, 1953). The inland silverside, Menidia beryllina, is a serial
spawner, and will spawn under controlled laboratory conditions. Spawning can
be induced by diurnal interruption in the circulation of water in the culture
tanks (Middaugh et al., 1986; USEPA, 1987a). The eggs are demersal,
approximately 0.75 mm in diameter (Hildebrand and Schroeder, 1928), and adhere
to vegetation in the wild, or to filter floss in laboratory culture tanks.
The larvae hatch in 6-7 days when incubated at 25°C and maintained in seawater
ranging from 5-30 /oo (USEPA, 1987a). Newly hatched larvae are 3.5-4.0 mm in
total length (Hildebrand, 1922).
170
-------
6.17.2 Inland silverside, Menidia beryllina, adults (see USEPA, 1987a and
USEPA, 1991c for detailed culture methods) may be cultured in the laboratory
or obtained from the Gulf of Mexico or Atlantic coast estuaries throughout the
year (Figure 2). Gravid females can be collected from low salinity waters
along the Atlantic coast during April to July, depending on the latitude. The
most productive and protracted spawning stock can be obtained from adults
brought into the laboratory. Broodstocks, collected from local estuaries
twice each year (in April and October), will become sexually active after 1-2
months and will generally spawn for 4-6 months.
6.17.3 The fish can be collected easily with a beach siene (3-6 mm mesh), but
the seine should not be completely landed onto the beach. Silversides are
very sensitive to handling and should never be removed from the water by net
-- only by beaker or bucket.
6.17.4 Samples may contain a mixture of inland silverside, Menidia beryllina,
and Atlantic silverside, Menidia menidia, on the Atlantic coast or inland
silverside and tidewater silverside, Menidia peninsulae, on the Gulf Coast
(see USEPA, 1987a for additional information on morphological differences for
identification). Johnson (1975) and Chernoff et al. (1981) have attempted to
differentiate these species. In the northeastern United States, M. beryllina
juveniles and adults are usually considerably smaller than M. menidia
juveniles and adults (Bengtson, 1984), and can be separated easily in the
field on that basis.
6.17.5 Record the water temperature and salinity at each collection site.
Aerate (portable air pump, battery operated) the fish and transport to the
laboratory as quickly as possible after collection. Upon arrival at the
laboratory, the fish and the water in which they were collected are
transferred to a tank at least 0.9 m in diameter. A filter system should be
employed to maintain water quality (see USEPA, 1987a). Laboratory water is
added to the tank slowly, and the fish are acclimated at the rate of 2°C per
day, to a final temperature of 25°C, and about 5 °/oo salinity per day, to a
final salinity in the range of 20 - 32 °/oo. The seawater in each tank should
be brought to a minimum volume of 150 L. A density of about 50 fish/tank is
appropriate. Maintain a photoperiod of 16 h light/8 h dark. Feed the adult
fish flake food or frozen brine shrimp twice daily and Artemia nauplii once
daily. Siphon the detritus from the bottom of the tanks weekly.
6.17.6 Larvae for a toxicity test can be obtained from the broodstock by
spawning onto polyester aquarium filter-fiber substrates, 15 cm long X 10 cm
wide X 10 cm thick, which are suspended with a string 8-10 cm below the
surface of the water and in contact with the side of the holding tanks for
24-48 h, 14 days prior to the beginning of a test. The floss should be gently
aerated by placing it above an airstone, and weighted down with a heavy
non-toxic object. The embryos, which are light yellow in color, can be seen
on the floss, and are round and hard to the touch compared to the soft floss.
6.17.7 Remove as much floss as possible from the embryos. The floss should
be stretched and teased to prevent the embryos from clumping. The embryos
should be incubated at the test salinity and lightly aerated. At 25°C, the
embryos will hatch in about 6-8 days. Larvae are fed about 500
171
-------
A. Adult, ca. 64 mm SI
Figure 2. Inland silverside, Mem'dia beryllina: A. Adult, ca. 64 mm SL; B.
Egg (diagrammatic), only bases of filaments shown; C. Egg, 2-cell stage; D
Egg, morula stage; E. Advanced embryo, 2 1/2 days after fertilization. From
Martin and Drewry (1978).
172
-------
rotifers/larva/day from hatch through four days post-hatch. On Days 5 and 6,
newly hatched (less than 12-h old) Artemia nauplii are mixed with the
rotifers, to provide a transition period. After Day 7, only nauplii are fed,
and the age range for the nauplii can be increased from 12-h old to 24-h old.
6.17.8 Silverside larvae are very sensitive to handling and shipping during
the first week after hatching. For this reason, if organisms must be shipped
to the test laboratory, it may be impractical to use larvae less than 11 days
old because the sensitivity of younger organisms may result in excessive
mortality during shipment. If organisms are to be shipped to a test site,
they should be shipped only as (1) early embryos, so that they hatch after
arrival, or (2) after they are known to be feeding well on Artemia nauplii (8
10 days of age). Larvae shipped at 8 - 10 days of age would be 9 - 11 days
old when the test is started. Larvae that are hatched and reared in the test
laboratory can be used at seven days of age.
6.17.9 If four replicates of 15 larvae are used at each effluent
concentration and in the control, 360 larvae will be needed for each test.
7. EFFLUENT AND RECEIVING WATER SAMPLE COLLECTION, PRESERVATION, AND STORAGE
7.1 See Section 8, Effluent and Receiving Water Sampling, Sample Handling,
and Sample Preparation for Toxicity Tests.
8. CALIBRATION AND STANDARDIZATION
8.1 See Section 4, Quality Assurance.
9. QUALITY CONTROL
9.1 See Section 4, Quality Assurance.
10. TEST PROCEDURES
10.1 TEST SOLUTIONS
10.1.1 Receiving Waters
10.1.1.1 The sampling point is determined by the objectives of the test. At
estuarine and marine sites, samples are usually collected at mid-depth.
Receiving water toxicity is determined with samples used directly as collected
or with samples passed through a 60 ^m NITEXR filter and compared without
dilution, against a control. Using four replicate chambers per test, each
containing 500-750 mL, and 400 mL for chemical analysis, would require
approximately 3400 mL or more of sample per test.
10.1.2 Effluents
10.1.2.1 The selection of the effluent test concentrations should be based on
the objectives of the study. A dilution factor of 0.5 is commonly used. A
dilution factor of 0.5 provides precision of + 100%, and testing of
concentrations between 6.25% and 100% effluent using only five effluent
173
-------
concentrations (6.25%, 12.5%, 25%, 50%, and 100%). Test precision shows
little improvement as dilution factors are increased beyond 0.5 and declines
rapidly if smaller dilution factors are used. Therefore, USEPA
recommends a dilution factor of 0.5. If hypersaline brine is used to adjust
salinities, the maximum effluent concentration will be 80% at 20 °/oo
salinity, and 70% at 30 °/oo salinity.
10.1.2.2 If the effluent is known or suspected to be highly toxic, a lower
range of effluent concentrations should be used (such as 25%, 12.5%, 6.25%,
3.12%, and 1.56%). If a high rate of mortality is observed during the first 1
to 2 h of the test, additional dilutions at the lower range of effluent
concentrations should be added.
10.1.2.3 The volume of effluent required to start the test and for daily
renewal of four replicates per treatment, each containing 750 ml of test
solution, is approximately 5 L. Prepare enough test solution at each effluent
concentration to provide 400 mL additional volume for chemical analyses.
10.1.2.4 Tests should begin as soon as possible after sample collection,
preferably within 24 h. The maximum holding time following retrieval of the
sample from the sampling device should not exceed 36 h for off-site toxicity
studies unless permission is granted by the permitting authority. In no case
should the test be started more than 72 h after sample collection.
10.1.2.5 Just prior to test initiation (approximately one h), the temperature
of a sufficient quantity of the sample to make the test solution should be
adjusted to the test temperature (25 + 1°C) and maintained at that temperature
during the addition of dilution waters.
10.2 START OF THE TEST
10.2.1 Inland silverside larvae seven to 11 days old can be used to start the
survival and growth test. At this age, the inland silverside feed on
newly-hatched Artemia nauplii. At 25°C, tests with inland silverside larvae
can be performed at salinities ranging from 5 °/oo to 32 °/oo. If the test
salinity ranges from 16 to 32 °/oo, the salinity for spawning, incubation, and
culture of the embryos and larvae should be maintained within this salinity
range. If the test salinity is in the range of 5 °/oo to 15 °/oo, the embryos
may be spawned at 30 °/oo, but egg incubation and larval rearing should be at
the test salinity. If the specific salinity required for the test differs
from the rearing salinity, adjustments of 5 °/oo daily should be made over the
three days prior to start of test.
10.2.2 One Day Prior to Beginning of Test.
10.2.2.1 Set up the Artemia culture so that newly hatched nauplii will be
available on the day the test begins. (See Subsection 7.16 above).
10.2.2.2 Increase the temperature of water bath, room, or incubator to the
required test temperature (25 + 1°C).
10.2.2.3 Label the test chambers with a marking pen and identify each
174
-------
concentration and replicate with various colored coded tape. A minimum of
five effluent concentrations and a control should be selected for each test.
Glass test chambers, such as crystallization dishes, beakers, or chambers with
a sump area (Figure 1), with a capacity for 500 750 ml of test solution,
should be used.
10.2.2.4 Randomize the position of test chambers in the temperature-
controlled water bath, room, or incubator at the beginning of the test, using
a position chart. Assign numbers for the position of each test chamber using
a table of random numbers or similar process (see Appendix A for an example of
randomization). Maintain the chambers in this configuration throughout the
test, using a position chart.
10.2.2.5 Because inland silverside larvae are very sensitive to handling, it
is advisable to distribute them to their respective test chambers which
contain control water on the day before the test is to begin. Each test
chamber should contain 15 larvae (minimum 10) and it is recommended that there
be four replicates (minimum of three) for each concentration and control.
10.2.2.6 Seven-to-eleven day old larvae are active and difficult to capture
and are subject to handling mortality. Carefully remove larvae (2-3 at a
time) by concentrating them in a corner of the aquarium or culture vessel, and
capture them with a wide-bore pipette, small petri dish, crystallization dish,
3-4 cm in diameter, or small pipette. They are active and will readily escape
from a pipette. Randomly transfer the larvae (2-3 at a time) into each test
chamber until the desired number (15) is attained. See Appendix A for an
example of randomization. After the larvae are dispensed, use a light table
to verify the number in each chamber.
10.2.3 Before beginning the test remove and replace any dead larvae from each
test chamber. The test is started by removing approximately 90% of the clean
sea water from each test chamber and replacing with the appropriate test
solution.
10.3 LIGHT, PHOTOPERIOD, SALINITY, AND TEMPERATURE
10.3.1 The light quality and intensity should be at ambient laboratory
levels, which is approximately 10-20 nE/m2/s, or 50 to 100 foot candles
(ft-C), with a photoperiod of 16 h of light and 8 h of darkness. The test
salinity should be in the range of 5 °/oo to 32 °/oo, and the salinity should
not vary by more than + 2 °/°° among the chambers on a given day. If effluent
and receiving water tests are conducted concurrently, the salinities of these
tests should be similar. The water temperature in the test chambers should be
maintained at 25 ± 2°C.
10.4 DISSOLVED OXYGEN (DO) CONCENTRATION
10.4.1 Aeration may affect the toxicity of effluents and should be used only
as a last resort to maintain satisfactory DO. The DO should be measured on
new solutions at the start of the test (Day 0) and before daily renewal of
test solutions on subsequent days. The DO should not fall below 4.0 mg/L. If
it is necessary to aerate, all concentrations and the control should be
175
-------
aerated. The aeration rate should not exceed 100 bubbles/min, using
a pipet with an orifice of approximately 1.5 mm, such as a 1-mL, Kimax
serological pipet, No. 37033, or equivalent. Care should be taken to ensure
that turbulence resulting from aeration does not cause undue stress to the
fish.
10.5 FEEDING
10.5.1 Artemia nauplii are prepared as described above.
10.5.2 The test larvae are fed newly-hatched (less than 24-h-old) Artemia
nauplii once a day from Day 0 through Day 6; larvae are not fed on Day 7.
Equal amounts of Artemia must be fed to each replicate test chamber to
minimize the variability of larval weight. Sufficient numbers of nauplii
should be fed to ensure that some remain alive overnight in the test chambers.
An adequate, but not excessive amount of Artemia nauplii, should be provided
to each replicate on a daily basis. Feeding excessive amounts of Artemia
nauplii will result in a depletion in DO to below an acceptable level. Siphon
as much of the uneaten Artemia nauplii as possible from each chamber daily to
ensure that the larvae principally eat newly hatched nauplii.
10.5.3 On days 0, 1, and 2, transfer 4 g wet weight or 4 ml of concentrated,
rinsed Artemia nauplii to seawater in a 100 ml beaker, and bring to a volume
of 80 ml. Aerate or swirl the suspension to equally distribute the nauplii
while withdrawing individual 2 ml portions of the Artemia nauplii suspension
by pipette or adjustable syringe to transfer to each replicate test chamber.
Because the nauplii will settle and concentrate at the tip of the pipette
during the transfer, limit the volume of concentrate withdrawn each time to a
2-mL portion for one test chamber helps ensure an equal distribution to the
replicate chambers. Equal distribution of food to the replicates is critical
for successful tests.
10.5.4 On Days 3-6, transfer 6 g wet weight or 6 ml of the Artemia nauplii
concentrate to sea water in a 100 ml beaker. Bring to a volume of 80 ml and
dispense as described above.
10.5.5 If the larvae survival rate in any replicate on any day falls below
50%, reduce the volume of Artemia nauplii suspension added to that test
chamber by one-half (i.e., reduce from 2 ml to 1 ml) and continue feeding
one-half the volume through Day 6. Record the time of feeding on the data
sheets.
10.6 DAILY CLEANING OF TEST CHAMBERS
10.6.1 Before the daily renewal of test solutions, uneaten and dead brine
shrimp, and other debris are removed from the bottom of the test chambers with
a siphon hose. Alternately, a large pipet (50 mL), fitted with a safety pipet
filler or rubber bulb, can be used. If the test chambers illustrated in
Figure 1 are used, remove only as much of the test solution from the chamber
as is necesssary to clean, and siphon the remainder of the test solution from
the sump area. Because of their small size during the first few days of the
test, larvae are easily drawn into a siphon tube when cleaning the test
176
-------
chambers. By placing the test chambers on a light box, inadvertent removal of
larvae can be greatly reduced because they can be more easily seen. If the
water siphoned from the test chambers is collected in a white plastic tray,
the larvae caught up in the siphon can be carefully retrieved, gently returned
by pipette to the appropriate test chamber and noted on data sheet. When
handling mortality occurs, record the number and chamber position of dead
larvae.
10.7 TEST SOLUTION RENEWAL
10.7.1 The test solutions are renewed daily, immediately after cleaning the
test chambers. The water level in each chamber is lowered to a depth of
7-to-10 mm, leaving 10 to 15% of the test solution. New test solution is
added slowly by refilling each chamber with the appropriate amount of test
solution without excessively disturbing the larvae. If the modified chamber
is used (Figure 1), renewals should be poured into the sump area using a
narrow bore (approximately 9 mm ID) funnel.
10.7.2 The effluent or receiving water used in the test is stored in an
incubator or refrigerator at 4°C. Plastic containers such as 8-20 L
cubitainers have proven suitable for effluent collection and storage. For
on-site toxicity studies no more than 24 h should elapse between collection of
the effluent and use in a toxicity test (see Section 8).
10.7.3 Approximately 1 h before test initiation, a sufficient quantity of
effluent or receiving water sample is warmed to 25 ± 1°C to prepare the test
solutions. A sufficient quantity of effluent should be warmed to make the
daily test solutions.
10.7.3.1 An illustration of the quantities of effluent and sea water needed
to prepare test solution at the appropriate salinity is provided in Table 2.
10.7.3.2 The higher effluent concentrations (i.e., 25% and above) may require
aeration to maintain adequate dissolved oxygen concentrations. However, if
one solution is aerated, then all concentrations must be aerated. Aerate the
test solutions in the test chambers gently, so that larvae are not disturbed.
10.8 ROUTINE CHEMICAL AND PHYSICAL ANALYSIS
10.8.1 At a minimum, the following measurements are made and recorded
(Figure 11):
10.8.1.1 DO is measured at the beginning and end of each 24-h exposure period
in one test chamber at all test concentrations and in the control.
10.8.1.2 Temperature, pH, and salinity are measured at the end of each 24-h
exposure period in one test chamber at all test concentrations and in the
control. The pH is measured in the effluent sample each day.
10.9 OBSERVATIONS DURING THE TEST
177
-------
10.9.1 The number of live larvae in each test chamber are recorded daily
(Figure 11), and the dead larvae are discarded.
10.9.2 Daily test observations, solution renewals, and removal of dead
larvae, should be carried out carefully to protect the larvae from unnecessary
disturbance during the test. Care should be taken to see that the larvae
remain immersed at all times during the performance of the above operations.
10.10 TERMINATION OF THE TEST
10.10.1 The test is terminated after seven days of exposure. At termination,
the number of surviving larvae in each test chamber are counted and as a group
are immediately prepared for drying and weighing, or are preserved in 4%
formalin or 70% ethanol for drying and weighing at a later date. For
immediate drying and weighing, siphon or pour live larvae onto a 500 urn mesh
screen in a large beaker to retain the larvae and allow Artemia to be rinsed
away. Rinse the larvae with cold deionized water to remove salts that might
contribute to the dry weight. Sacrifice the larvae in an ice bath of
deionized water. Small aluminum weighing boats can be used to dry and weigh
larvae. An appropriate number of aluminum weigh boats (one per replicate) are
marked for identification and weighed to 0.01 mg, and the weights are recorded
(Figure 12) on the data sheets.
10.10.2 Immediately before drying, the preserved larvae are rinsed in
distilled water. The rinsed larvae from each test chamber are transferred,
using forceps, to a tared weighing boat and dried at 60°C for 24 h, or at
105°C for a minimum of 6 h. Immediately upon removal from the drying oven,
the weighing boats are placed in a desiccator to cool and to prevent the
adsorption of moisture from the air until weighed. Weigh all weighing boats
containing the dried larvae to 0.01 mg, subtract the tare weight to determine
dry weight of larvae in each replicate, and record (Figure 12) of data
sheets). Divide the dry weight by the number of larvae per replicate to
determine the average dry weight, and record (Figures 12 and 13) on the data
sheets. Complete the summary data sheet (Figure 13) after calculating the
average measurements and statistically analyzing the dry weights and percent
survival for the entire test.
11. ACCEPTABILITY OF TEST RESULTS
11.1 Test results are acceptable if (1) the average survival of control
larvae is equal to or greater than 80%, and (2) where the test starts with
7-day old larvae, the average dry weight of the control larvae, when dried
immediately after test termination, is equal to or greater than 0.50 mg, or
the average dry weight of the control larvae preserved in 4% formalin or 70%
ethanol is equal to or greater than 0.43 mg.
12. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA
12.1 A summary of test conditions and test acceptability criteria is listed
in Table 3.
13. DATA ANALYSIS
178
-------
13.1 GENERAL
13.1.1 Tabulate and summarize the data.
13.1.2 The endpoints of toxicity tests using the inland silverside are based
on the adverse effects on survival and growth. The LC50, the IC25, and the
IC50 are calculated using point estimation techniques (see Section 9, Chronic
Toxicity Test End Endpoints and Data Analysis). LOEC and NOEC values, for
survival and growth, are obtained using a hypothesis test approach such as
Dunnett's Procedure (Dunnett, 1955) or Steel's Many-one Rank Test (Steel,
1959; Miller, 1981). See the Appendices for examples of the manual
computations, program listings, and examples of data input and program output.
13.1.3 The statistical tests described here must be used with a knowledge of
the assumptions upon which the tests are contingent. The assistance of a
statistician is recommended for analysts who are not proficient in statistics.
13.2 EXAMPLE OF ANALYSIS OF INLAND SILVERSIDE, MENIDIA BERYLLINA, SURVIVAL
DATA
13.2.1 Formal statistical analysis of the survival data is outlined in Fiqure
3. The response used in the analysis is the proportion of animals surviving
in each test or control chamber. Separate analyses are performed for the
estimation of the NOEC and LOEC endpoints and for the estimation of the LC50
endpoint. Concentrations at which there is no survival in any of the test
chambers are excluded from statistical analysis of the NOEC and LOEC, but
included in the estimation of the LC50.
13.2.2 For the case of equal numbers of replicates across all concentrations
and the control, the evaluation of the NOEC and LOEC endpoints is made via a
parametric test, Dunnett's Procedure, or a nonparametric test, Steel's
Many-one Rank Test, on the arc sine transformed data. Underlying assumptions
of Dunnett's Procedure, normality and homogeneity of variance, are formally
tested. The test for normality is the Shapiro-Wilk's Test, and Bartlett's
Test is used to determine the homogeneity of variance. If either of these
tests fail, the nonparametric test, Steel's Many-one Rank Test, is used to
determine the NOEC and LOEC endpoints. If the assumptions of Dunnett's
Procedure are met, the endpoints are estimated by the parametric procedure.
13.2.3 If unequal numbers of replicates occur among the concentration levels
tested, there are parametric and nonparametric alternative analyses. The
parametric analysis is a t-test with the Bonferroni adjustment. The Wilcoxon
Rank Sum Test with the Bonferroni adjustment is the nonparametric alternative.
For detailed information on the Bonferroni adjustment, see Appendix D.
13.2.4 Probit Analysis (Finney, 1971) is used to estimate the LC50. In this
analysis, the total mortality data from all test replicates at a given
concentration are combined. If the data do not fit the Probit model, the
Spearman-Karber method, the trimmed Spearman-Karber method, or the Graphical
method may be used (USEPA, 1991c).
179
-------
TABLE 3. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA FOR THE
INLAND SILVERSIDE, HENIDIA BERYLLINA, LARVAL SURVIVAL AND GROWTH
TEST WITH EFFLUENTS AND RECEIVING WATERS
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
concentrations:
10. Age of test organisms:
11. Larvae/test chamber
and control:
12. Replicate
chambers/concentration:
13. Source of food:
Static renewal
5 °/oo to 32 °/oo (+ 2 °/oo Of
the selected test salinity)
25 ± 1°C
Ambient laboratory illumination
10-20 nE/mz/s (50-100 ft-c) (ambient
laboratory levels)
16 h light, 8 h darkness
600 mL - 1 L containers
500-750 mL/replicate (loading and
DO restrictions must be met)
Daily
7-11 days post hatch; 24-h range in age
15 (minimum of 10)
4 (minimum of 3)
Newly hatched Artemia nauplii (survival
of 7-9 days old Menidia beryTlina larvae
improved by feeding < 24 hours old Artemia)
14. Feeding regime:
15. Cleaning:
Feed 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
180
-------
TABLE 3. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA FOR THE
INLAND SILVERSIDE, MENIDIA BERYLLINA, LARVAL SURVIVAL AND GROWTH
TEST WITH EFFLUENTS AND RECEIVING WATERS (CONTINUED)
16. Aeration:
17. Dilution water:
18. Test concentrations:
19. Dilution factor:
20. Test duration:
21. Effects measured:
22. Test acceptability
criteria:
23. Sampling requirement:
24. Estimated maximum sample
volume required:
None, unless DO concentration falls
below 4.0 mg/L, then aerate all chambers.
Rate should be less than 100 bubbles/min.
Uncontaminated source of natural
sea water, artificial seawater (GP2, FORTY
FATHOMSR, or equivalent) or hypersaline
brine mixed with deionized water.
Effluents: Minimum of five effluent
concentrations and a control
Receiving Waters: 100% receiving water and
a control
Effluents: > 0.5 series
Receiving waters: None, or > 0.5 series
7 days
Survival and growth (weight)
80% or greater survival in controls,
0.50 mg average dry weight of control
larvae where test starts with 7-days old
larvae and dried immediately after test
termination, or 0.43 mg or greater average
dry weight of control larvae preserved in
4% formalin or 70% ethanol
For on-site tests, samples are collected
daily, and used within 24 h of the time
they are removed from the sampling device.
For off-site tests, a minimum of three
samples are collected on days one, three,
and five with a maximum holding time of
36 h before first use (see Section 13,
Subsection 10.1.2.4).
7 L per day
181
-------
STATISTICAL ANALYSIS OF INLAND SILVERSIDE LARVAL
SURVIVAL AND GROWTH TEST
SURVIVAL
SURVIVAL DATA
PROPORTION SURVIVING
ENDPOINT ESTIMATE
LC50
ARC SINE
TRANSFORMATION
i
SHAPIRO-WILK'S TEST
NORMAL DISTRIBUTION
NON-NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
NO
BARTLETT'S TEST
HETEROGENEOUS
VARIANCE
EQUAL NUMBER OF
REPLICATES?
YES
T-TEST WITH
BONFERRONI
ADJUSTMENT
EQUAL NUMBER OF
REPLICATES?
NO
YES
DUNNETT'S
TEST
STEEL'S MANY-ONE
RANK TEST
WILCOXON RANK SUM
TEST WITH
BONFERRONI ADJUSTMENT
ENDPOINT ESTIMATES
NOEC.LOEC
Figure 3. Flowchart for statistical analysis of the inland silverside,
Menidia beryl Una, survival data.
182
-------
13.2.5 The proportion surviving in each replicate in this example must first
be transformed by the arc sine transformation procedure described in Appendix
B. The raw and transformed data, means and standard deviations of the
transformed observations at each effluent concentration and control are listed
in Table 4. A plot of the data is provided in Figure 4. Since there is 100%
mortality in all three replicates for the 50% and 100% concentrations, they
are not included in this statistical analysis and are considered a qualitative
mortality effect.
TABLE 4. INLAND SILVERSIDE, MENIDIA BERRYLINA, LARVAL SURVIVAL DATA
Effluent Concentration^)
Replicate Control
6.25 12.5
25.0
50.0 100.0
RAW
ARC SINE
TRANS-
FORMED
M|AN(Y,)
i
1
A
B
C
A
B
C
0.80
0.87
0.93
1.107
1.202
1.303
1.204
0.010
1
0.73
0.80
0.87
1.024
1.107
1.202
1.111
0.008
2
0.80
0.33
0.60
1.107
0.612
0.886
0.868
0.061
3
0.40 0.0
0.53 0.0
0.07 0.0
0.685
0.815
0.268
0.589
0.082
4
0.0
0.0
0.0
-
-
13.2.6 Test for Normality
13.2.6.1 The first step of the test for normality is to center the
observations by subtracting the mean of all observations within a
concentration from each observation in that concentration. The centered
observations are summarized in Table 5.
TABLE 5. CENTERED OBSERVATIONS FOR SHAPIRO-MILK'S EXAMPLE
Replicate
Control
Effluent Concentration (%)
6.25 12.5 25.0
A
B
C
-0.097
-0.002
0.099
-0.087
-0.004
0.091
0.239
-0.256
0.018
0.096
0.226
-0.321
183
-------
1.0 -
0.9
0.8
0.7 -
g
£
0.5-
00
-pk
C/3
0.3 -
0.2 -
0.1
0.0
0.00
T
CONNBCTS THE MEAN VALUE FOR EACH CONCENTRATION
REPRESENTS THE CRITICAL VALUE FOR DUNNBTTS TEST
(ANY PROPORTION BELOW THIS VALUE WOULD BE
SIGNIFICANTLY DIFFERENT FROM THE CONTROL)
~T
6.25 12.50
EFFLUENT CONCENTRATION (%)
T
25.00
Figure 4. Plot of mean survival proportion of the inland silverside, Menidia beryllina, larvae,
-------
13.2.6.2 Calculate the denominator, D, of the statistic:
n
D = S (X, - X)2
Where: X( = the ith centered observation
X = the overall mean of the centered observations
n = the total number of centered observations
13.2.6.3 For this set of data, n = 12
J_(0.002) = 0.0
12
D = 0.3214
13.2.6.4 Order the centered observations from smallest to largest:
X(1) X<2) - - XCn>
where X(1) denotes the ith ordered observation. The ordered
observations for this example are listed in Table 6.
TABLE 6. ORDERED CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S EXAMPLE
i
1
2
3
4
5
6
Xd)
-0.321
-0.256
-0.097
-0.087
-0.004
-0.002
i
7
8
9
10
11
12
x«>
0.018
0.091
0.096
0.099
0.226
0.239
13.2.6.5 From Table 4, Appendix B, for the number of observations, n, obtain
the coefficients a.,, a2, ... ak where k is n/2 if n is even and (n-l)/2 if n
is odd. For the data in this example, n = 12 and k = 6. The a,- values are
listed in Table 7.
13.2.6.6 Compute the test statistic, W, as follows:
k
W = 1 [ S a,. (X(n'l>1) X(i)) ]2
D i-1
185
-------
The differences x(tvi+1) - X(i> are listed in Table 7. For the data
in this example,
W = _J (0.5513)2 = 0.945
0.3214
TABLE 7. COEFFICIENTS AND DIFFERENCES FOR SHAPIRO-MILK'S EXAMPLE
a,.
1
2
3
4
5
6
0.5475
0.3325
0.2347
0.1586
0.0922
0.0303
0.560
0.482
0.196
0.183
0.095
0.020
X( >
Xd>
xcio>
X(9)
X(8,
X(7)
X(D
- X<2)
- X(3>
- X<4)
- X(5)
X(6)
13.2.6.7 The decision rule for this test is to compare W as calculated in
Subsection 13.2.6.6 to a critical value found in Table 6, Appendix B. If the
computed W is less than the critical value, conclude that the data are not
normally distributed. For the data in this example, the critical value at a
significance level of 0.01 and n = 12 observations is 0.805. Since W = 0.945
is greater than the critical value, conclude that the data are normally
distributed.
13.2.7 Test for Homogeneity of Variance
13.2.7.1 The test used to examine whether the variation in survival is the
same across all effluent concentrations including the control, is Bartlett's
Test (Snedecor and Cochran, 1980). The test statistic is as follows:
[ ( E V,) In S2 - E V, In S,2 ]
B = ' -
Where: Vf = degrees of freedom for each effluent concen-
tration and control, V,- = (n; - 1)
p = number of levels of effluent concentration
including the control
186
-------
, 2
S = i=l
,2 Vi
C = 1 + [ 3(p-l)r1 [ i l/M. - ( 2 V,.)'1 ]
i=l i=l
In = loge
i = 1, 2, ..., p where p is the number of concentrations
including the control
ni = the number of replicates for concentration i.
13.2.7.2 For the data in this example (See Table 4), all effluent
concentrations including the control have the same number of replicates (n- =
3 for all i). Thus, V, = 2 for all i.
13.2.7.3 Bartlett's statistic is therefore:
P ,
B = [(8)ln(0.0402) 2 I! ln(S2)]/1.2083
1-1
= [8(-3.21391) 2(-14.731)]/1.2083
= 3.7508/1.2083
= 3.104
13.2.7.4 B is approximately distributed as chi-square with p - 1 degrees of
freedom, when the variances are in fact the same. Therefore, the appropriate
critical value for this test, at a significance level of 0.01 with three
degrees of freedom, is 11.345. Since B = 3.104 is less than the critical
value of 11.345, conclude that the variances are not different.
13.2.8 Dunnett's Procedure
13.2.8.1 To obtain an estimate of the pooled variance for the Dunnett's
Procedure, construct an ANOVA table as described in Table 8.
187
-------
TABLE 8. ANOVA TABLE
Source
Between
Within
Total
df Sum of Squares
(SS)
p - 1 SSB
N - p SSW
N - 1 SST
Mean Square(MS)
(SS/df)
SB = SSB/(p-l)
S* = SSW/(N-p)
Where: p = number of effluent concentrations including the control
N = total number of observations n,, + n2 ... +np
n = number of observations in concentration i
SSB = Z T,. 2/ii - G2/N Between Sum of Squares
P ni , ,
SST = Z Z Yfj. - GZ/N Total Sum of Squares
SSW = SST - SSB Within Sum of Squares
P
G - the grand total of all sample observations, G = z T,
T,- = the total of the replicate measurements for
concentration "i"
YJJ = the jth observation for concentration "i" (represents
the proportion surviving for effluent concentration
i in test chamber j)
n. - n, - rij - n4 - 3
13.2.8.2 For the data in this example:
S1
T1 - YH + Y12-+ Y13 = 3.612
T2 = Y2. + Y22 + Y23 = 3.333
T: : fe: fe: %: f :S
G = T, + T2 + T3 + T4 = 11.318
SSB = Z T,2/n, - G2/N
188
-------
_J. (34.067) - (11.31812 = 0.681
3 12
P n,-
SST = S S Y?s - G2/N
= 11.677 - (11.318)2 = 1.002
12
SSW = SST - SSB = 1.002 - 0.681 = 0.321
B = SSB/(p-l) = 0.681/(4-l) = 0.227
y = SSW/(N-p) = 0.321/(12-4) = 0.040
13.2.8.3 Summarize these calculations in the ANOVA table (Table 9)
TABLE 9. ANOVA TABLE FOR DUNNETT'S PROCEDURE EXAMPLE
Source
Between
Within
Total
df
3
8
11
Sum of Squares
(SS)
0.681
0.321
1.002
Mean Square(MS)
(SS/df)
0.227
0.040
13.2.8.4 To perform the individual comparisons, calculate the t statistic for
each concentration, and control combination as follows:
Sw
Where: Y,- = mean proportion surviving for effluent concentration i
Y, = mean proportion surviving for the control
Su = square root of within mean square
n1 = number of replicates for the control
n,- = number of replicates for concentration i.
189
-------
13.2.8.5 Table 10 includes the calculated t values for each concentration and
control combination. In this example, comparing the 1.0% concentration with
the control the calculation is as follows:
( 1.204 - 1.111 )
t2 = = 0.570
[ 0.20 / (1/3) + (1/3) ]
TABLE 10. CALCULATED T-VALUES
Effluent Concentration(%) i t{
6.25
12.5
25.0
2
3
4
0.570
2.058
3.766
13.2.8.6 Since the purpose of this test is to detect a significant reduction
in survival, a one-sided test is appropriate. The critical value for this
one-sided test is found in Table 5, Appendix C. For an overall alpha level of
0.05, eight degrees of freedom for error and three concentrations (excluding
the control) the critical value is 2.42. The mean proportion surviving for
concentration "i" is considered significantly less than the mean proportion
surviving for the control if t,- is greater than the critical value.
Therefore, only the 25.0% concentration has a significantly lower mean
proportion surviving than the control. Hence the NOEC is 12.5% and the LOEC
is 25.0%.
13.2.8.7 To quantify the sensitivity of the test, the minimum significant
difference (MSD) that can be detected statistically may be calculated.
MSD = d Sw / (1/n,) + (1/n)
Where: d = the critical value for the Dunnett's procedure
Su = the square root of the within mean square
n = the common number of replicates at each concentration
(this assumes equal replication at each concentration)
n, - the number of replicates in the control.
13.2.8.8 In this example:
MSD = 2.42 (0.20) / (1/3) + (1/3)
= 2.42 (0.20)(0.817)
= 0.395
13.2.8.9 The MSD (0.395) is in transformed units. To determine the MSD in
terms of percent survival, carry out the following conversion.
190
-------
1. Subtract the MSD from the transformed control mean.
1.204 - 0.395 = 0.809
2. Obtain the untransformed values for the control mean and the difference
calculated in step 1.
[Sine (1.204) ]* = 0.871
[Sine (0.809) ]2 = 0.524
3. The untransformed MSD (MSDJ is determined by subtracting the
untransformed values from step 2.
MSDU = 0.871 0.524 = 0.347
13.2.8.10 Therefore, for this set of data, the minimum difference in mean
proportion surviving between the control and any effluent concentration that
can be detected as statistically significant is 0.347.
13.2.8.11 This represents a 40% decrease in survival from the control.
13.2.9 Probit Analysis
13.2.9.1 The data used for the Probit Analysis is summarized in Table 11.
For the Probit Analysis, the two effluent concentrations with 100% mortality
in all three replicates were included. To perform the Probit Analysis, run
the USEPA Probit Analysis Program. An example of the output is supplied in
Table 12 and Figure 5. The current version of the Probit program makes no
distinction between EC and LC endpoints, and labels all results as EC values.
Because the response here is mortality, the EC values output by the program
should be treated as the corresponding LC values.
13.2.9.2 For this example, the chi-square test for heterogeneity was not
significant. Thus Probit Analysis appears to be appropriate for this set of
data.
TABLE 11. DATA FOR PROBIT ANALYSIS
Effluent Concentration (%)
Control 6.25 12.5 25.0 50.0 100,0
Number Dead
Number Exposed
6
45
9
45
19
45
30
45
45
45
45
45
191
-------
TABLE 12. OUTPUT FOR USEPA PROBIT ANALYSIS PROGRAM,
VERSION 1.4
EPA PROBIT ANALYSIS PROGRAM
USED FOR CALCULATING EC VALUES
Version 1.4
Probit Analysis of Inland Silverside Larval Survival Data
Cone.
Control
6.2500
12.5000
25.0000
50.0000
100.0000
Number
Exposed
45
45
45
45
45
45
Number
Resp.
6
9
19
30
45
45
Observed
Proportion
Responding
0.1333
0.2000
0.4222
0.6667
1.0000
1.0000
Adjusted
Proportion
Responding
0.0000
0.0488
0.3130
0.6037
1.0000
1.0000
Predicted
Proportion
Responding
0.
0.
,1589
.0274
0.2478
0.7114
0.9638
0.9988
Chi - Square Heterogeneity = 4.149
Mu
Sigma
Parameter
1.262502
0.242988
Estimate
Std. Err.
95% Confidence Limits
Intercept
Slope
-0.195731
4.115425
1.009659
0.740050
( -2.174661,
( 2.664926,
1.783200)
5.565923)
Spontaneous
Response Rate
0.158944
0.044766
0.071203,
0.246686)
Estimated EC Values and Confidence Limits
Point
EC 1.
EC 5,
00
00
EC10.00
EC15.00
EC50.00
EC85.00
EC90.00
EC95.00
EC99.00
Cone.
4.9801
7.2914
8.9349
10.2487
18.3021
32.6840
37.4900
45.9402
67.2613
Lower Upper
95% Confidence Limits
2.0233
3.6128
4.9096
6.0293
13.8859
26.9977
30.5931
36.3826
49.3731
7.7892
10.4188
12.1961
13.5852
22.1751
42.8767
51.7593
69.2336
121.8529
192
-------
Probit Analysis of Inland Silverside Larval Survival Data
PLOT OF ADJUSTED PROBITS AND PREDICTED REGRESSION LINE
Probit
10+
9+
8+
7 +
6+
o
5+
4 +
O
3+
1+
0+
_+ + + + + + + _
EC01 EC10 EC25 EC50 EC75 EC90 EC99
Figure 5. Plot of adjusted Probits and predicted regression line.
193
-------
13.3 ANALYSIS OF GROWTH DATA
13.3.1 Formal statistical analysis of the growth data is outlined in Figure
6. The response used in the statistical analysis is mean weight per
replicate. The IC25 and IC50 can be calculated for the growth data via a
point estimation technique (see Section 9, Chronic Toxicity Test Endpoints and
Data Analysis). Hypothesis testing can be used to obtain an NOEC and LOEC for
growth. Concentrations above the NOEC for survival are excluded from the
hypothesis test for growth effects.
13.3.2 The statistical analysis using hypothesis tests consists of a
parametric test, Dunnett's Procedure, and a nonparametric test, Steel's
Many-one Rank Test. The underlying assumptions of the Dunnett's Procedure,
normality and homogeneity of variance, are formally tested. The test for
normality is the Shapiro-Wilk's Test and Bartlett's Test is used to test for
homogeneity of variance. If either of these tests fail, the nonparametric
test, Steel's Many-one Rank Test, is used to determine the NOEC and LOEC
endpoints. If the assumptions of Dunnett's Procedure are met, the endpoints
are determined by the parametric test.
13.3.3 Additionally, if unequal numbers of replicates occur among the
concentration levels tested there are parametric and nonparametric alternative
analyses. The parametric analysis is a t-test with the Bonferroni adjustment.
The Wilcoxon Rank Sum Test with the Bonferroni adjustment is the nonparametric
alternative. For detailed information on the Bonferroni adjustment, see
Appendix D.
13.3.4 The data, mean and standard deviation of the growth observations at
each concentration including the control are listed in Table 13. A plot of
the data in Table 13 is provided in Figure 7. Since there was no survival in
the 50% and 100% concentrations, these are not considered in the growth
analysis. Additionally, since there is significant mortality in the 25%
effluent concentration, its effect on growth is not considered.
TABLE 13. INLAND SILVERSIDE, MENIDIA BERYLLINA, GROWTH DATA
Effluent Concentration (%)
Replicate Control 6.25 12.5 25.0 50.0 100.0
A
B
C
MeanfY,)
sf
i
0.939
0.976
0.975
0.963
0.0004
1
0.996
1.152
1.066
1.071
0.006
2
0.903
0.864
1.197
0.988
0.033
3
0.491
0.589
1.131
0.737
0.119
4 5
-
6
194
-------
STATISTICAL ANALYSIS OF INLAND SILVERSIDE LARVAL
SURVIVAL AND GROWTH TEST
GROWTH
GROWTH DATA
MEAN WEIGHT
POINT ESTIMATION
HYPOTHESIS TESTING
(EXCLUDING CONCENTRATIONS
ABOVE NOEC FOR SURVIVAL)
ENDPOINT ESTIMATE
IC25, IC50
SHAPIRO-WILK'S TEST
NON-NORMAL DISTRIBUTION
NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
BARTLETT'S TEST
HETEROGENEOUS
VARIANCE
EQUAL NUMBER OF
REPLICATES?
YES
i
*
EQUAL NUMBER OF
REPLICATES?
YES
T-TEST WITH
BONFERRONI
ADJUSTMENT
DUNNETT'S
TEST
STEEL'S MANY-ONE
RANK TEST
WILCOXON RANK SUM
TEST WITH
BONFERRONI ADJUSTMENT
ENDPOINT ESTIMATES
NOEC.LOEC
Figure 6. Flowchart for statistical analysis of the inland silverside,
Hem'dia beryllina, growth data.
195
-------
CO
1.2
1.1
1.O
0.9 -
0.8 -
0.7 -
0.6 -
0.00
CONNECTS MEAN VALUE FOR EACH CONCENTRATION
REPRESENTS THE CRITICAL VALUE FOR DUNNBTTS TEST
(ANY WEIGHT BELOW THIS VALUE WOULD BE
SIGNIFICANTLY DIFFERENT FROM THE CONTROL)
6.25
EFFLUENT CONCENTRATION (%)
r
12.50
Figure 7. Plot of mean weights of the inland silverside, Mem'dia beryllina, larvae
at each treatment.
-------
13.3.5 Test for Normality
13.3.5.1 The first step of the test for normality is to center the
observations by subtracting the mean of all the observations within a
concentration from each observation in that concentration. The centered
observations are summarized in Table 14.
TABLE 14. CENTERED OBSERVATIONS FOR SHAPIRO-MILK'S EXAMPLE
Effluent Concentration (%)
Replicate Control 6.25 12.5
A
B
C
-0.024
0.013
0.012
-0.075
0.081
-0.005
-0.085
-0.124
0.209
13.3.5.2 Calculate the denominator, D, of the test statistic:
n _
D = E (X, X)2
Where: X1 = the ith centered observation
X = the overall mean of the centered observations
n = the total number of centered observations.
For this set of data: n = 9
X =_1_(0.002) = 0.000
9
D = 0.0794
13.3.5.3 Order the centered observations from smallest to largest:
v(D _ v(2> Y(n)
A ~"A ~ * ~ A
Where X(1) is the ith ordered observation. These ordered observations are
listed in Table 15.
197
-------
TABLE 15. ORDERED CENTERED OBSERVATIONS FOR SHAPIRO-MILK'S EXAMPLE
1
2
3
4
5
-0.124
-0.085
-0.075
-0.024
-0.005
6
7
8
9
0.012
0.013
0.081
0.209
13.3.5.4 From Table 4, Appendix B, for the number of observations, n, obtain
the coefficients a^ a2, ..., ak where k is n/2 if n is even and (n-l)/2 if n
is odd. For the data in this example, n = 9, k = 4. The ai values are listed
in Table 16.
13.3.5.5 Compute the test statistic, W, as follows:
k
W = 1 [ E a,- (X(rvi+1) X(1)) ]2
D i=l
The differences x(n"1+1) - X(i) are listed in Table 15. For this set of data:
W = 1 (0.2707)2 = 0.923
0.0794
TABLE 16. COEFFICIENTS AND DIFFERENCES FOR SHAPIRO-MILK'S EXAMPLE
i a,. X(n-i+1) - X(1)
1
2
3
4
0.5888
0.3244
0.1976
0.0947
0.333
0.166
0.088
0.036
X(9)
X<8)
X(7>
x<6>
- X(1)
- x(2)
X(3)
X<4>
13.3.5.6 The decision rule for this test is to compare M with the critical
value found in Table 6, Appendix B. If the computed M is less than the
critical value, conclude that the data are not normally distributed. For this
example, the critical value at a significance level of 0.01 and nine
observations (n) is 0.764. Since M = 0.923 is greater than the critical
value, the conclusion of the test is that the data are normally distributed.
198
-------
13.3.6 Test for Homogeneity of Variance
13.3.6.1 The test used to examine whether the variation in mean dry weight is
the same across all effluent concentrations including the control, is
Bartlett's Test (Snedecor and Cochran, 1980). The test statistic is as
follows:
P P
[ ( Z V,) In S2 - 2 V, In S? ]
B 1-1 i = l
Where: V,- = degrees of freedom for each effluent concen
tration and control, V; = (n( - 1)
p = number of levels of effluent concentration
including the control
( 2 V,- S')
S = 1=1
C = 1 + ( 3(p-l))-lP[ I 1/V, \ S V,.)'1 ]
i=l i=l
In = loge
i = 1, 2, ..., p where p is the number of concentrations
including the control
n,- = the number of replicates for concentration i.
13.3.6.2 For the data in this example, (See Table 13) all effluent
concentrations including the control have the same number of replicates (n-
3 for all i). Thus, V, = 2 for all i.
13.3.6.3 Bartlett's statistic is therefore:
B = [(6)ln(0.0132) - 2 S ln(S?)]/1.25
1 = 1
= [6(-4.3275) 2(ln(0.0004)+ln(0.0061)+ln(0.0331))]/1.25
= [-25.965 - (-32.664)1/1.25
= 5.359
199
-------
13.3.6.4 B is approximately distributed as chi-square with p 1 degrees' of
freedom, when the variances are in fact the same. Therefore, the appropriate
critical value for this test, at a significance level of 0.01 with 2 degrees
of freedom, is 9.210. Since B = 5.359 is less than the critical value of
9.210, conclude that the variances are not different.
13.3.7 Dunnett's Procedure
13.3.7.1 To obtain an estimate of the pooled variance for the Dunnett's
Procedure, construct an ANOVA table as described in Table 17.
TABLE 17. ANOVA TABLE
Source
Between
Within
Total
df Sum of Squares
(SS)
p - 1 SSB
N - p SSW
N 1 SST
Mean Square(MS)
(SS/df)
S* = SSB/(p-l)
Sy = SSW/(N-p)
Where: p = number of effluent concentrations including the control
N = total number of observations n1 + n? ... +n
n,- = number of observations in concentration i
P , .
SSB = s T, /n,- - GZ/N Between Sum of Squares
P
*2 r.2
1-1 J-l
SST = s Z Y2j - G2/N Total Sum of Squares
SSW = SST - SSB Within Sum of Squares
p
G = the grand total of all sample observations, G = s Tf
i=l
T, = the total of the replicate measurements for
concentration "i"
YJJ- = the jth observation for concentration "i" (represents
the mean dry weight of the fish for effluent
concentration i in test chamber j)
200
-------
13.3.7.2 For the data in this example:
n1 = n2 = n3 = 3
N = 9
T1 " Yn + Y12 + Y13 " °-939 + °-976 + °-975 = 2.890
T2 = Y21 + Y22 + Y23 = 0.996 + 1.152 + 1.066 = 3.214
T3 - Y31 + Y32 + Y33 « °-903 + 0.864 + 1.197 = 2.964
G = T, + T2 + T3 = 9.068
SSB = Z TjVn,. - G2/N
= J^(27.467) - (9.068)2 = 0.019
3 9
SST = S S'Y?, - G2/N
1-1 j-1
= 9.235 - (9.068)2 = 0.098
SSW = SST - SSB = 0.098 - 0.019 = 0.079
S2 = SSB/(p-l) = 0.019/(3-l) = 0.009
S2 = SSW/(N-p) = 0.079/(9-3) = 0.013
13.3.7.3 Summarize these calculations in the ANOVA table (Table 18)
TABLE 18. ANOVA TABLE FOR DUNNETT'S PROCEDURE EXAMPLE
Source
Between
Within
df
2
6
Sum of Squares
(SS)
0.019
0.079
Mean Square(MS)
(SS/df)
0.0095
0.013
Total 8 0.098
201
-------
13.3.7.4 To perform the individual comparisons, calculate the t statistic for
each concentration and control combination as follows:
( Y, Y, )
t; =
Sw / (1/n,) + (1/n,-)
Where: Y,- = mean dry weight for effluent concentration i
Y1 = mean dry weight for the control
Sw = square root of within mean square
n,, = number of replicates for the control
n,- = number of replicates for concentration i.
13.3.7.5 Table 19 includes the calculated t values for each concentration and
control combination. In this example, comparing the 6.25% concentration with
the control the calculation is as follows:
( 0.963 1.071 )
t2 = = -1.16
[ 0.114 / (1/3) + (1/3) ]
TABLE 19. CALCULATED T-VALUES
Effluent Concentration(%) i t,-
6.25
12.5
2
3
-1.16
-0.27
13.3.7.6 Since the purpose of this test is to detect a significant reduction
in mean weight, a one-sided test is appropriate. The critical value for this
one-sided test is found in Table 5, Appendix C. For an overall alpha level of
0.05, six degrees of freedom for error and two concentrations (excluding the
control) the critical value is 2.34. The mean weight for concentration "i" is
considered significantly less than mean weight for the control if t- is
greater than the critical value. Therefore, all effluent concentrations in
this example do not have significantly lower mean weights than the control.
Hence the NOEC and the LOEC for growth cannot be calculated.
13.3.7.7 To quantify the sensitivity of the test, the minimum significant
difference (USD) that can be detected statistically may be calculated.
MSD = d Su / (1/n,) + (1/n)
202
-------
Where: d = the critical value for the Dunnett's procedure
Sw = the square root of the within mean square
n = the common number of replicates at each concentration
(this assumes equal replication at each concentration)
n, = the number of replicates in the control.
13.3.7.8 In this example:
MSD = 2.34 (0.114) / (1/3) + (1/3)
= 2.34 (0.114)(0.8165)
= 0.218
13.3.7.9 Therefore, for this set of data, the minimum difference that can be
detected as statistically significant is 0.218 mg.
13.3.7.10 This represents a 22.6% reduction in mean weight from the control.
13.3.8 Calculation of the 1C
13.3.8.1 The growth data in Table 13 are utilized in this example. As can be
seen, the observed means are not monotonically non-increasing with respect to
concentration. Therefore, it is necessary to smooth the means prior to
calculating the_IC. In the following discussion, the observed means are
represented by Yf and the smoothed means by M,..
13.3.8.2 Starting with the control mean, YT = 0.963 and Y2 = 1.071, we see
that Y1 < Y2. Calculate the smoothed means:
M, = M2 = (V, + Y2)/2 = 1.017
13.3.8.3 Since Y4 = 0.737 < Y, = 0.988 < M,, set M3 = 0.988 and M4 = 0.737.
Table 20 contains the smoothed means and Figure 7 gives a plot of the smoothed
response curve.
TABLE 20. INLAND SILVERSIDE MEAN GROWTH
RESPONSE AFTER SMOOTHING
Effluent
Cone.
(%)
Control
6.25
12.50
25.00
i
1
2
3
4
Mi
(nig)
1.017
1.017
0.988
0.737
203
-------
13.3.8.4 An IC25 and IC50 can be estimated using the Linear Interpolation
Method. A 25% reduction in mean weight, compared to the controls, would
result in a mean weight of 0.763 mg, where M^l-p/100) = 1.017(1-25/100). A
50% reduction in mean weight, compared to the controls, would result in a mean
weight of 0.509 mg. Examining the means and their associated concentrations
(Table 20), the response, 0.763 mg, is bracketed by C3 = 12.5% effluent and C4
= 25.0 % effluent. The response (0.737 mg) at the highest effluent
concentration (25.0%) is greater than 50% of the control (0.509 mg). Thus,
the IC50 is specified as greater than 25% effluent.
13.3.8.5 Using Equation 1 from Appendix J, the estimate of the IC25 is
calculated as follows:
ICp = Cj + [M^l - p/100) - MJ (C^ - C,)
(MJ+1 MJ
IC25 = 12.5 + [1.017(1 - 25/100) - 0.988] (25.00 - 12.50)
(0.737 0.988)
= 23.7%.
13.3.8.6 When the Bootstrap program (BOOTSTRP) was used to analyze this set
of data, requesting 80 resamples, the mean estimate of the IC25 was 20.4622%,
with a standard deviation of 2.5075% (coefficient of variation = 12.3%). The
empirical 96.4% confidence interval for the true mean was (16.1877%,
24.7514%). The BOOTSTRP computer program output for the IC25 for this data
set is shown in Figure 9.
13.3.8.7 When the Bootstrap program (BOOTSTRP) was used to analyze this set
of data for the IC50, requesting 80 resamples, the output indicated that the
response of the highest effluent concentration exceeded 50% of the control.
Thus, the IC50 could not be calculated. The BOOTSTRP computer program output
is shown in Figure 10.
204
-------
f\J
o
en
1-4 -
1.3 -
1.2
1.1
1.0
0.9
0.8 -
0.7
0.6 -
0.5
0.4
0.00
INDIVIDUAL REPLICATE MEAN WEIGHT
CONNECTS THE OBSERVED MEAN VALUB
CONNECTS THE SMOOTHED MEAN VALUE
6.25 12.50
EFFLUENT CONCENTRATION (%)
Figure 8. Plot of the raw data, the observed means and the smooth means
from Tables 13 and 20.
25.00
-------
THE NUMBER OF RESAMPLES IS 80
*** LISTING OF GROUP CONCENTRATIONS (% EFF.) AND RESPONSE MEANS ***
CONC. (%EFF) RESPONSE MEAN MEAN AFTER POOLING
0.000 0.963 1.017
6.250 1.071 1.017
12.500 0.988 0.988
25.000 0.737 0.737
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
FROM THE INPUT SAMPLE IS 23.7052.
* BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY *
* OF THE ESTIMATED ICp *
************************************************************
THE MEAN OF THE BOOTSTRAP ESTIMATES IS 20.4622.
THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS 2.5075.
AN EMPIRICAL 96.4% CONFIDENCE INTERVAL FOR THE
BOOTSTRAP ESTIMATE IS (16.1877, 24.7514).
*** NOTE: THE ABOVE BOOTSTRAP CALCULATIONS WERE BASED ON 56
INSTEAD OF 80 RESAMPLINGS. THOSE RESAMPLES NOT
USED HAD ESTIMATES ABOVE THE HIGHEST CONCENTRATION / % EFF.
Figure 9. BOOTSTRP program output for the IC25.
206
-------
THE NUMBER OF RESAMPLES IS 80
*** LISTING OF GROUP CONCENTRATIONS (% EFF.) AND RESPONSE MEANS ***
CONC. (%EFF) RESPONSE MEAN MEAN AFTER POOLING
0.000 0.963 1.017
6.250 1.071 1.017
12.500 0.988 0.988
25.000 0.737 0.737
*** NO LINEAR INTERPOLATION ESTIMATE CAN BE CALCULATED FROM THE INPUT
DATA, SINCE NONE OF THE (POSSIBLY POOLED) GROUP RESPONSE MEANS
WERE LESS THAN 50.0% OF THE CONTROL RESPONSE MEAN.
* BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY *
* OF THE ESTIMATED ICp *
************************************************************
THE MEAN OF THE BOOTSTRAP ESTIMATES IS 24.5144.
THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS .2122.
AN EMPIRICAL ****% CONFIDENCE INTERVAL FOR THE
BOOTSTRAP ESTIMATE IS (24.2058, 24.8765).
*** NOTE: THE ABOVE BOOTSTRAP CALCULATIONS WERE BASED ON 9
INSTEAD OF 80 RESAMPLINGS. THOSE RESAMPLES NOT
USED HAD ESTIMATES ABOVE THE HIGHEST CONCENTRATION / % EFF.
Figure 10. BOOTSTRP program output for the IC50,
207
-------
14. PRECISION AND ACCURACY
14.1 PRECISION
14.1.1 Data on the single-laboratory precision of the inland silverside
larval survival and growth test using copper (CU) sulfate and sodium dodecyl
sulfate (SDS) as reference toxicants, in natural seawater and GP2 are provided
in Tables 21-23. In Tables 21-22, the coefficient of variation for copper
based on the IC25 is 43.2 and for SDS is 43.2% indicating acceptable
precision. In the five tests with each reference toxicant, the NOEC's varied
by only one concentration interval, indicating good precision. The
coefficient of variation for all reference toxicants based on the IC50 in two
types of seawater (GP2 and natural) ranges from 1.8% to 50.7% indicating
acceptable precision. Data in Table 23 show no detectable differences between
tests conducted in natural and artificial seawaters.
14.1.2 Data on the multilaboratory precision of the inland silverside larval
survival and growth test are not yet available.
14.2 ACCURACY
14.2.1 The accuracy of toxicity tests cannot be determined.
208
-------
TABLE 21. SINGLE-LABORATORY PRECISION OF THE INLAND SILVERSIDE, HENIDIA
BERYLLINA, SURVIVAL AND GROWTH TEST PERFORMED IN NATURAL SEAWATER,
USING LARVAE FROM FISH MAINTAINED AND SPAWNED IN NATURAL SEAWATER,
AND COPPER (CU) AS A REFERENCE TOXICANT1'2'3'*'5'6'7
Test NOEC
Number (ng/L)
1 63
2 125
3 63
4 125
5 31
n: 5
Mean: NA
CV(%): NA
IC25
(ng/L)
96.2
207.2
218.9
177.5
350.1
5
209.9
43.7
IC50
(H9/L)
148.6
NC*
493.4
241.4
479.8
4
340.8
50.7
Most
Sensitive
Endpoint
S
S
G
S
G
1Data from USEPA (1989a) and USEPA (1991a)
2Tests performed by George Morrison and Elise Torello,
Environmental Research Laboratory, U. S. Environmental Protection
Agency, Narragansett, Rhode Island.
3Three replicate exposure chambers with 10-15 larvae were used for
the control and each copper concentration. Copper concentrations
were: 31, 63, 125, 250, and 500 jig/L.
4Adults collected in the field.
5S = Survival effects. G = Growth data at these toxicant
concentrations were disregarded because there was a significant
reduction in survival.
6NOEC Range: 31 - 125 ng/L (this represents a difference of two exposure
concentrations).
7For a discussion of the precision of data from chronic toxicity
tests see Section 4, Quality Assurance.
*No linear interpolation estimate could be calculated from the data, since
none of the group response means were less than 50 percent of the control
response mean.
209
-------
TABLE 22. SINGLE-LABORATORY PRECISION OF THE INLAND SILVERSIDE, MENIDIA
BERYLLINA, SURVIVAL AND GROWTH TEST PERFORMED IN NATURAL SEAWATER,
USING LARVAE FROM FISH MAINTAINED AND SPAWNED IN NATURAL SEAWATER,
AND SODIUM DODECYL SULFATE (SDS) AS A REFERENCE TOXICANT1'2'3'4'5'6'r
Test
Number
1
2
3
4
5
n:
Mean:
CV(%):
NOEC
(mg/L)
1.3
1.3
1.3
1.3
1.3
5
NA
NA
IC25
(mg/L)
0.3
1.6
1.5
1.5
1.6
5
1.3
43.2
IC50
(mg/L)
1.7
1.9
1.9
1.9
2.2
5
1.9
9.4
Most
Sensitive
Endpoint
S
S
S
S
S
'Data from USEPA (1989a) and USEPA (1991a)
2Tests performed by George Morrison and Elise Torello, Environmental
Research Laboratory, U. S. Environmental Protection Agency,
Narragansett, Rhode Island.
3Three replicate exposure chambers with 10-15 larvae were used for
the control and each SDS concentration. SDS concentrations were:
0.3, 0.6, 1.3, 2.5, and 5.0 mg/L.
4Adults collected in the field.
5S = Survival Effects. Growth data at these toxicant
concentrations were disregarded because there was a significant
reduction in survival.
6NOEC Range 1.3 mg/L.
7For a discussion of the precision of data from chronic toxicity
tests see Section 4, Quality Assurance.
210
-------
TABLE 23. COMPARISON OF THE SINGLE-LABORATORY PRECISION OF THE INLAND
SILVERSIDE, MENIDIA BERYLLINA, LARVAL SURVIVAL (LC50) AND GROWTH
(IC50) VALUES EXPOSED TO SODIUM DODECYL SULFATE (SDS) OR COPPER
(CU) SULFATE IN GP2 ARTIFICIAL SEAWATER MEDIUM OR NATURAL SEAWATER
(NSW)1'2'3'4
Survival
SDS (mg/L)
GP2
NSW
Copper (fj.g/1)
GP2
NSW
Growth
GP2
GP2
NSW
3.59
4.87
5.95
3.69
4.29
8.05
3.60
5.54
6.70
3.55
5.27
8.53
MEAN
CV (%)
4.81
24.6
5.34
44.2
5.28
29.6
5.79
43.8
NSW
247
215
268
256
211
240
260
236
a
277
223
238
MEAN
CV (%)
243
10.9
236
9.8
248
6.9
246
11.2
1Tests performed by George Morrison and Glen Modica.
2Three replicate exposure chambers with 10-15 larvae per treatment.
3Adults collected in the field.
4For a discussion of the precision of data from chronic toxicity tests see
Section 4, Quality Assurance.
a - No linear interpolation estimate could be calculated from the data, since
none of the group response means were less than 50 percent of the control
response mean.
211
-------
Figure 11 Data forms for inland silverside, Henidia beryl Una,^larval survival and growth test.
Daily record of larval survival and test conditions.
Test Dates:
Type Effluent:
Species:
Field
Lab
Test
Effluent Tested:
CONCENTRATION:
REPLICATE:
DAYS
LIVE
LARVAE
TEMP
rci
SALINITY
r/_i
DO
(mg/l|
LARVAE/
DRY WT
0
1
2
3
4
5
MEAN WEIGHT/
LARVAE (mfll 1 S D
6
7
REPLICATE:
0
1
LARVAE
DRY WT
2
3
4
5
6
MEAN WEIGHT
LARVAE |mg) ' S 0
7
REPLICATE
0
1
LARVAE
DRV WT
2
3
4
5
6
7
REPLICATE
0
1
LARVAE
DRYWT
2
3
4
5
6
MEAN WEIGHT
LARVAE Imgl ' S D
7
CONCENTRATION:
LIVE
LARVAE
TEMP
rci
SALINITY
(''-I
DO
(mg/ll
LARVAE/
DRY WT
MEAN WEIGHT/
LARVAE (mgl t S D
LARVAE
DRY WT
MEAN WEIGHT'
LARVAE imgl ' S D
LARVAE
DRV WT
LARVAE
DRY WT
MEAN WEIGHT'
LARVAE (mgl I S D
CONCENTRATION:
LIVE
LARVAE
TEMP
rci
SALINITY
1* -1
DO
lmg/11
LARVAE/
DRY WT
MEAN WEIGHT/
LARVAE (Tigl t S 0
LARVAE
DRYWT
MEAN WEIGHT
LARVAE lmH| I S D
LARVAE
DRY WT
LARVAE
DRYWT
MEAN WEIGHT/
LARVAE (mgl I S D
PO
I'
PO
TIME
FED
COMMENTS:
Adapted from USEPA (1987c)
-------
Figure 11.
Data forms for inland silverside, Menldl*
Daily record of larval survival and test
Mnnu
. (Continued)
Test Dates:
Type Effluent:
Species:
Field
Lab
Test
Effluent Tested:
CONCENTRATION:
REPLICATE:
DAYS
LIVE
LARVAE
TEMP
rci
SALINITY
I''-)
DO
(mg/ll
LARVAE/
DRV WT
0
1
2
3
4
5
MEAN WEIGHT/
LARVAE |mg» t S D
6
7
REPLICATE
0
1
LARVAE
DRY WT
2
3
4
5
6
MEAN WEIGHT
LARVAE |mg| ' S D
7
REPLICATE
0
1
LARVAE
DRY WT
2
3
4
5
6
7
REPLICATE
0
1
LARVAE
DRY WT
2
3
4
5
6
MEAN WEIGHT
LARVAE (mg| I S D
7
CONCENTRATION:
LIVE
LARVAE
TEMP
rci
SALINITY
r/j
00
|Hlg/l|
LARVAE/
MY WT
MEAN WEIGHT/
LARVAE |mg| t S O
LARVAE
DRY WT
MEAN WEIGHT
LARVAE |mg| i S D
LARVAE
DRV WT
LARVAE
DRY WT
MEAN WEIGHT-
LARVAE (mgl I S D
CONCENTRATION:
LIVE
LARVAE
TEMP
(°CI
SALINITY
C -1
DO
|mg/M
LARVAE/
DRY WT
MEAN WEIGHT/
LARVAE Intgl 1 S D
LARVAE
DRY WT
MEAN WEIGHT
LARVAE Img) t S D
LARVAE
DRY WT
LARVAE
DRY WT
MEAN WEIGHT/
LARVAE (mgl I S D
TIME
FED
COMMENTS:
-------
Figure 12.
Test Dates:
Data forms for inland silverside, Mem'dia beryllina, larval
survival and growth test. Dry weights of larvae.
Species:
Pan
#
Cone.
&
Rep.
Initial
Wt.
(mg)
Final
Wt.
(mg)
Diff.
(mg)
#
Larvae
Av. V
Larv
Adapted from USEPA (1987c).
214
-------
Figure 13. Data forms for inland silverside, Menidia beryllina. larval
survival and growth test. Summary of test results.
Test Dates:
Species:
Effluent Tested:
TREATMENT
#LIVE
LARVAE
SURVIVAL
(%)
MEAN DRY WT./
LARVAE (mg)
±S.D.
SIGNIF. DIFF.
FROM CONTROL
(o)
MEAN
TEMPERATURE
(oC)
±S.D.
MEAN SALINITY
0/00
±S.D.
AV. DISSOLVED
OXYGEN
(mg./L) ±S.D.
COMMENTS:
Adapted from USEPA (1987c).
215
-------
SECTION 14
TEST METHOD1'2
MYSID, HYSIDOPSIS BAHIA, SURVIVAL, GROWTH, AND FECUNDITY TEST
METHOD 1007
1. SCOPE AND APPLICATION
1.1 This method estimates the chronic toxicity of effluents and receiving
waters to the mysid, Mysidopsis bahia, during a seven-day, static-renewal
exposure. The effects include the synergistic, antagonistic, and additive
effects of all the chemical, physical, and additive components which adversely
affect the physiological and biochemical functions of the test organisms.
1.2 Daily observations on mortality make it possible to also calculate acute
toxicity for desired exposure periods (i.e., 24-h, 48-h, 96-h LCSOs).
1.3 Detection limits of the toxicity of an effluent or pure substance are
organism dependent.
1.4 Single or multiple excursions in toxicity may not be detected using 24-h
composite samples. Also, because of the long sample collection period
involved in composite sampling and because the test chambers are not sealed,
highly volatile and highly degradable toxicants in the source may not be
detected in the test.
1.5 This test is commonly used in one of two forms: (1) a definitive test,
consisting of a minimum of five effluent concentrations and a control, and (2)
a receiving water test(s), consisting of one or more receiving water
concentrations and a control.
2. SUMMARY OF METHOD
2.1 This rapid-chronic test consists of an exposure of 7-day old Mysidopsis
bahia juveniles to different concentrations of effluent, or to receiving water
in a static system, during the period of egg development. The test endpoints
are survival, growth (measured as dry weight), and fecundity (measured as the
percentage of females with eggs in the oviduct and/or brood pouch).
3. INTERFERENCES
3.1 Toxic substances may be introduced by contaminants in dilution water,
glassware, sample hardware, and testing equipment (see Section 5, Facilities,
Equipment, and Supplies).
3.2 Improper effluent sampling and handling may adversely affect test results
1The format used for this method was taken from USEPA (1983).
2This method was adapted from USEPA (1987f).
216
-------
(see Section 8, Effluent and Receiving Water Sampling, Sample Handling, and
Sample Preparation for Toxicity Tests).
3.3 The test results can be confounded by (1) the presence of pathogenic
and/or predatory organisms in the dilution water, effluent, and receiving
water, (2) the condition of the brood stock from which the test animals were
taken, (3) the amount and type of natural food in the effluent, receiving
water, or dilution water, (4) nutritional value of the brine shrimp, Artemia
nauplii, fed during the test, and (5) the quantity of brine shrimp, Artemia
nauplii, or other food added during the test, which may sequester metals and
other toxic substances, and lower the DO.
4. SAFETY
4.1 See Section 3, Health and Safety.
5. APPARATUS AND EQUIPMENT
5.1 Facilities for holding and acclimating test organisms.
5.2 Brine shrimp, Artemia, culture unit -- see Subsection 6.12 below and
Section 4, Quality Assurance.
5.3 Mysid, Mysidopsis bahia, culture unit -- see Subsection 6 below. This
test requires a minimum of 240 7-day old (juvenile) mysids. It is preferable
to obtain the test organisms from an inhouse culture unit. If it is not
feasible to culture mysids inhouse, juveniles can be obtained from other
sources, if shipped in well oxygenated saline water in insulated containers.
5.4 Samplers -- automatic sampler, preferably with sample cooling capability,
that can collect a 24-h composite sample of 5 L.
5.5 Environmental chamber or equivalent facility with temperature control
(26 ± 1°C).
5.6 Water purification system -- Millipore Milli-QR, deionized water or
equivalent.
5.7 Balance -- capable of accurately weighing to 0.0001 g.
5.8 Reference weights, Class S -- for checking performance of balance.
Reference weights should bracket the expected weights of the weighing boats
and weighing boats plus organisms.
5.9 Drying oven -- 105°C, for drying organisms.
5.10 Desiccator -- for holding dried organisms.
5.11 Air pump -- for supplying air.
5.12 Air lines, and air stones -- for aerating cultures, brood chambers, and
holding tanks, and supplying air to test solutions with low DO.
217
-------
5.13 pH and DO meters -- for routine physical and chemical measurements.
Unless the test is being conducted to specifically measure the effect of one
of the above parameters, a portable, field-grade instrument is acceptable.
5.14 Tray -- for test vessels; approximately 90 X 48 cm to hold 56 vessels.
5.15 Standard or micro-Winkler apparatus -- for determining DO and checking
DO meters.
5.16 Dissecting microscope (350-400X magnification) -- for examining
organisms in the test vessels to determine their sex and to check for the
presence of eggs in the oviducts of the females.
5.17 Light box -- for illuminating organisms during examination.
5.18 Refractometer or other method-- for determining salinity.
5.19 Thermometers, glass or electronic, laboratory grade -- for measuring
water temperatures.
5.20 Thermometers, bulb-thermograph or electronic-chart type -- for
continuously recording temperature.
5.21 Thermometer, National Bureau of Standards Certified (see USEPA METHOD
170.1, USEPA, 1979b) -- to calibrate laboratory thermometers.
5.22 Test vessels -- 200 ml borosilicate glass beakers or 8 oz disposable
plastic cups (manufactured by Falcon Division of Becton, Dickinson Co., 1950
Williams Dr., Oxnard, CA 93030) or other similar containers. Cups must be
rinsed thoroughly in distilled or deionized water and then pre-soaked
(conditioned) overnight in dilution water before use. Forty-eight (48) test
vessels are required for each test (eight replicates at each of five effluent
concentrations and a control).
5.23 Beakers or flasks -- six, borosilicate glass or non-toxic plasticware,
2000 ml for making test solutions.
5.24 Wash bottles -- for deionized water, for washing organisms from
containers and for rinsing small glassware and instrument electrodes and
probes.
5.25 Volumetric flasks and graduated cylinders -- Class A, borosilicate glass
or non-toxic plastic labware, 50-2000 ml for making test solutions.
5.26 Separatory funnels, 2-L -- Two-four for culturing Artemia.
5.27 Pipets, volumetric -- Class A, 1-100 ml.
5.28 Pipets, automatic -- adjustable, 1-100 ml.
5.29 Pipets, serological -- 1-10 ml, graduated.
218
-------
5.30 Pipet bulbs and fillers -- PROPIPET", or equivalent.
5.31 Droppers, and glass tubing with fire polished edges, 4mm ID -- for
transferring organisms.
5.32 Forceps -- for transferring organisms to weighing boats.
5.33 NITEXR mesh sieves (< 150 \im and 1000 \im) -- for concentrating
organisms.
5.34 Depression glass slides or depression spot plates -- two, for observing
organisms.
6. REAGENTS AND CONSUMABLE MATERIALS
6.1 Sample containers -- for sample shipment and storage (see Section 8,
Effluent and Receiving Water Sampling, Sample Handling, and Sample Preparation
for Toxicity Tests).
6.2 Data sheets (one set per test) -- for data recording (Figures 14, 15, and
16).
6.3 Tape, colored and markers, water-proof -- for labelling and marking test
chambers, containers, etc.
6.4 Weighing boats, aluminum -- to determine the dry weight of organisms.
6.5 Buffers, pH 4, 7, and 10 (or as per instructions of instrument
manufacturer) -- for standards and calibration check (see USEPA Method 150.1,
USEPA, 1979b).
6.6 Membranes and filling solutions for dissolved oxygen probe (see USEPA
Method 360.1, USEPA, 1979b), or reagents for modified Winkler analysis.
6.7. Laboratory quality assurance samples and standards for the above
methods.
6.8 Reference toxicant solutions (see Section 4, Quality Assurance,
Subsections 4.7, 4.14, 4.15, 4.16, and 4.17).
6.9 Reagent water -- defined as distilled or deionized water that does not
contain substances which are toxic to the test organisms (see Subsection 5.6
above).
6.10 Effluent, receiving water, and dilution water -- see Section 7, Dilution
Water, and Section 8, Effluent and Receiving Water Sampling, Sample Handling,
and Sample Preparation for Toxicity Tests. Dilution water containing
organisms that might prey upon or otherwise interfere with the test organisms
should be filtered through a fine mesh net (with 150 \im or smaller openings).
6.10.1 Saline test and dilution water -- The salinity of the test water must
be in the range of 20 °/oo to 30 °/oo. The salinity should vary by no more
219
-------
than + 2 °/oo among the chambers on a given day. If effluent and receiving
water tests are conducted concurrently, the salinities of these tests should
be similar.
6.10.1.1 The overwhelming majority of industrial and sewage treatment
effluents entering marine and estuarine systems contain little or no
measurable salts. Exposure of mysids to these effluents will require
adjustments in the salinity of the test solutions. It is important to
maintain a constant salinity across all treatments. In addition, it may be
desirable to match the test salinity with that of the receiving water. This
salinity adjustment may be accomplished by the use of a modified GP2
artificial seawater formulation (Table 1) or hypersaline brine. FORTY
FATHOMS" and HW MARINEMIX" artificial sea salts have been found suitable for
culturing and life cycle toxicity tests with mysids (Home et al., 1983; ASTM,
1990; EMSL-Cincinnati, and USEPA, Region 6). Hypersaline brine derived from
natural seawater should be used to adjust salinities. However, it should be
noted that with (100 °/oo) hypersaline brine, the maximum concentration of
effluent that can be tested is 80% effluent at 20 °/oo salinity and 70%
effluent at 70 °/oo salinity.
6.10.1.2 Hypersaline brine (HSB) has several advantages that make it
desirable for use in toxicity testing. It can be made from any high quality,
filtered seawater by evaporation, and can be added to the effluent or to
deionized water to increase the salinity. Brine derived from natural seawater
contains the necessary trace metals, biogenic colloids, and some of the
microbial components necessary for adequate growth, survival, and/or
reproduction of marine and estuarine organisms, and may be stored for
prolonged periods without any apparent degradation.
6.10.1.3 The ideal container for making brine from natural seawater is one
that (1) has a high surface to volume ratio, (2) is made of a non-corrosive
material, and (3) is easily cleaned (fiberglass containers are ideal).
Special care should be used to prevent any toxic materials from coming in
contact with the seawater being used to generate the brine. If a heater is
immersed directly into the seawater, ensure that the heater materials do not
corrode or leach any substances that would contaminate the brine. One
successful method used is a thermostatically controlled heat exchanger made
from fiberglass. For aeration, use only oil-free air compressors to prevent
contamination.
6.10.1.4 Before adding seawater to the brine generator, thoroughly clean the
generator, aeration supply tube, heater, and any other materials that will be
in direct contact with the brine. A good quality biodegradable detergent
should be used, followed by several thorough deionized water rinses. High
quality (and preferably high salinity) seawater should be filtered to at least
10 [im before placing into the brine generator. Water should be collected on
an incoming tide to minimize the possibility of contamination.
6.10.1.5 The temperature of the seawater is increased slowly to 40°C. The
water should be aerated to prevent temperature stratification and to increase
water evaporation. The brine should be checked daily (depending on the volume
being enerated) to ensure that the salinity does not exceed 100 °/oo and that
220
-------
TABLE 1. REAGENT GRADE CHEMICALS USED IN THE PREPARATION OF GP2 ARTIFICIAL
SEAWATER FOR THE MYSID, MYSIDOPSIS BAHIA, TOXICITY TEST1'2'3
Compound
1.
2.
3.
4.
5.
6.
7.
8.
9.
NaCl
Na2S04
KC1
KBr
Na2B407 . 10 H20
MgCl2 . 6 H20
CaCl2 . 2 H20
SrCl2 . 6 H20
NaHC03
Concentration
(9/L)
21.03
3.52
0.61
0.088
0.034
9.50
1.32
0.02
0.17
Amount (g)
Required for
20 L
420.6
70.4
12.2
1.76
0.68
190.0
26.4
0.400
3.40
Modified GP2 from Spotte et al. (1984).
2The constituent salts and concentrations were taken from
USEPA (1990b). The salinity is 30.89 g/L.
3GP2 can be diluted with deionized (DI) water to the desired test salinity.
221
-------
the temperature does not exceed 40°C. Additional seawater may be added to the
brine to obtain the volume of brine required.
6.10.1.6 After the required salinity is attained, the MSB should be filtered
a second time through a 1 \im filter and poured directly into portable
containers (20-L cubitainers or polycarbonate water cooler jugs are suitable).
The containers should be capped and labelled with the date the brine was
generated and its salinity. Containers of MSB should be stored in the dark
and maintained under room temperature until used.
6.10.1.7 If a source of MSB is available, test solutions can be made by
following the directions below:
6.10.1.8 Thoroughly mix together the deionized water and brine before mixing
in the effluent. Divide the salinity of the MSB by the expected test salinity
to determine the proportion of deionized water to brine. For example, if the
salinity of the brine is 100 °/oo and the test is to be conducted at 20 °/oo,
100 °/°° divided by 20 °/oo = 5.0. The proportion of brine is 1 part in 5
(one part brine to four parts deionized water). To make 1 L of sea water at
20 °/
-------
TABLE 2. QUANTITIES OF EFFLUENT, DEIONIZED WATER, AND HYPERSALINE BRINE
(100 °/oo) NEEDED TO PREPARE 1800 ML VOLUMES OF TEST SOLUTION
WITH A SALINITY OF 20 °/oo.
Effluent
Concentration
(%)
80.0
40.0
20.0
10.0
5.0
Control
Volume of
Effluent
(0 °/oo)
(ml)
1440.0
720.0
360.0
180.0
90.0
0.0
Volume of
Deionized
Water
(mL)
0.0
720.0
1080.0
1260.0
1350.0
1440.0
Volume of
Hypersaline
Brine
(mL)
360.0
360.0
360.0
360.0
360.0
360.0
Total
Volume
(mL)
1800.0
1800.0
1800.0
1800.0
1800.0
1800.0
Total
2790.0
5850.0
2160.0
10800.0
223
-------
shrimp cysts, the Brazilian or Colombian strains are preferred because the
supplies examined have had low concentrations of chemical residues and produce
nauplii of suitably small size. For commercial sources of brine shrimp,
Artemia, cysts, see Table 2 (Section 5, Facilities, Equipment, and Supplies)
and Section 4, Quality Assurance, 4.8 Food Quality. Each new batch of Artemia
cysts must be evaluated for size (Vanhaecke and Sorgeloos, 1980, and Vanhaecke
et al., 1980) and nutritional suitability (see Leger et al., 1985, 1986)
against known suitable reference cysts by performing a side-by-side larval
growth test using the "new" and "reference" cysts. The "reference" cysts used
in the suitability test may be a previously tested and acceptable batch of
cysts, or may be obtained from the Quality Assurance Research Division,
Environmental Monitoring Systems Laboratory, Cincinnati, Ohio 45268. A sample
of newly-hatched Artemia nauplii from each new batch of cysts should be
chemically analyzed. If the concentration of total organic chlorine exceeds
0.15 \ig/g wet weight, or the total concentration of organochlorine pesticides
plus PCBs exceeds 0.3 pg/g wet weight, the Artemia cysts should not be used
(For analytical method see USEPA, 1982).
6.12.2 Artemia nauplii are obtained as follows:
1. Add 1 L of seawater, or an aqueous uniodized salt (NaCl) solution
prepared with 35 g salt or artificial sea salts per liter, to a 2-L
separatory funnel, or equivalent.
2. Add 10 ml Artemia cysts to the separatory funnel and aerate for 24 h
at 27°C. Hatching time varies with incubation temperature and the
geographic strain of Artemia used. See USEPA (1985d) and USEPA
(1991c) for details on Artemia culture and quality control.
3. After 24 h, cut off the air supply in the separatory funnel.
Artemia nauplii are phototactic, and will concentrate at the bottom of
the funnel if it is covered for 5-10 min with a dark cloth or paper
towel. Caution: if the concentrated nauplii are left on the bottom
much longer than 10 min without aeration, excessive mortality will
result.
4. Drain the nauplii into a funnel fitted with a < 150 urn Nitex screen,
and rinse with seawater or equivalent before use.
5. Resuspend the nauplii on the funnel in a small amount of water for
feeding.
6.12.3 Testing the acceptability of Artemia nauplii as food for mysids.
6.12.3.1 The primary criteria for acceptability of each new supply of brine
shrimp, Artemia, cysts is adequate survival, growth, and reproduction of the
mysids. The mysids used to evaluate the acceptability of the brine shrimp
nauplii must be of the same geographical origin and stage of development (7
days old) as those used routinely in the toxicity tests. Two 7-day chronic
tests are performed side-by-side, each consisting of eight replicate test
vessels containing five juveniles (40 organisms per test, total of 80
organisms). The juveniles in one set of test chambers is fed reference
(acceptable) nauplii and the other set is fed Artemia nauplii from the "new"
source of Artemia cysts.
6.12.3.2 The feeding rate and frequency, test vessels, volume of control
224
-------
water, duration of the test, and age of the Artemia nauplii at the start of
the test, should be the same as used for the routine toxicity tests.
6.12.3.3 Results of the brine shrimp, Artemia, nauplii nutrition assay, where
there are only two treatments, can be evaluated statistically by use of a
t-test. The "new" food is acceptable if there are no statistically
significant differences in the survival, growth, and reproduction of the
mysids fed the two sources of Artemia nauplii.
6.13 MYSIDS, Mysidopsis bahia (See Rodgers et al., 1986 and USEPA, 1991c for
information on mysid ecology).
6.13.1 Brood Stock
6.13.1.1 To provide an adequate supply of juveniles for a test, mysid,
Mysidopsis bahia, cultures should be started at least four weeks before the
test animals are needed. At least 200 mysids, Mysidopsis bahia, should be
placed in each culture tank to ensure that 1500 to 2000 animals will be
available by the time preparations for a test are initiated.
6.13.1.2 Mysids, Mysidopsis bahia, may be shipped or otherwise transported in
polyethylene bottles or CUBITAINERS*. Place 50 animals in 700 ml of seawater
in a 1-L shipping container. To control bacterial growth and prevent DO
depletion during shipment, do not add food. Before closing the shipping
container, oxygenate the water for 10 min. The mysids, Mysidopsis bahia, will
starve if not fed within 36 h, therefore, they should be shipped so that they
are not in transit more than 24 h.
6.13.1.3 The identification of the Mysidopsis bahia stock culture should be
verified using the key from Heard (1982), Price (1978, 1982), and Stuck et al.
(1979a,b). Records of the verification should be retained along with a few of
the preserved specimens.
6.13.1.4 Glass aquaria (120- to 200-L) are recommended for cultures. Other
types of culture chambers may also be convenient. Three or more separate
cultures should be maintained to protect against loss of the entire culture
stock in case of accident, low DO, or high nitrite levels, and to provide
sufficient numbers of juvenile mysids, Mysidopsis bahia, for toxicity tests.
Fill the aquaria about three-fourths full of seawater. A flow-through system
is recommended if sufficient natural seawater is available, but a closed,
recirculating or static renewal system may be used if proper water
conditioning is provided and care is exercised to keep the pH above 7.8 and
nitrite levels below 0.05 mg/L.
6.13.1.5 Standard aquarium undergravel filters should be used with either the
flow-through or recirculating system to provide aeration and a current
conducive to feeding (Gentile et al., 1983). The undergravel filter is covered
with a prewashed, coarse (2-5 mm) dolomite substrate, 2.5 cm deep for
flow-through cultures or 10 cm deep for recirculating cultures.
6.13.1.6 The recirculating culture system is conditioned as follows:
225
-------
1. After the dolomite has been added, the filter is attached to the air
supply and operated for 24 h.
2. Approximately 4 L of seed water obtained from a successfully operating
culture is added to the culture chamber.
3. The nitrite level is checked daily with an aquarium test kit or with
EPA Method 354.1 (USEPA, 1979b).
4. Add about 30 ml of concentrated Artemia nauplii every other day until
the nitrite level reaches at least 2.0 mg/L. The nitrite will continue
to rise for several days without adding more Artemia nauplii and will
then slowly decrease to less than 0.05 mg/L.
5. After the nitrite level falls below 0.05 mg/L, add another 30 mL of
Artemia nauplii concentrate and check the nitrite concentration every
day.
6. Continue this cycle until the addition of Artemia nauplii does not
cause a rise in the nitrite concentration. The culture chamber is
then conditioned and is ready to receive mysids.
7- Add only a few (5-20) mysids at first, to determine if conditions are
favorable. If these mysids are still doing well after a week, several
hundred more can be added.
6.13.1.7 It is important to add enough food to keep the adult animals from
cannibalizing the young, but not so much that the DO is depleted or that there
is a buildup of toxic concentrations of ammonia and nitrite. Just enough
newly-hatched Artemia nauplii are fed twice a day so that each feeding is
consumed before the next feeding.
6.13.1.8 Natural sea water is recommended as the culture medium, but MSB may
be used to make up the culture water if natural sea water is not available.
EMSL-Cincinnati has successfully used FORTY FATHOMS" artificial sea salts for
culturing and toxicity tests of mysids, and USEPA, Region 6 has used HW
MARINEMIXR artificial sea salts.
6.13.1.9 Mysids, Mysidopsis bahia, should be cultured at a temperature of 26
+ 1°C. No water temperature control equipment is needed if the ambient
laboratory temperature remains in the recommended range, and if there are no
frequent, rapid, large temperature excursions in the culture room.
6.13.1.10 The salinity should be maintained at 30 + 2 °/oo, or at a lower
salinity (but not less than 20 °/oo) if m0st of the tests will be conducted at
a lower salinity.
6.13.1.11 Day/night cycles prevailing in most laboratories will provide
adequate illumination for normal growth and reproduction. A 16-h/8-h day/night
cycle in which the light is gradually increased and decreased to simulate dawn
and dusk conditions, is recommended.
6.13.1.12 Mysid, Mysidopsis bahia, culture may suffer if DOs fall below 5
mg/L for extended periods. The undergravel filter will usually provide
sufficient DO. If the DO drops below 5 mg/L at 25°C and 30 °/oo, additional
aeration should be provided. Measure the DO in the cultures daily the first
week and then at least weekly thereafter.
226
-------
6.13.1.13 Suspend a clear glass or plastic panel over the cultures, or use
some other means of excluding dust and dirt, but leave enough space between
the covers and culture tanks to allow circulation of air over the cultures.
6.13.1.14 If hydroids or worms appear in the cultures, remove the mysids and
clean the chambers thoroughly, using soap and hot water. Rinse once with acid
(10% HC1) and three times with distilled or deionized water. Mysids with
attached hydroids should be discarded. Those without hydroids should be
transferred by hand pipeting into three changes of clean seawater before
returning them to the cleaned culture chamber. To guard against predators,
natural sea water should be filtered through a net with 30 \im mesh openings
before entering the culture vessels.
6.13.1.15 Mysids, Mysidopsis bahia, are very sensitive to low pH and sudden
changes in temperature. Care should be taken to maintain the pH at 8.0 ± 0.3,
and to limit rapid changes in water temperature to less than 3 C.
6.13.1.16 Mysids, Mysidopsis bahia, should be handled carefully and as little
as possible so that they are not unnecessarily stressed or injured. They
should be transferred between culture chambers with long handled cups with
netted bottoms. Animals should be transferred to the test vessels with a
large bore pipette (4-mm), taking care to release the animals under the
surface of the water. Discard any mysids that are injured during handling.
6.13.1.17 Culture Maintenance (Also See USEPA, 1991c)
6.13.1.17.1 Cultures in closed, recirculating systems are fed twice a day.
If no nauplii are present in the culture chamber after four hours, the amount
of food should be increased slightly. In flow-through systems, excess food
can be a problem by promoting bacterial growth and low dissolved oxygen.
6.13.1.17.2 Careful culture maintenance is essential. The organisms should
not be allowed to become too crowded. The cultures should be cropped as often
as necessary to maintain a density of about 20 mysids per liter. At this
density, at least 70% of the females should have eggs in their brood pouch.
If they do not, the cultures are probably under stress, and the cause should
be found and corrected. If the cause cannot be found, it may be necessary to
restart the cultures with a clean culture chamber, a new batch of culture
water, and clean gravel.
6.13.1.17.3 In closed, recirculating systems, about one third of the culture
water should be replaced with newly prepared seawater every week. Before
siphoning the old media from the culture, it is recommended that the sides of
the vessel be scraped and the gravel carefully turned over to prevent
excessive buildup of algal growth. Twice a year the mysids should be removed
from the recirculating cultures, the gravel rinsed in clean seawater, the
sides of the chamber washed with clean seawater, and the gravel and animals
returned to the culture vessel with newly conditioned seawater. No detergent
should be used, and care should be taken not to rinse all the bacteria from
the gravel.
6.13.2 Test Organisms
227
-------
6.13.2.1 The test is begun with 7-day old juveniles. To have the test
animals available and acclimated to test conditions at the start of the test,
they must be obtained from the stock culture eight days in advance of the
test. Whenever possible, brood stock should be obtained from cultures having
similar salinity, temperature, light regime, etc., as are to be used in the
toxicity test.
6.13.2.2 Eight days before the test is to start, sufficient gravid females
are placed in brood chambers. Assuming that 240 juveniles are needed for each
test, approximately half this number (120) of gravid females should be
transferred to brood chambers. The mysids are removed from the culture tank
with a net or netted cup and placed in 20-cm diameter finger bowls. The
gravid females are transferred from the finger bowls to the brood chambers
with a large-bore pipette or, alternatively, are transferred by pouring the
contents of the finger bowls into the water in the brood chambers.
6.13.2.3 The mysid juveniles may be collected for the toxicity tests by
transferring gravid females from the stock cultures to netted (1000 \im)
flow-through containers (Figure 1) held within 4-L glass, wide-mouth
separatory funnels. Newly released juveniles can pass through the netting,
whereas the females are retained. The gravid females are fed newly hatched
Artemia nauplii, and are held overnight to permit the release of young. The
juvenile mysids are collected by opening the stopcock on the funnel and
collecting them in another container from which they are transferred to
holding tanks using a wide bore (4 mm ID) pipette. The brood chambers usually
require aeration to maintain sufficient DO and to keep the food in suspension.
6.13.2.4 The temperature in the brood chamber should be maintained at the
upper acceptable culture limit (26 27°C), or 1°C higher than the cultures,
to encourage faster brood release. At this temperature, sufficient juveniles
should be produced for the test.
6.13.2.5 The newly released juveniles (age = 0 days) are transferred to 20-L
glass aquaria (holding vessels) which are gently aerated. Smaller holding
vessels may be used, but the density of organisms should not exceed 10 mysids
per liter. The test animals are held in the holding vessel for six days prior
to initiation of the test. The holding medium is renewed every other day.
6.13.2.6 During the holding period, the mysids are acclimated to the salinity
at which the test will be conducted, unless already at that salinity. The
salinity should be changed no more than 2 °/oo per 24 h to minimize stress on
the juveniles.
6.13.2.7 The temperature during the holding period is critical to mysid
development, and must be maintained at 26 + 1°C. If the temperature cannot be
maintained in this range, it is advisable to hold the juveniles an additional
day before beginning the test.
6.13.2.8 During the holding period, just enough newly-hatched Artemia nauplii
are fed twice a day (a total of at least 150 nauplii per mysid per day) so
that some food is constantly present.
228
-------
6.13.2.9 If the test is to be performed in the field, the juvenile mysids,
Mysidopsis bahia, should be gently siphoned into 1-L polyethylene wide-mouth
jars with screw-cap lids filled two-thirds full with clean seawater from the
holding tank. The water in these jars is aerated for 10 min, and the jars are
capped and packed in insulated boxes for shipment to the test site. Food
should not be added to the jars to prevent the development of excessive
bacterial growth and a reduction in DO.
6.13.2.10 Upon arrival at the test site (in less than 24 h) the mysids,
Mysidopsis bahia, are gently poured from the jars into 20-cm diameter glass
culture dishes. The jars are rinsed with salt water to dislodge any mysids
that may adhere to the sides. If the water appears milky, siphon off half of
it with a netted funnel (to avoid siphoning the mysids) and replace with clean
salt water of the same salinity and temperature. If no Artemia nauplii are
present in the dishes, feed about 150 Artemia nauplii per mysid.
7. SAMPLE COLLECTION, PRESERVATION AND HANDLING
7.1 See Section 8, Effluent and Receiving Water Sampling, Sample Handling,
Sample Preparation for Toxicity Tests.
8. CALIBRATION
8.1 See Section 4, Quality Assurance.
9. QUALITY CONTROL
9.1 See Section 4, Quality Assurance.
9.2 The reference toxicant recommended for use with the mysid 7-day test is
copper sulfate or soduim dodecyl sulfate.
10. TEST PROCEDURES
10.1 TEST DESIGN
10.1.1 The test consists of at least five effluent concentrations plus a site
water control and a reference water treatment (natural seawater or seawater
made up from hypersaline brine, or equivalent).
10.1.2 Effluent concentrations are expressed in percent effluent.
10.1.3 Eight replicate test vessels, each containing five 7-day old animals,
are used per effluent concentration and control.
10.2 TEST SOLUTIONS
10.2.1 Receiving waters
10.2.1.1 The sampling point(s) is determined by the objectives of the test.
At estuarine and marine sites, samples are usually collected at mid-depth.
Receiving water toxicity is determined with samples used directly as collected
229
-------
INFLOW
OUTFLOW
.NETTED
CHAMBER
.SEPARATORY
FUNNEL
NETTED
CHAMBER
^.CULTURE DISH
Figure 1. Apparatus (brood chamber) for collection of juvenile
mysids, Mysidopsis bahia. From USEPA (1987f).
230
-------
or with samples passed through a 60 \im NITEXR filter and compared without
dilution, against a control. Using four replicate chambers per test, each
containing 150 ml, and 400 ml for chemical analysis, would require
approximately 1 L or more of sample per test.
10.2.2 Effluents
10.2.2.1 The selection of the effluent test concentrations should be based on
the objectives of the study. A dilution factor of 0.5 is commonly used. A
dilution factor of 0.5 provides precision of + 100%, and testing of
concentrations between 6.25% and 100% effluent using only five effluent
concentrations (6.25%, 12.5%, 25%, 50%, and 100%). Test precision shows
little improvement as dilution factors are increased beyond 0.5 and declines
rapidly if smaller dilution factors are used. Therefore, USEPA recommends a
dilution factor of 0.5.
10.2.2.2 If the effluent is known or suspected to be highly toxic, a lower
range of effluent concentrations should be used (such as 25%, 12.5%, 6.25%,
3.12%, and 1.56%). If high mortality is observed during the first l-to-2 h of
the test, additional dilutions at the lower range of effluent concentrations
should be added.
10.2.2.3 The volume of effluent required for daily renewal of eight
replicates per concentration, each containing 150 ml of test solution, is
approximately 1200 ml. Prepare enough test solution (approximately 1600 ml)
at each effluent concentration to provide 400 ml additional volume for
chemical analyses.
10.2.2.4 The test should begin as soon as possible, preferably within 24 h
after sample collection. The maximum holding time following retrieval of the
sample from the sampling device should not exceed 36 h for off-site toxicity
tests unless permission is granted by the permitting authority. In no case
should the test be started more than 72 h after sample collection.
10.2.2.5 Just prior to test initiation (approximately one h), a sufficient
quantity of the sample to make the test solutions should be adjusted to the
test temperature (26 ± 1°C) and maintained at that temperature during the
addition of dilution water.
10.2.2.6 Effluent dilutions should be prepared for all replicates in each
treatment in one flask to minimize variability among the replicates. The test
chambers (cups) are labelled with the test concentration and replicate number.
Dispense 150 mL of the appropriate effluent dilution to each cup.
10.3 START OF THE TEST
10.3.1 Begin the test by randomly placing five animals (one at a time) in
each test cup of each treatment using a large bore (4 mm ID) pipette (see
Appendix A for an example of randomization). It is easier to capture the
animals if the volume of water in the dish is reduced and the dish is placed
on a light table. It is recommended that the transfer pipette be rinsed
frequently because mysids tend to adhere to the inside surface.
231
-------
10.4 LIGHT, PHOTOPERIOD, DO, AND TEMPERATURE
10.4.1 The light quality and intensity under ambient laboratory conditions
are generally adequate. Light intensity of 10-20 uE/m/s, or 50 to 100 foot
candles (ft.c), with a 16 h light and 8 h dark cycle is recommended. It is
critical that the test water temperature be maintained at 26 ± 1°C. It is
recommended that the test water temperature be continuously recorded.
10.4.1.1 If a water bath is used to maintain the test temperature, the water
depth surrounding the test cups should be at least 2.5 cm deep.
10.4.1.2 Rooms or incubators with high volume ventilation should be used with
caution because the volatilization of the test solutions and evaporation of
dilution water may cause wide fluctuations in salinity. Covering the test
cups with clear polyethylene plastic may help prevent volatilization and
evaporation of the test solutions.
10.4.2 Low DOs may be a problem when conducting effluent toxicity tests.
However, test cups should not be aerated unless the DO falls below 4.0 mg/L.
The higher concentrations of some effluents will require aeration to maintain
adequate DO concentrations. If one solution is aerated then all the treatments
and the controls must also be gently aerated.
10.5 FEEDING
10.5.1 During the test, the mysids in each test chamber should be fed Artemia
nauplii, which are less than 24-h old, at the rate of 150 nauplii per mysid
per day. Adding the entire daily ration at a single feeding immediately after
test solution renewal may result in a significant DO depression. Therefore,
it is preferable to feed half of the daily ration immediately after test
solution renewal, and the second half 8 - 12 h later. Increase the feeding if
the nauplii are consumed in less than 4 h. It is important that the nauplii
be washed before introduction to the test vessels.
10.6 TEST SOLUTION RENEWAL
10.6.1 Test solutions are renewed daily. Slowly pour off all but 10 mL of
the old test medium into a 20 cm diameter culture dish on a light table. Be
sure to check for animals that may have adhered to the sides of the test
vessel. Rinse them back into the test cups. Add 150 mL of new test solution
slowly to each cup. Check the culture dish for animals that may have been
poured out with the old media, and return them to the test vessel.
10.7 ROUTINE CHEMICAL AND PHYSICAL DETERMINATIONS
10.7.1 At a minimum, the following measurements should be made in at least
one replicate in the control and the high and low test concentrations at the
beginning of the test: temperature, dissolved oxygen, pH, and salinity (see
Figure 14).
10.7.2 DO should be measured in at least one replicate in the control and the
high and low test concentrations before renewing the test medium and after the
232
-------
medium is renewed each day. In addition to the daily calibrations, the DO
meter should be checked at least once a week against a standard Winkler
titration.
10.7.3 pH, temperature, and salinity should be measured in at least one
replicate for each treatment at the beginning of each 24-h exposure period.
10.7.4 It may be advisable to measure the ammonia and nitrite in the controls
before each renewal to be certain that toxicity from these sources is not
confounding the test results.
10.8 OBSERVATIONS DURING THE TEST
10.8.1 The number of live mysids are counted and recorded each day when the
test solutions are renewed (see Figure 15). Dead animals and excess food
should be removed with a pipette before the test solutions are renewed.
10.9 TERMINATION OF THE TEST
10.9.1 After measuring the DO, pH, temperature, and salinity and recording
survival, terminate the test by pouring off the test solution in all the cups
to a one cm depth and refilling the cups with clean seawater. This will keep
the animals alive, but not exposed to the toxicant, while waiting to be
examined for sex and the presence of eggs.
10.9.2 The live animals must be examined for eggs and the sexes determined
within 12 h of the termination of the test. If the test was conducted in the
field, and the animals cannot be examined on site, the live animals should be
shipped back to the laboratory for processing. Pour each replicate into a
labelled 100 mL plastic screw capped jar, and send to the laboratory
immediately.
10.9.3 If the test was conducted in the laboratory, or when the test animals
arrive in the laboratory from the field test site, the test organisms must be
processed immediately while still alive as follows:
10.9.3.1 Examine each replicate under a stereomicroscope (240X) to determine
the number of immature animals, the sex of the mature animals, and the
presence or absence of eggs in the oviducts or brood sacs of the females (see
Figures 2-5). This must be done while the mysids are alive because they turn
opaque upon dying. This step should not be attempted by a person who has not
had specialized training in the determination of sex and presence of eggs in
the oviduct. NOTE: Adult females without eggs in the oviduct or brood sac
look like immature mysids (see Figure 5).
10.9.3.2 Record the number of immatures, males, females with eggs and females
without eggs on data sheets.
10.9.3.3 Rinse the mysids by pipetting them into a small netted cup and
dipping the cup into a dish containing deionized water. Using forceps, place
the mysids from each replicate cup on tared weighing boats and dry at 60 C for
24 h or at 105°C for at least 6 h.
233
-------
MATURE FEMALE,EGGS IN OVIDUCTS
eyestalk
carapace
antennule
developing
brood
sac Pleopods
antenna
statocyst
telson
uropod
^developing brood sac
oviducts with developing ova
Figure 2. Mature female, mysid, Mysidopsis bahia, with eggs in
oviducts. From USEPA (1987f).
234
-------
MATURE FEMALE, EGGS IN BROOD SAC
eyestolk
coropace
antennule
antenna
pleopods
brood sac with
developing embryos
statocyst
telson
uropod'
brood sac with
developing embryos
oviducts with developing ova
Figure 3. Mature female mysid, Mysidopsis. bahia, with eggs in oviducts and
developing embryos in the brood sac. Above: lateral view. Below:
dorsal view. From USEPA (1987f).
235
-------
ontennule
antenna
MATURE MALE
eyestalk
carapace
statocyst
telson
uropod'
gonad
oil globules
Figure 4. Mature male mysid, Mysidopsis bahia. From USEPA (1987f).
236
-------
IMMATURE
antennule
antenna
eyestalk
carapace
telson
uropod'
statocyst
Figure 5. Immature mysid, Mysidopsis bahia, (A) lateral view, (B) dorsal
view. From USEPA (1987f).
237
-------
10.9.3.3.1 Pieces of aluminum foil (1-cm square) or small aluminum weighing
boats can be used for dry weight analyses. The weighing pans or boats should
not exceed 10 mg in weight.
10.9.3.3.2 Number each pan with a waterproof pen with the treatment
concentration and replicate number. Forty-eight (48) weighing pans are
required per test if all the organisms survive.
10.9.3.3.3 Remove the trays with the pans on from the oven and transfer
immediately to a dessicator before weighing. After cooling for 1 h, handle
pans with forceps only and weigh to the nearest microgram.
11. ACCEPTABILITY OF TEST RESULTS
11.1 The minimum requirements for an acceptable test are 80% survival and an
average weight of at least 0.20 mg/mysid in the controls. If fecundity in the
controls is adequate (egg production by 50% of females), fecundity should be
used as a criterion of effect in addition to survival and growth.
12. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA
12.1 A summary of test conditions and test acceptability criteria is listed
in Table 3.
13. DATA ANALYSIS
13.1 GENERAL
13.1.1 Tabulate and summarize the data. Table 4 presents a sample set of
survival, growth, and fecundity data.
13.1.2 The endpoints of the mysid 7-day rapid chronic test are based on the
adverse effects on survival, growth, and egg development. The LC50, the IC25,
and the IC50 are calculated using point estimation techniques (see Section 9,
Chronic Toxicity Test Endpoints and Data Analysis). LOEC and NOEC values, for
survival, growth, and reproduction are obtained using a hypothesis test
approach such as Dunnett's Procedure (Dunnett, 1955) or Steel's Many-one Rank
Test (Steel, 1959; Miller, 1981). See the Appendices for examples of the
manual computations, program listings, and examples of data input and program
output.
13.1.3 The statistical tests described here must be used with a knowledge of
the assumptions upon which the tests are contingent. The assistance of a
statistician is recommended for analysts who are not proficient in statistics.
13.2 EXAMPLE OF ANALYSIS OF MYSID SURVIVAL DATA
13.2.1 Formal statistical analysis of the survival data is outlined in Fiqure
6. The response used in the analysis is the proportion of animals surviving
in each test or control chamber. Separate analyses are performed for the
estimation of the NOEC and LOEC endpoints and for the estimation of the LC50
238
-------
TABLE 3. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA
FOR THE MYSID, MYSIDOPSIS BAHIA, SEVEN DAY SURVIVAL, GROWTH,
AND FECUNDITY TEST WITH EFFLUENTS AND RECEIVING WATERS
1. Test type:
2. Salinity:
3. Temperature:
4. Photoperiod:
5. Light-intensity:
6. Test chamber:
7-
8.
9.
10.
11.
12.
13.
14.
Test solution volume:
Renewal of test solutions:
Age of test organisms:
Test concentrations:
Number of organisms per test
chamber:
Number of replicate chambers
per treatment:
Source of food:
Feeding regime:
15. Aeration:
16. Dilution water:
Static renewal
20 °/oo to 30 °/oo ± 2 °/oo
26 ± 1°C
16 h light, 8 h dark, with phase
in/out period
10-20 \i£/m2/s (50-100 ft.c.)
8 oz plastic disposable cups, or
400 mL glass beakers
150 mL per replicate cup
Daily
7 days
Effluents: Minimum of five effluent
concentrations and a control
Receiving waters: 100% receiving
water and a control
Artemia nauplii
Feed 150 24-h old nauplii per mysid
daily, half after test solution
renewal and half after 8 - 12 h.
None unless DO falls below 4.0 mg/L,
then gently aerate in all cups
Natural sea water, hypersaline brine,
GP2, FORTY FATHOMS", or HW MARINEMIX*
artifical sea salts
239
-------
TABLE 3. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA
FOR THE MYSID, HYSIDOPSIS BAHIA, SEVEN DAY SURVIVAL, GROWTH,
AND FECUNDITY TEST WITH EFFLUENTS AND RECEIVING WATERS
(CONTINUED)
17. Test duration:
18. Dilution factor:
7 days
Effluents:
> 0.5 series
19. Effects measured:
20. Cleaning:
21. Test acceptability
criteria:
22. Sampling requirements:
23. Estimated maximum sample
volume required:
Receiving waters: None, or > 0.5
series
Survival, growth, and egg development
Pipette excess food from cups daily
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
For on-site tests, samples are
collected daily, and used within 24 h
of the time they are removed from the
sampling device. For off-site tests,
a minimum of three samples are
collected on days one, three, and five
with a maximum holding time of 36 h
before first use (see Section 14,
Subsection 10.2.2.4).
3 L per day
240
-------
TABLE 4. DATA FOR MYSIDOPSIS BAHIA 7-DAY SURVIVAL, GROWTH, AND FECUNDITY TEST1
Treatment Replicate
Chamber
1
2
3
Control 4
5
6
7
8
1
2
3
50 ppb 4
5
6
7
8
1
2
3
100 ppb 4
5
6
7
8
1
2
3
210 ppb 4
5
6
7
8
1
2
3
450 ppb 4
5
6
7
8
Total
Mysids
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
4
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
No.
Alive
4
4
5
5
5
5
5
4
4
5
4
4
5
5
4
5
3
5
5
5
5
3
4
4
5
4
1
4
3
4
4
4
0
1
0
1
0
0
0
2
Total
Females
1
2
3
1
2
5
2
3
2
3
3
0
5
2
4
3
3
2
1
2
3
1
4
0
1
2
1
3
1
2
1
3
0
0
0
0
0
0
0
1
Females
w/Eggs
1
2
2
1
2
4
2
3
1
1
2
0
2
1
1
1
1
1
0
1
2
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Mean
Weight
0.183
0.148
0.216
0.199
0.176
0.243
0.213
0.180
0.192
0.193
0.237
0.237
0.256
0.191
0.152
0.177
0.190
0.172
0.160
0.199
0.165
0.241
0.259
0.186
0.153
0.117
0.085
0.153
0.086
0.193
0.137
0.129
_ _
0.060
-
0.009
- -
- -
- -
0.203
1Data provided by Lussier, Kuhn and Sewall, ERL
241
Narragansett, Rhode Island.
-------
endpoint. Concentrations at which there is no survival in any of the test
chambers are excluded from statistical analysis of the NOEC and LOEC, but
included in the estimation of the LC50 endpoint.
13.2.2 For the case of equal numbers of replicates across all concentrations
and the control, the evaluation of the NOEC and LOEC endpoints is made via a
parametric test, Dunnett's Procedure, or a nonparametric test, Steel's
Many-one Rank Test, on the arc sine transformed data. Underlying assumptions
of Dunnett's Procedure, normality and homogeneity of variance, are formally
tested. The test for normality is the Shapiro-Milk's Test, and Bartlett's
Test is used to determine the homogeneity of variance. If either of these
tests fail, the nonparametric test, Steel's Many-one Rank Test, is used to
determine the NOEC and LOEC endpoints. If the assumptions of Dunnett's
Procedure are met, the endpoints are estimated by the parametric procedure.
13.2.3 If unequal numbers of replicates occur among the concentration levels
tested, there are parametric and nonparametric alternative analyses. The
parametric analysis is a t-test with the Bonferroni adjustment. The Wilcoxon
Rank Sum Test with the Bonferroni adjustment is the nonparametric alternative.
For detailed information on the Bonferroni adjustment, see Appendix D.
13.2.4 Probit Analysis (Finney, 1971) is used to estimate the LC50. In this
analysis, the total mortality data from all test replicates at a given
concentration are combined. If the data do not fit the Probit model, the
Spearman-Karber method, the trimmed Spearman-Karber method, or the Graphical
method may be used (USEPA, 1991c).
13.2.5 The proportion of survival in each replicate must first be transformed
by the arc sine transformation procedure decribed in Appendix B. The raw and
transformed data, means and standard deviation of the transformed observations
at each concentration including the control are listed in Table 5. A plot of
the mean survival is provided in Figure 7.
13.2.6 Test for Normality
13.2.6.1 The first step of the test for normality is to center the
observations by subtracting the mean of all observations within a
concentration from each observation in that concentration. The centered
observations are listed in Table 6.
242
-------
TABLE 5. MYSID, MYSIDOPSIS BAHIA, SURVIVAL DATA
Concentration
Repl
Raw
Arc sine
Trans
formed
Mean(Y,.)
s?
i
icate
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
Control
0.80
0.80
1.00
1.00
1.00
1.00
1.00
0.80
1.107
1.107
1.345
1.345
1.345
1.345
1.345
1.107
1.256
0.015
1
50.0
0.80
1.00
0.80
0.80
1.00
1.00
0.80
1.00
1.107
1.345
1.107
1.107
1.345
1.345
1.107
1.345
1.226
0.016
2
100.0
0.60
1.00
1.00
1.00
1.00
0.60
0.80
0.80
0.886
1.345
1.345
1.345
1.345
0.886
1.107
1.107
1.171
0.042
3
210.0
1.00
0.80
0.20
0.80
0.60
0.80
0.80
0.80
1.345
1.107
0.464
1.107
0.886
1.107
1.107
1.107
1.029
0.067
4
450.0
0.00
0.20
0.00
0.20
0.00
0.00
0.00
0.40
0.225
0.464
0.225
0.464
0.225
0.225
0.225
0.685
0.342
0.031
5
243
-------
STATISTICAL ANALYSIS OF MYSIDOPSIS BAHIA
SURVIVAL, GROWTH, AND FECUNDITY TEST
SURVIVAL
SURVIVAL DATA
PROPORTION SURVIVING
ARC SINE
TRANSFORMATION
ENDPOINT ESTIMATE
LC50
SHAPIRO-WILK'S TEST
NORMAL DISTRIBUTION
NON-NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
NO
BARTLETTS TEST
HETEROGENEOUS
VARIANCE
EQUAL NUMBER OF
REPLICATES?
YES
T-TESTWITH
BONFERRONI
ADJUSTMENT
EQUAL NUMBER OF
REPLICATES?
i
YES
>
DUNNETTS
TEST
STEEL'S MANY-ONE
RANK TEST
WILCOXON RANK SUM
TEST WITH
BONFERRONI ADJUSTMENT
ENDPOINT ESTIMATES
NOEC.LOEC
Figure 6. Flowchart for analysis of mysid, Mysidopsis bahia, survival
data.
244
-------
CONNECTS THE MEAN VALUE FOR EACH CONCENTRATION
1.0
0.9-
0.8
0.7-
o
§0.61
o
a.
0.4-
0.3-
0.2-
0.1
o.o-l
50
100
CONCENTRATION (PPB)
210
450
Figure 7. Plot of mean survival of mysids, Mysidopsis bahia, at each treatment level.
-------
TABLE 6. CENTERED OBSERVATIONS FOR SHAPIRO-MILK'S EXAMPLE
Concentration
Replicate
1
2
3
4
5
6
7
8
Control
(Site Water)
-0.149
-0.149
0.089
0.089
0.089
0.089
0.089
-0.149
50.0
-0.119
0.119
-0.119
-0.119
0.119
0.119
-0.119
0.119
100.0
-0.285
0.174
0.174
0.174
0.174
-0.285
-0.064
-0.064
210.0
0.316
0.078
-0.565
0.078
-0.142
0.078
0.078
0.078
450.0
-0.117
0.121
-0.117
0.121
-0.117
-0.117
-0.117
0.342
13.2.6.2 Calculate the denominator, D, of the test statistic:
2
D = E (X,. - X)
Where: Xj = the ith centered observation
X = the overall mean of the centered observations
n = the total number of centered observations.
For this set of data,
n = 40
X = _1_(-0.006) = 0.0
40
D = 1.197
13.2.6.3 Order the centered observations from smallest to largest:
v(1) v(2) j/(n)
Where X(1) is the ith ordered observation. These ordered observations are
listed in Table 7.
246
-------
TABLE 7. ORDERED CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S EXAMPLE
xci>
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
-0.565
-0.285
-0.285
-0.149
-0.149
-0.149
-0.143
-0.119
-0.119
-0.119
-0.119
-0.117
-0.117
-0.117
-0.117
-0.117
-0.064
-0.064
0.078
0.078
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
0.078
0.078
0.078
0.089
0.089
0.089
0.089
0.089
0.119
0.119
0.119
0.119
0.121
0.121
0.174
0.174
0.174
0.174
0.316
0.342
13.2.6.4 From Table 4, Appendix B, for the number of observations, n, obtain
the coefficients a,, a2, , ak where k is n/2 if n is even and (n-l)/2 if n
is odd. For the data in this example, n = 40 and k = 20. The a{ values are
listed in Table 8.
13.2.6.5 Compute the test statistic, W, as follows:
W = 1 [ E a, (X(n-i+1) X(i)) f
D i = l
The differences x(n"i+1) - X(i) are listed in Table 7. For this set
of data:
W = 1 (1.0475)2 = 0.9167
0 1.197
247
-------
13.2.6.6 The decision rule for this test is to compare W with the critical
value found in Table 6, Appendix B. If the computed W is less than the
critical value, conclude that the data are not normally distributed. For this
set of data, the critical value at a significance level of 0.01 and n = 40
observations is 0.919. Since W = 0.9167 is less than the critical value, the
conclusion of the test is that it is reasonable to assume the data are not
normally distributed.
13.2.6.7 Since the data do not meet the assumption of normality, Steel's
Many-one Rank Test will be used to analyze the survival data.
TABLE 8. COEFFICIENTS AND DIFFERENCES FOR SHAPIRO-WILK'S EXAMPLE
i
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
ai
0.3964
0.2737
0.2368
0.2098
0.1878
0.1691
0.1526
0.1376
0.1237
0.1108
0.0986
0.0870
0.0759
0.0651
0.0546
0.0444
0.0343
0.0244
0.0146
0.0049
«(n-i+1) «(i)
0.907
0.601
0.459
0.323
0.323
0.323
0.264
0.240
0.238
0.238
0.238
0.236
0.206
0.206
0.206
0.206
0.153
0.142
0.0
0.0
X<40)
x<39>
X(38)
x<37>
X(36)
x<35)
X(34)
v(33)
A
X(32)
X(31)
x<30)
X(29)
X(28)
X(27)
X(26)
X(25)
x<2*>
X(23)
X(22)
X(21>
x<1)
x<2)
X(3)
x<4)
X(5>
x<6)
X(7)
x<8)
X(9)
X(10)
- x(11>
X(12)
X(13)
X(H)
X(15>
x<16)
- x(17)
- x(18)
x<19)
v(20)
A
248
-------
13.2.7 Steel's Many-one Rank Test
13.2.7.1 For each control and concentration combination, combine the data and
arrange the observations in order of size from smallest to largest. Assign
the ranks (1,2, ... ,16) to the ordered observations with a rank of 1 assigned
to the smallest observation, rank of 2 assigned to the next larger
observation, etc. If ties occur when ranking, assign the average rank to each
tied observation.
13.2.7.2 An example of assigning ranks to the combined data for the control
and 50.0 ppb concentration is given in Table 9. This ranking procedure is
repeated for each control/concentration combination. The complete set of
rankings is summarized in Table 10. The ranks are then summed for each
concentration level, as shown in Table 11.
13.2.7.3 For this example, we want to determine if the survival in any of the
concentrations is significantly lower than the survival in the control. If
this occurs, the rank sum at that concentration would be significantly lower
than the rank sum of the control. Thus we are only concerned with comparing
the rank sums for the survival at each of the various concentration levels
with some "minimum" or critical rank sum, at or below which the survival would
be considered significantly lower than the control. At a significance level
of 0.05, the minimum rank sum in a test with four concentrations (excluding
the control) and eight replicates is 47 (See Table 5, Appendix E).
13.2.7.4 Since the rank sum for the 450 ppb concentration level is less than
the critical value, the proportion surviving in that concentration is
considered significantly less than that in the control. Since no other rank
sums are less than or equal to the critical value, no other concentrations
have a significantly lower proportion surviving than the control. Hence, the
NOEC and the LOEC are assumed to be 210.0 ppb and 450.0 ppb, respectively.
249
-------
TABLE 9. ASSIGNING RANKS TO THE CONTROL AND 50 PPB CONCENTRATION LEVEL
FOR STEEL'S MANY-ONE RANK TEST
Transformed Proportion
Rank of Total Mortality Concentration
4
4
4
4
4
4
4
12
12
12
12
12
12
12
12
12
1.107
1.107
1.107
1.107
1.107
1.107
1.107
1.571
1.571
1.571
1.571
1.571
1.571
1.571
1.571
1.571
Control
Control
Control
50 ppb
50 ppb
50 ppb
50 ppb
Control
Control
Control
Control
Control
50 ppb
50 ppb
50 ppb
50 ppb
250
-------
TABLE 10. TABLE OF RANKS1
Repli- Control
50
100
210
450
cate
1
2
3
4
5
6
7
8
1.107(4,5,6.5,10)
1.107(4,5,6.5,10)
1.345(12,12,13.5,14)
1.345(12,12,13.5,14)
1.345(12,12,13.5,14)
1.345(12,12,13.5,14)
1.345(12,12,13.5,14)
1.107(4,5,6.5,10)
1
1
1
1
1
1
1
1
.107(4)
.345(12)
.107(4)
.107(4)
.345(12)
.345(12)
.107(4)
.345(12)
0
1
1
1
1
0
1
1
.886(1.5)
.345(12)
.345(12)
.345(12)
.345(12)
.886(1.5)
.107(5)
.107(5)
1
1
0
1
0
1
1
1
.345(13
.107(6.
.464(1)
.107(6.
.886(2)
.107(6.
.107(6.
.107(6.
.5)
5)
5)
5)
5)
5)
0
0
0
0
0
0
0
0
.225(3)
.464(6.
.225(3)
.464(6.
.225(3)
.225(3)
.225(3)
.685(8)
5)
5)
1Control ranks are given in the order of the concentration with which
they were ranked.
TABLE 11. RANK SUMS
Concentration
Rank Sum
50
100
210
450
64
61
49
36
251
-------
13.2.8 Probit Analysis
13.2.8.1 The data used for the Probit Analysis is summarized in Table 12. To
perform the Probit Analysis, run the EPA Probit Analysis Program. An example
of the program output is provided in Table 13 and Figure 8. The current
version of the Probit program makes no distinction between EC and LC
endpoints, and labels all results as EC values. Because the response here is
mortality, the EC values output by the program should be treated as the
corresponding LC values.
13.2.8.2 For this example, the chi-square test for heterogeneity was not
significant. Thus Probit Analysis appears to be appropriate for this set of
data.
TABLE 12. DATA FOR PROBIT ANALYSIS
Concentration
Control 50.0 100.0 210.0 450.0
No Dead
No Exposed
3
40
4
40
6
40
11
40
36
40
13.3 EXAMPLE OF ANALYSIS OF MYSID GROWTH DATA
13.3.1 Formal statistical analysis of the growth data is outlined in Figure
9. The response used in the statistical analysis is mean weight of males and
females combined per replicate. The IC25 and IC50 can be calculated for the
growth data via a point estimation technique (see Section 9). Hypothesis
testing can be used to obtain an NOEC and LOEC for growth. Concentrations
above the NOEC for survival are excluded from the hypothesis test for growth
effects.
13.3,2 The statistical analysis using hypothesis tests consists of a
parametric test, Dunnett's Procedure, and a nonparametric test, Steel's
Many-one Rank Test. The underlying assumptions of the Dunnett's Procedure,
normality and homogeneity of variance, are formally tested. The test for
normality is the Shapiro-Wilk's Test and Bartlett's Test is used to test for
homogeneity of variance. If either of these tests fail, the nonparametric
test, Steel's Many-one Rank Test, is used to determine the NOEC and LOEC
endpoints. If the assumptions of Dunnett's Procedure are met, the endpoints
are determined by the parametric test.
252
-------
TABLE 13. OUTPUT FOR USEPA PROBIT ANALYSIS PROGRAM,
VERSION 1.4
EPA PROBIT ANALYSIS PROGRAM
USED FOR CALCULATING EC VALUES
Version 1.4
Probit Analysis of Mysidopsis bahia Survival Data
Cone.
Control
50.0000
100.0000
210.0000
450.0000
Number
Exposed
40
40
40
40
40
Number
Resp.
4
4
6
11
36
Observed
Proportion
Responding
0.1000
0.1000
0.1500
0.2750
0.9000
Adjusted
Proportion
Responding
0.0000
-.0179
0.0387
0.1801
0.8869
Predicted
Proportion
Responding
0.1158
0.0000
0.0015
0.1827
0.8861
Chi - Square Heterogeneity
Mu
Sigma
Parameter
2.464117
0.156802
Estimate
0.517
Std. Err.
95% Confidence Limits
Intercept
Slope
-10.714873
6.377486
3.194243
1.274491
( -16.975590,
( 3.879483,
-4.454156)
8.875488)
Spontaneous
Response Rate
0.115787
0.029555
0.057859,
0.173715)
Estimated EC Values and Confidence Limits
Point
.00
.00
EC 1.
EC 5.
EC10.00
EC15.00
EC50.00
EC85.00
EC90.00
EC95.00
EC99.00
Cone.
125.7042
160.7661
183.2992
200.2664
291.1503
423.2791
462.4602
527.2789
674.3496
Lower Upper
95% Confidence Limits
66.2441
98.3363
121.0638
139.0534
240.4525
361.8588
391.0961
436.0173
528.6572
169.
203.
225.
242.
338.
545.
621.
760.
1121.
1211
7259
5930
0966
8918
0946
6678
1824
1471
253
-------
Probit Analysis of Mysidopsis bahia Survival Data
PLOT OF ADJUSTED PROBITS AND PREDICTED REGRESSION LINE
Probit
10+
8+
O..
4 +
-O
3+
2+
1+
0+0
_+ + + + + + +_
EC01 EC10 EC25 EC50 EC75 EC90 EC99
Figure 8. Plot of adjusted Probits and predicted regression line.
254
-------
STATISTICAL ANALYSIS OF MYSIDOPSIS BAHIA
SURVIVAL, GROWTH, AND FECUNDITY TEST
GROWTH
GROWTH DATA
MEAN WEIGHT
POINT ESTIMATION
HYPOTHESIS TESTING
(EXCLUDING CONCENTRATIONS
ABOVE NOEC FOR SURVIVAL)
ENDPOINT ESTIMATE
IC25, IC50
i
SHAPIRO-WILK'S TEST
NORMAL DISTRIBUTION
NON-NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
BARTLETTS TEST
HETEROGENEOUS
VARIANCE
NO
-
T-TESTWITH
BONFERRONI
ADJUSTMENT
EQUAL NUMBER OF
REPLICATES?
YES
i
EQUAL NL
REPLIC
YES
i i '
DUNNETTS STEEL'S MANY-ONE
TEST RANK TEST
jMBER OF
ATES?
\
'
WILCOXON RANK SUM
TEST WITH
BONFERRONI ADJUSTMENT
ENDPOINT ESTIMATES
NOEC.LOEC
Figure 9. Flowchart for statistical analysis of mysid, Mysidopsis bahia, growth
data.
255
-------
13.3.3 Additionally, if unequal numbers of replicates occur among the
concentration levels tested, there are parametric and nonparametric
alternative analyses. The parametric analysis is a t-test with the Bonferroni
adjustment. The Wilcoxon Rank Sum Test with the Bonferroni adjustment is the
nonparametric alternative. For detailed information on the Bonferroni
adjustment, see Appendix D.
13.3.4 The data, mean and standard deviation of the observations at each
concentration including the control for this example are listed in Table 14.
A plot of the data is provided in Figure 10. Since there is significant
mortality in the 450 ppb concentration, its effect on growth is not
considered.
TABLE 14. MYSID, MYSIDOPSIS BAHIA, GROWTH DATA
Concentration (ppb)
Replicate Control
50.0
100.0
210.0
450.0
Mean(Yj)
s?
i
1
2
3
4
5
6
7
8
0.183
0.148
0.216
0.199
0.176
0.243
0.213
0.180
0.195
0.0008
1
0.192
0.193
0.237
0.237
0.256
0.191
0.152
0.177
0.204
0.0012
2
0.190
0.172
0.160
0.199
0.165
0.241
0.259
0.186
0.197
0.0013
3
0.153
0.117
0.085
0.153
0.086
0.193
0.137
0.129
0.132
0.0013
4
0.060
0.009
-
-
-
0.203
5
256
-------
CONNECTS THE MEAN VALUE FOR EACH CONCENTRATION
REPRESENTS THE CRITICAL VALUE FOR DUNNETT'S TEST
(ANY MEAN WEIGHT BELOW THIS VALUE WOULD 3E
SIGNIFICANTLY DIFFERENT FROM THE CONTROL)
ro
en
100
210
CONCENTRATION (PPB)
Figure 10. Plot of mean values for mysid, Mysidopsis bahia, growth
-------
13.3.5 Test for Normality
13.3.5.1 The first step of the test for normality is to center the
observations by subtracting the mean of all observations within a
concentration from each observation in that concentration. The centered
observations are listed in Table 15.
TABLE 15. CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S EXAMPLE
Concentration (ppb)
Replicate
Control
50.0
100.0
210.0
1 -0.012
2 -0.047
3 0.021
4 0.004
5 -0.019
6 0.048
7 0.018
8 -0.015
-0.012
-0.011
0.033
0.033
0.052
-0.013
-0.052
-0.027
-0.006
-0.024
-0.036
0.002
-0.032
0.044
0.062
-0.010
0.021
-0.015
-0.047
0.021
-0.046
0.061
0.005
-0.003
13.3.5.2 Calculate the denominator, D, of the statistic:
2
D = S (X, X)
Where: X,- = the ith centered observation
X = the overall mean of the centered observations
n = the total number of centered observations
13.3.5.3 For this set of data: n = 32
X = _1_ (-0.000) = 0.000
32
D = 0.0329
13.3.5.4 Order the centered observations from smallest to largest
^(1) _ v(2) _ _ v(n)
where X(1) denotes the ith ordered observation. The ordered observations for
this example are listed in Table 16.
258
-------
TABLE 16. ORDERED CENTERED OBSERVATIONS FOR SHAPIRO-MILK'S EXAMPLE
cn
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
-0.052
-0.047
-0.047
-0.046
-0.036
-0.032
-0.027
-0.024
-0.019
-0.015
-0.015
-0.013
-0.012
-0.012
-0.011
-0.010
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
-0.006
-0.003
0.002
0.004
0.005
0.018
0.021
0.021
0.021
0.033
0.033
0.044
0.048
0.052
0.061
0.062
13.3.5.5 From Table 4, Appendix B, for the number of observations, n, obtain
the coefficients a,,, a2, ... ak where k is n/2 if n is even and (n-l)/2 if n
is odd. For the data in this example, n = 32 and k = 16. The a( values are
listed in Table 17.
13.3.5.6 Compute the test statistic, W, as follows:
k
£
D i-1
W = 1 [ E a, (X(n-i+1) - X(i)) ]*
The differences x(n"i+1) - X(i> are listed in Table 17. For the data in this
example,
W = _J (0.1765)2 = 0.9469
0.0329
259
-------
TABLE 17. COEFFICIENTS AND DIFFERENCES FOR SHAPIRO-MILK'S EXAMPLE
a, Xn'+ - X(0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
0.4188
0.2898
0.2462
0.2141
0.1878
0.1651
0.1449
0.1265
0.1093
0.0931
0.0777
0.0629
0.0485
0.0344
0.0206
0.0068
0.114
0.108
0.099
0.094
0.080
0.065
0.060
0.045
0.040
0.036
0.033
0.018
0.016
0.014
0.008
0.004
v(32)
A_,,_
v(31)
A
v(30)
A
v<29>
A
v<28)
A
Y(27>
A
v<26)
A
v(25)
A.-, .
Y<24)
A
X
X
X
X(20)
X(19)
X(18)
X(17)
Y
-------
-2 ( 2 V, S,-2)
S = 1=1
C = 1 + ( 3(p-l))-1 [ L 1/V, - ( I VJ'1 ]
1=1 i=l
In = loge
i = 1, 2, ..., p where p is the number of concentrations
including the control
nf = the number of replicates for concentration i.
13.3.6.2 For the data in this example (See Table 13), all concentrations
including the control have the same number of replicates (n- = 8 for all i).
Thus, V, = 7 for all i.
13.3.6.3 Bartlett's statistic is therefore:
P .
B = [(28)ln(0.0012) 7 E ln(S*)]/1.06
i=l
= [28(-6.7254) - 7(-27.1471)]/1.06
= [-188.3112 - (-190.0297)1/1.06
= 1.621
13.3.6.4 B is approximately distributed as chi-square with p 1 degrees of
freedom, when the variances are in fact the same. Therefore, the appropriate
critical value for this test, at a significance level of 0.01 with three
degrees of freedom, is 9.210. Since B = 1.621 is less than the critical value
of 9.210, conclude that the variances are not different.
13.3.7 Dunnett's Procedure
13.3.7.1 To obtain an estimate of the pooled variance for the Dunnett's
Procedure, construct an ANOVA table as described in Table 18.
261
-------
TABLE 18. ANOVA TABLE
Source
Between
Within
Total
Where: p
N
n,-
df
P - 1
N - p
N - 1
= number
= total
= number
Sum of Squares
(SS)
SSB
SSW
SST
Mean Square(MS)
(SS/df)
S* = SSB/(p-l)
S* = SSW/(N-p)
of concentration levels including the control
number of observations n., + n, ... +np
of observations in concentration i
SSB
T,2/nf - G2/N
Between Sum of Squares
SST = S s'
- G2/N
Total Sum of Squares
SSW = SST - SSB
Within Sum of Squares
G = the grand total of all sample observations, G = E T,.
i=l
T,- = the total of the replicate measurements for
concentration "i"
Yjj = the jth observation for concentration "i" (represents
the mean dry weight of the mysids for concentration
i in test chamber j)
13.3.7.2 For the data in this example:
.I ' n
'ft
n,
n,
8
N
T
T, = Y,; + Y22 + ... + Y2~ = 1.635
T, = Yn + Y12 + ... + Y18 = 1.558
,2 - Y21
T3 - Y31
T4 = Y41
32
Y38 ' 1-572
42 48 1-053
G = T, + T2 + T3 + T4 = 5.818
262
-------
SSB = S T,2/n, G2/N
= _i_(8.680) - (5.818)2 = 0.027
8 32
P n,-
SST = S S Y?: - G2/N
= 1.118 - (5.818)2 = 0.060
32
SSW = SST - SSB = 0.060 - 0.027 = 0.033
S2 = SSB/(p-l) = 0.027/(4-l) = 0.009
Sj = SSW/(N-p) = 0.033/(32-4) = 0.001
13.3.7.3 Summarize these calculations in the ANOVA table (Table 19)
TABLE 19. ANOVA TABLE FOR DUNNETT'S PROCEDURE EXAMPLE
Source
Between
Within
Total
df
3
28
31
Sum of Squares
(SS)
0.027
0.033
0.060
Mean Square(MS)
(SS/df)
0.009
0.001
13.3.7.4 To perform the individual comparisons, calculate the t statistic for
each concentration, and control combination as follows:
Sw / (1/n,) + (1/n,)
Where: Y,- = mean dry weight for concentration i
Y, = mean dry weight for the control
Sw = square root of within mean square
n, = number of replicates for the control
n,- = number of replicates for concentration i
263
-------
13.3.7.5 Table 20 includes the calculated t values for each concentration and
control combination. In this example, comparing the 50.0 ppb concentration
with the control the calculation is as follows:
( 0.195 - 0.204 )
[ 0.032 / (1/8) + (1/8) ]
= -0.562
TABLE 20. CALCULATED T-VALUES
Concentration (ppb) i t{
50.0
100.0
210.0
2
3
4
-0.562
-0.125
3.938
13.3.7.6 Since the purpose of this test is to detect a significant reduction
in mean weight, a one-sided test is appropriate. The critical value for this
one-sided test is found in Table 5, Appendix C. For an overall alpha level of
0.05, 28 degrees of freedom for error and three concentrations (excluding the
control) the approximate critical value is 2.15. The mean weight for
concentration "i" is considered significantly less than the mean weight for
the control if t,- is greater than the critical value. Therefore, the 210.0
ppb concentration has significantly lower mean weight than the control. Hence
the NOEC and the LOEC for growth are 100.0 ppb and 210.0 ppb, respectively.
13.3.7.7 To quantify the sensitivity of the test, the minimum significant
difference (MSD) that can be detected statistically may be calculated.
MSD = d Su / (1/n,) + (1/n)
Where: d = the critical value for the Dunnett's procedure
Su = the square root of the within mean square
n = the common number of replicates at each concentration
(this assumes equal replication at each concentration)
n, = the number of replicates in the control.
14.3.7.8 In this example:
MSD = 2.15 (0.032) / (1/8) + (1/8)
= 2.15 (0.032)(0.5).
= 0.034
13.3.7.9 Therefore, for this set of data, the minimum difference that can be
detected as statistically significant is 0.034 mg.
264
-------
13.3.7.10 This represents a 17.4% reduction in mean weight from the control.
13.3.8 Calculation of the 1C
13.3.8.1 The growth data in Table 14 are utilized in this example. As can be
seen, the observed means are not monotonically nonincreasing with respect to
concentration. - Therefore, it .is necessary to smooth the means prior to
calculating the_IC. In the following discussion, the observed means are
represented by Y,- and the smoothed means by M^
13.3.8.2 Starting with the control mean, Y, = 0.195 and Y2 = 0.204, we see
that Y, < Y2. Calculate the smoothed means:
M, = M2 = (Y, + Y2)/2 = 0.200
13.3.8.3 Since Y5 = 0.091 < Y4 = 0.132 < Y, = 0.197 < M2, set M3 = 0.197 and
M, = 0.132, and M5 =0.091. Table 21 contains the smoothed means and Figure
11 gives a plot of the smoothed response curve.
TABLE 21. MYSID, MYSIDOPSIS BAHIA, MEAN GROWTH
RESPONSE AFTER SMOOTHING
Toxicant
Cone.
(ppb)
Control
50.0
100.0
210.0
450.0
i
1
2
3
4
5
Mi
(mg)
0.200
0.200
0.197
0.132
0.091
13.3.8.4 An IC25 and IC50 can be estimated using the Linear Interpolation
Method. A 25% reduction in mean weight, compared to the controls, would
result in a mean weight of 0.150 mg, where M^l-p/lOO) = 0.200(1-25/100). A
50% reduction in mean weight,' compared to the controls, would result in a mean
weight of 0.100 mg. Examining the means and their associated concentrations
(Table 21), the response, 0.150 mg, is bracketed by C, = 100 ppb and C4 = 210
ppb. The response, 0.100 mg, is bracketed by C4 = 210 ppb and C5J= 450 ppb.
13.3.8.5 Using Equation 1 from Appendix I, the estimate of the IC25 is
calculated as follows:
ICp = Cj + [M^l - p/100) - MJ (CjCJ)
265
-------
IC25 = 100 + [0.200(1 - 25/100) 0.19/1 (210 - 100)
(0.132 - 0.197)
= 179 ppb.
13.3.8.6 Using Equation 1 from Appendix I, the estimate of the IC50 is
calculated as follows:
ICp = Cj + [M^l - p/100) - MJ (CJ41 - C,)
(MJ+1 - MJ
IC50 = 210 + [0.200(1 - 50/100) 0.132] (450 - 210)
(0.091 - 0.132)
= 397 ppb.
13.3.8.7 When the Bootstrap program (BOOTSTRP) was used to analyze this set
of data, requesting 80 resamples, the mean estimate of the IC25 was 183.0960
ppb, with a standard deviation of 22.0142 ppb (coefficient of variation =
12.0%). The empirical 95.0% confidence interval for the true mean was
(149.2220 ppb, 239.4104 ppb). The BOOTSTRP computer program output for the
IC25 for this data set is shown in Figure 12.
13.3.8.8 When the Bootstrap program (BOOTSTRP) was used to analyze this set
of data for the IC50, requesting 80 resamples, the mean estimate of the IC50
was 336.9059 ppb, with a standard deviation of 51.5538 ppb (coefficient of
variation = 15.3%). The empirical 95.7% confidence interval for the true mean
was (259.7452 ppb, 424.8817 ppb). The BOOTSTRP computer program output for
the IC50 for this data set is shown in Figure 13.
13.4 EXAMPLE OF ANALYSIS OF MYSID, MYSIDOPSIS BAHIA, FECUNDITY DATA
13.4.1 Formal statistical analysis of the fecundity data is outlined in
Figure 14. The response used in the statistical analysis is the proportion of
females with eggs in each test or control chamber. If no females were present
in a replicate, a response of zero should not be used. Instead there are no
data available for that replicate and the number of replicates for that level
of concentration or the control should be reduced by one. Separate analyses
are performed for the estimation of the NOEC and LOEC endpoints, and for the
estimation of the IC25 and IC50 endpoints. The data for a concentration are
excluded from the statistical analysis of the NOEC and LOEC endpoints if no
eggs were produced in all of the replicates in which females existed.
However, all data are included in the estimation of the IC25 and IC50.
266
-------
* * *
ro
en
INDIVIDUAL REPLICATE MEAN WEIGHT
CONNECTS THE OBSERVED MEAN VALUE
CONNECTS THE SMOOTHED MEAN VALUE
50
100
TOXICANT CONCENTRATION (PPB)
210
450
Figure 11.
Plot of raw data, observed means, and smoothed means for the mysid, Mysidopsis
bahia, growth data from Tables 14 and 21.
-------
THE NUMBER OF RESAMPLES IS 80
*** LISTING OF GROUP CONCENTRATIONS (% EFF.) AND RESPONSE MEANS ***
CONC. (%EFF) RESPONSE MEAN MEAN AFTER POOLING
0.000 0.195 0.200
50.000 0.204 0.200
100.000 0.196 0.196
210.000 0.132 0.132
450.000 0.091 0.091
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
FROM THE INPUT SAMPLE IS 179.4003.
* BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY *
* OF THE ESTIMATED ICp *
THE MEAN OF THE BOOTSTRAP ESTIMATES IS 183.0960.
THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS 22.0142.
AN EMPIRICAL 95.0% CONFIDENCE INTERVAL FOR THE
BOOTSTRAP ESTIMATE IS (149.2220,239.4104).
Figure 12. BOOTSTRP program output for the IC25.
268
-------
THE NUMBER OF RESAMPLES IS 80
*** LISTING OF GROUP CONCENTRATIONS (X EFF.) AND RESPONSE MEANS ***
CONC. (XEFF) RESPONSE MEAN MEAN AFTER POOLING
0.000 0.195 0.200
50.000 0.204 0.200
100.000 0.196 0.196
210.000 0.132 0.132
450.000 0.091 0.091
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
FROM THE INPUT SAMPLE IS 396.5921.
BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY
OF THE ESTIMATED ICp
THE MEAN OF THE BOOTSTRAP ESTIMATES IS 336.9059.
THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS 51.5538.
AN EMPIRICAL 95.7% CONFIDENCE INTERVAL FOR THE
BOOTSTRAP ESTIMATE IS (259.7452,424.8817).
*** NOTE: THE ABOVE BOOTSTRAP CALCULATIONS WERE BASED ON 47
INSTEAD OF 80 RESAMPLINGS. THOSE RESAMPLES NOT
USED HAD ESTIMATES ABOVE THE HIGHEST CONCENTRATION / % EFF.
Figure 13. BOOTSTRP program output for the IC50.
269
-------
13.4.2 For the case of equal numbers of replicates across all concentrations
and the control, the evaluation of the NOEC and LOEC endpoints is made via a
parametric test, Dunnett's Procedure, or a nonparametric test, Steel's
Many-one Rank Test, on the arc sine transformed data. Underlying assumptions
of Dunnett's Procedure, normality and homogeneity of variance, are formally
tested. The test for normality is the Shapiro-Wilk's Test, and Bartlett's
Test is used to determine the homogeneity of variance. If either of these
tests fail, the nonparametric test, Steel's Many-one Rank Test, is used to
determine the NOEC and LOEC endpoints. If the assumptions of Dunnett's
Procedure are met, the endpoints are estimated by the parametric procedure.
13.4.3 If unequal numbers of replicates occur among the concentration levels
tested, there are parametric and nonparametric alternative analyses. The
parametric analysis is a t-test with the Bonferroni adjustment. The Wilcoxon
Rank Sum Test with the Bonferroni adjustment is the nonparametric alternative.
For detailed information on the Bonferroni adjustment see Appendix D.
13.4.4 In this example, the proportion of female mysids, Mysidopsis bahia,
with eggs in each replicate is first transformed by the arc sine
transformation procedure described in Appendix B. Since the denominator of
the proportion of females with eggs varies with the number of females
occurring in that replicate, the adjustment of the arc sine transformation for
0% and 100% is not used for this data. The raw and transformed data, means
and standard deviations of the transformed observations at each test
concentration including the control are listed in Table 22. Since there is
significant mortality in the 450 ppb concentration, its effect on reproduction
is not considered. Additionally, since no eggs were produced by females in
any of the replicates for the 210 ppb concentration, it is not included in
this statistical analysis and is considered a qualitative reproductive effect.
A plot of the mean proportion of female mysids with eggs is illustrated in
Figure 15.
270
-------
STATISTICAL ANALYSIS OF MYSIDOPSIS BAHIA
SURVIVAL, GROWTH, AND FECUNDITY TEST
FECUNDITY
FECUNDITY DATA
PROPORTION OF FEMALES WITH EGGS
POINT ESTIMATION
I
HYPOTHESIS TESTING
(EXCLUDING CONCENTRATIONS
ABOVE NOEC FOR SURVIVAL)
ENDPOINT ESTIMATE
IC25, IC50
I
ARC SINE TRANSFORMATION
I
SHAPIRO-WILK'S TEST
NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
NON-NORMAL DISTRIBUTION
BARTLETT'S TEST
EQUAL NUMBER OF
REPLICATES?
HETEROGENEOUS
VARIANCE
NO
EQUAL NUMBER OF
REPLICATES?
YES
T-TESTWITH
BONFERRONI
ADJUSTMENT
DUNNETTS
TEST
YES
STEEL'S MANY-ONE
RANK TEST
WILCOXON RANK SUM
TEST WITH
BONFERRONI ADJUSTMENT
ENDPOINT ESTIMATES
NOEC.LOEC
Figure 14. Flowchart for statistical analysis of mysid, Mysidopsis
bahia, fecundity data.
271
-------
TABLE 22. MYSID, MYSIDOPSIS BAHIA, FECUNDITY DATA: PERCENT FEMALES
WITH EGGS
Test Concentration (ODD)
Repl
RAW
ARC SINE
TRANS-
FORMED
Mean(Yi)
s?
i
icate
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
Control
1.00
1.00
0.67
1.00
1.00
0.80
1.00
1.00
1.57
1.57
0.96
1.57
1.57
1.12
1.57
1.57
1.44
0.064
1
50.0
0.50
0.33
0.67
-
0.40
0.50
0.25
0.33
0.78
0.61
0.96
-
0.68
0.78
0.52
0.61
0.71
0.021
2
100.0
0.33
0.50
0.00
0.50
0.67
0.00
0.25
-
0.61
0.78
0.00
0.78
0.96
0.00
0.52
-
0.52
0.147
3
210.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
-
-
-
-
-
-
-
-
4
272
-------
CONNECTS THE MEAN VALUE FOR EACH CONCENTRATION
ro
->j
CO
1.0
0.9-
0.8 »
{ft
80.7-
0.6
2 0.51
I 0.41
o
Q.
O
£ 0.3
0.2
0.1
o.o-
REPRESENTS THE CRITICAL VALUE FOR DUNNETT'S TEST
(ANY PROPORTION BELOW THIS VALUE WOULD BE
SIGNIFICANTLY DIFFERENT FROM THE CONTROL)
r
50
*
100
CONCENTRATION (PPB)
Figure 15. Proportion of females mysids, Mysidopsis bahia, with eggs.
-------
13.4,5 Test for Normality
13.4.5.1 The first step of the test for normality is to center the
observations by subtracting the mean of all observations within a
concentration from each observation in that concentration. The centered
observations are listed in Table 23.
TABLE 23. CENTERED OBSERVATIONS FOR SHAPIRO-MILK'S EXAMPLE
Test Concentration (oob)
Replicate
1
2
3
4
5
6
7
8
Control
0.13
0.13
-0.48
0.13
0.13
-0.32
0.13
0.13
50.0
0.07
-0.10
0.25
-
-0.03
0.07
-0.19
-0.10
100.0
0.09
0.26
-0.52
0.26
0.44
-0.52
0.00
-
13.4.5.2 Calculate the denominator, D, of the statistic:
D = 2 (X, - X)2
i = l
Where: X,- = the ith centered observation
X = the overall mean of the centered observations
n = the total number of centered observations
13.4.5.3 For this set of data: n = 22
X = _L_ (0.000) = 0.000
22
D = 1.4412
13.4.5.4 Order the centered observations from smallest to largest:
X<1) X<2) - ... X(n)
where XO) denotes the ith ordered observation. The ordered observations for
this example are listed in Table 24.
274
-------
TABLE 24. ORDERED CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S EXAMPLE
1
2
3
4
5
6
7
8
9
10
11
-0.52
-0.52
-0.48
-0.32
-0.19
-0.10
-0.10
0.03
0.00
0.07
0.07
12
13
14
15
16
17
18
19
20
21
22
0.09
0.13
0.13
0.13
0.13
0.13
0.13
0.25
0.26
0.26
0.44
13.4.5.5 From Table 4, Appendix B, for the number of observations, n, obtain
the coefficients a,, a2, ... ak where k is n/2 if n is even and (n-l)/2 if n
is odd. For the data in this example, n = 22 and k = 11. The ai values are
listed in Table 25.
13.4.5.6 Compute the test statistic, W, as follows:
W
1 [
D
a, (X(rv
The differences x
-------
TABLE 25. COEFFICIENTS AND DIFFERENCES FOR SHAPIRO-WILK'S EXAMPLE
/(n-i+1)
- X
(i)
1
2
3
4
5
6
7
8
9
10
11
0.4590
0.3156
0.2571
0.2131
0.1764
0.1443
0.1150
0.0878
0.0618
0.0368
0.0122
0.96
0.78
0.74
0.57
0.32
0.23
0.23
0.16
0.13
0.06
0.02
x<22>
X(21)
X(20>
XT
X(18)
XC17>
x< )
X
X( )
X(S)
X(12)
- x<1)
- x<2)
- x<3>
- x(4)
- x<5>
- x<6)
- x<7>
- x(8)
- x(9)
- x(10)
- x(11>
13.4.5.7 The decision rule for this test is to compare W as calculated in
13.4.5.6 to a critical value found in Table 6, Appendix B. If the computed W
is less than the critical value, conclude that the data are not normally
distributed. For this set of data, the critical value at a signficance level
of 0.01 and n = 22 observations is 0.878. Since W = 0.900 is greater than the
critical value, it is reasonable to assume that the data are normally
distributed.
13.4.6 Test for Homogeneity of Variance
13.4.6.1 The test used to examine whether the variation in proportion of
female mysids with eggs is the same across all concentration levels including
the control, is Bartlett's Test (Snedecor and Cochran, 1980). The test
statistic is as follows:
B =
P P
[ ( S V,.) In S2 - S V,- In S,2 ]
Where: V,- = degrees of freedom for each copper concen-
tration and control, V1 = (n,- - 1)
p = number of concentration levels including the control
276
-------
C - 1 + ( 3(p-l))-1 [ SP1/V,- - ( A,)'1 ]
i=l i=l
In = loge
i = 1, 2, ..., p where p is the number of concentrations
including the control
n( = the number of replicates for concentration i.
13.4.6.2 For the data in this example, (See Table 21) n, = 8, n2 = 7 and n3
7. Thus, the respective degrees of freedom are 7, 6 and 6.
13.4.6.3 Bartlett's statistic is therefore:
B = [(19)ln(0.077) - (7 ln(0.064) + 6 ln(0.021) + 6 ln(0.147))]/1.07
= [19(-2.564) - (-53.925)]/1.07
= [-48.716 - (-53.925)1/1.07
= 4.868
13.4.6.4 B is approximately distributed as chi-square with p - 1 degrees of
freedom, when the variances are in fact the same. Therefore, the appropriate
critical value for this test, at a significance level of 0.01 with two degrees
of freedom, is 9.210. Since B = 4.868 is less than the critical value of
9.210, conclude that the variances are not different.
13.4.7 T-test with the Bonferroni Adjustment
13.4.7.1 A t-test with the Bonferroni adjustment is used as an alternative to
Dunnett's Procedure when, as in this set of data, the number of replicates is
not the same for all concentrations. Like Dunnett's Procedure, it uses a
pooled estimate of the variance, which is equal to the error value calculated
in an analysis of variance. To obtain an estimate of the pooled variance,
construct an ANOVA table as described in Table 26.
277
-------
Table 26 ANOVA TABLE
Source
df
Sum of Squares
(SS)
Mean Square(MS)
(SS/df)
Between
Within
P 1
N - p
SSB
SSW
2
SB = SSB/(p-l)
2
Sw = SSW/(N-p)
Total
N - 1
SST
Where: p = number concentration levels including the control
N = total number of observations n, + n? ... +np
n; = number of observations in concentration i
P
SSB = X
G2/N
Between Sum of Squares
SST =
? ,
S Y2: - G2/N
:
Total Sum of Squares
SSW = SST - SSB
Within Sum of Squares
G = the grand total of all sample observations, G = 2 T,-
i=l
TJ = the total of the replicate measurements for
concentration "i"
13.4.7.2
n.
YJJ = the jth observation for concentration "i" (represents
the proportion of females with eggs for concentration
i in test chamber j)
For the data in this example:
,
N
T3 =
8 n2 = 7 n, = 7
22
Y
11 T '12
Y21 + Y22
Y 4- V
Y31 + '32
11.5
... + Y27 = 4.94
... + Y37= 3.65
G =
+ T2 + T3 + T4 = 20.09
278
-------
p - -
SSB = X T,-2/n, G2/N
i = l
= 132.25 + 24.40 + 13.32 - 403.61 = 3.57
877 22
P ni , ,
SST = S S Y2j - G2/N
= 23.396 403.61 = 5.05
22
SSW = SST - SSB = 5.05 - 3.57 = 1.48
S2 = SSB/(p-l) = 3.57/(3-l) = 1.785
Sj = SSW/(N-p) = 1.48/(22-3) = 0.078
13.4.7.3 Summarize these calculations in the ANOVA table (Table 27).
Table 27. ANOVA TABLE FOR THE T-TEST WITH BONFERRONI'S ADJUSTMENT EXAMPLE
Source df Sum of Squares Mean Square(MS)
(SS) (SS/df)
Between
Within
Total
2
19
21
3.57
1.48
5.05
1.785
0.078
13.4.7.4 To perform the individual comparisons, calculate the t statistic for
each concentration, and control combination as follows:
( Y, - Y- )
t; =
Sw / (1/n,) + (1/n,.)
Where: Y,. = mean proportion of females with eggs for concentration i
Y! = mean proportion of females with eggs for the control
Sw = square root of within mean square
n, = number of replicates for the control
n,- = number of replicates for concentration i.
279
-------
13.4.7.5 Table 28 includes the calculated t values for each concentration and
control combination. In this example, comparing the 50.0 ppb concentration
with the control the calculation is as follows:
( 1.44 0.52 )
[ 0.279 / (1/8) + (1/7) ]
= 5.05
TABLE 28. CALCULATED T-VALUES
Test Concentration (ppb) i t,-
50.0
100.0
2
3
5.05
6.37
13.4.7.6 Since the purpose of this test is to detect a significant reduction
in mean proportion of females with eggs, a one-sided test is appropriate. The
critical value for this one-sided test is found in Table 5, Appendix D,
Critical Values for the t-test with Bonferroni's adjustment. For an overall
alpha level of 0.05, 19 degrees of freedom for error and two concentrations
(excluding the control) the approximate critical value is 2.094. The mean
proportion for concentration "i" is considered significantly less than the
mean proportion for the control if t, is greater than the critical value.
Therefore, the 50.0 ppb and the 100.0 ppb concentrations have significantly
lower mean proportion of females with eggs than the control. Hence the LOEC
for fecundity is 50.0 ppb.
13.4.7.7 To quantify the sensitivity of the test, the minimum significant
difference (MSD) that can be detected statistically may be calculated.
MSD = t Sw / (1/n,) + (1/n)
Where: t - the critical value for the t-test with Bonferroni's adjustment
Su = the square root of the within mean square
n - the common number of replicates at each concentration
(this assumes equal replication at each concentration)
n1 = the number of replicates in the control.
13.4.7.8 In this example:
MSD = 2.094 (0.279) / (1/8) + (1/7)
= 2.094 (0.279)(0.518)
= 0.303
280
-------
13.4.7.9 Therefore, for this set of data, the minimum difference that can be
detected as statistically significant is 0.30.
13.4.7.10 The MSD (0.30) is in transformed units. To determine the MSD in
terms of percent of females with eggs, carry out the following conversion.
1. Subtract the MSD from the transformed control mean.
1.44 0.30 = 1.14
2. Obtain the untransformed values for the control mean and the
difference calculated in 4.10.1.
[Sine (1.44) ]* = 0.983
[Sine (1.14) ]2 = 0.823
3. The untransformed MSD (MSD ) is determined by subtracting the
untransformed values from 14.4.8.10.2.
MSDU = 0.983 - 0.823 = 0.16
13.4.7.11 Therefore, for this set of data, the minimum difference in mean
proportion of females with eggs between the control and any copper
concentration that can be detected as statistically significant is 0.16.
13.4.7.12 This represents a 17% decrease in proportion of females with eggs
from the control.
13.4.8 Calculation of the 1C
13.4.8.1 The fecundity data in Table 4 are utilized in this example. Table
29 contains the mean proportion of females with eggs for each toxicant
concentration. As can be seen, the observed means are monotonically
nonincreasing with respect to concentration. Therefore, it is not necessary
to smooth the means prior to calculating the 1C. Figure 16 gives a plot of
the response curve.
TABLE 29. MYSID, MYSIDOPSIS BAHIA, MEAN PROPORTION
OF FEMALES WITH EGGS
Toxicant
Cone.
(ppb)
Control
50.0
100.0
210.0
450.0
i
1
2
3
4
5
M,-
0.934
0.426
0.321
0.000
0.000
281
-------
13.4.8.2 An IC25 and IC50 can be estimated using the Linear Interpolation
Method. A 25% reduction in mean proportion of females with eggs, compared to
the controls, would result in a mean proportion of 0.701, where M^l-p/100) =
0.934(1-25/100). A 50% reduction in mean proportion of females with eggs,
compared to the controls, would result in a mean proportion of 0.467.
Examining the means and their associated concentrations (Table 29), the
response, 0.701, is bracketed by C, = 0 ppb and C2 = 50 ppb. The response,
0.467, is bracketed by C, = 0 ppb and C2 = 50 ppb.
13.4.8.3 Using Equation 1 from Appendix I, the estimate of the IC25 is
calculated as follows:
ICp = Cj + [M,(l - p/100) - MJ (CJ+1 - CJ
IC25 = 0 + [0.934(1 - 25/100) - 0.934] (50 - 0)
(0.426 - 0.934)
= 23 ppb.
13.4.8.4 Using Equation 1 from Appendix I, the estimate of the IC50 is
calculated as follows:
ICp = Cd + [M,(l - p/100) - MJ (CM - CJ
/u MA
* J-f 1 *" J /
IC50 = 0 + [0.934(1 - 50/100) - 0.934] (50 - 0)
(0.426 - 0.934)
= 46 ppb.
13.4.8.5 When the Bootstrap program (BOOTSTRP) was used to analyze this set
of data, requesting 80 resamples, the mean estimate of the IC25 was 23.3831
ppb, with a standard deviation of 2.3878 ppb (coefficient of variation =
10.2%). The empirical 95.0% confidence interval for the true mean was
(19.6188 ppb, 28.4061 ppb). The BOOTSTRP computer program output for the IC25
for this data set is shown in Figure 17.
13.4.8.6 When the Bootstrap program (BOOTSTRP) was used to analyze this set
of data for the IC50, requesting 80 resamples, the mean estimate of the IC50
was 47.6949 ppb, with a standard deviation of 7.1464 ppb (coefficient of
variation = 15.0%). The empirical 95.0% confidence interval for the true mean
was (39.2377 ppb, 69.0272 ppb). The BOOTSTRP computer program output for the
IC50 for this data set is shown in Figure 18.
282
-------
po
00
CO
1.2 -
1.1
1.0 H
o.g
0.8
0.7 :
fo 0.6 :
Q 0.5 H
§ °-4
£
Z 0.3 H
0.2 '-_
0.1 :
0.0
INDIVIDUAL PROPORTION OF FEMALES WITH EGGS
CONNECTS THE OBSERVED MEAN VALUE
50 100
TOXICANT CONCENTRATION (PPB)
210
450
Figure 16. A plot of the mean proportion of female mysids, Mysidopsis bahia, with eggs.
-------
THE NUMBER OF RESAMPLES IS 80
*** LISTING OF GROUP CONCENTRATIONS (% EFF = ) AND RESPONSE MEANS ***
CONC. (%EFF) RESPONSE MEAN MEAN AFTER POOLING
0.000 0.934 0.934
50.000 0.426 0.426
100.000 0.321 0.321
210.000 0.000 0.000
450.000 0.000 0.000
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
FROM THE INPUT SAMPLE IS 22.9745.
************************************************************
* BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY *
* OF THE ESTIMATED ICp *
THE MEAN OF THE BOOTSTRAP ESTIMATES IS 23.3831.
THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS 2.3878.
AN EMPIRICAL 95.0% CONFIDENCE INTERVAL FOR THE
BOOTSTRAP ESTIMATE IS (19.6188, 28.4061).
Figure 17. BOOTSTRP program output for the IC25.
284
-------
THE NUMBER OF RESAMPLES IS 80
*** LISTING OF GROUP CONCENTRATIONS (% EFF.) AND RESPONSE MEANS ***
CONC. (%EFF) RESPONSE MEAN MEAN AFTER POOLING
0.000 0.934 0.934
50.000 0.426 0.426
100.000 0.321 0.321
210.000 0.000 0.000
450.000 0.000 0.000
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
FROM THE INPUT SAMPLE IS 45.9490.
* BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY *
* OF THE ESTIMATED ICp *
************************************************************
THE MEAN OF THE BOOTSTRAP ESTIMATES IS 47.6949.
THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS 7.1464.
AN EMPIRICAL 95.0% CONFIDENCE INTERVAL FOR THE
BOOTSTRAP ESTIMATE IS (39.2377, 69.0272).
Figure 18. BOOTSTRP program output for the IC50,
285
-------
14. PRECISION AND ACCURACY
14.1 PRECISION
14.1.1 Data on the single-laboratory precision of the mysid survival, growth,
and fecundity using copper (CU) sulfate and sodium dodecyl sulfate (SDS) in
natural seawater and in artificial seawater (GP2) are shown in Tables 30-34.
Survival NOEC/LOEC pairs showed good precision, and were the same in four of
the six tests with CU and SDS. Growth and fecundity were generally not
acceptable endpoints in either sets of tests. In Tables 30-31 the coefficient
of variation for the IC25, ranges from 18.0 to 35.0 and the IC50, ranges from
5.8 to 47.8, indicating acceptable test precision. Data in Tables 32-34 show
no detectable differences between tests conducted in natural or artificial
seawaters.
14.1.2 The multilaboratory precision of the test has not yet been determined.
14.2 ACCURACY
14.2.1 The accuracy of toxicity tests cannot be determined.
286
-------
TABLE 30. SINGLE-LABORATORY PRECISION OF THE MYSID, MYSIDOPSIS BAHIA,
SURVIVAL, GROWTH, AND FECUNDITY TEST PERFORMED IN NATURAL SEAWATER,
USING JUVENILES FROM MYSIDS CULTURED AND SPAWNED IN NATURAL
SEAWATER, AND COPPER (CU) SULFATE AS A REFERENCE TOXICANT1'2'3'4'5'6
Test
Number
NOEC
(MA)
IC25
(H9/L)
IC50
(M/L)
Most
Sensitive
Endpoint
1
2
3
4
5
63
125
125
125
125
96,
138,
156,
143,
,5
,5
157.7
175,
187,
179.9
200.3
S
S
S
S
S
n:
Mean:
5
NA
NA
5
138.3
18.0
4
185.8
5.8
]Data from USEPA (1989a) and USEPA (1991a).
2Tests performed by Randy Cameleo, Environmental Research Laboratory,
U. S. Environmental Protection Agency, Narragansett, Rhode Island.
3Eight replicate exposure chambers, each with five juveniles, were
used for the control and each toxicant concentration. The temperature
of the test solutions was maintained at 26 + 1°C.
ACopper concentrations in Tests 1-2 were: 8, 16, 31, 63, and 125 ng/L.
Copper concentrations in Tests 3-6 were, 16, 31, 63, 125, and 250 ng/L.
5NOEC Range: 63 - 125 ng/L (this represents a difference of two exposure
concentrations).
6For a discussion of the precision of data from chronic toxicity
tests see Section 4, Quality Assurance.
*No linear interpolation estimate could be calculated from the data, since
none of the group response means were less than 50 percent of the control
concentrations.
287
-------
TABLE 31. SINGLE-LABORATORY PRECISION OF THE MYSID, MYSIDOPSIS BAHIA,
SURVIVAL, GROWTH, AND FECUNDITY TEST PERFORMED IN NATURAL
SEAWATER USING JUVENILES FROM MYSIDS CULTURED AND SPAWNED IN
NATURAL SEAWATER, AND SODIUM DODECYL SULFATE (SDS) AS A
REFERENCE TOXICANT1'2'3'4'5'6
Test NOEC
Number (mg/L)
1 2.5
2 <0.3
3 <0.6
4 5.0
5 2.5
6 5.0
n: 4
Mean: NA
CV(%): NA
IC25
(mg/L)
4.5
NC7
NC7
7.8
3.6
7.0
4
5.7
35.0
IC50
(mg/L)
NC8
NC8
NC8
NC8
4.6
9.3
2
6.9
47.8
Most
Sensitive
Endpoint
S
S
S
S
S
S
1Data from USEPA (1989a) and USEPA (1991a).
2Tests performed by Randy Cameleo, Environmental Research Laboratory,
U. S. Environmental Protection Agency, Narragansett, Rhode Island.
3Eight replicate exposure chambers, each with five juveniles, were
used for the control and each toxicant concentration. The temperature
of the test solutions was maintained at 26 + 1°C.
4SDS concentrations in Tests 1-2 were: 0.3, 0.6, 1.3, 2.5, and
5.0 mg/L. SDS concentrations in Tests 3-4 were: 0.6, 1.3, 2.5, 5.0 and
10.0 mg/L. SDS concentrations in Tests 5-6 were: 1.3, 2.5, 5.0, 10.0,
and 20.0 mg/L.
5NOEC Range: <0.3 5.0 mg/L (this represents a difference of four exposure
concentrations).
6For a discussion of the precision of data from chronic toxicity
tests see Section 4, Quality Assurance.
NC7 = No linear interpolation estimate could be calculated from the data,
since none of the group response means were less than 75 percent of the
control response mean.
NC8 = No linear interpolation estimate could be calculated from the data,
since none of the group response means were less than 50 percent of the
control response mean.
288
-------
TABLE 32. COMPARISON OF SURVIVAL (LC50)a, GROWTH AND FECUNDITY (IC50)a
RESULTS FROM 7-DAY TESTS WITH THE MYSID, MYSIDOPSIS BAHIA, USING
NATURAL SEAWATER (NSW) AND ARTIFICIAL SEAWATER (GP2) AS DILUTION
WATER AND SODIUM DODECYL SULFATE (SDS) AS A REFERENCE TOXICANT.
Survival LC5Q Growth IC50 Fecundity IC50
Test
1
2
3
NSW GP2
16.2 16.3
20.5 19.2
21.9
NSW GP2 NSW
16.8 16.3 12.0
24.2 23.3 20.1
24.4
GP2
10.9
18.5
21.7
aAll LC50/IC50 values in mg/L.
No test performed.
TABLE 33. COMPARISON OF SURVIVAL (LC50)a, GROWTH AND FECUNDITY (IC50)a
RESULTS FROM 7-DAY TESTS WITH THE MYSID, MYSIDOPSIS BAHIA, USING
NATURAL SEAWATER (NSW) AND ARTIFICIAL SEAWATER (GP2) AS DILUTION
WATER AND COPPER (CU) SULFATE AS A REFERENCE TOXICANT.
Survival LC50 Growth IC50 Fecundity IC50
Test
1
2
3
NSW
177
-
190
GP2
182
173
174
NSW
208
-
195
GP2
186
210
179
NSW
177
-
168
GP2
125
142
186
aAll LC50/IC50 values in jig/L
- No test performed.
289
-------
Table 34. CONTROL RESULTS FROM 7-DAY SURVIVAL, GROWTH, AND FECUNDITY TESTS
WITH THE MYSID, MYSIDOPSIS BAHIA, USING NATURAL SEAWATER AND
ARTIFICIAL SEAWATER (GP2) AS A DILUTION WATER.
Control8
Survival
Test NSW
1 98
2 80
3
4 94
5
6 80
(%)
GP2
93
90
95
84
94
75
Growth (mq)
NSW GP2
0.32 0.32
0.40 0.43
0.40
0.34 0.37
0.36
0.40 0.41
Fecundity 1
NSW
73
100
89
79
^ /Of
GP2
77
95
100
83
83
93
"Survival as percent of mysids alive after 7 days; growth as mean individual
dry weight; fecundity as percent females with eggs.
- No test performed.
290
-------
Figure 19. Data sheet for the mysid, Mysidopsis bahia, water quality
measurements. From USEPA (1987f).
TEST:
START DATE:_
SALINITY:
DAY 1
DAY 2
DAYS
DAY 4
DAYS
DAY 6
DAY 7
DAY 1
DAY 2
DAY 3
DAY 4
DAYS
DAY 6
DAY 7
TRTMT
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
TRTMT
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
TEMP
TEMP
SALINITY
SALINITY
DO
DO
pH
pH
TRTMT
TRTMT
TEMP
TEMP
SALINITY
SALINITY
DO
DO
pH
PH
291
-------
TEST:
Fiqure 20. Data sheet for the mysid, Mysidopsis bahia, survival
and fecundity data. From USEPA (1987f).
START DATE:
SALINITY:
TREATMENT/
REPLICATE
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
DAY 1
# ALIVE
DAY 2
» ALIVE
DAY 3
* ALIVE
DAY 4
* ALIVE
DAYS
* ALIVE
DAY 6
t ALIVE
DAY 7
t ALIVE
FEMALES
W/EGGS
FEMALES
NO EGGS
MALES
IMMATURES
292
-------
TEST:
Figure 20. Data sheet for the mysid, Mysidopsis bahia, survival
and fecundity data (Continued). From USEPA (1987f).
START DATE:
SALINITY:
TREATMENT/
REPLICATE
1
2
3
4
5
6
7
8
1
2
3
4
4,
5
6
7
8
1
2
3
4
C
5
6
7
8
'DAY 1
* ALIVE
DAY 2
# ALIVE
DAY 3
» ALIVE
DAY 4
t ALIVE
DAYS
* ALIVE
DAY 6
* ALIVE
DAY 7
# ALIVE
FEMALES
W/EGGS
FEMALES
NO EGGS
MALES
IMMATURES
293
-------
Figure 21. Data sheet for the mysid, Mysidopsis bahia, dry weight
measurements. From USEPA (1987f).
TEST:
START DATE:_
SALINITY:
TREATMENT
REPLICATE
1
2
3
4
5
6
7
8
1
2
3
4
1 5
6
7
8
1
2
3
4
" 5
6
7
8
PAN*
TARE
WT.
TOTAL
wrr.
ANIMAL
WT.
#OF
ANIMALS
XWT./
ANIMAL
294
-------
Figure 21. Data sheet for the mysid, Mysidopsis bahia, dry weight
measurements (Continued). From USEPA (1987f).
TEST:
START DATE:_
SALINITY:
TREATMENT
REPLICATE
1
2
3
4
6
7
8
1
2
3
4
5
6
7
8
1
2
3
4
6
7
8
PAN*
TARE
WT.
TOTAL
WT.
ANIMAL
WT.
f OF
ANIMALS
X WT./
ANIMAL
295
-------
SECTION 15
TEST METHOD1'2
SEA URCHIN, ARBACIA PUNCTULATA, FERTILIZATION TEST
METHOD 1008
1. SCOPE AND APPLICATION
1.1 This method measures the toxicity of effluents and receiving water to the
gametes of the sea urchin, Arbacia punctulata, during a 1 h and 20 min
exposure. The purpose of the sperm cell toxicity test is to determine the
concentration of a test substance that reduces fertilization of exposed
gametes relative to that of the control.
1.2 Detection limits of the toxicity of an effluent or chemical substance are
organism dependent.
1.3 Single or multiple excursions in toxicity may not be detected using 24-h
composite samples. Also, because of the long sample collection period
involved in composite sampling and because the test chambers are not sealed,
highly volatile and highly degradable toxicants in the source may not be
detected in the test.
1.4 This test is commonly used in one of two forms: (1) a definitive test,
consisting of a minimum of five effluent concentrations and a control, and (2)
a receiving water test(s), consisting of one or more receiving water
concentrations and a control.
2. SUMMARY OF METHOD
2.1 The method consists of exposing dilute sperm suspensions to effluents or
receiving waters for one hour. Eggs are then added to the sperm suspensions.
Twenty minutes after the eggs are added, the test is terminated by the
addition of preservative. The percent fertilization is determined by
microscopic examination of an aliquot from each treatment. The test results
are reported as the concentration of the test substance which causes a
statistically significant reduction in fertilization, compared to the
controls.
3. INTERFERENCES
3.1 Toxic substances may be introduced by contaminants in dilution water,
glassware, sample hardware, and testing equipment (see Section 5, Facilities,
Equipment, and Supplies).
3.2 Improper effluent sampling and handling may adversely affect test results
1The format used for this method was taken from USEPA (1983).
2This method was adapted from USEPA (1987g).
296
-------
(see Section 8, Effluent and Receiving Water Sampling, Sample Handling, and
Sample Preparation for Toxicity Tests).
4. SAFETY
4.1 See Section 3, Health and Safety.
5. APPARATUS AND EQUIPMENT
5.1 Facilities for holding and acclimating test organisms.
5.2 Laboratory sea urchins, Arbacia punctulata, culture unit -- See
Subsection 6.17, culturing methods below. To test effluent or receiving water
toxicity, sufficient eggs and sperm must be available.
5.3 Samplers -- automatic sampler, preferably with sample cooling capability,
that can collect a 24-h composite sample of 1 L.
5.4 Environmental chamber or equivalent facility with temperature control
(20 + 1°C) for controlling temperature during exposure.
5.5 Water purification system -- Millipore Milli-QR, Deionized water (DI) or
equivalent.
5.6 Balance -- Analytical, capable of accurately weighing to 0.0001 g.
5.7 Reference weights, Class S -- for checking performance of balance.
5.8 Air pump -- for supplying air.
5.9 Air lines, and air stones -- for aerating water containing adults.
5.10 Vacuum suction device -- for washing eggs.
5.11 pH and DO meters -- for routine physical and chemical measurements.
Unless the test is being conducted to specifically measure the effect of one
of these two parameters, portable, field-grade instruments are acceptable.
5.12 Standard or micro-Winkler apparatus -- for determining DO (optional).
5.13 Transformer, 10-12 Volt, with steel electrodes -- for stimulating
release of eggs and sperm.
5.14 Centrifuge, bench-top, slant-head, variable speed -- for washing eggs.
5.15 Fume hood -- to protect the analyst from formaldehyde fumes.
5.16 Dissecting microscope -- for counting diluted egg stock.
5.17 Compound microscope -- for examining and counting sperm cells and
fertilized eggs.
297
-------
5.18 Sedgwick-Rafter counting chamber -- for counting egg stock and examining
fertilized eggs.
5.19 Hemacytometer, Neubauer -- for counting sperm.
5.20 Count register, 2-place -- for recording sperm and egg counts.
5.21 Refractometer -- for determining salinity.
5.22 Thermometers, glass or electronic, laboratory grade -- for measuring
water temperatures.
5.23 Thermometers, bulb-thermograph or electronic-chart type -- for
continuously recording temperature.
5.24 Thermometer, National Bureau of Standards Certified (see USEPA METHOD
170.1, USEPA, 1979b) -- to calibrate laboratory thermometers.
5.25 Ice bucket, covered -- for maintaining live sperm.
5.26 Centrifuge tubes, conical -- for washing eggs.
5.27 Cylindrical glass vessel, 8-cm diameter -- for maintaining dispersed egg
suspension.
5.28 Beakers -- six Class A, borosilicate glass or non-toxic plasticware,
1000 mL for making test solutions.
5.29 Glass dishes, flat bottomed, 20-cm diameter -- for holding urchins
during gamete collection.
5.30 Wash bottles -- for deionized water, for rinsing small glassware and
instrument electrodes and probes.
5.31 Volumetric flasks and graduated cylinders -- Class A, borosilicate glass
or non-toxic plastic labware, 10-1000 ml for making test solutions.
5.32 Syringes, 1-mL, and 10-mL, with 18 gauge, blunt-tipped needles (tips cut
off) -- for collecting sperm and eggs.
5.33 Pipets, volumetric -- Class A, 1-100 ml.
5.34 Pipets, automatic -- adjustable, 1-100 ml.
5.35 Pipets, serological -- 1-10 ml, graduated.
5.36 Pipet bulbs and fillers -- PROPIPET", or equivalent.
5.37 Tape, colored -- for labelling tubes.
5.38 Markers, water-proof -- for marking containers, etc.
298
-------
6. REAGENTS AND CONSUMABLE MATERIALS
6.1 Sea Urchins, Arbacia punctulata, (minimum 12 of each sex).
6.2 Food -- kelp, Laminan'a sp., or romaine lettuce for the sea urchin,
Arbacia punctulata.
6.3 Standard salt water aquarium or Instant Ocean Aquarium (capable of
maintaining sea water at 15°C), with appropriate filtration and aeration
system.
6.4 Sample containers -- for sample shipment and storage (see Section 8,
Effluent and Receiving Water Sampling, Sample Handling, and Sample Preparation
for Toxicity Tests).
6.5 Scintillation vials, 20 ml, disposable -- to prepare test concentrations.
6.6 Parafilm -- to cover tubes and vessels containing test materials.
6.7 Gloves, disposable; labcoat and protective eyewear -- for personal
protection from contamination.
6.8 Data sheets (one set per test) -- for data recording (see Figures 4, 5,
and 6).
6.9 Acetic acid, 10%, reagent grade, in sea water -- for preparing killed
sperm dilutions.
6.10 Formalin, 1%, in 2 ml of seawater -- for preserving eggs (see Subsection
10.7 Termination of the Test).
6.11 pH buffers 4, 7, and 10 (or as per instructions of instrument
manufacturer) for standards and calibration check (see USEPA Method 150.1,
USEPA, 1979b).
6.12 Membranes and filling solutions for dissolved oxygen probe (see USEPA
Method 360.1, USEPA, 1979b), or reagents for modified Winkler analysis.
6.13 Laboratory quality assurance samples and standards for the above
methods.
6.14 Reference toxicant solutions (see Section 4, Quality Assurance,
Subsections 4.7, 4.14, 4.15, 4.16, and 4.17).
6.15 Reagent water -- defined as distilled or deionized water that does not
contain substances which are toxic to the test organisms.
6.16 Effluent, receiving water, and dilution water -- see Section 7, Dilution
Water, and Section 8, Effluent and Receiving Water Sampling, Sample Handling,
and Sample Preparation for Toxicity Tests.
6.16.1 Saline test and dilution water -- The salinity of the test water must
299
-------
be 30 °/oo. The salinity should vary by no more than + 2 °/oo among the
replicates.
6.16.2 The overwhelming majority of industrial and sewage treatment effluents
entering marine and estuarine systems contain little or no measurable salts.
Exposure of sea urchin eggs and sperm to these effluents will require
adjustments in the salinity of the test solutions. It is important to
maintain a constant salinity across all treatments. Two methods are available
to adjust salinities--hypersaline brine derived from natural seawater or
artificial sea salts. Use of hypersaline brine will limit the maximum
concentration of effluent tested to 70%.
6.16.3 Hypersaline brine: Hypersaline brine (HSB) has several advantages
that make it desirable for use in toxicity testing. It can be made from any
high quality, filtered seawater by evaporation, and can be added to the
effluent or to deionized water to increase the salinity. HSB derived from
natural seawater contains the necessary trace metals, biogenic colloids, and
some of the microbial components necessary for adequate growth, survival,
and/or reproduction of marine and estuarine organisms, and may be stored for
prolonged periods without any apparent degradation.
6.16.3.1 The ideal container for making HSB from natural seawater is one that
(1) has a high surface to volume ratio, (2) is made of a noncorrosive
material, and (3) is easily cleaned (fiberglass containers are ideal).
Special care should be used to prevent any toxic materials from coming in
contact with the seawater being used to generate the brine. If a heater is
immersed directly into the seawater, ensure that the heater materials do not
corrode or leach any substances that would contaminate the brine. One
successful method used is a thermostatically controlled heat exchanger made
from fiberglass. If aeration is utilized, use only oil-free air compressors
to prevent contamination.
6.16.3.2 Before adding seawater to the brine generator, thoroughly clean the
generator, aeration supply tube, heater, and any other materials that will be
in direct contact with the brine. A good quality biodegradable detergent
should be used, followed by several thorough deionized water rinses. High
quality (and preferably high salinity) seawater should be filtered to at least
10 jim before placing into the brine generator. Water should be collected on
an incoming tide to minimize the possibility of contamination.
6.16.3.3 The temperature of the seawater is increased slowly to 40°C. The
water should be aerated to prevent temperature stratification and to increase
water evaporation. The brine should be checked daily (depending on the volume
being generated) to ensure that the salinity does not exceed 100 °/oo and that
the temperature does not exceed 40°C. Additional seawater may be added to the
brine to obtain the volume of brine required.
6.16.3.4 After the required salinity is attained, the HSB should be filtered
a second time through a 1-^m filter and poured directly into portable
containers (20-L cubitainers or polycarbonate water cooler jugs are suitable).
The containers should be capped and labelled with the date the brine was
generated and its salinity. Containers of HSB should be stored in the dark
300
-------
and maintained under room temperature until used.
6.16.3.5 If a source of MSB is available, test solutions can be made by
following the directions below. Thoroughly mix together the deionized water
and brine before mixing in the effluent.
6.16.3.6 Divide the salinity of the MSB by the expected test salinity to
determine the proportion of deionized water to brine. For example, if the
salinity of the brine is 100 °/oo and the test is to be conducted at 30 °/oo,
100 °/oo divided by 30 °/oo = 3.3. The proportion of brine is 1 part in 3.3
(one part brine to 2.3 parts deionized water).
6.16.3.7 To make 1 L of seawater at 30 °/°° salinity from a hypersaline brine
of 100 °/oo, 300 mL of brine and 700 ml of deionized water are required.
6.16.3.8 Table 1 illustrates the preparation of test solutions at 30 °/oo if
they are made by combining effluent (0 °/oo), deionized water and HSB (100
°/oo), or FORTY FATHOMS" sea salts.
6.16.4 Artificial sea salts: FORTY FATHOMS" brand sea salts have been used
successfully at the EMSL-Cincinnati, Newtown Facility for long-term (6-12
months) maintenance of stock cultures of sexually mature sea urchins and to
perform the sea urchin fertilization test. GP2 seawater formulation (Table 2)
has also been used successfully at the Environmental Research Laboratory, U.S.
Environmental Protection Agency, Narragansett, Rhode Island.
6.17 SEA URCHINS
6.17.1 Adult sea urchins, Arbacia punctulata, can be obtained from commercial
suppliers. After acquisition, the animals are sexed by briefly stimulating them
with current from a 12 V transformer. Electrical stimulation causes the
immediate release of masses of gametes that are readily identifiable by color --
the eggs are red, and the sperm are white.
6.17.2 The sexes are separated and maintained in 20-L, aerated fiberglass tanks,
each holding about 20 adults. The tanks are supplied continuously (approximately
5 L/min) with filtered natural seawater, or salt water prepared from commercial
sea salts is recirculated. The animals are checked daily and any obviously
unhealthy animals are discarded.
6.17.3 The culture unit should be maintained at 15°C ± 3°C, with a water
temperature control device.
6.17.4 The food consists of kelp, Laminaria sp., gathered from known
uncontaminated zones or obtained from commercial supply houses whose kelp comes
from known uncontaminated areas, or romaine lettuce. Fresh food is introduced
into the tanks at approximately one week intervals. Decaying food is removed as
necessary. Ample supplies of food should always be available to the sea urchins.
6.17.5 Natural or artificial seawater with a salinity of 30 °/oo is used to
maintain the adult animals, for all washing and dilution steps, and as the
control water in the tests (see Subsection 6.16).
301
-------
TABLE 1. PREPARATION OF TEST SOLUTIONS AT A SALINITY OF 30 °/oo
USING NATURAL SEAWATER, HYPERSALINE BRINE, OR ARTIFICIAL
SEA SALTS
Solutions To Be Combined
Effluent
Solution
1
2
3
4
5
Control
Effluent
Cone.
(%)
1001
50
25
12.5
6.25
0.0
Volume of Volume of Diluent
Effluent Seawater (30 °/oo)
Solution
840 mL
420 mL Solution 1
420 mL Solution 2
420 mL Solution 3
420 mL Solution 4
+ 420 mL
+ 420 mL
+ 420 mL
+ 420 mL
420 mL
Total
2080 mL
1This illustration assumes: (1) the use of 5 mL of test
solution in each of four replicates (total of 20 mL) for the control
and five concentrations of effluent, (2) an effluent dilution factor
of 0.5, (3) the effluent lacks appreciable salinity, and (4) 400 mL of
each test concentration is used for chemical analysis. A sufficient
initial volume (840 mL) of effluent is prepared by adjusting the
salinity to 30 /oo. In this example, the salinity is adjusted by
adding artificial sea salts to the 100% effluent, and preparing a
serial dilution using 30 °/oo seawater (natural seawater,
hypersaline brine, or artificial seawater). Stir solutions 1 h to
assure that the salts dissolve. The salinity of the initial 840 mL of
100% effluent is adjusted to 30 °/oo by adding 25.2 g of dry
artificial sea salts (FORTY FATHOMS"). Test concentrations are then
made by mixing appropriate volumes of salinity adjusted effluent and
30 °/oo salinity dilution water to provide 840 mL of solution for
each concentration. If hypersaline brine alone (100 °/oo) is used
to adjust the salinity of the effluent, the highest concentration of
effluent that could be tested would be 70% at 30 °/oo salinity.
302
-------
TABLE 2. REAGENT GRADE CHEMICALS USED IN THE PREPARATION OF GP2 ARTIFICIAL
SEAWATER FOR THE SEA URCHIN, ARBACIA PUNCTULATA, TOXICITY TEST1'2'3
Compound Concentration Amount (g)
(g/L) Required for
20 L
1.
2.
3.
4.
5.
6.
7.
8.
9.
NaCl
Na2S04
KC1
KBr
Na2B407 . 10 H20
MgCl2 . 6 H20
CaCl2 . 2 H20
SrCl2 . 6 H20
NaHCO,
21.03
3.52
0.61
0.088
0.034
9.50
1.32
0.02
0.17
420.6
70.4
12.2
1.76
0.68
190.0
26.4
0.400
3.40
Modified GP2 from Spotte et al. (1984).
2The constituent salts and concentrations were taken from
USEPA (1990b). The salinity is 30.89 g/L.
3GP2 can be diluted with deionized (DI) water to the desired test salinity.
6.17.6 Adult male and female animals used in field studies are transported in
separate or partitioned insulated boxes or coolers packed with wet kelp or paper
toweling. Upon arrival at the field site, aquaria (or a single partitioned
aquarium) are filled with control water, loosely covered with a styrofoam sheet
and allowed to equilibrate to 15°C before animals are added. Healthy animals
will attach to the kelp or aquarium within hours.
6.17.7 To successfully maintain about 25 adult animals for seven days at a field
site, a screen-partitioned, 40-L glass aquarium using aerated, recirculating,
clean saline water (30 °/oo) and a gravel bed filtration system, is housed within
a water bath, such as FORTY FATHOMS11 or INSTANT OCEANR Aquarium (15°C). The
inner aquarium is used to avoid contact of animals and water bath with cooling
coils.
7. SAMPLE COLLECTION, PRESERVATION AND HANDLING
303
-------
7.1 See Section 8, Effluent and Receiving Water Sampling, Sample Handling, and
Sampling Preparation for Toxicity Tests.
8. CALIBRATION AND STANDARDIZATION
8.1 See Section 4, Quality Assurance.
9. QUALITY CONTROL
9.1 See Section 4, Quality Assurance.
10. TEST PROCEDURES
10.1 TEST SOLUTIONS
10.1.1 Receiving Waters
10.1.1.1 The sampling point is determined by the objectives of the test. At
estuarine and marine sites, samples are usually collected at mid-depth.
Receiving water toxicity is determined with samples used directly as collected or
with samples passed through a 60 jim NITEXR filter and compared without dilution
against a control. Using four replicate chambers per test, each containing 5 ml,
and 400 ml for chemical analysis, would require approximately 420 ml or more of
sample per test.
10.1.2 Effluents
10.1.2.1 The selection of the effluent test concentrations should be based on
the objectives of the study. A dilution factor of 0.5 is commonly used. A
dilution factor of 0.5 provides precision of ± 100%, and testing of
concentrations between 6.25% and 100% effluent using only five effluent
concentrations (6.25%, 12.5%, 25%, 50%. and 100%). Test precision shows little
improvement as dilution factors are increased beyond 0.5 and declines rapidly if
smaller dilution factors are used. Therefore, USEPA recommends a dilution factor
of 0.5.
10.1.2.2 If the effluent is known or suspected to be highly toxic, a lower range
of effluent concentrations should be used (such as 25%, 12.5%, 6.25%, 3.12%, and
1.56%).
10.1.2.3 The test should begin as soon as possible, preferably within 24 h of
sample collection. The maximum holding time following retrieval of the sample
from the sampling device should not exceed 36 h for off-site toxicity tests
unless permission is granted by the permitting authority. In no case should the
sample be used in a test more than 72 h after sample collection.
10.1.2.4 Just prior to test initiation (approximately one h), a sufficient
quantity of the sample to make the test solutions should be adjusted to the test
temperature (20 ± 1 C) and maintained at that temperature during the addition of
dilution water.
304
-------
10.1.3 Control Water
10.1.3.1 Prepare 3 L of control water at 30 °/oo using hypersaline brine or
artificial sea salts (see Table 1). This water is used in all washing and
diluting steps and as control water in the test. Natural sea water and local
waters may be used as additional controls.
10.1.4 Effluent Dilutions
10.1.4.1 Effluent/receiving water samples are adjusted to salinity of 30 °/oo.
10.1.4.2 Four replicates (minimum of three) are prepared for each test
concentration, using 5 ml of solution in disposable liquid scintillation vials.
A 50% (0.5) concentration series can be prepared by serially diluting test
concentrations with control water. Sufficient test solution is prepared at each
effluent concentration to provide additional volume for chemical analyses, at the
high, medium, and low test concentrations.
10.1.4.3 All test samples are equilibrated at 20°C + 1°C before addition of
sperm.
10.2 REFERENCE TOXICANT TEST
10.2.1 A reference toxicant test using copper sulfate, SDS (sodium dodecyl
sulfate), or equivalent is performed on a monthly basis with the fertilization
test (see Section 4, Quality Assurance, Subsections 4.14, 4.16, and 4.17).
10.3 COLLECTION OF GAMETES FOR THE TEST
10.3.1 Select four females and place in shallow bowls, barely covering the
shell with seawater. Stimulate the release of eggs by touching the shell with
steel electrodes connected to a 10-12 volt transformer (about 30 seconds each
time). Collect the eggs from each female using a 10 mL disposable syringe
fitted with an 18-gauge, blunt-tipped needle (tip cut off). Remove the needle
from the syringe before adding the eggs to a conical centrifuge tube. Pool
the eggs. The egg stock may be held at room temperature for several hours
before use. Note: Eggs should be collected first to eliminate possibility of
pre-fertilization.
10.3.2 Select four males and place in shallow bowls, barely covering the
animals with seawater. Stimulate the release of sperm as described above.
Collect the sperm (about 0.25 mL) from each male, using a 1-3 mL disposable
syringe fitted with an 18-gauge, blunt-tipped needle. Pool the sperm.
Maintain the pooled sperm sample on ice. The sperm must be used in a toxicity
test within 1 h of collection.
10.4 PREPARATION OF SPERM DILUTION FOR USE IN THE TEST
10.4.1 Using control water, dilute the pooled sperm sample to a concentration
of about 5 X 107
described below:
of about 5 X 107 sperm/mL (SPM). Estimate the sperm concentration as
305
-------
1. Make a sperm dilutions of 1:50, 1:100, 1:200, and 1:400, using 30'°/oo
seawater, as follows:
a. Add 400 uL of collected sperm to 20 ml of sea water in Vial A.
Mix by gentle pipetting using a 5-mL pipetter, or by inversion;
b. Add 10 ml of sperm suspension from Vial A to 10 ml of seawater
in Vial B. Mix by gentle pipetting using a 5-mL pipetter, or
by inversion;
c. Add 10 ml of sperm suspension from Vial B to 10 ml of seawater
in Vial C. Mix by gentle pipetting using a 5-mL pipetter, or
by inversion;
d. Add 10 mL of sperm suspension from Vial C to 10 mL of seawater
in Vial D. Mix by gentle pipetting using a 5-mL pipetter, or by
inversion;
e. Discard 10 mL from Vial D. (The volume of all suspensions is
10 mL).
2. Make a 1:2000 killed sperm suspension and determine the SPM.
a. Add 10 mL 10% acetic acid in seawater to Vial C. Cap Vial C and
mix by inversion.
b. Add 1 mL of killed sperm from Vial C to 4 mL of seawater in
Vial E. Mix by gentle pipetting with a 4-mL pipetter.
c. Add sperm from Vial E to both sides of the Neubauer
hemacytometer. Let the sperm settle 15 min.
d. Count the number of sperm in the central 400 squares on both
sides of the hemacytometer using a compound microscope (100X).
Average the counts from the two sides.
e. SPM in Vial E = 104 x average count.
3. Calculate the SPM in all other suspensions using the SPM in Vial E
above:
SPM in Vial A = 40 x SPM in Vial E
SPM in Vial B = 20 x SPM in Vial E
SPM in Vial D = 5 x SPM in Vial E
SPM in original sperm sample = 2000 x SPM in Vial E
4. Dilute the sperm suspension with a SPM greater than 5 x 107 SPM to 5
107 SPM.
Actual SPM/(5 x 107) = dilution factor (DF)
[(DF) x 10] - 10 = mL of seawater to add to vial.
5. Confirm the sperm count by sampling from the test stock. Add 0.1 mL
of test stock to 9.9 mL of 10% acetic acid in seawater, and count
with the hemacytometer. The count should average 50 + 5.
10.5 PREPARATION OF EGG SUSPENSION FOR USE IN THE TEST Note: The egg
suspension may be prepared during the 1-h sperm exposure.
306
-------
10,5.1 Wash the pooled eggs three times using control water with gentle
centrifugation (500xg for 3 min using a tabletop centrifuge). If the wash
water becomes red, the eggs have lysed and must be discarded.
10.5.2 Dilute the egg stock, using control water, to about 2000 eggs/ml.
1. Transfer the eggs to a glass beaker containing 200 mL of control water
("egg stock").
2. Mix the egg stock using an air-bubbling device. Using a wide-mouth
pi pet tip, transfer 1 mi of eggs from the egg stock to a vial
containing 9 ml of control water. (This vial contains an egg
suspension diluted 1:10 from egg stock).
3. Mix the contents of the vial by inversion. Using a wide-mouth pipet
tip, transfer 1 ml of eggs from the vial to a Sedgwick-Rafter
counting chamber. Count all eggs in the chamber using a dissecting
microscope at 24X "egg count".
4. Calculate the concentration of eggs in the stock. Eggs/mL = 10X (egg
count). Dilute the egg stock to 2000 eggs/ml by the formula below.
a. If the egg count is equal to or greater than 200:
(egg count) - 200 = volume (mL) of control water to
add to egg stock.
-i
b. If the egg count is less than 200, allow the eggs to settle and
remove enough control water to concentrate the eggs to greater
than 200, repeat the count, and dilute the egg stock as in a.
above.
NOTE: It requires 24 mL of a egg stock solution for each test
with a control and five exposure concentrations.
c. Transfer 1 ml of the diluted egg stock to a vial containing 9 ml
of control water. Mix well, then transfer 1 mL from the vial to
a Sedgwick-Rafter counting chamber. Count all eggs using a
dissecting microscope. Confirm that the final egg count =
2000/mL (± 200).
10.6 START OF THE TEST
10.6.1 Within 1 h of collection add 100 uL of appropriately diluted sperm to
each test vial. Record the time of sperm addition.
10.6.2 Incubate all test vials at 20 ± 1°C for 1 h.
10.6.3 Mix the diluted egg suspension (2000 eggs/mL), using gentle bubbling.
Add 1 mL of diluted egg suspension to each test vial using a wide mouth pipet
tip. Incubate 20 min at 20 ± 1°C.
307
-------
10.7 TERMINATION OF THE TEST
10.7.1 Terminate the test and preserve the samples by adding 2 ml of 1%
formalin in seawater to each vial.
10.7.2 Vials should be evaluated within 48 hours.
10.7.3 To determine fertilization, transfer about 1 ml eggs from the bottom
of a test vial to a Sedgwick-Rafter counting chamber. Observe the eggs using
a compound microscope (100 X). Count between 100 and 200 eggs/sample. Record
the number counted and the number unfertilized. Fertilization is indicated by
the presence of a fertilization membrane surrounding the egg. Note:
adjustment of the microscope to obtain proper contrast may be required to
observe the fertilization membrane. Because samples are fixed in formalin, a
ventilation hood is setup surrounding the microscope to protect the analyst
from prolonged exposure to formaldehyde fumes.
11. ACCEPTABILITY OF TEST RESULTS
11.1 The sperm:egg ratio routinely employed should result in fertilization of
>50% of the eggs in the control chambers.
12. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA
12.1 A summary of test conditions and test acceptability criteria is listed
in Table 3.
13. DATA ANALYSIS
13.1 GENERAL
13.1.1 Tabulate and summarize the data. Calculate the proportion of
fertilized eggs for each replicate. A sample set of test data is listed in
Table 4.
13.1.2 The endpoints of toxicity tests using the sea urchin are based on the
reduction in proportion of eggs fertilized. The IC25 and the IC50 are
calculated using the Linear Interpolation Method (see Section 9). See the
Appendices for examples of the manual computations, program listings, and
examples of data input and program output.
13.1.3 Formal statistical analysis of the fertilization data is outlined in
Figure 1. The response used in the analysis is the proportion of fertilized
eggs in each test or control chamber. Separate analyses are performed for the
estimation of the NOEC and LOEC endpoints and for the estimation of the IC25
and IC50 endpoints. Concentrations at which there are no eggs fertilized in
any of the test chambers are excluded from statistical analysis of the NOEC
and LOEC, but included in the estimation of the IC25 and IC50.
13.1.4 For the case of equal numbers of replicates across all concentrations
and the control, the evaluation of the NOEC and LOEC endpoints is made via a
parametric test, Dunnett's Procedure, or a nonparametric test, Steel's
308
-------
Many-one Rank Test, on the arc sine transformed data. Underlying assumptions
of Dunnett's Procedure, normality and homogeneity of variance, are formally
tested. The test for normality is the Shapiro-Milk's Test, and Bartlett's
Test is used to determine the homogeneity of variance. Tests for normality
and homogeneity of variance are included in Appendix B. If either of these
tests fail, the nonparametric test, Steel's Many-one Rank Test, is used to
determine the NOEC and LOEC endpoints. If the assumptions of Dunnett's
Procedure are met, the endpoints are estimated by the parametric procedure.
13.1.5 If unequal numbers of replicates occur among the concentration levels
tested, there are parametric and nonparametric alternative analyses. The
parametric analysis is a t-test with the Bonferroni adjustment. The Wilcoxon
Rank Sum Test with the Bonferroni adjustment is the nonparametric alternative.
For detailed information on the Bonferroni adjustment see Appendix D.
13.1.6 The statistical tests described here must be used with a knowledge of
the assumptions upon which the tests are contingent. The assistance of a
statistician is recommended for analysts who are not proficient in statistics.
13.2 EXAMPLE OF ANALYSIS OF SEA URCHIN, ARBACIA PUNCTULATA, FERTILIZATION
DATA.
13.2.1 This example uses toxicity data from a sea urchin, Arbacia punctulata,
fertilization test performed with copper. The response of interest is the
proportion of fertilized eggs, thus each replicate must first be transformed
by the arc sine transformation procedure described in Appendix B. The raw and
transformed data, means and standard deviations of the transformed
observations at each copper concentration and control are listed in Table 5.
The data are plotted in Figure 2.
13.2.2 Test for Normality
13.2.2.1 The first step of the test for normality is to center the
observations by subtracting the mean of all observations within a
concentration from each observation in that concentration. The centered
observations are summarized in Table 6.
309
-------
TABLE 3. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA FOR
SEA URCHIN, ARBACIA PUNCTULATA, FERTILIZATION TEST WITH EFFLUENT
AND RECEIVING WATERS
1. Test type
2. Salinity:
3. Temperature
4. Light quality:
5. Light intensity:
6. Test vessel size:
7. Test solution volume:
8. Number of sea urchins
9.
10.
11.
Number of egg and sperm cells
per chamber:
Number of replicate
chambers per treatment:
Dilution water:
12. Test concentrations:
13. Dilution factor:
Static
30 °/oo ± 2 °/oo
20 ± 1°C
Ambient laboratory light
during test preparation;
10-20 |iE/m2/s, or 50-100 ft-c
(Ambient laboratory levels)
Disposable (glass) liquid
scintillation vials (20 mL
capacity), not precleaned,
but presoaked in control
water
5 mL
Pooled sperm from four males
and pooled eggs from four
females are used per test
About 2000 eggs and 5,000,000
sperm cells per vial
4 (minimum of 3)
Uncontaminated source of
natural seawater; deionized
water mixed with hypersaline
brine or artificial sea salts
(GP2, FORTY FATHOMS", or equivalent)
Effluents: Minimum of five effluent
concentrations and a control
Receiving waters: 100% receiving
water and a control
Effluents: > 0.5 series
Receiving waters: None, or > 0.5
series
310
-------
TABLE 3. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA FOR
SEA URCHIN, ARBACIA PUNCTULATA, FERTILIZATION TEST WITH EFFLUENT
AND RECEIVING WATERS (CONTINUED)
14. Test duration:
15. Effects measured:
16. Test acceptability
criteria:
17. Sampling requirements:
18. Estimated maximum sample
volume required:
1 h and 20 min
Fertilization of sea urchin
eggs
50% or greater egg fertilization
in controls
One sample collected at test
initiation, and preferably used
within 24 h of the time it is
removed from the sampling device (see
Section 15, Subsection 10.1.2.3).
1 L per test (see Subsection 10.1)
311
-------
TABLE 4. DATA FROM SEA URCHIN, AKBACIA PUNCTULATA, FERTILIZATION TEST1
Copper
Concentration
(ugA)
0.0
2.5
5.0
10.0
20.0
40.0
Replicate
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
No. of Eggs
Counted
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
No. of Eggs
Fertilized
85
78
87
81
65
71
63
74
78
63
66
51
41
41
37
12
30
26
Proportion
Fertilized
.85
.78
.87
.81
.65
.71
.63
.74
.78
.63
.66
.51
.41
.41
.37
.12
.30
.26
1Tests performed by Dennis M. McMullen, Technology Applications, Inc., Newtown
Facility, Environmental Monitoring Systems Laboratory - Cincinnati.
TABLE 5. SEA URCHIN, ARBACIA PUNCTULATA, FERTILIZATION DATA
Copper Concentration
Repl
RAW
ARC SINE
TRANSFORMED
MfAN (7,)
S)
i
icate
A
B
C
A
B
C
Control
0.85
0.78
0.87
1.173
1.083
1.202
1.153
0.004
1
2.5
0.81
0.65
0.71
1.120
0.938
1.002
1.020
0.009
2
5.0
0.63
0.74
0.78
0.917
1.036
1.083
1.012
0.007
3
10.0
0.63
0.66
0.51
0.917
0.948
0.795
0.887
0.007
4
(M/L)
20.0
0.41
0.41
0.37
0.695
0.695
0.654
0.681
0.001
5
40.0
0.12
0.30
0.26
0.354
0.580
0.535
0.490
0.014
6
312
-------
STATISTICAL ANALYSIS OF SEA URCHIN FERTILIZATION TEST
FERTILIZATION DATA
PROPORTION OF FERTILIZED EGGS
POINT ESTIMATION
ARC SINE
TRANSFORMATION
ENDPOINT ESTIMATE
IC25, IC50
SHAPIRO-WILK'S TEST
NORMAL DISTRIBUTION
NON-NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
BARTLETT'S TEST
HETEROGENEOUS
VARIANCE
NO
EQUAL NUMBER OF
REPLICATES?
EQUAL NUMBER OF
REPLICATES?
NO
YES
YES
T-TEST WITH
BONFERRONI
ADJUSTMENT
DUNNETT'S
TEST
STEEL'S MANY-ONE
RANK TEST
WILCOXON RANK SUM
TEST WITH
BONFERRONI ADJUSTMENT
ENDPOINT ESTIMATES
NOEC.LOEC
Figure 1. Flowchart for statistical analysis of the sea urchin,
Arbacia punctulata, data.
313
-------
GO
1.0
0.9
0.8
0.7
0.6 -
&
0.4 -
Q 0.3
0.2
0.1
0.0-
T
0.0
2.5
CONNECTS THE MEAN VALUE FOE EACH CONCENTRATION
REPRESENTS THE CRITICAL VALUE FOR DUNNETTS TEST
(ANY PROPORTION BELOW THIS VALUE WOULD BE
SIGNIFICANTLY DIFFERENT FROM THE CONTROL)
5.0 10.0
COPPER CONCENTRATION (UG/L)
20.0
Figure 2. Plot of mean percent of fertilized sea urchin, Arbacia punctulata, eggs.
40.0
-------
TABLE 6. CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S EXAMPLE
Copper Concentration (ng/L)
Replicate
A
B
C
Control
0.020
-0.070
0.049
2.5
0.100
-0.082
-0.018
5.0
-0.095
0.024
0.071
10.0
0.030
0.061
-0.092
20.0
0.014
0.014
-0.027
40.0
-0.136
0.090
0.045
13.2.2.2 Calculate the denominator, D, of the statistic:
n
D = 2 (X,- X)2
Where: X,- = the ith centered observation
X = the overall mean of the centered observations
n = the total number of centered observations
13.2.2.3 For this set of data, n = 18
X - 1 (0) - 0
18
D = 0.0822
13.2.2.4 Order the centered observations from smallest to largest
y(D v(2) Y(n)
A A A
where X(1) denotes the ith ordered observation. The ordered observations for
this example are listed in Table 7.
TABLE 7. ORDERED CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S EXAMPLE
i
1
2
3
4
5
6
7
8
9
X<|)
-0.136
-0.095
-0.092
-0.082
-0.070
-0.027
-0.018
0.014
0.014
i
10
11
12
13
14
15
16
17
18
x
0.020
0.024
0.030
0.045
0.049
0.061
0.071
0.090
0.100
315
-------
13.2.2.5 From Table 4, Appendix B, for the number of observations, n, obtain
the coefficients ar az, ... ak where k is n/2 if n is even and (n-l)/2 if n
is odd. For the data in this example, n = 18 and k = 9. The a,- values are
listed in Table 8.
13.2.2.6 Compute the test statistic, W, as follows:
W - 1 [ J a, (X(n-'"+1) X(i)) ]2
D i = l
The differences, x X(i), are listed in Table 7. For the data in this
example:
W = _J (0.2782)2 = 0.942
0.0822
TABLE 8. COEFFICIENTS AND DIFFERENCES FOR SHAPIRO-MILK'S EXAMPLE
i X(rvi+1) - X(i)
1
2
3
4
5
6
7
8
9
0.4886
0.3253
0.2553
0.2027
0.1587
0.1197
0.0837
0.0496
0.0163
0.236
0.185
0.163
0.143
0.119
0.072
0.048
0.010
0.006
XC18)
X(17>
X(16)
X(15)
v(U)
A
x
xcio>
- x(1)
- X(2)
- X<3)
X(4)
- x(5)
x<6)
X(7)
- x(8)
X(9)
13.2.2.7 The decision rule for this test is to compare W as calculated in 2.6
to a critical value found in Table 6, Appendix B. If the computed W is less
than the critical value, conclude that the data are not normally distributed.
For the data in this example, the critical value at a significance level of
0.01 and n = 18 observations is 0.858. Since W = 0.942 is greater than the
critical value, conclude that the data are normally distributed.
13.2.3 Test for Homogeneity of Variance
13.2.3.1 The test used to examine whether the variation in the proportion of
fertilized eggs is the same across all copper concentrations including the
control, is Bartlett's Test (Snedecor and Cochran, 1980). The test statistic
is as follows:
P P
[ ( S V,) In S2 - 31 V,. In S,-2 ]
B = 1-1 i-1
316
-------
Where: V1 = degrees of freedom for each copper concen
tration and control, V,- = (n, - 1)
p = number of levels of copper concentration
including the control
z v,.
C = 1 + ( 3(p-l))'1 [ Z l/V, - ( Z Vj)'1 ]
i=l i=l
In = loge
i = 1, 2, ..., p where p is the number of concentrations
including the control
n,- = the number of replicates for concentration i.
13.2.3.2 For the data in this example, (See Table 5) all copper
concentrations including the control have the same number of replicates (n, =
3 for all i). Thus, V,- = 2 for all i.
13.2.3.3 Bartlett's statistic is, therefore:
P ,
B = [(12)ln(0.007) - 2 Z ln(Sj)]/1.194
i=l
= [12(-4.962) - 2(-31.332)]/1.194
= 3.122/1.194
= 3.615
13.2.3.4 B is approximately distributed as chi-square with p-1 degrees of
freedom, when the variances are in fact the same. Therefore, the appropriate
critical value for this test, at a significance level of 0.01 with 5 degrees
of freedom, is 15.09. Since B = 2.615 is less than the critical value of
15.09, conclude that the variances are not different.
13.2.4 Dunnett's Procedure
13.2.4.1 Calculations
To obtain an estimate of the pooled variance for the Dunnett's Procedure,
construct an ANOVA table as described in Table 9.
317
-------
TABLE 9. ANOVA TABLE
Source
Between
Within
Total
df Sum of Squares
(SS)
p - 1 SSB
N - p SSW
N - 1 SST
Mean Square(MS)
(SS/df)
Sg = SSB/(p-l)
Sy = SSW/(N-p)
Where: p = number of copper concentrations including the control
N = total number of observations n1 + n? . . . +np
n = number of observations in concentration i
> ,
SSB = 2 T, /n, GYM Between Sum of Squares
i = l
P n,-
SST = S X Y2: G2/N Total Sum of Squares
1-1 j=l
SSW = SST - SSB Within Sum of Squares
p
G = the grand total of all sample observations, G = s T,.
i=l
T,. = the total of the replicate measurements for
concentration "i"
Y,-J = the jth observation for concentration "i" (represents
the proportion of fertilized eggs for copper
concentration i in test chamber j)
13.2.4.2 For the data in this example:
n1 = n, = n3 = n4 = n5 = n6 = 3
N - 18
T3
Yn -
' Y21 '
= Y31 -
= Y41 H
= Y51 H
' Y61 H
h Y12H
h Y22n
hY32H
Y42^
h Y52H
> Y62^
h Y13 = 3.458
h Y23 = 3.060
h Y33 = 3.036
h Y43 = 2.660
h Y53 = 2.044
h Y63 = 1.469
T, + T2 + T3 + T4 = 15.727
318
-------
p , ?
SSB = Z T,-2/n, - G2/N
= (43.950)/3 - (15.727)2/18 = 0.909
P ni
SST = S E Y2, - G2/N
i=l j=l
= 14.732 - (15.727)2/18 = 0.991
SSW = SST SSB = 0.991 - 0.909 = 0.082
Sg = SSB/(p-l) = 0.909/(6-l) = 0.182
y = SSW/(N-p) = 0.082/(18-6) = 0.007
13.2.4.3 Summarize these calculations in the ANOVA table (Table 10)
TABLE 10. ANOVA TABLE FOR DUNNETT'S PROCEDURE EXAMPLE
Source df Sum of Squares Mean Square(MS)
(SS) (SS/df)
Between
Within
Total
5
12
17
0.909
0.082
0.991
0.182
0.007
13.2.4.4 To perform the individual comparisons, calculate the t statistic
for each concentration, and control combination as follows:
SH / (1/n,) + (1/n,)
Where: Y,. = mean proportion fertilized eggs for copper concentration i
Y, = mean proportion fertilized eggs for the control
Sw = square root of within mean square
n1 = number of replicates for the control
n,- = number of replicates for concentration i.
319
-------
Since we are looking for a decreased response from the control in the
proportion of fertilized eggs, the mean concentration is subtracted from the
control mean.
13.2.4.5 Table 11 includes the calculated t values for each concentration and
control combination. In this example, comparing the 2.5 |ig/L concentration
with the control the calculation is as follows:
( 1.153 1.020 )
t2 = = 1.939
[ 0.084 / (1/3) + (1/3) ]
TABLE 11. CALCULATED T-VALUES
Copper Concentration (|ig/L) i tt-
2.5
5.0
10.0
20.0
40.0
2
3
4
5
6
1.939
2.056
3.878
6.882
9.667
13.2.4.6 Since the purpose of this test is to detect a significant decrease
in the proportion of fertilized eggs, a one-sided test is appropriate. The
critical value for this one-sided test is found in Table 5, Appendix D. For
an overall alpha level of 0.05, 12 degrees of freedom for error and five
concentrations (excluding the control) the critical value is 2.50. The mean
proportion of fertilized eggs for concentration "i" is considered
significantly less than the mean proportion of fertilized eggs for the control
if tj is greater than the critical value. Therefore, the 10.0, 20.0 and 40.0
|ig/L concentrations have a significantly lower mean proportion of fertilized
eggs than the control. Hence the NOEC is 5.0 jig/L and the LOEC is 10.0 |ig/L.
13.2.4.7 To quantify the sensitivity of the test, the minimum significant
difference (MSD) that can be statistically detected may be calculated.
MSD = d Su / (1/n,) + (1/n)
Where: d = the critical value for the Dunnett's procedure
Sw = the square root of the within mean square
n = the common number of replicates at each concentration
(this assumes equal replication at each concentration)
n1 = the number of replicates in the control.
13.2.4.8 In this example,
320
-------
MSD = 2.50 (0.084) / (1/3) + (1/3)
= 2.50 (0.084)(0.816)
= 0.171
13.2.4.9. The MSD (0.171) is in transformed units. To determine the MSD in
terms of proportion of fertilized eggs, carry out the following conversion.
1. Subtract the MSD from the transformed control mean.
1.153 - 0.171 = 0.982
2. Obtain the untransformed values for the control mean and the
difference calculated in step 1 of 13.2.4.9.
[Sine (1.153) ]* = 0.835
[Sine (0.982) ]2 = 0.692
3. The untransformed MSD (MSDU) is determined by subtracting the
untransformed values from step 2 in 14.2.4.9.
MSDU = 0.835 - 0.692 = 0.143
13.2.4.10 Therefore, for this set of data, the minimum difference in mean
proportion of fertilized eggs between the control and any copper concentration
that can be detected as statistically significant is 0.143.
13.2.4.11 This represents a 17% decrease in the proportion of fertilized eggs
from the control.
13.2.5 Calculation of the 1C
13.2.5.1 The fertilization data in Table 4 are utilized in this example.
Table 12 contains the mean proportion of fertilized eggs for each toxicant
concentration. As can be seen, the observed means are monotonically non-
increasing with respect to concentration. Therefore, it is not necessary to
smooth the means prior to calculating the 1C. Refer to Figure 2 for a plot of
the response curve.
321
-------
TABLE 12. SEA URCHIN, ARBACIA PUNCTULATA,
MEAN PROPORTION OF FERTILIZED EGGS
Copper
Cone.
(M/L)
Control
2.5
5.0
10.0
20.0
40.0
i
1
2
3
4
5
6
M
0.833
0.723
0.717
0.600
0.397
0.227
13.2.5.2 An IC25 and IC50 can be estimated using the Linear Interpolation
Method. A 25% reduction in mean proportion of fertilized eggs, compared to
the controls, would result in a mean proportion of 0.625, where M^l-p/lOO)
0.833(1-25/100). A 50% reduction in mean proportion of fertilized eggs,
compared to the controls, would result in a mean proportion of 0.417.
Examining the means and their associated concentrations (Table 12), the
response, 0.625, is bracketed by C3 = 5.0 jig/L copper and C4 = 10.0 |ig/L
copper. The response, 0.417, is bracketed by C4 = 10.0 (ig/L copper and C5 =
20.0 |ig/L copper.
13.2.5.3 Using Equation 1 from Appendix I, the estimate of the IC25 is
calculated as follows :
ICp = Cj + [M^l - p/100) MJ (CJ+1 - C,)
(MM - Mj)
IC25 =5.0 + [0.833(1 - 25/100) - 0.717] (10.0 - 5.0)
(0.600 - 0.717)
= 8.9 ug/L.
13.2.5.4 Using Equation 1 from Appendix I, the estimate of the IC50 is
calculated as follows:
ICp = Cj + [M^l - p/100) - MJ (CJ+1 - CJ
(MJ+1 - MJ
322
-------
IC50 = 10.0 + [0.833(1 50/100) 0.600] (20.0 - 10.0)
(0.397 0.600)
= 19.0 jig/L.
13.2.5.5 When the Bootstrap program (BOOTSTRP) was used to analyze this set
of data, requesting 80 resamples, the mean estimate of the IC25 was 8.8874
lig/L, with a standard deviation of 1.4390 jig/L (coefficient of variation =
16.2%). The empirical 95.0% confidence interval for the true mean was (5.8984
ng/L, 11.3291 |ig/L). The BOOTSTRP computer program output for the IC25 for
this data set is shown in Figure 3.
13.2.5.6 When the Bootstrap program (BOOTSTRP) was used to analyze this set
of data, requesting 80 resamples, the mean estimate of the IC50 was 19.0418
|ig/L, with a standard deviation of 0.7996 |ig/L (coefficient of variation =
4.2%). The empirical 95.0% confidence interval for the true mean was (17.7174
jig/L, 20.8475 |ig/L). The BOOTSTRP computer program output for the IC50 for
this data set is shown in Figure 4.
323
-------
THE NUMBER OF RESAMPLES IS 80.
*** LISTING OF GROUP CONCENTRATIONS (% EFF.) AND RESPONSE MEANS ***
CONC. (%EFF) RESPONSE MEAN MEAN AFTER POOLING
0.000 0.833 0.833
2.500 0.723 0.723
5.000 0.717 0.717
10.000 0.600 0.600
20.000 0.397 0.397
40.000 0.227 0.227
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
FROM THE INPUT SAMPLE IS 8.9286.
* BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY *
* OF THE ESTIMATED ICp *
************************************************************
THE MEAN OF THE BOOTSTRAP ESTIMATES IS 8.8874.
THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS 1.4390.
AN EMPIRICAL 95.0% CONFIDENCE INTERVAL FOR THE
BOOTSTRAP ESTIMATE IS (5.8984, 11.3291).
Figure 3. BOOTSTRP program output for the IC25.
324
-------
THE NUMBER OF RESAMPLES IS 80.
*** LISTING OF GROUP CONCENTRATIONS (% EFF.) AND RESPONSE MEANS ***
CONC. (%EFF) RESPONSE MEAN MEAN AFTER POOLING
0.000 0.833 0.833
2.500 0.723 0.723
5.000 0.717 0.717
10.000 0.600 0.600
20.000 0.397 0.397
40.000 0.227 0.227
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
FROM THE INPUT SAMPLE IS 19.0164.
***********************************************************
* BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY *
* OF THE ESTIMATED ICp *
************************************************************
THE MEAN OF THE BOOTSTRAP ESTIMATES IS 19.0418.
THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS .7796.
AN EMPIRICAL 95.0% CONFIDENCE INTERVAL FOR THE
BOOTSTRAP ESTIMATE IS (17.7174, 20.8475).
Figure 4. BOOTSTRP program output for the IC50.
325
-------
14. PRECISION AND ACCURACY
14.1 PRECISION
14.1.1 Single-laboratory precision data for the reference toxicants, copper
(CU) and sodium dodecyl sulfate (SDS), tested in FORTY FATHOMS" artificial
seawater, GP2 artificial seawater, and natural seawater are provided in Tables
13-18. The test results were similar in the three types of seawater. The
IC25 and IC50 for the reference toxicants (copper and sodium dodecyl sulfate)
are reported in Tables 13-16. The coefficient of variation, based on the
IC25, is 28.7% to 54.6% for natural and FORTY FATHOMS" seawater, indicating
acceptable precision. The IC50 ranges from 23.3% to 48.2%, showing acceptable
precision.
14.1.2 No data are available on the multilaboratory precision of the test.
14.2 ACCURACY
14.2.1 The accuracy of toxicity tests cannot be determined.
326
-------
TABLE 13. SINGLE-LABORATORY PRECISION OF THE SEA URCHIN, ARBACIA PUNCTULATA,
FERTILIZATION TEST PERFORMED IN FORTY FATHOMS* ARTIFICIAL SEAWATER,
USING GAMETES FROM ADULTS MAINTAINED IN FORTY FATHOMSR ARTIFICIAL
SEAWATER, OR OBTAINED DIRECTLY FROM NATURAL SOURCES, AND COPPER
(CU) SULFATE AS A REFERENCE TOXICANT1'2'3'4'5
Test
Number
NOEC
(ugA)
IC25
(ng/L)
IC50
(M/L)
1
2
3
4
5
n:
Mean:
CV(%):
5.0
12.5
<6.2
6.2
12.5
4
NA
NA
8.92
26.35
11.30
34.28
36.67
5
23.51
54.60
29.07
38.96
23.93
61.75
75.14
5
45.77
47.87
1Data from USEPA (1989a) and USEPA (1991a)
2Tests performed by Dennis M. McMullen, Technology Applications, Inc.,
Newtown Facility, Environmental Monitoring Systems Laboratory -
Cincinnati.
2A11 tests were performed using Forty Fathoms" synthetic seawater.
3Copper test solutions were prepared with copper sulfate. Copper
concentrations in Test 1 were: 2.5, 5.0, 10.0, 20.0, and 40.0 |ig/L.
Copper concentrations in Tests 2-5 were: 6.25, 12.5, 25.0,
50.0, and 100.0 |ig/L.
4NOEC Range: <5.0 - 12.5 |ig/L (this represents a difference of one exposure
concentrations).
5For a discussion of the precision of data from chronic toxicity
tests see Section 4, Quality Assurance.
327
-------
TABLE 14. SINGLE-LABORATORY PRECISION OF THE SEA URCHIN, ARBACIA PUNCTULATA,
FERTILIZATION TEST PERFORMED IN FORTY FATHOMS* ARTIFICIAL SEAWATER,
USING GAMETES FROM ADULTS MAINTAINED IN FORTY FATHOMS" ARTIFICIAL
SEAWATER, OR OBTAINED DIRECTLY FROM NATURAL SOURCES, AND SODIUM
DODECYL (SDS) AS A REFERENCE TOXICANT1'2'3'4'5'6
Test
Number
1
2
3
4
5
n:
Mean:
CV(%):
NOEC
(mg/L)
<0.9
0.9
1.8
0.9
1.8
4
NA
NA
IC25
(mg/L)
1.11
1.27
2.26
1.90
2.11
5
1.73
29.7
IC50
(mg/L)
1.76
1.79
2.87
2.69
2.78
5
2.38
23.3
1Data from USEPA (1989a) and USEPA (1991a)
Tests performed by Dennis M. McMullen, Technology Applications, Inc.,
Newtown Facility, Environmental Monitoring Systems Laboratory -
Cincinnati.
All tests were performed using Forty Fathoms" synthetic seawater.
NOEC Range: 1.2 -3.3 mg/L (this represents a difference of one exposure
concentration).
5SDS concentrations for all tests were: 0.9, 1.8, 3.6, 7.2, and 14.4 mg/L.
6For a discussion of the precision of data from chronic toxicity
tests see Section 4, Quality Assurance.
328
-------
TABLE 15. SINGLE-LABORATORY PRECISION OF THE SEA URCHIN, ARBACIA PUNCTULATA,
FERTILIZATION TEST PERFORMED IN NATURAL SEAWATER, USING GAMETES
FROM ADULTS MAINTAINED IN NATURAL SEAWATER AND COPPER (CU) SULFATE
AS A REFERENCE TOXICANT 1-2-3'4-5'6
Test
Number
1
2
3
4
5
n:
Mean:
CV(%):
NOEC
(M/L)
12.2
12.2
24.4
<6.1
6.1
4
NA
NA
IC25
(ng/L)
14.2
32.4
30.3
26.2
11.2
5
22.8
41.9
IC50
(W/L)
18.4
50.8
46.3
34.1
17.2
5
29.9
48.2
1Data from USEPA (1989a) and USEPA (1991a)
2Tests performed by Ray Walsh and Wendy Greene, Environmental Research
Laboratory, U. S. Environmental Protection Agency, Narragansett, Rhode Island
3Copper concentrations were: 6.1, 12.2, 24.4, 48.7, and 97.4 jig/L.
4NOEC Range: <6.1 - 24.4 |ig/L (this represents a difference of two exposure
concentrations).
'Adults collected
'For a discussion
Section 4, Quality Assurance.
5Adults collected in the field.
6For a discussion of the precision of data from chronic toxicity tests see
329
-------
TABLE 16. SINGLE-LABORATORY PRECISION OF THE SEA URCHIN, ARBACIA PUNCTULATA,
FERTILIZATION TEST PERFORMED IN NATURAL SEAWATER, USING GAMETES
FROM ADULTS MAINTAINED IN NATURAL SEAWATER AND SODIUM DODECYL
SULFATE (SDS) AS A REFERENCE TOXICANT 1'2'3'4-5-6
Test NOEC
Number (mg/L)
1 1.8
2 1.8
3 1.8
4 0.9
5 1.8
n: 5
Mean: NA
CV(%): NA
IC25
(mg/L)
2.3
3.9
2.3
2.1
2.3
5
2.58
28.7
IC50
(mg/L)
2.7
5.1
2.9
2.6
2.7
5
3.2
33.3
1Data from USEPA (1989a) and USEPA (1991a).
2Tests performed by Ray Walsh and Wendy Greene, Environmental Research
Laboratory, U. S. Environmental Protection Agency, Narragansett, Rhode Island
SDS concentrations were: 0.9, 1.8, 3.6, 7.3, and 14.5 mg/L.
NOEC Range: 0.9 - 1.8 mg/L (this represents a difference of one exposure
concentration).
5Adults collected in the field.
6For a discussion of the precision of data from chronic toxicity tests see
Section 4, Quality Assurance.
330
-------
TABLE 17. SINGLE-LABORATORY PRECISION OF THE SEA URCHIN, ARBACIA PUNCTULATA,
FERTILIZATION TEST PERFORMED IN GP2, USING GAMETES FROM ADULTS
MAINTAINED IN GP2 ARTIFICIAL SEAWATER AND COPPER (CU) SULFATE AND
SODIUM DODECYL SULFATE (SDS) AS REFERENCE TOXICANTS1'*'3'4'5
CU
SDS
Test
1
2
3
4
5
LC50
CI
(MA) (|ig/L)
29.1
47.6
32.7
78.4
45.6
Mean
SD
CV
27.3-31
44.6-50
29.8-35
73.3-83
41.0-50
46.7
19.5
41.8
.1
.8
.8
.9
.7
NOEC
LOEC
(99/1) (pg/L)
6.3
25.0
6.3
50.0
12.5
12
50
12
100
25
.5
.0
.5
.0
.0
LC50
CI
(mg/L) (mg/L)
2.1
1.8
2.2
2.3
1.8
2.0-2
1.8-1
2.1-2
2.2-2
1.7-2
2.0
0.2
10.0
.1
.9
.2
.4
.8
NOEC
(mg/L)
1.3
1.3
1.3
1.3
1.3
LOEC
(mg/L)
2.5
2.5
2.5
2.5
2.5
1Tests performed by Pamela Comeleo, Science Application International Corp.,
Environmental Research Laboratory, Narragansett, Rhode Island.
2A11 tests were performed using GP2 artificial seawater.
Copper concentrations were: 6.25, 12,5, 25.0, 50.0 and 100 |ig/L.
4SDS concentrations were: 0.6, 1.25, 2.5, 5.0, and 10.0 mg/L. SDS stock
(14.645 mg/mL) provided by Environmental Monitoring Systems Laboratory, U.S.
Environmental Protection Agency, Cincinnati, Ohio.
5For a discussion of the precision of data from chronic toxicity
tests see Section 4, Quality Assurance.
331
-------
TABLE 18. SINGLE-LABORATORY PRECISION OF THE SEA URCHIN, ARBACIA PUNCTULATA,
FERTILIZATION TEST PERFORMED IN NATURAL SEAWATER, USING GAMETES
FROM ADULTS MAINTAINED IN NATURAL SEAWATER AND COPPER (CU) SULFATE
AND SODIUM DODECYL SULFATE (SDS) AS REFERENCE TOXICANTS1'2'3'4
CU
Test
1
2
3
4
5
LC50
(M/L)
28.6
13.0
67.8
36.7
35.6
Mean
SD
CV
CI
(H9/L)
26.7-30.6
11.9-14.2
63.2-72.6
33.9-39.8
33.6-37.7
36.3
20.0
55.1
NOEC
(ng/L)
6.3
6.3
6.3
< 6.3
< 6.3
LOEC
(M/L)
12.5
12.5
12.5
6.3
6.3
LC50
(mg/L)
2.2
1.9
2.2
3.3
2.9
SDS
CI
(mg/L)
2.1-2.2
1.9-2.0
2.1-2.3
3.1-3.4
2.8-3.1
2.5
0.58
23.2
NOEC
(mg/L)
1.3
1.3
1.3
< 0.6
< 0.6
LOEC
(mg/L)
2.5
2.5
2.5
0.6
0.6
Tests performed by Anne Kuhn-Hines, Catherine Sheehan, Glen Modica, and
Pamela Comeleo, Science Application International Corp., Environmental
Research Laboratory, U.S. Environmental Protection Agency, Narragansett,
Rhode Island.
2Copper concentrations were prepared with copper sulfate. Concentrations were
6.25, 12.5, 25.0, 50.0, and 100 |ig/L.
3SDS concentrations were: 0.6, 1.25, 2.5, 5.0, and 10.0 mg/L. SDS stock
(14.64 mg/mL) provided by Environmental Monitoring Systems Laboratory, U.S.
Environmental Protection Agency, Cincinnati, Ohio.
For a discussion of the precision of data from chronic toxicity
tests see Section 4, Quality Assurance.
332
-------
Figure 5. Data sheet for (1) fertilization test using the sea urchin, Arbacia
punctulata.
TEST DATE:
SAMPLE:
COMPLEX EFFLUENT SAMPLE:
COLLECTION DATE:
SALINITY/ADJUSTMENT:
PH/ADJUSTMENT REQUIRED:
PHYSICAL CHARACTERISTICS:
STORAGE:
COMMENTS:
SINGLE COMPOUND:
SOLVENT (CONC):
TEST CONCENTRATIONS:
DILUTION WATER:
CONTROL WATER:
TEST TEMPERATURE:
TEST SALINITY: _
COMMENTS:
333
-------
Figure 6. Data sheet (2) for fertilization test using the sea urchin,
Arbacia punctulata.
TEST DATE:
SAMPLE:
SPERM DILUTIONS:
HEMACYTOMETER COUNT, E: x 104 = SPM SOLUTION E
SPERM CONCENTRATIONS: SOLUTION E x 40 = SOLUTION A = SPM
SOLUTION E x 20 = SOLUTION B = SPM
SOLUTION E x 5 = SOLUTION D = SPM
SOLUTION SELECTED FOR TEST ( = 5 x 107 SPM):
DILUTION: SPM/(5 x 107) = DF
[(DF) x 10) 10 = + SW, mL
FINAL SPERM COUNTS =
EGG DILUTIONS:
INITIAL EGG COUNT =
ORIGINAL EGG STOCK CONCENTRATION = 10X (INITIAL
EGG COUNT) =
VOLUME OF SW TO ADD TO DILUTE EGG STOCK TO 2000/mL:
(EGG COUNT) - 200 =
CONTROL WATER TO ADD EGG STOCK, mL =
FINAL EGG COUNT =
TEST TIMES:
SPERM COLLECTED:
EGGS COLLECTED:
SPERM ADDED:
EGGS ADDED:
FIXATIVE ADDED:
SAMPLES READ:
334
-------
Figure 7. Data sheet (3) for fertilization test using the sea urchin, Arbacia
punctulata.
DATE TESTED:
SAMPLE:
TOTAL AND UNFERTILIZED EGG COUNT AT END OF TEST:
EFFLUENT
CONC (%)
REPLICATE VIAL
1
TOTAL-UNFERT
2
TOTAL-UNFERT
3
TOTAL-UNFERT
4
TOTAL-UNFERT
STATISTICAL ANALYSIS:
ANALYSIS OF VARIANCE:
CONTROL:
DIFFERENT FROM CONTROL (P)
COMMENTS:
335
-------
SECTION 16
TEST METHOD1'2
RED MACROAL6A, CHANPIA PARVULA, SEXUAL REPRODUCTION TEST
METHOD 1009
1. SCOPE AND APPLICATION
1.1 This method measures the effects of toxic substances in effluents and
receiving water on the sexual reproduction of the marine red macroalga,
Champia parvuTa. The method consists of exposing male and female plants to
test substances for two days, followed by a 5-7 day recovery period in control
medium, during which the cystocarps mature.
1.2 Detection limits of the toxicity of an effluent or chemical substance are
organism dependent.
1.3 Single or multiple excursions in toxicity may not be detected using 24-h
composite samples. Also, because of the long sample collection period
involved in composite sampling, highly volatile and highly degradable
toxicants in the source may not be detected in the test.
1.4 This test is commonly used in one of two forms: (1) a definitive test,
consisting of a minimum of five effluent concentrations and a control, and (2)
a receiving water test(s), consisting of one or more receiving water
concentrations and a control.
2. SUMMARY OF METHOD
2.1 Sexually mature male and female branches of the red macroalga, Champia
parvula, are exposed in a static system for two days to different
concentrations of effluent, or to receiving water, followed by a 5- to 7-day
recovery period in control medium. The recovery period allows time for the
development of cystocarps resulting from fertilization during the exposure
period. The test results are reported as the concentration of the test
substance which causes a statistically significant reduction in the number of
cystocarps formed compared to control organisms.
3. INTERFERENCES
3.1 Toxic substances may be introduced by contaminants in dilution water,
glassware, sample hardware, and testing equipment (see Section 5, Facilities
and Equipment).
3.2 Improper effluent sampling and handling may adversely affect test
1The format used for this method was taken from USEPA (1983).
2This method was adapted from USEPA (1987e).
336
-------
results (see Section 8, Effluent and Receiving Water Sampling, Sample
Handling, and Sample Preparation for Toxicity Tests).
3.3. Adverse effects of high concentrations of suspended and/or dissolved
solids, and extremes of pH, may mask the presence of toxic substances.
3.4 Pathogenic and/or predatory organisms in the dilution water and
effluent may affect test organism survival, and confound test results.
4. SAFETY
4.1 See Section 3, Safety and Health.
5. APPARATUS AND EQUIPMENT
5.1 Facilities for holding and acclimating test organisms.
5.2 Laboratory red macroalga, Champia parvula, culture unit -- See culturing
methods below. To test effluent or receiving water toxicity, sufficient
numbers of sexually mature male and female plants must be available.
5.3 Samplers -- automatic samplers, preferably with sample cooling
capability, that can collect a 24-h composite sample of 1 L.
5.4 Environmental chamber or equivalent facility with temperature
(23 ± 1°C) and light (75 |iE/m2/s, or 500 ft-c) control.
5.5 Water purification system -- Millipore Milli-QR, deionized water (DI) or
equivalent (see Section 5, Facilities, Equipment, and Supplies).
5.6 Air pump -- for supplying air.
5.7 Air lines, and air stones -- for aerating cultures.
5.8 Balance -- Analytical, capable of accurately weighing to 0.0001 g.
5.9 Reference weights, Class S -- for checking performance of balance.
5.10 pH meter -- for routine physical and chemical measurements. Unless the
test is being conducted to specifically measure the effect of this
parameter, a portable, field-grade instrument is acceptable.
5.11 Dissecting (stereoscope) microscope -- for counting cystocarps.
5.12 Compound microscope -- for examining the condition of plants.
5.13 Count register, 2-place -- for recording cystocarp counts.
5.14 Rotary shaker -- for incubating exposure chambers (hand-swirling twice
a day can be substituted).
5.15 Drying oven -- to dry glassware.
337
-------
5.16 Filtering apparatus -- for use with membrane filters (47mm).
5.17 Refractometer -- for determining salinity.
5.18 Thermometers, glass or electronic, laboratory grade -- for measuring
water temperatures.
5.19 Thermometers, bulb-thermograph or electronic-chart type -- for
continuously recording temperature.
5.20 Thermometer, National Bureau of Standards Certified (see USEPA METHOD
170.1, USEPA, 1979b) -- to calibrate laboratory thermometers.
5.21 Beakers -- Class A, borosilicate glass or non-toxic plasticware, 1000
ml for making test solutions.
5.22 Erlenmeyer flasks, 250 ml, or 200 mL disposable polystyrene cups, with
covers -- for use as exposure chambers.
5.23 Bottles -- borosilicate glass or disposable polystyrene cups (200-400
mL) for use as recovery vessels.
5.24 Wash bottles -- for deionized water, for rinsing small glassware and
instrument electrodes and probes.
5.25 Volumetric flasks and graduated cylinders -- Class A, borosilicate
glass or non-toxic plastic labware, 10-1000 mL for making test solutions.
5.26 Micropipetters, digital, 200 and 1000 [iL -- to make dilutions.
5.27 Pipets, volumetric -- Class A, 1-100 mL.
5.28 Pipetter, automatic -- adjustable, 1-^100 mL.
5.29 Pipets, serological -- 1-10 mL, graduated.
5.30 Pipet bulbs and fillers -- PROPIPET", or equivalent.
5.31 Forceps, fine-point, stainless steel -- for cutting and handling
branch tips.
6. REAGENTS AND CONSUMABLE MATERIALS
6.1 Mature red macroalga, Champia parvula, plants -- see Subsection 6.14
below.
6.2 Sample containers -- for sample shipment and storage (see Section 8,
Effluent and Receiving Water Sampling, Sample Handling, and Sample Preparation
for Toxicity Tests).
6.3 Petri dishes, polystyrene -- to hold plants for cystocarp counts and to
cut branch tips. Other suitable containers may be used.
338
-------
6.4 Disposable tips for micropipetters.
6.5 Aluminum foil, foam stoppers, or other closures -- to cover culture and
test flasks.
6.6 Tape, colored -- for labelling test chambers.
6.7 Markers, water-proof -- for marking containers, etc.
6.8 Data sheets (one set per test) -- for data recording.
6.9 Buffers, pH 4, 7, and 10 (or as per instructions of instrument
manufacturer) for standards and calibration check (see USEPA Method 150.1,
USEPA, 1979b).
6.10 Laboratory quality assurance samples and standards for the above
methods.
6.11 Reference toxicant solutions (see Section 4, Quality Assurance,
Subsections 4.7, 4.14, 4.15. 4.16, and 4.17).
6.12 Reagent water -- defined as distilled or deionized water that does not
contain substances which are toxic to the test organisms (see Subsection 5.5
above).
6.13 Effluent, receiving water, and dilution water -- see Section 7, Dilution
Water, and Section 8, Effluent and Receiving Water Sampling, Sample Handling,
and Sample Preparation for Toxicity Tests.
6.13.1 Saline test and dilution water -- The use of natural seawater is
recommended for this test. A recipe for the nutrients that must be added to
the natural sea water is given in Table 1. The salinity of the test water
must be 30 °/°o> and vary no more than + 2 °/oo among the replicates.
6.13.2 The overwhelming majority of industrial and sewage treatment
effluents entering marine and estuarine systems contain little or no
measurable salts. Therefore, exposure of the red macroalga, Champia parvula,
to effluents will usually require adjustments in the salinity of the test
solutions. Although the red macroalga, Champia parvula, cannot be cultured in
100% artificial seawater, 100% artificial seawater can be used during the two
day exposure period. This allows 100% effluent to be tested. It is important
to maintain a constant salinity across all treatments. The salinity of the
effluent can be adjusted by adding brine prepared from natural seawater (100
°/oo), concentrated (triple strength) salt solution (GP2 described in Table
2), or dry GP2 salts (Table 2), to the effluent to provide a salinity of 30
°/oo. Control solutions should be prepared with the same percentage of
natural seawater and at the same salinity (using deionized water adjusted with
dry salts, or brine) as used for the effluent dilutions.
6.13.3 Artificial seawater -- The preparation of artificial seawater (GP2)
is described in Table 2.
339
-------
6.14 THE RED MACROALGA, CHAMPIA PARVULA, CULTURES
6.14.1 Mature plants (Figure 1) are available from the Environmental
Research Laboratory, U.S. Environmental Protection Agency, 27 Tarzwell Drive
Road, Narragansett, Rhode Island, 02882 (401-782-3000). The adult plant body
(thallus) is hollow, septate, and highly branched. New cultures can be
propagated asexually from excised branches, making it possible to maintain
clonal material indefinitely.
tctrosporongia?
spwmatia
fertilization
TETRASPOROPHYTC
cyttocorp
Figure 1. Life history of the red macroalga, Champia parvula. Upper left:
Size and degree of branching in female branch tips used for toxicity tests.
From USEPA (1987e).
340
-------
6.14.1.1 Unialgal stock cultures of both males and females are maintained in
separate, aerated 1000 ml Erlenmeyer flasks containing 800 ml of the culture
medium. All culture glass must be acid-stripped in 15% HC1 and rinsed in
deionized water after washing. This is necessary since some detergents can
leave a residue that is toxic to the red macroalga, Champia parvula.
Periodically (at least every 6 months) culture glassware should be baked in a
muffle furnace to remove organic material that may build up on its surface.
Alternately, a few ml of concentrated sulfuric acid can be rolled around the
inside of wet glassware. Caution the addition of acid to the wet glassware
generates heat.
6.14.1.2 The culture medium is made from natural seawater to which additional
nutrients are added. The nutrients added are listed in Table 1. Almost any
nutrient recipe can be used for the red macroalga, Champia parvula, cultured
in either natural seawater or a 50-50 mixture of natural and artificial
seawaters. Healthy, actively growing plants are the goal, not a standard
nutrient recipe for cultures.
6.14.1.3 Several cultures of both males and females should be maintained
simultaneously to keep a constant supply of plant material available. To
maintain vigorous growth, initial stock cultures should be started
periodically with about twenty 0.5 to 1.0 cm branch tips. Cultures are gently
aerated through sterile, cotton-plugged, disposable, polystyrene 1 ml
pipettes. Cultures are capped with foam plugs and aluminum foil and
illuminated with ca. 75 //Em s/1 (ca 500 ft candles) of cool-white fluorescent
light on a 16:8, light:dark cycle. Depending on the type of culture chamber
or room used, i.e., the degree of reflected light, the light levels may have
to be adjusted downward. The temperataure is 22 to 24°C and the salinity 28-
30 °/oo. Media are changed once a week.
6.14.1.4 Prior to use in toxicity tests, stock cultures should be examined to
determine their condition. Females can be checked by examining a few branch
tips under a compound microscope (100 X or greater). Several trichogynes
(reproductive hairs to which the spermatia attach) should be easily seen near
the apex (Figure 2).
6.14.1.5 Male plants should be visibly producing spermatia. This can be
checked by placing some male tissue in a petri dish, holding it against a
dark background and looking for the presence of spermatial sori. Mature
sori can also be easily identified by looking along the edge of the thallus
under a compound microscope (Figures 3 and 4).
6.14.1.6 A final, quick way to determine the relative "health" of the male
stock culture is to place a portion of a female plant into some of the water
from the male culture for a few seconds. Under a compound microscope
numerous spermatia should be seen attached to both the sterile hairs and the
trichogynes (Figure 5).
6.14.2 Culture medium prepared from natural seawater is preferred
(Table 1). However, as much as 50% of the natural seawater may be replaced
by the artificial seawater (GP2) described in Table 2.
341
-------
sterile hairs
Irichogynes
1 mm
Figure 2. Apex of branch of female plant, showing sterile hairs and
reproductive hairs (trichogynes). Sterile hairs are wider and generally much
longer than trichogynes, and appear hollow except at the tip. Both types of
hairs occur on the entire circumference of the thallus, but are seen easiest
at the "edges." Receptive trichogynes occur only near the branch tips. From
USEPA (1987e).
1 cm
spermatial
sorus
Figure 3. A portion of the male thallus showing spermatial sori. The
sorus areas are generally slightly thicker and somewhat lighter in color.
From USEPA (1987e).
342
-------
-cuticle
thallus surface
\QQjjm
spermatia
Figure 4. A magnified portion of a spermatial sorus. Note the rows of cells
that protrude from the thallus surface. From USEPA (1987e).
Figure 5. Apex of a branch on a mature female plant that was exposed to
spermatia from a male plant. The sterile hairs and trichogynes are covered
with spermatia. Note that few or no spermatia are attached to the older hairs
(those more than 1 mm from the apex). From USEPA (1987e).
343
-------
6.14.2.1 Seawater for cultures is filtered at least to 0.45 ^m to remove most
particulates and then autoclaved for 30 min. at 15 psi (120°C). Carbon
stripping the seawater may be necessary before autoclaving to enhance its
water quality (USEPA, 1990c). This is done by adding 2 g activated carbon per
liter of seawater and stirring on a stir plate for 2 hours. After stirring
filter through Whatman number 2 filter, then through 0.45 membrane filter.
The culture flasks are capped with aluminum foil and autoclaved dry, for 10
min. Culture medium is made up by dispensing seawater into sterile flasks and
adding the appropriate nutrients from a sterile stock solution.
6.14.2.2 Alternately, 1-L flasks containing seawater can be autoclaved.
Sterilization is used to prevent microalgal contamination, and not to keep
cultures bacteria free.
7. SAMPLE COLLECTION, PRESERVATION, AND HANDLING
7.1 See Section 8, Effluent and Receiving water Sampling, Sample Handling,
and Sample Preparation for Toxicity Tests
8. CALIBRATION AND STANDARDIZATION
8.1 See Section 4, Quality Assurance.
9. QUALITY CONTROL
9.1 See Section 4, Quality Assurance.
10 TEST PROCEDURES
10.1 TEST SOLUTIONS
10.1.1 Receiving Waters
10.1.1.1. The sampling point is determined by the objectives of the test. At
estuarine and marine sites, samples are usually collected at mid-depth.
Receiving water toxicity is determined with samples used directly as collected
or with samples passed through a 60 ^m NITEXR filter and compared without
dilution, against a control. Using four replicate chambers per test, each
containing 100 mL, and 400 mL for chemical analysis, would require
approximately 800 ml or more of sample per test.
10.1.2 Effluents
10.1.2.1. Effluents can be tested at 100%. A 100% concentration of effluent
can be achieved if the salinity of the effluent is adjusted to 30 °/°° by
adding the GP2 dry salt formulation described in Table 2.
10.1.2.2 The selection of the effluent test concentrations should be based on
the objectives of the study. A dilution factor of 0.5 is commonly used. A
dilution factor of 0.5 provides precision of ± 100%, and testing of
concentrations between 6.25% and 100% effluent using only five effluent
concentrations (6.25%, 12.5%, 25%, 50%, and 100%). Test precision shows
344
-------
little improvement as dilution factors are increased beyond 0.5 and declines
rapidly if smaller dilution factors are used. Therefore, USEPA recommends a
dilution factor of 0.5.
10.1.2.3 If the effluent is known or suspected to be highly toxic, a lower
range of effluent concentrations should be used (such as 25%, 12.5%, 6.25%,
3.12%, and 1.56%).
10.1.2.4 The volume of effluent required for the test using a 0.5 dilution
series is approximately 1800 ml. Prepare enough test solution at each
effluent concentration (approximately 800 ml) to provide 100 ml of test
solution for each of four (minimum of three) replicate test chambers and 400
ml for chemical analyses.
10.1.3 DILUTION WATERS
10.1.3.1. The formula for the enrichment for natural seawater is listed in
Table 1. Both EDTA and trace metals have been omitted. This formula should
be used for the 2-day exposure period, but it is not critical for the recovery
period. Since natural seawater quality can vary among laboratories, a more
complete nutrient medium (e.g., + EDTA) may result in faster growth (and
therefore faster cystocarp development) during the recovery period.
10.2 PREPARATION OF PLANTS FOR TEST
10.2.1 Once cultures are determined to be usable for toxicity testing (have
trichogynes and sori with spermatia), plant cuttings should be prepared for
the test, using fine-point forceps, with the plants in a little seawater in
a petri dish. For female plants, five cuttings, severed 7- to 10-mm from the
ends of the branch, should be prepared for each treatment chamber. Try to
be consistent in the number of branch tips on each cutting. For male
plants, one cutting, severed 2.0 to 3.0 cm from the end of the branch, is
prepared for each test chamber. Prepare the female cuttings first, to
minimize the chances of contaminating them with water containing spermatia
from the male stock cultures.
10.3. START OF TEST
10.3.1 On-site tests should be initiated within 24 h of sample collection.
The maximum holding time following retrieval of the sample from the sampling
device should not exceed 36 h for off-site toxicity tests unless permission is
granted by the permitting authority. In no case should the sample be used in
a test more than 72 h after sample collection.
10.3.2 Just prior to test initiation (approximately one h), a sufficient
quantity of the sample to make the test solution should be adjusted to the
test temperature (23 ± 1°C) and maintained at that temperature during the
addition of dilution water.
10.3.3 Set up and label four test chambers (minimum of three) per treatment
and controls.
345
-------
TABLE 1. NUTRIENTS TO BE ADDED TO NATURAL SEAWATER AND TO ARTIFICIAL
SEAWATER (GP2) DESCRIBED IN TABLE 2. THE CONCENTRATED NUTRIENT
STOCK SOLUTION IS AUTOCLAVED FOR 15 MIN (VITAMINS ARE
AUTOCLAVED SEPARATELY FOR 2 MIN AND ADDED AFTER THE NUTRIENT
STOCK SOLUTION IS AUTOCLAVED). THE pH OF THE SOLUTION IS
ADJUSTED TO APPROXIMATELY pH 2 BEFORE AUTOCLAVING TO MINIMIZE
THE POSSIBILITY OF PRECIPITATION.
Amount of Reagent Per Liter of Concentrated
Nutrient Stock Solution
Stock Solution For Stock Solution For
Culture Medium Test Medium
Nutrient Stock Solution1
Sodium Nitrate
(NaN03) 6.35 g 1.58 g
Sodium Phosphate
(NaH2P04 . H20) 0.64 g 0.16 g
Na2EDTA . 2 H20 133 mg
Sodium Citrate
(Na3C6H507 . 2 H20) 51 mg 12.8 mg
Iron2 9.75 mL 2.4 mL
Vitamins3 10 mL 2.5 mL
Add 10 mL of appropriate nutrient stock solution per liter of culture or
test medium.
2A stock solution of iron is made that contains 1 mg iron/mL. Ferrous or
ferric chloride can be used.
3A vitamin stock solution is made by dissolving 4.88 g thiamine HC1,
2.5 mg biotin, and 2.5 mg B12 in 500 mL deionized water. Adjust to
approximately pH 4 before autoclaving 2 min. It is convenient to
subdivide the vitamin stock into 10 mL volumes in test tubes prior to
autoclaving.
346
-------
TABLE 2. REAGENT GRADE CHEMICALS IN THE PREPARATION OF GP2 ARTIFICIAL SEAWATER
FOR USE IN CONJUNCTION WITH NATURAL SEAWATER FOR THE RED MACROALGA,
CHAMPIA PARVULA, CULTURING AND TOXICITY TESTING1'2'3'4'5'6'7
Compound Concentration Amount (g)
(g/L) Required for
20 L
1.
2.
3.
4.
5.
6.
7.
8.
9.
NaCl
Na2S04
KC1
KBr
Na2B407 . 10 H20
MgCl2 . 6 H20
CaCl2 . 2 H20
SrCl2 . 6 H20
NaHCO,
21.03
3.52
0.61
0.088
0.034
9.50
1.32
0.02
0.17
420.6
70.4
12.2
1.76
0.68
190.0
26.4
0.400
3.40
Modified GP2 from Spotte et al. (1984).
2The constituent salts and concentrations were taken from USEPA (1990b).
3The original formulation calls for autoclaving anhydrous and hydrated
salts separately to avoid precipitation. However, if the sodium
bicarbonate is autoclaved separately (dry), all of the other salts can be
autoclaved together. Since no nutrients are added until needed, autoclaving
is not critical for effluent testing. To minimize microalgal contamination,
the artificial seawater should be autoclaved when used for stock cultures.
Autoclaving (120°C) should be for a least 10 min for 1-L volumes, and 20 min
for 10-to-20 L volumes.
4Prepare in 10-L to 20-L batches.
A stock solution of 68 mg/mL sodium bicarbonate is prepared by autoclaving
it as a dry powder, and then dissolving it in sterile deionized water. For
each liter of GP2, use 2.5 mL of this stock solution.
6Effluent salinity adjustment to 30 °/oo can be made by adding the
appropriate amount of dry salts from this formulation, by using a
triple-strength brine prepared from this formulation, or by using a
100 °/oo salinity brine prepared from natural seawater.
7Nutrients listed in Table 1 should be added to the artificial seawater in
the same concentration described for natural seawater.
347
-------
10.3.4 Randomize the position of test chambers at the beginning of the test.
10.3.5 Fill the test chambers with 100 ml of control or treatment water
(28-to-32 °/oo). For reference toxicant tests, all test chambers can be
filled with control water and the toxicant added with a pipet. For toxicant
volumes exceeding 1 ml, adjust the amount of dilution water to give a final
volume of 100 ml.
10.3.6 Add five female branches and one male branch to each test chamber.
The toxicant must be present before the male plant is added.
10.3.7 Place the test chambers under cool white light (approx.
75 f£/m2/s, or 500 ft-c), with a photoperiod of 16 h light and 8 h
darkness. Maintain the temperature between 22 and 24°C. Check the
temperature by placing a laboratory or recording thermometer in a flask of
water among the test chambers. Record the temperature daily.
10.3.8 Gently hand swirl the chambers twice a day, or shake continuously at
100\rpm on a rotary shaker.
10.3.9 If desired, the media can be changed after 24 h.
10.4 TRANSFER OF PLANTS TO CONTROL WATER AFTER 48 H
10.4.1 Label the recovery vessels. These vessels can be almost any type of
container or flask containing 100 to 200 mL of seawater and nutrients (see
Tables 1 and 2). Smaller volumes can be used, but should be checked to make
sure that adequate growth will occur without having to change the medium.
10.4.2 With forceps, gently remove the female branches from test chambers
and place into recovery bottles. Add aeration tubes and foam stoppers.
10.4.3 Place the vessels under cool white light (at the same irradiance as
the stock cultures) and aerate for the 5-7 day recovery period. If a
shaker is used, do not aerate the solutions (this will enhance the water
motion).
10.5 TERMINATION OF THE TEST
10.5.1 At the end of the recovery period, count the number of cystocarps
(Figs. 6, 7, and 8) per female and record the data (Figure 12). Cystocarps
may be counted by placing females between the inverted halves of a
polystyrene petri dish or other suitable containers with a small amount of
seawater (to hold the entire plant in one focal plane). Cystocarps can be
easily counted under a stereomicroscope, and are distinguished from young
branches because they possess an apical opening for spore release (ostiole)
and darkly pigmented spores.
10.5.2 One advantage of this test procedure is that if there is uncertainty
about the identification of an immature cystocarp, it is necessary only to
aerate the plants a little longer in the recovery bottles. Within 24 to
48 h, the presumed cystocarp will either look more like a mature cystocarp or
348
-------
a young branch, or will have changed very little, if at all (i.e., an
aborted cystocarp). No new cystocarps will form since the males have been
removed, and the plants will only get larger. Occasionally, cystocarps will
abort, and these should not be included in the counts. Aborted cystocarps
are easily identified by their dark pigmentation and, often, by the
formation of a new branch at the apex.
11. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA
11.1 A summary of test conditions and test acceptability criteria is listed
in Table 3.
12. ACCEPTABILITY OF TEST RESULTS
12.1 A test is not acceptable if there is control mortality.
12.2 If plants fragment in the controls or lower exposure concentrations,
it may be an indication that they are under stress.
12.3 Control plants should average 10 or more cystocarps.
349
-------
1 mm
Figure 6. A mature cystocarp. In the controls and lower effluent
concentrations, cystocarps often occur in clusters of 10 or 12. From USEPA
(1987e).
young branch
cells
immature
cystocarp
Figure 7. Comparison of a very young branch and an immature cystocarp. Both
can have sterile hairs. Trichogynes might or might not be present on a young
branch, but are never present on an immature cystocarp. Young branches are
more pointed at the apex and are made up of larger cells than immature
cystocarps, and never have ostioles. From USEPA (1987e)-
1mm
Figure 8. An aborted cystocarp.
apex. From USEPA (1987e).
A new branch will eventually develop at the
350
-------
TABLE 3. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA FOR
THE RED MACROALGA, CHAHPIA PARVULA, SEXUAL REPRODUCTION TEST
WITH EFFLUENTS AND RECEIVING WATERS
1. Test type:
2. Salinity:
3. Temperature:
4. Photoperiod:
5. Light intensity:
6. Light source:
7. Test chamber:
8. Test solution volume:
9. Dilution water:
10. Test concentrations:
11. Dilution factor:
12. Number of replicate Chambers
per treatment:
13. Number of organisms
per test chamber:
14. Test duration:
15. Effects measured:
Static, non-renewal
30 °/oo ± 2 °/oo
23 ± 1°C
16 h light, 8 h dark
75 \i£/m2/s (500 ft-c)
Cool-white fluorescent lights
200 mL polystyrene cups, or
250 mL Erlenmeyer flasks
100 mL (minimum)
30 °/oo salinity natural seawater,
or a combination of 50% 30 °/oo
salinity natural seawater and 50%
30 °/oo salinity GP2 artificial
seawater
Effluents: Minimum of five effluent
concentrations and a control
Receiving waters: 100% receiving water
and a control
Effluents: > 0.5 series
Receiving waters: None, or > 0.5
series
4 (minimum of 3)
5 female branch tips and
1 male plant
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
351
-------
TABLE 3. SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA FOR
THE RED MACROALGA, CHAHPIA PARVULA, SEXUAL REPRODUCTION TEST
WITH EFFLUENTS AND RECEIVING WATERS (CONTINUED)
16. Test acceptability
criteria:
17. Sampling requirements:
18. Estimated maximum sample
volume required:
No mortality in controls; average 10
or more cystocarps of plants in
controls
One sample collected at test
initiation, and preferably used within
24 h of the time it is removed from the
sampling device (see Section 16,
Subsection 10.3.1).
2 L per test
352
-------
13. DATA ANALYSIS
13.1 GENERAL
13.1.1 Tabulate and summarize the data. A sample set of reproduction data
is listed in Table 4.
13.1.2 The endpoints of the red macroalga, Champia parvula, toxicity test are
based on the adverse effects on sexual reproduction. Statistically
significant differences in the mean number of cystocarps, yielding NOEC and
LOEC endpoints, are determined in most cases by a hypothesis test such as
Dunnett's Procedure (Dunnett, 1955) or Steel's Many-one Rank Test (Steel,
1959; Miller, 1981). The IC25 and IC50 are calculated using point estimation
techniques (see Section 9, Chronic Toxicity Test Endpoints and Data Analysis).
13.1.3 Formal statistical analysis of the data is outlined in Figure 9.
The response used in the analysis is the mean number of cystocarps per
replicate chamber. Separate analyses are performed for the estimation of the
NOEC and LOEC endpoints and for the estimation of the IC25 endpoint and the
IC50 endpoint. Concentrations that have exhibited no sexual reproduction
(less than 5% of controls) are excluded from the statistical analysis of the
NOEC and LOEC, but included in the estimation of the IC25 and IC50.
13.1.4 For the case of equal numbers of replicates across all concentrations
and the control, the evaluation of the NOEC and LOEC endpoints is made via a
parametric test, the Dunnett's Procedure, or a nonparametric test, the Steel's
Many-one Rank Test. The test for normality is the Shapiro-Wilk's Test and
Bartlett's Test is used to test for homogeneity of variance. Tests for
normality and homogeneity of variance are included in Appendix B. If either
of these tests fail, the nonparametric test, Steel's Many-one Rank Test is
used to determine the NOEC and LOEC endpoints. If the assumptions of
Dunnett's Procedure are met, the endpoints are determined by the parametric
test.
13.1.5 Additionally, if unequal numbers of replicates occur among the
concentration levels tested there are parametric and nonparametric
alternative analyses. The parametric analysis is a t-test with the Bonferroni
adjustment. The Wilcoxon Rank Sum Test with the Bonferroni adjustment is the
nonparametric alternative. For detailed information on the Bonferroni
adjustment, see Appendix D.
13.1.6 The statistical tests described here must be used with a knowledge
of the assumptions upon which the tests are contingent. The assistance of a
statistician is recommended for analysts who are not proficient in statistics.
353
-------
TABLE 4. DATA FROM THE RED MACROALGA, CHAMPIA PARVULA, EFFLUENT
TOXICITY TEST. CYSTOCARP COUNTS FOR INDIVIDUAL PLANTS
AND MEAN COUNT PER TEST CHAMBER FOR EACH EFFLUENT
CONCENTRATION1
Effluent Replicate
Concentration Test
(%) Chamber
A
Control B
C
A
0.8% B
C
A
1.3% B
C
A
2.2% B
C
A
3.6% B
C
A
6.0% B
C
A
10.0% B
C
Plant
1
19
19
17
10
12
12
10
6
4
1
7
3
2
3
0
1
1
0
0
1
2
2
20
12
25
16
10
9
0
4
4
2
9
2
1
4
4
0
2
4
0
0
1
3
24
21
18
11
6
9
3
4
2
5
9
2
1
6
3
0
1
3
0
0
0
4
7
11
20
12
9
13
5
8
6
4
4
0
5
4
1
0
0
1
0
0
0
5
18
23
16
11
10
8
4
4
4
0
6
0
0
2
3
0
0
3
_
0
0
Mean
Cystocarp
Count
17
17
19
12
9
10
4
5
4
2
7
1
1
3
2
0
0
2
0
0
0
.60
.20
.20
.00
.40
.20
.40
.20
.00
.40
.00
.40
.80
.80
.20
.20
.80
.20
.00
.20
.60
1Data provided by the Environmental Research Laboratory, U. S. Environmental
Protection Agency, Narragansett, Rhode Island.
354
-------
13.2
DATA
EXAMPLE OF ANALYSIS OF THE RED MACROALGA, CHAHPIA PARVULA, REPRODUCTION
13.2.1 In this example, the data, mean and standard deviation of the
observations at each concentration including the control are listed in Table
5. The data are plotted in Figure 10. As can be seen from the data in the
table, mean reproduction per chamber in the 10% effluent concentration is less
than 5% of the control. Therefore the 10% effluent concentration is not
included in the subsequent analysis.
TABLE 5. RED MACROALGA, CHAMPIA PARVULA, SEXUAL REPRODUCTION
DATA
Effluent Concentration (%)
Replicate Control
0.8
1.3
2.2
3.6
6.0 10.0
A
6
C
Meant?,)
sf
i
17.60
17.20
19.20
18
1.12
1
12.00 4.40 2.40 1.80 0.20 0.00
9.40 5.20 7.00 3.80 0.80 0.20
10.20 4.00 1.40 2.20 2.20 0.60
10.53 4.53 3.60 2.60 1.07 0.27
1.77 0.37 8.92 1.12 1.05 0.09
234567
13.2.2 Test for Normality
13.2.2.1 The first step of the test for normality is to center the
observations by subtracting the mean of all the observations within a
concentration from each observation in that concentration. The centered
observations are summarized in Table 6.
TABLE 6. CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S EXAMPLE
Effluent Concentration^)
Replicate Control
0.8
1.3
2.2
3.6
6.0
A
B
C
-0
-0
1
.40
.80
.20
1
-1
-0
.47
.13
.33
-0
0
-0
.13
.67
.53
-1
3
-2
.20
.40
.20
-0
1
-0
.80
.20
.40
-0.87
-0.27
1.13
355
-------
STATISTICAL ANALYSIS OF CHAMPIA PARVULA
SEXUAL REPRODUCTION TEST
REPRODUCTION DATA
MEAN CYSTOCARP COUNT
|
POINT ESTIMATION
"
ENPOINT ESTIMATE f
IC25, IC50 QHAPIRn-WILK
NORMAL DISTRIBUTION
i '
MOM NJORMA
'S TEST 1
BAH 1 Lh 1 1 S 1 tST ^- yt
HOMOGENEOUS VARIANCE
v
NO EQUAL NUMBER OF
REPLICATES?
YES
» 1
..
EQUAL NUMBER OF
REPLICATES?
YES
oT^rtl^TiVi DUNNETTS STEEL'S MANY-ONE WILCOXO
DUNrtnnL/NI TCOT DAK.II/ TCCT Tec
ADJUSTMENT TCST RANK TEST BrtMMB;^
1
ENDPOINT ESTIMATES
NOEC.LOEC
L DISTRIBUTION
30GENEOUS
\RIANCE
NO
I
N RANK SUM
.TWITH
SJI ADJUSTMENT
Figure 9. Flowchart for statistical analysis of the red macroalga,
Champia parvula, data.
356
-------
CONNECTS THE MEAN VALUE FOR EACH CONCENTRATION
to
en
0.0
REPRESENTS THE CRITICAL VALUE FOR DUNNETT'S TEST
(ANY MEAN REPRODUCTION BELOW THIS VALUE WOULD BE
SIGNIFICANTLY DIFFERENT FROM THE CONTROL)
0.8
1.3 2.2 3.6
EFFLUENT CONCENTRATION (*)
6.0
10.0
Figure 10. Plot of mean number of cystocarps per plant.
-------
13.2.2.2 Calculate the denominator, D, of the test statistic:
D = S (X,- - X)2
Where: X; = the ith centered observation
X = the overall mean of the centered observations
n = the total number of centered observations.
For this set of data, n = 18
X = J_(0.01) = 0.00
18
D = 28.7201
13.2.2.3 Order the centered observations from smallest to largest
YO) v(2) v
A A - ... A
Where XO) is the ith ordered observation. These ordered observations
are listed in Table 7.
TABLE 7. ORDERED CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S EXAMPLE
i
1
2
3
4
5
6
7
8
9
Xd>
-2.20
-1.20
-1.13
-0.87
-0.80
-0.80
-0.53
-0.40
-0.40
i
10
11
12
13
14
15
16
17
18
x
-0.33
-0.27
-0.13
0.67
1.13
1.20
1.20
1.47
3.40
13.2.2.4 From Table 4, Appendix B, for the number of observations, n,
obtain the coefficients ar a2, ..., ak where k is n/2 if n is even and
(n-l)/2 if n is odd. For the data in this example, n = 18, k = 9. The
values are listed in Table 8.
13.2.2.5 Compute the test statistic, W, as follows:
358
-------
W = 1 [ S a, (X(nM+1) X0) ]2
D i = l
The differences x(rvi+1) - X(i) are listed in Table 8. For this set
of data,
W = 1
(5.1425)2 = 0.921
28.7201
TABLE 8. COEFFICIENTS AND DIFFERENCES FOR SHAPIRO-MILK'S EXAMPLE
u(n-
1
2
3
4
5
6
7
8
9
0.4886
0.3253
0.2553
0.2027
0.1587
0.1197
0.0837
0.0496
0.0163
5.60
2.67
2.33
2.07
1.93
1.47
0.40
0.13
0.07
x<18)
X(17)
x<17)
X(13)
X(14)
X03)
X(12)
X<11>
X(10)
X(1)
- X(2)
- X(3)
- X(4)
X(5)
- X(6>
X(7)
X(8)
- x<9)
13.2.2.6 The decision rule for this test is to compare W with the
critical value found in Table 6, Appendix B. If the computed W is less
than the critical value, conclude that the data are not normally
distributed. For this example, the critical value at a significance
level of 0.01 and 18 observations (n) is 0.858. Since W = 0.921 is
greater than the critical value, the conclusion of the test is that the
data are normally distributed.
13.2.3 Test for Homogeneity of Variance
13.2.3.1 The test used to examine whether the variation in mean
cystocarp production is the same across all effluent concentrations
including the control, is Bartlett's Test (Snedecor and Cochran, 1980).
The test statistic is as follows:
B =
[ (
In S2 - E V,- In S*
Where: V,- =
degrees of freedom for each effluent concen-
tration and control, Vf = (n,. - 1)
359
-------
p = number of levels of effluent concentration
including the control
P
2
1=1
2 V,
C = 1 + [ 3(p-l)]'1 [ S 1/V, - ( S V,)'1 ]
In = loge
i = 1, 2, ..., p where p is the number of concentrations
including the control
n,- = the number of replicates for concentration i.
13.2.3.2 For the data in this example (See Table 5) all effluent
concentrations including the control have the same number of replicates
(n, = 3 for all i). Thus, V; = 2 for all i.
13.2.3.3 Bartlett's statistic is therefore:
P ,
B = [(12)ln(2.3917) - 2 I ln(S?)]/1.1944
= [12(0.8720) - 2(ln(1.12)+ln(1.77)+...+ln(1.05))]/1.1944
= (10.4640 - 4.0809)/1.1944
= 5.34
13.2.3.4 B is approximately distributed as chi-square with p - 1 degrees
of freedom, when the variances are in fact the same. Therefore, the
appropriate critical value for this test, at a significance level of 0.01
with five degrees of freedom, is 15.09. Since B = 5.34 is less than the
critical value of 15.09, conclude that the variances are not different.
13.2.4 Dunnett's Procedure
13.2.4.1 Calculations
To obtain an estimate of the pooled variance for the Dunnett's Procedure,
construct an ANOVA table as described in Table 9.
360
-------
TABLE 9. ANOVA TABLE
Source df Sum of Squares
(SS)
Between p - 1 SSB
Within N - p SSW
Mean Square(MS)
(SS/df)
S2 = SSB/(p 1)
S2 = SSW/(N - p)
Total N 1 SST
Where: p = number effluent concentrations including the control
N = total number of observations n1 + n? ... +n
n,- = number of observations in concentration i
P , ,
SSB = S T,/n, - GYM Between Sum of Squares
P ni > ,
SST = S X Yf: - GYN Total Sum of Squares
1-1 j-1
SSW = SST - SSB Within Sum of Squares
P
G = the grand total of all sample observations, G = S T,.
1 = 1
Tj = the total of the replicate measurements for
concentration "i"
Y,-: = the jth observation for concentration "i" (represents
the mean (across plants) number of cystocarps for
effluent concentration i in test chamber j)
13.2.4.2 For the data in this example:
n, = n, = n3 = n4 = n5 = n6 = n7 = 3
N
T. = Y.. + Y12 + Y13 - 17.6 + 17.2 + 19.2 = 54
T2 = Y, + Ylf + Y* = 12.0 + 9.4 + 10.2 = 31.6
TS = Y, + Y" + Y" = 4.4 + 5.2 + 4.0 = 13.6
TA = Yf. + Y,? + Y,, = 2.4 + 7.0 + 1.4 = 10.8
TS = Y* + Y^ + Y5, = 1.8 + 3.8 + 2.2 = 7.8
T6 = Y« + Y62 + Y63 - 0.2 + 0.8 + 2.2 = 3.2
G = T, + T2 + T3+ T4 + T5 + T6= 121.0
361
-------
p , ,
SSB = S T,-2/,. G2/N
i = l
« _L_ (4287.24) - (121.O)2 = 615.69
3 18
"
>
2
SST = 2 S Y, - G/N
i-1 j-1
= 1457.8 (121. O)2 = 644.41
18
SSW = SST SSB = 644.41 - 615.69 = 28.72
Sg = SSB/(p-l) = 615.69/(6-l) = 123.14
Sy = SSW/(N-p) = 28.72/(18-6) = 2.39
13.2.4.3 Summarize these calculations in the ANOVA table (Table 10)
TABLE 10. ANOVA TABLE FOR DUNNETT'S PROCEDURE EXAMPLE
Source
Between
Within
Total
df
5
12
17
Sum of Squares
(SS)
615.69
28.72
644.41
Mean Square(MS)
(SS/df)
123.14
2.39
13.2.4.4 To perform the individual comparisons, calculate the t
statistic for each concentration, and control combination as follows;
*f
Su
362
-------
Where: Y_. = mean number of cystocarps for effluent concentration i
Y1 = mean number of cystocarps for the control
Su = square root of within mean square
n1 = number of replicates for the control
ni = number of replicates for concentration i
13.2.4.5 Table 11 includes the calculated t values for each
concentration and control combination. In this example, comparing the
0.8% concentration with the control the calculation is as follows:
( 18 - 10.53 )
t2 = = 5.90
[ 1.55 / (1/3) + (1/3) ]
TABLE 11. CALCULATED T-VALUES
Effluent Concentration(%) i t,-
0.8
1.3
2.2
3.6
6.0
2
3
4
5
6
5.90
10.64
11.38
12.17
13.38
13.2.4.6 Since the purpose of this test is to detect a significant
reduction in cystocarp production, a one-sided test is appropriate.
The critical value for this one-sided test is found in Table 5,
Appendix C. For an overall alpha level of 0.05, 12 degrees of freedom
for error and five concentrations (excluding the control) the critical
value is 2.50. Mean cystocarp production for concentration "i" is
considered significantly less than mean cystocarp production for the
control if t,- is greater than the critical value. Therefore, all
effluent concentrations in this example have significantly lower
cystocarp production than the control. Hence the NOEC is 0% and the LOEC
is 0.8%.
13.2.4.7 To quantify the sensitivity of the test, the minimum
significant difference (MSD) that can be statistically detected may be
calculated.
MSD = d Sw / (1/n,) + (1/n)
Where: d = the critical value for the Dunnett's procedure
Sw = the square root of the within mean square
n = the common number of replicates at each concentration
(this assumes equal replication at each concentration)
n, = the number of replicates in the control.
363
-------
13.2.4.8 In this example,
MSD = 2.50 (1.55) / (1/3) + (1/3)
= 2.50 (1.55)(.8165)
= 3.16
13.2.4.9 Therefore, for this set of data, the minimum difference that
can be detected as statistically significant is 3.16 cystocarps.
13.2.4.10 This represents a 17.6% reduction in cystocarp production from
the control.
13.3 Calculation of the 1C
13.3.1 The sexual reproduction data in Table 5 are utilized in this example.
Table 12 contains the mean number of cystocarps for each effluent
concentration. As can be seen, the observed means are monotonically non-
increasing with respect to concentration. Therefore, it is not necessary to
smooth the means prior to calculating the 1C. Refer to Figure 10 for a plot
of the response curve.
TABLE 12. RED MACROALGA, CHAMPIA PARVULA,
MEAN NUMBER OF CYSTOCARPS
Effluent
Cone. MJ
i (mg)
Control
0.8
1.3
2.2
3.6
6.0
10.0
1
2
3
4
5
6
7
18.00
10.53
4.53
3.60
2.60
1.07
0.27
13.3.2 An IC25 and IC50 can be estimated using the Linear Interpolation
Method. A 25% reduction in mean number of cystocarps, compared to the
controls, would result in a mean number of 13.50 cystocarps, where M^l-p/100)
= 18.00(1-25/100). A 50% reduction in mean number of cystocarps, compared to
the controls, would result in a mean number of 9.00 cystocarps. Examining the
means and their associated concentrations (Table 12), the response, 13.50, is
bracketed by C, = 0.0% effluent and C, = 0.8% effluent. The response, 9.00,
is bracketed by C2 = 0.8% effluent and C3 = 1.3% effluent.
13.3.3 Using Equation 1 from Appendix I, the estimate of the IC25 is
calculated as follows:
364
-------
ICp = Cd + [M^l p/100) Mj] (CJt1 - C,)
(MJ+1 Mj)
IC25 = 0.0 + [18.00(1 - 25/100) 18.00] (0.8 0.0)
(10.53 18.00)
= 0.5%.
13.3.4 Using Equation 1 from Appendix I, the estimate of the IC50 is
calculated as follows:
ICp = Cj + [Mjd p/100) - MJ (CJ41 - C,)
(MJ+1 - MJ
IC50 = 0.8 + [18.00(1 50/100) 10.53] (1.3 - 0.8)
(4.53 - 10.53)
= 0.9 %.
13.3.5 When the Bootstrap program (BOOTSTRP) was used to analyze this set of
data, requesting 80 resamples, the mean estimate of the IC25 was .4882%, with
a standard deviation of 0.0440% (coefficient of variation = 9.0%). The
empirical 95.0% confidence interval for the true mean was (0.4267%, 0.5758%).
The BOOTSTRP computer program output for the IC25 for this data set is shown
in Figure 11.
13.3.6 When the Bootstrap program (BOOTSTRP) was used to analyze this set of
data, requesting 80 resamples, the mean estimate of the IC50 was .9301%, with
a standard deviation of 0.0437% (coefficient of variation = 4.7%). The
empirical 95.0% confidence interval for the true mean was (0.8529 %, 0.9990%)
The BOOTSTRP computer program output for the IC50 for this data set is shown
in Figure 12.
365
-------
THE NUMBER OF RESAMPLES IS 80.
*** LISTING OF GROUP CONCENTRATIONS (% EFF.) AND RESPONSE MEANS ***
CONC. (%EFF) RESPONSE MEAN MEAN AFTER POOLING
0.000 18.000 18.000
0.800 10.533 10.533
1.300 4.533 4.533
2.200 3.600 3.600
3.600 2.600 2.600
6.000 1.067 1.067
10.000 0.267 0.267
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
FROM THE INPUT SAMPLE IS 0.4821.
************************************************************
* BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY *
* OF THE ESTIMATED ICp *
************************************************************
THE MEAN OF THE BOOTSTRAP ESTIMATES IS 0.4882.
THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS 0.0440.
AN EMPIRICAL 95.0% CONFIDENCE INTERVAL FOR THE
BOOTSTRAP ESTIMATE IS (0.4267, 0.5758).
Figure 11. BOOTSTRP program output for the IC25.
366
-------
THE NUMBER OF RESAMPLES IS 80.
*** LISTING OF GROUP CONCENTRATIONS (X EFF=) AND RESPONSE MEANS ***
CONC. (XEFF) RESPONSE MEAN MEAN AFTER POOLING
0.000 18.000 18.000
0.800 10.533 10.533
1.300 4.533 4.533
2.200 3.600 3.600
3.600 2.600 2.600
6.000 1.067 1.067
10.000 0.267 0.267
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
FROM THE INPUT SAMPLE IS 0.9278.
* BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY *
* OF THE ESTIMATED ICp *
************************************************************
THE MEAN OF THE BOOTSTRAP ESTIMATES IS 0.9301.
THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS 0.0437.
AN EMPIRICAL 95.0% CONFIDENCE INTERVAL FOR THE
BOOTSTRAP ESTIMATE IS (0.8529, 0.9990).
Figure 12. BOOTSTRP program output for the IC50.
367
-------
14. PRECISION AND ACCURACY
14.1 PRECISION
14.1.1 The single-laboratory precision data from six tests with copper
sulfate (CU) and six tests with sodium dodecyl sulfate (SDS) are listed
in Tables 13-16. The NOECs with CU differed by only one concentration
interval (factor of two), showing good precision. The precision of the
first four tests with SDS was somewhat obscured by the choice of toxicant
concentrations, but appeared similar to that of CU in the last two tests.
The IC25 and IC50 are indicated in Tables 13-16. The coefficient of
variation, based on the IC25 for these two reference toxicants in natural
seawater and a mixture of natural seawater and GP2, ranged from 59.6% to
69.0%, and for the IC50, ranged from 22.9% to 43.7%.
14.1.2 The multilaboratory precision of the test has not yet been
determined.
14.2 ACCURACY
14.2.1 The accuracy of toxicity tests cannot be determined.
368
-------
TABLE 13. SINGLE-LABORATORY PRECISION OF THE RED MACROALGA, CHAMPIA PARVULA,
REPRODUCTION TEST PERFORMED IN A 50/50 MIXTURE OF NATURAL SEAWATER
AND GP2 ARTIFICIAL SEAWATER, USING GAMETES FROM ADULTS CULTURED IN
NATURAL SEAWATER. THE REFERENCE TOXICANT USED WAS COPPER (CU)
SULFATE1'2'3'4'5
Test
Number
1
2
3
4
5
6
n:
Mean:
CV(%):
NOEC
(H9/L)
1.0
1.0
1.0
1.0
0.5
0.5
6
NA
NA
IC25
(H9/L)
1.67
1.50
0.69
0.98
0.38
0.38
6
0.93
59.6
IC50
(ugA)
2.37
1.99
1.53
1.78
0.76
0.75
6
1.5
43.7
1Data from USEPA (1989a) and USEPA (1991a).
2Tests performed by Glen Thursby and Mark Tagliabue, Environmental
Research Laboratory, U. S. Environmental Protection Agency,
Narragansett, Rhode Island. Tests were conducted at a temperature
of 22 C, in 50/50 GP2 and natural seawater at a salinity of
30 °/oo.
3Copper concentrations were: 0.5, 1.0, 2.5, 5.0, 7.5, and
1.0 jig/L.
NOEC Range: 0.5 1.0 jig/L (this represents a difference of one exposure
concentration).
5For a discussion of the precision of data from chronic toxicity
tests see Section 4, Quality Assurance.
369
-------
TABLE 14. SINGLE-LABORATORY PRECISION OF THE RED MACROALGA, CHAHPIA PARVULA,
REPRODUCTION TEST PERFORMED IN A 50/50 MIXTURE OF NATURAL SEAWATER
AND GP2 ARTIFICIAL SEAWATER, USING GAMETES FROM ADULTS CULTURED IN
NATURAL SEAWATER. THE REFERENCE TOXICANT USED WAS SODIUM DODECYL
SULFATE (SOS)1'2'3'4'5
Test
Number
1
2
3
4
5
6
7
8
9
n:
Mean:
CV(%):
NOEC
(mg/L)
<0.80
0.48
<0.48
<0.48
0.26
0.09
0.16
0.09
<0.29
5
NA
NA
IC25
(mg/L)
0.6
0.7
0.4
0.2
0.2
0.1
0.2
0.1
0.3
9
0.31
69.0
IC50
(mg/L)
0.3
0.6
0.2
0.4
0.5
0.3
0.3
0.2
0.4
9
0.36
37.0
1Data from USEPA (1989a) and USEPA (1991a).
2Tests performed by Glen Thursby and Mark Tagliabue, Environmental
Research Laboratory, U. S. Environmental Protection Agency,
Narragansett, Rhode Island. Tests were conducted at a temperature
of 22*C, in 50/50 GP2 and natural seawater at a salinity of
30 °/oo.
3SDS concentrations for Test 1 were: 0.8, 1.3, 2.2, 3.6, 6.0, and
10.0 mg/L. SDS concentrations for Tests 2, 3, and 4 were: 0.48,
0.8, 1.3, 2.2, 3.6, and 6.0 mg/L. SDS concentrations for Tests 5
and 6 were: 0.09, 0.16, 2.26, 0.43, 0.72, and 1.2 mg/L.
NOEC Range: 0.09 0.48 mg/L (this represents a difference of two exposure
concentrations).
5For a discussion of the precision of data from chronic toxicity
tests see Section 4, Quality Assurance.
370
-------
TABLE 15. SINGLE-LABORATORY PRECISION OF THE RED MACROALGA, CHAMPIA PARVULA,
REPRODUCTION TEST IN NATURAL SEAWATER (30 °/oo SALINITY). THE
REFERENCE TOXICANT USED WAS COPPER (CU) SULFATE1'2'3
Test
1
2
3
4
n:
Mean:
CV(%) :
NOEC
1.00
0.50
0.50
0.50
4
NA
NA
Cu (uQ/L)
IC25
2.62
0.71
2.83
0.99
4
1.79
61.09
IC50
4.02
1.66
3.55
4.15
4
3.35
34.45
1Data from USEPA (1991a).
2Copper concentrations were 0.5, 1.0, 2.5, 5.0, 7.5, and 10 jig/L.
Concentrations Cu were made from a 100 jig/mL CuS04 standard obtained fron
Inorganic Ventures, Inc., Brick, New Jersey.
'Prepared by Steven H. Ward and Glen Thursby, Environmental Research
Laboratory, U. S. Environmental Protection Agency, Narragansett, Rhode
Island.
371
-------
TABLE 16. SINGLE-LABORATORY PRECISION OF THE RED MACROALGA, CHAMPIA PARVULA,
REPRODUCTION TEST IN NATURAL SEAWATER (30 °/oo SALINITY). THE
REFERENCE TOXICANT USED WAS SODIUM DODECYL SULFATE (SDS)1'2'3
SDS fma/L)
Test NOEC
1 0.60
2 0.60
3 0.30
4 0.15
n: 4
Mean: NA
CV(%): NA
IC25
0.05
0.48
0.69
0.60
4
0.46
62.29
IC50
0.50
0.81
0.89
0.81
4
0.75
22.92
1Data from USEPA (1991a).
SDS concentrations were 0.0375, 0.075, 0.15, 0.03, 0.60, and 1.20 mg/L.
Concentrations of SDS were made from a 44.64 + 3.33 mg/mL standard obtained
from the Environmental Monitoring Systems Laboratory, U. S. Environmental
Protection Agency, Cincinnati, Ohio.
Prepared by Steven H. Ward and Glen Thursby, Environmental Research
Laboratory, U. S. Environmental Protection Agency, Narragansett, Rhode
Island.
372
-------
Figure 13. Data sheet for the red macroalga, Champia parvula. sexual
reproduction test. Receiving water summary sheet .
SITE
COLLECTION DATE_
TEST DATE
LOCATION
INITIAL
SALINITY
FINAL
SALINITY
SOURCE OF SALTS FOR
SALINITY ADJUSTMENT*
I.e. natural seawater, GP2 brine, GP2 salts, etc.
(include some indication of amount)
COMMENTS:
1From USEPA (1987e).
373
-------
Figure 14. Data sheet for the red macroalga, Champia parvula, sexual
reproduction test. Cystocarp data sheet1.
COLLECTION DATE
EXPOSURE BEGAN (date)_
RECOVERY BEGAN (date)_
COUNTED (date)
EFFLUENT OR TOXICANT
TREATMENTS (% EFFLUENT, ^G/L, or REC. WATER SITES)
REPLICATES
CONTROL
A 1
2
3
4
MEAN
B 1
2
3
4
MEAN
C 1
2
3
4
MEAN
OVERALL
MEAN
Temperature
Salinity
Light
Source of Dilution Water
Vrom USEPA (1987e)
374
-------
SELECTED REFERENCES
Anderson, J. W., 0. M. Neff, B. A. Cox, H. E. Tatem, and G. H. Hightower.
1974. Characteristics of dispersions and water-soluble extracts of
crude and refined oils and their toxicity to estuarine crustaceans and
fish. Mar. Biol. 27:75-88.
Anonymous. 1979. Test 6: Mysidopsis bahia life cycle. Federal Register
44(53):16291.
APHA. 1989. Standard Methods for the Examination of Water and Wastewater.
17th Ed. American Public Health Association, Washington, DC.
ASTM. 1987. Standard practice for using brine shrimp nauplii as food for
test animals in aquatic toxicology. ASTM E 1203-87. American Society
for Testing and Materials, Philadelphia, Pennsylvania.
ASTM. 1990. Standard guide for conducting life cycle toxicity tests with
salt water mysids. ASTM E 1191-90. American Society for Testing and
Materials, Philadelphia, Pennsylvania.
Bartlett, M. S. 1937. Some examples of statistical methods of research in
agriculture and applied biology. J. Royal Statist. Soc. Suppl.
4:137-183.
Battelle. 1988. A robust statistical method for estimating effects
concentrations in short-term fathead minnow toxicity tests.
Battelle Washington Environmental Program Office, Washington, DC.
Bay, S. M., P. S. Oshida, and K. D. Jenkins. 1983. A simple new bioassay
based on echinochrome synthesis by larval sea urchins. Mar. Environ.
Res. 8:29-39.
Beck, A. D., and D. A. Bengtson. 1982. International study on Artemia XXII:
Nutrition in aquatic toxicology Diet quality of geographical strains
of the brine shrimp, Artemia. In: Pearson, J.G., R.B. Foster, and W.E.
Bishop, eds., ASTM STP 766, American Society for Testing and Materials,
Philadelphia, Pennsylvania, pp. 161-169.
Beck, A. D., D. A. Bengtson, and W. H. Howell. 1980. International study on
Artemia. V. Nutritional value of five geographical strains of Artemia:
Effects of survival and growth of larval Atlantic silversides, Menidia
menidia. In: Persoone, G., P. Sorgeloos, D. A. Roels, and E. Jaspers.
eds., The brine shrimp, Artemia. Vol. 3, Ecology, culturing, use in
aquaculture. Universa Press, Wetteren, Belgium, pp. 249-259.
Beckett, D. C., and P. A. Lewis. 1982. An efficient procedure for slide
mounting of larval chironomids. Trans. Amer. Fish. Soc. 101(1):96-99.
Bengtson, D. A. 1984. Resource partitioning by Menidia menidia and Menidia
beryllina (Osteichthyes: Atherinidae). Mar. Ecol. Prog. Ser. 18:21-32.
375
-------
Bengtson, D. A. S., A. D. Beck, S. M. Lussier, D. Migneault, and C. E. Olney.
1984. International study on Artemia. XXXI. Nutritional effects in
toxicity tests: use of different Artemia geographical strains.
IniPersoone, G., E. Jaspers, and C. Claus, eds., Ecotoxicological
testing for the marine environment, Vol. 2. State Univ. Ghent and
Inst. Mar. Sci. Res., Bredene, Belgium, pp. 399-417.
Benoit, D. A., F. A. Puglisi, and D. L. Olson. 1982. A fathead minnow,
Pimephales promelas, early life state toxicity test method evaluation
and exposure to four organic chemicals. Environ. Pollut. (Series A)
28:189-197.
Bidwell, J.P.,
methods.
and S. Spotte. 1985. Artificial Seawaters: formulas and
Jones and Barlett, Publ., Boston, Massachusetts. 349 pp.
Bigelow, H. B., and W. C. Schroeder.
U.S. Fish Wildl. Serv., Fish.
1953. Fishes of the Gulf of Maine.
Bull. 53:1-577.
Birge, W. J., J. A. Black, and B. A. Ramey. 1981. The reproductive
toxicology of aquatic contaminants. Hazard assessments of chemicals,
current developments, Vol. 1, Academic Press, Inc., p. 59-114.
Birge, W. J., J. A. Black, and A. G. Westerman. 1979. Evaluation of aquatic
pollutants using fish and amphibian eggs as bioassay organisms.
National Academy of Sciences, Washington, DC, p. 108-118.
Birge, W. J., J. A. Black, and A. G. Westerman. 1985. Short-term fish and
amphibian embryo-larval tests for determining the effects of toxicant
stress on early life stages and estimating chronic values for single
compounds and complex effluents. Environ. Tox. Chem. (4):807-821.
Birge, W. J., J. A., Black, A. G., Westerman, and B. A. Ramey. 1983. Fish
and amphibian embryos - A model system for evaluating teratogenicity.
Fundam. Appl. Toxicol. 3: 237-242.
Birge, W. J., and R. A. Cassidy. 1983. Importance of structure-activity
relationships in aquatic toxicology. Fundam. Appl. Toxicol. 3:359-368.
Black, J. A., W. J. Birge, A. G. Westerman, and P. C. Francis. 1983.
Comparative aquatic toxicology of aromatic hydrocarbons. Fundam. Appl.
Toxicol. 3: 353-358.
Blaxter, J. H. S., F. S. Russell, and M. Yonge, eds. 1980. The biology of
mysids and euphausiids. Part 1. The biology of the mysids. Adv. Mar.
Biol. 18:1-319.
Borthwick, P. W., J. M. Patrick, Jr., and D. P. Middaugh. 1985. Comparative
acute sensitivities of early life stages of atherinid fishes to
chloropyrifos and thiobencarb. Arch. Environm. Contam. Toxicol.
14:465-473.
376
-------
Bower, C. E. 1983. The basic marine aquarium. Charles C. Thomas, Publ.,
Springfield, Illinois.
Breteler, R. J., J.W. Williams, and R.L. Buhl. 1982. Measurement of chronic
toxicity using the opossum shrimp Mysidopsis bahia. Hydrobiol.
93:189-194.
Buikema, A. L., Jr., B. R. Niederlehner, and J. Cairns, Jr. 1981. The
effects of simulated refinery effluent and its components on the
estuarine crustacean, Mysidopsis bahia. Arch. Environm. Contam.
Toxicol. 10:231-240.
Buikema, A. L., B. R. Niederlehner, and J. Cairns. 1982. Biological
monitoring. Part IV. Toxicity Testing. Water Res. 16:239-262.
Chernoff, B., J. V. Conner, and C. F. Byran. 1981. Systematics of the
Mem'dia beryllina complex (Pices:Atherinidae) from the Gulf of Mexico
and its tributaries. Copeia 2:319-335.
Conover, W. J. 1980. Practical nonparametric statistics. Second edition.
John Wiley and Sons, New York. pp. 466-467.
Cripe, G. M., D. R. Nimmo, and T. L. Hamaker. 1981. Effects of two
organophosphate pesticides on swimming stamina of the mysid Mysidopsis
bahia. In: Vernberg, F. J., A. Calabrese, F. P. Thurberg, and W. B.
Vernberg, eds., Biological monitoring of marine pollutants. Academic
Press, New York, pp 21-36.
Davey, E. W., J. H. Gentile, S. J. Erickson, and P. Betzer. 1970. Removal of
trace metals from marine culture media. Limnol. Oceanogr. 15:486-488.
DeGraeve, G. M., J. D. Cooney, T. L. Pollock, N. G. Reichenbach, J. H. Dean,
M. D. Marcus, and D. 0. Mclntyre. 1989. Fathead minnow 7-day test:
round robin study. Intra- and interlaboratory study to determine the
reproducibility of the seven-day fathead minnow larval survival and
growth test. Battelle Columbus Division, Columbus, Ohio.
DeGraeve, G. M., W.H. Clement, and M.F. Arthur. 1989. A method for
conducting laboratory toxicity degradation evaluations of complex
effluents. Battelle Columbus Division, Columbus, Ohio. 22 pp.
DeWoskin, R. S. 1984. Good laboratory practice regulations: a comparison.
Research Triangle Institute, Research Triangle Park, North Carolina, 63
PP.
Dinnel, P., Q. Stober, J. Link, M. Letourneau, W. Roberts, S. Felton, and
R. Nakatani. 1983. Methodology and validation of a sperm cell toxicity
test for testing toxic substances in marine waters. Final Report. Grant
R/TOX. FRI-UW-83. University of Washington Sea Grant Program in
cooperation with U. S. Environmental Protection Agency. 208 pp.
377
-------
Dixon, W. J., and F. J. Massey, Jr. 1983. Introduction to statistical
analysis. Fourth edition. McGraw Hill, New York.
Dorn, P.B., and J.H. Rogers. 1989. Variability associated with
identification of toxics in National Pollutant Discharge Elimination
System (NPDES) effluent toxicity tests. Environ. Toxicol. Chem. 8:893-
902.
Downs, T. R., and W. W. Wilfinger. 1983. Fluorometric quantification of DNA
in cells and tissue. Anal. Biochem. 131: 538-547.
Draper, N. R., and J. A. John. 1981. Influential observations and outliers
in regression. Technometrics 23:21-26.
Dunnett, C. W. 1955. Multiple comparison procedure for comparing several
treatments with a control. J. Amer. Statist. Assoc. 50:1096-1121.
Dunnett, C. W. 1964. New table for multiple comparisons with a control.
Biometrics 20:482.
Dye, J. E. 1980. The production and efficient use of freshly hatched brine
shrimp nauplii (Artemia) in the larval rearing of marine fish at the
hatcheries of the British White Fish Authority. In: Persoone, G., P.
Sorgeloos, D. A. Roels, and E. Jaspers, eds., The brine shrimp, Artemia.
Vol. 3, Ecology, culturing, use in aquaculture. Universa Press,
Wetteren, Belgium, pp. 271-276.
Efron, B. 1982. The Jackknife, the Bootstrap, and other resampling
plans. CBMS 38, Soc. Industr. Appl. Math., Philadelphia, Pennsylvania.
Emerson, K., R. C. Russo, R.E. Lund, and R.V. Thurston. 1975. Aqueous
ammonia equilibrium calculations; effect of pH and temperature. J.
Fish. Res. Bd. Can. 32(12): 2379-2383.
Farrell, D. H. 1979. Guide to the shallow-water mysids from Florida. Fla.
Dept. Environ. Reg., Techn. Ser. 4(l):l-69.
FDA. 1978. Good laboratory practices for nonclinical laboratory studies.
Part 58, Fed. Reg. 43(247):60013-60020, December 22, 1978.
Finney, D. J. 1948. The Fisher-Yates test of significance in 2X2 contingency
tables. Biometrika 35:145-156.
Finney, D. J. 1971. Probit analysis. Third Edition. Cambridge Press, New
York. 668 pp.
Finney, D. J. 1978. Statistical method in biological assay. 3rd ed. Charles
Griffin & Co. Ltd, London. 508 pp.
Finney, D. J. 1985. The median lethal dose and its estimation. Arch.
Toxicol. 56:215-218.
378
-------
Fotheringham, N., and S. L. Brunenmeister. 1975.
of the northwestern Gulf coast. Gulf Publ.
Common marine invertebrates
Co., Houston, Texas.
Fujita, S., T. Watanabe, and C. Kitajima. 1980. Nutritional quality of
Artemia from different localities as a living feed for marine fish
from the viewpoint of essential fatty acids. In: Persoone, G., P.
Sorgeloos, D. A. Roels, and E. Jaspers, eds., The brine shrimp,
Artemia. Vol. 3, Ecology, culturing, use in aquaculture. Universa
Press, Wetteren, Belgium, pp. 277-290.
Gast, M. H., and W. A. Brungs. 1973. A procedure for separating eggs of the
fathead minnow. Prog. Fish. Cult. 35:54.
Gentile, J. H., S. M. Gentile, N. G. Hairston, Jr., and B. K. Sullivan. 1982.
The use of life-tables for evaluating the chronic toxicity of pollutants
to Mysidopsis bahia. Hydrobiol. 93:179-187.
Gentile, J. H., S. M. Gentile, G. Hoffman, J. F. Heltshe, and N. G. Hairston,
Jr. 1983. The effects of chronic mercury expsoure on the survival,
reproduction and population dynamics of Mysidopsis bahia. Environ.
Toxicol. Chem. 2:61-68.
Gentile, S. M., J. H. Gentile, J. Walker, and J. F. Heltshe. 1982. Chronic
effects of cadmium on two species of mysid shrimp: Mysidopsis bahia and
M. bigelowi. Hydrobiol. 93:195-204.
Goodman, L. R., D. J. Hansen, D. P. Middaugh, G. M. Cripe, and J. C. Moore.
1985. Method for early life-stage toxicity tests using three atherinid
fishes and results with chloropyrifos. In: Cardwell, R. D., R. Purdy,
and R. C. Bahner, eds., Aquatic Toxicology and Hazard Evaluation, ASTM
STP 854, American Society for Testing and Materials, Philadelphia,
Pennsylvania, pp. 145-154.
Hall, W.S., J.B. Patoczka, R.J. Mirenda, B.A. Porter, and E. Miller. 1989.
Acute toxicity of industrial surfactants to Mysidopsis bahia. Arch.
Environ. Contam. Toxicol. 18:765-772.
Hamilton, M.A., R.C. Russo, and R.V. Thurston. 1977. Trimmed Spearman-Karber
method for estimating median lethal concentrations. Environ. Sci. Tech.
11(7):714-719.
Heard, R. W. 1982. Guide to the common tidal marsh invertebrates of the
northeastern Gulf of Mexico. Publ. No. MASGP-79-004,
Mississippi-Alabama Sea Grant Consortium, Ocean Springs, Mississippi.
Hildebrand, S. F. 1922. Notes on habits and development of eggs and larvae
of the silversides Menidia menidia and Menidia beryllina. Bull. U. S.
Bur. Fish. 38:113-120.
Hildebrand, S. F., and W. C. Schroeder. 1928. Fishes of Chesapeake Bay.
Bull. U. S. Fish. 43(l):366pp.
379
-------
Hodges, J. L., Jr., and E. L. Lehmann. 1956. The efficiency of some
nonparametric competitors of the t-test. Ann. Math. Statist.
31:625-642.
Home, J. D., M. A. Swirsky, T. A. Hollister, B. R. Oblad, and J. H. Kennedy.
1983. Aquatic toxicity studies of five priority pollutants, Contract
Report, EPA Contract No. 68-01-6201, U. S. Environmental Protection
Agency, Washington, DC.
Hutton, C. H., P. F. DeLisle, M. H. Roberts, and D. A. Hepworth. 1986.
Chrysaora quinquecirrha: a predator on mysids (Mysidopsis bahia) in
culture. Progr. Fish-Cult. 48:154-155.
Jackim, E., and D. Nacci. 1986. Improved sea urchin DMA-based embryo growth
toxicity test. Environ. Toxicol. Chem. 5:561-565.
Jensen, J. P. 1958. The relation between body size and number of eggs in
marine malacostrakes. Meddr. Danm. Fisk.-og Havunders 2:1-25.
Jensen, A. L. 1972. Standard error of LC50 and sample size in fish
bioassays. Water. Res. 6:85-89.
Jensen, A. 1984a. Marine ecotoxicological tests with seaweeds. In:
Personne, G., E. Jaspers, and C. Claus, eds., Ecotoxicological testing
for the marine environment. Vol. 1. State University of Ghent and
Institute for Marine Scientific Research, Bredene, Belgium, pp.
181-193.
Jensen, A. 1984b. Marine ecotoxicological tests with phytoplankton. In:
Personne, G., E. Jaspers, and C. Claus, eds., Ecotoxicological testing
for the marine environment. Vol. 1. State University of Ghent and
Institute for Marine Scientific Research, Bredene, Belgium, pp.
195-213.
Johns, D. M., W. J. Berry, and W. Walton. 1981. International study on
Artemia. XVI. Survival, growth and reproductive potential of the mysid
Mysidopsis bahia Molenock fed various geographical strains of the
brine shrimp Artemia. J. Exp. Mar. Biol. Ecol. 53:209-219.
Johns, D. M., M. E. Peters, and A. D. Beck. 1980. International study on
Artemia. VI. Nutritional value of geographical and temporal strains of
Artemia: effects on survival and growth of two species of Brachyuran
larvae. In: Persoone, G., P. Sorgeloos, D. A. Roels, and E. Jaspers.
eds., The brine shrimp, Artemia. Vol. 3, Ecology, culturing, use in
aquaculture. Universa Press, Wetteren, Belgium, pp. 291-304.
Johnson, M. S. 1975. Biochemical systematics of the atherinid genus Menidia
Copeia 4:662-691.
Jop, K.M., J. H. Rogers, Jr., P. B. Dorn, and K. L. Dickson. 1986. Use of
hexavalent chromium as a reference toxicant in aquatic toxicity tests.
In: T.M. Poston, and R. Purdy, eds., Aquatic Toxicology and
380
-------
Environmental Fate, ASTM STP 921, American Society for Testing and
Materials, Philadelphia, Pennsylvania, pp. 390-403.
Kenaga, E. E. 1982. The use of environmental toxicology and chemistry data
in hazard assessment: Progress, needs, challenges. Environ. Toxicol.
Chem. 1:69-79.
Kenaga, E. E., and R. J. Moolenaar. 1976. Fish and Daphnia toxicity as
surrogates for aquatic vascular plants and algae. Environ. Sci.
Technol. 13:1479-1480.
Kester, D. R., I. W. Dredall, D. N. Connors, and R. M. Pytokowicz. 1967.
Preparation of artificial seawater. Limnol. Oceanogr. 12:176-179.
Klein-MacPhee, G. W. H. Howell, and A. D. Beck. 1980. International study on
Artemia. VII. Nutritional value of five geographical strains of Artemia
to winter flounder Pseudopleuronectes americanus larvae. In:
Persoone, G., P. Sorgeloos, D. A. Reels, and E. Jaspers, eds., The brine
shrimp, Artemia. Vol. 3, Ecology, culturing, use in aquaculture.
Universa Press, Wetteren, Belgium, pp. 305-312.
Klein-MacPhee, G., W. H. Howell, and A. D. Beck. 1982. International study
on Artemia. XX. Comparison of a reference and four geographical strains
of Artemia as food for winter flounder (Pseudopleuronectes americanus)
larvae. Aquacult. 29:279-288.
Kuntz, A. 1916. Notes on the embryology and larval development of five
species of teleostean fishes. Bull. U.S. Bur. Fish. 34(831):409-429.
Lawler, A. R., and S. L. Shepard. 1978. Procedures for eradication of
hydrozoan pests in closed-system mysid culture. Gulf Res. Rept.
6:177-178.
Leger, P., and P. Sorgeloos. 1982. Automation in stock-culture maintenance
and juvenile separation of the mysid Hysidopsis bahia (Molenock).
Aquacult. Eng. 1:45-53.
Leger, P., Sorgeloos, P., Millamena, 0. M., and Simpson, K. L. 1985.
"International Study on Artemia. XXV. Factors determining the
nutritional effectiveness of Artemia: The relative impact of
chlorinated hydrocarbons and essential fatty acids in San Francisco Bay
and San Pablo Bay Artemia." J. Exp. Mar. Biol. Ecol., Vol. 93, 1985.
pp. 71-82.
Leger, P., Bengtson, D. A., Simpson, K. L. and Sorgeloos, P. 1986. "The use
and nutritional value of Artemia as a food source," M. Barnes (ed.),
Oceanography and Marine Biology Annual Review, Vol. 24, Aberdeen
University Press, Aberdeen, Scotland, 1986, pp. 521-623.
Lussier, S. M., J. H. Gentile, and J. Walker. 1985. Acute and chronic
effects of heavy metals and cyanide on Mysidopsis bahia (Crustacea:
Mysidacea). Aquat. Toxicol. 7:25-35.
381
-------
Macek, K. J., and B. H. Sleight. 1977. Utility of toxicity tests with
embryos and fry of fish in evaluating hazards associated with the
chronic toxicity of chemicals to fishes. In: Mayer, F. L., and J. L.
Hamelink, eds., Aquatic Toxicology and Hazard Evaluation, ASTM STP 634,
American Society for Testing and Materials, Philadelphia, Pennsylvania.
pp. 137-146.
Marcus, A. H., and A. P. Holtzman. 1988. A robut statistical method
for estimating effects concentrations in short-term fathead minnow
toxicity tests. Manuscript submitted to the Criteria and
Standards Division, U. S. Environmental Protection Agency, by
Battelle Washington Environmental Program Office, Washington, DC,
June 1988, under EPA Contract No. 69-03-3534. 39 pp.
Martin, F.D. and G. E. Drewry. 1978. Development of fishes of the Mid-
Atlantic Bight. An atlas of eggs, larval, and juvenile stages.
Biological Service Program, Fish and Wildlife Service, U.S. Department
of the Interior, Washington, D.C. FWS/OBS-78/12.
Mauchline, J., and M. Murano. 1977. World list of the Mysidacea, Crustacea.
J. Tokyo Univ. Fish. 64:39-88.
McKim, J. M. 1977. Evaluation of tests with the early life stages of fish
for predicting long-term toxicity. J. Fish. Res. Bd. Can.
34:1148-1154.
Middaugh, D. P., and M. J. Hemmer. 1984. Spawning of the tidewater
silverside, Menidia peninsulae (Goode and Bean) in response to tidal and
lighting schedules in the laboratory. Estuaries 7:137-146.
Middaugh, D. P., M. J. Hemmer, and Y. Lamadrid-Rose. 1986. Laboratory
spawning cues in Menidia beryllina and H, pennisulae (Pices:
Atherinidae) with notes on survival and growth of larvae at different
salinities. Environ. Biol. Fish. 15(2) :107-117.
Middaugh, D. P., and T. Takita. 1983. Tidal and diurnal spawning cues in the
Atlantic Silverside, Menidia menidia. Environ. Biol. Fish.
8(2):97-104.
Miller, R. G. 1981. Simultaneous statistical inference. Springer-Verlag,
New York. 299 pp.
Molenock, J. 1969. Mysidopsis bahia, a new species of mysid (Crustacea:
Mysidacea) from Galveston Bay, Texas. Tulane Stud. Zool . Bot.
Morgan, M. D. 1982. The ecology of Mysidacea. Developments in hydrobiology
10. W. Junk, Publ., The Hague, Netherlands. 232 pp.
Mount, D. I., and C. E. Stephan. 1967. A method for establishing acceptable
toxicant limits for fish - Malathion and 2,4-D. Trans. Am. Fish. Soc.
96: 185-193.
382
-------
Nacci, D., and E. Jackim. 1985. Rapid aquatic toxicity assay using
incorporation of tritiated-thymidine into sea urchin, Arbacia
punctulata, embryo. In; Bahner, R. C., and D. J. Hansen. eds., Aquatic
Toxicology and Hazard Assessment: Eighth Symposium. STP 891. American
Society for Testing and Materials, Philadelphia, Pennsylvania, pp.
382-393.
Nacci, D., E. Jackim, and R. Walsh. 1986. Comparative evaluation of three
rapid marine toxicity tests: Sea urchin early embryo growth test, sea
urchin sperm cell toxicity test and Microtox. Environ. Toxicol. Chem.
5:521-525.
Nimmo, D. R., L. H. Bahner, R. A. Rigby, J. M. Sheppard, and A. J. Wilson, Jr.
1977. Mysidopsis bahia: an estuarine species suitable for life-cycle
toxicity tests to determine the effects of a pollutant. In: Mayer, F.
L., and J. L. Hamelin, eds., Aquatic Toxicology and Hazard Evaluation.
ASTM STP 634, American Society for Testing and Materials, Philadelphia,
Pennsylvania, pp. 109-116.
Nimmo, D. R.,
review.
and T. L. Hamaker. 1982.
Hydrobiol. 93:171-178.
Mysids in toxicity testing a
Nimmo, D. R., T. L. Hamaker, J. C. Moore, and C. A. Sommers. 1979. Effect of
diflubenzuron on an estuarine crustacean. Bull. Environm. Contam.
Toxicol. 22:767-770.
Nimmo, D. R., T. L. Hamaker, E. Matthews, and J. C. Moore. 1981. An overview
of the acute and chronic effects of first and second generation
pesticides on an estuarine mysid. In: Vernberg, F. J., A. Calabrese,
F. P. Thurberg, and W.B. Vernberg, eds., Biological Monitoring of Marine
Pollutants. Academic Press, New York. pp. 3-19.
Nimmo, D. R., T. L. Hamaker, E. Matthews, and W. T. Young. 1982. The long-
term effects of suspended particulates on survival and reproduction of
the mysid shrimp, Mysidopsis bahia, in the laboratory. In: Mayer, G.
F., ed., Ecological Stress and the New York Bight: Science and
Management. Estuarine Res. Found., Columbia, South Carolina, pp. 41-50.
Nimmo, D. R., T. L. Hamaker, J. C. Moore, and R. A. Wood. 1980. Acute and
chronic effects of Dimilin on survival and reproduction of Mysidopsis
bahia. In: Eaton, J. G., P. R. Parrish, and A. C. Hendricks, eds.,
ASTM STP 707, American Society for Testing and Materials, Philadelphia,
Pennsylvania, pp. 366-376.
Nimmo, D. R., T. L. Hamaker, and C. A. Sommers. 1978a. Culturing the mysid
(Mysidopsis bahia) in flowing sea water or a static system. In:
Bioassay Procedures for the Ocean Disposal Permit Program, U. S.
Environmental Protection Agency, Environmental Research Laboratory, Gulf
Breeze, Florida, EPA/600/9-78/010, pp. 59-60.
383
-------
Nimmo, D. R., T. L. Hamaker, and C. A. Sommers. 1978b. Entire life cycle
toxicity test using mysids (Mysidopsis bahia) in flowing water. In:
Bioassay Procedures for the Ocean Disposal Permit Program, U. S.
Environmental Protection Agency, Environmental Research Laboratory, Gulf
Breeze, Florida, EPA/600/9-78/010, pp. 64-68.
Nimmo, D, R., and E. S. Iley, Jr. 1982 Culturing and chronic toxicity of
Mysidopsis bahia using artificial seawater. Office of Toxic Substances,
U. S. Environmental Protection Agency, Washington, DC., Publ. PA 902.
Nimmo, D. R., R. A. Rigby, L. H. Bahner, and J. M. Sheppard. 1978. The acute
and chronic effects of cadmium on the estuarine mysid, Mysidopsis bahia.
Bull. Environm. Contam. Toxicol. 19(l):80-85.
Norberg-King, T. J. 1991. Calculations of ICp values of IC15, IC20, IC25,
IC30, and IC50 for Appendix A of the Revised Technical Support Document.
Memorandum to M. Heber, USEPA.
Norberg, T. J., and D. I. Mount. 1983. A seven-day larval growth test.
Presented at the Annual Meeting, Society of Environmental Toxicology and
Chemistry, November 6-9, 1983, Arlington, Virginia.
Norberg, T. J., and D. I. Mount. 1985. A new fathead minnow (Pimephales
promelas) subchronic toxicity test. Environ. Toxicol. Chem.
4(5):711-718.
Pearson, E. S., and T. 0. Hartley. 1962. Biometrika tables for
statisticians. Vol. 1. Cambridge Univ. Press, England, pp. 65-70.
Persoone, G., E. Jaspers, and C. Claus, eds. 1980. Ecotoxicological testing
for the marine environment. Vol. 1. State University of Ghent and
Institute for Marine Scientific Research, Bredene, Belgium.
Persoone, G., P. Sorgeloos, 0. Roels, and E. Jaspers, eds, 1980a. The brine
shrimp Artemia. Vol. 1. Morphology, genetics, radiobiology,
toxicology. Proceedings of the International Symposium on the brine
shrimp Artemia salina, Corpus Christi, Texas, 1979. Universa Press,
Wetteren, Belgium. 318 pp.
Persoone, G., P. Sorgeloos, 0. Roels, and E. Jaspers, eds. 1980b. The brine
shrimp Artemia. Vol. 2. Physiology, biochemistry, molecular biology.
Proceedings of the International Symposium on the brine shrimp Artemia
salina, Corpus Christi, Texas, 1979. Universa Press, Wetteren, Belgium.
636 pp.
Persoone, G., P- Sorgeloos, 0. Roels, and E. Jaspers, eds. 1980c. The brine
shrimp Artemia. Vol. 3. Ecology, culturing, use in aquaculture.
Proceedings of the International Symposium on the brine shrimp Artemia
salina, Corpus Christi, Texas, 1979. Universa Press, Wetteren, Belgium.
428 pp.
384
-------
Price, W. W. 1978. Occurrence of Mysidopsis almyra Bowman, M. bahia
Molenock, and Bowmam'ella brasiliensis Bacescu (Crustacea, Mysidacea)
from the eastern coast of Mexico. Gulf Res. Repts. 6:173-175.
Price, W. W. 1982. Key to the shallow water Mysidacea of the Texas coast
with notes on their ecology. Hydrobiol. 93(1/2):9-21.
Rand, G. M., and S. R. Petrocelli. 1985 Fundamentals of aquatic toxicology.
Hemisphere Publ. Corp., New York. 66 pp.
Reitsema, L. A. 1981. The growth, respiration, and energetics of Mysidopsis
almyra (Crustacea; Mysidacea) in relation to temperature, salinity, and
hydrocarbon exposure. Ph.D. thesis, Texas A & M University, College
Station, Texas.
Reitsema, L. A., and J. M. Neff. 1980. A recirculating artificial seawater
system for the laboratory culture of Mysidopsis almyra (Crustacea;
Pericaridea). Estuaries 3:321-323.
Renfro, W. C. 1960. Salinity relations of some fishes in the Aransas River.
Tulane Stud. Zool. 8:83-91.
Richards, F. A., and N. Corwin. 1956. Some oceanographic applications of
recent determinations of the solubility of oxygen in sea water. Limnol.
Oceanogr. 1(4):263-267.
Rodgers, J. H., Jr., P. B. Dorn, T. Duke, R. Parrish, and B. Venables
(rapporteur). 1986. Mysidopsis sp.: life history and culture. A
report from a workshop held at Gulf Breeze, Florida, October 15-16,
1986. American Petroleum Institute, Washington, DC.
Scheffe, H. 1959. The analysis of variance. John Wiley, New York. 477 pp.
Simmons, E. G. 1957. Ecological survey of the Upper Laguna Madre of Texas.
Publ. Inst. Mar. Sci. 4(2):156-200.
Smith, B. A. 1971. An ecological study of the Delaware River in the vicinity
of Artificial Island. Part V. The fish of four low-salinity tidal
tributaries of the Delaware River estuary. Progress Report to Public
Service Electric and Gas Co. Ichthyological Assoc. Ithaca, New York.
291 pp.
Snedecor, G. W., and W. G. Cochran. 1980. Statistical Methods. Seventh
edition. Iowa State University Press, Ames. 593 pp.
Sorgeloos, P. 1980. Life history of the brine shrimp Artemia. In: Persoone,
G., P. Sorgeloos, D.A. Roels, and E. Jaspers, eds., The brine shrimp,
Artemia. Vol. 1. Morphology, genetics, radiobiology, toxicology.
Proceedings of the International Symposium on the brine shrimp Artemia
saliha, Corpus Christi, Texas, 1979. Universa Press, Wetteren, Belgium.
pp. ixx-xxii.
385
-------
Spehar, R. L., D. K. Tanner, and B. R. Nordling. 1983. Toxicity of the
synthetic pyrethroids, permethrin, AC-222, 705, and their accumulation
in early life stages of fathead minnows and snails. Aquat. Toxicol.
3:171-182.
Spotte, S. 1973. Marine aquarium keeping. John Wiley and Sons, New York,
New York.
Spotte, S., G. Adams, and P. M. Bubucis. 1984. GP2 as an artificial
seawater for culture or maintenance of marine organisms. Zool. Biol.
3:229-240.
Sprague, J. B. 1969. Measurement of pollutant toxicity to fish. I. Bioassay
methods for acute toxicity. Water Res. 3:793-821.
Steel, R. G. 1959. A multiple comparison rank sum test: treatments versus
control. Biometrics 15:560-572.
Steel, R. G., and J. H. Torrie. 1960. Principles and procedures of
statistics with special reference to biological sciences. McGraw-Hill
Publ., New York.
Steele, R. L., and M. D. Hanisak. 1978. Sensitivity of some brown algal
reproductive stages to oil pollution. In: Jensen, A. and J. R. Stein,
eds., Proceedings of the Ninth International Seaweed Symposium, Vol. 9,
Science Press, Princeton, New Jersey, pp. 181-190.
Steele, R. L., and G. B. Thursby. 1983. A toxicity test using life stages of
Champi parvula (Rhodophyta). In: Bishop, W. E., R. D. Cardwell, and B.
B. Heidolph, eds., Aquatic Toxicology and Hazard Assessment: Sixth
Symposium. STP 802, American Society for Testing and Materials,
Philadelphia, Pennsylvania, pp. 73-89.
Stephens, M. A. 1974. EOF statistics for goodness of fit and come
comparisons. J. Amer. Stat. Assoc. (JASA) 69:730-7737.
Stuck, K. C., H. M. Perry, and R. W. Heard. 1979a. An annotated key to the
Hysidacea of the North Central Gulf of Mexico. Gulf Res. Rept.
6(3):225-238.
Stuck, K. C., H. M. Perry, and R. W. Heard. 1979b. Records and range
extensions of Mysidacea from coastal and shelf water of the Eastern Gulf
of Mexico. Gulf Res. Rept. 6(3):239-248.
Tagatz, M. E., and D. L. Dudley. 1961. Seasonal occurrence of marine fishes
in four shore habitats near Beaufort, N.C., 1957-1960. U.S. Fish.
Wildl. Serv. Sc. Rept. Fish. 390. 19 pp.
Tarzwell, C. M. 1971. Bioassays to determine allowable waste concentrations
in the aquatic environment. In: Cole, H. A. (Organizer), A discussion
on biological effects of pollution in the sea. Proc. Roy. Soc. Lond. B
177(1048):279-285.
386
-------
Tattersall, W. M., and 0. S. Tattersall. 1951. The British Mysidacea. Royal
Soc. London. 460 pp.
Taylor, J. K. 1987. Quality assurance of chemical measurements. Lewis Publ.,
Inc., Chelsea, Michigan.
Theilacker, G. H., and M. F. McMaster. 1971. Mass culture of the rotifer
Brachionus plicatilis and its evaluation as a food for larval anchovies.
Mar. Biol. 10:183-188.
Thompson, R. S., and E. M. Burrows. 1984. The toxicity of copper, zinc, and
mercury to the brown macroalga Laminan'a saccharina. In: Persoone, G.,
E. Jaspers, and C. Claus, eds., Ecotoxicological testing for the marine
environment, Vol. 2, State University of Ghent and Institute for Marine
Scientific Research, Bredene, Belgium, pp. 259-269.
Thursby, G. B., and R. L. Steele. 1984. Toxicity of arsenite and arsenate to
the marine macroalga Champia parvula. Environ. Toxicol. Chem.
3:391-397.
Thursby, G. B., and R. L. Steele. 1986. Comparison of short- and long-term
sexual reproduction tests with the marine red alga Champia parvula.
Environ. Toxicol. Chem. 5:1013-1018.
Thursby, G. B., R. L. Steele, and M. E. Kane. 1985. Effect of organic
chemicals on growth and reproduction in the marine red alga Champia
parvula. Environ. Toxicol. Chem. 4:797-805.
Thurston, R. V., R.C. Russo, and K. Emerson. 1974. Aqueous ammonia
equilibrium calculations. Tech. Rep. No. 741. Fisheries Bioassay
Laboratory, Montana State University, Bozeman. 18 pp.
USEPA. 1973. Biological field and laboratory methods for measuring the
quality of surface waters and effluents. C. I. Weber (ed.). U. S.
Environmental Protection Agency, Methods Development and Quality
Assurance Research Laboratory, Cincinnati, Ohio 45268, EPA 600/4-73-001.
USEPA. 1975. Methods for acute toxicity tests with fish, macroinvertebrates,
and amphibians. Environmental Research Laboratory, U. S. Environmental
Protection Agency, Duluth, Minnesota, EPA/660/3-75/009.
USEPA. 1977. Occupational health and safety manual. Office of Planning and
Management, U. S. Environmental Protection Agency, Washington, DC.
USEPA. 1978a. Methods for measuring the acute toxicity of effluents to
aquatic organisms. Second edition. Peltier, W., Environmental Monitoring
and Support Laboratory Cincinnati, U. S. Environmental Protection
Agency, Cincinnati, Ohio 45268, EPA/600/4-78/012.
USEPA. 1978b. Life-cycle toxicity test using sheepshead minnows (Cyprinodon
variegatus). Hansen, D. J., P. R. Parrish, S. C. Schimmel, and L. R.
Goodman. In: Bioassay procedures for the ocean disposal permit program,
387
-------
U. S. Environmental Protection Agency, Environmental Research
Laboratory, Gulf Breeze, Florida, EPA/600/9-78/010, pp. 109-117.
USEPA. 1979a. Handbook for analytical quality control in water and
wastewater laboratories. U. S. Environmental Protection Agency,
Environmental Monitoring and Support Laboratory, Cincinnati, Ohio 45268,
EPA/600/4-79/019.
USEPA. 1979b. Methods for chemical analysis of water and wastes.
Environmental Monitoring and Support Laboratory, U. S. Environmental
Protection Agency, Cincinnati, Ohio 45268, EPA-600/4-79/020, revised
March 1983.
USEPA. 1979c. Interim NPDES compliance biomonitoring inspection manual.
Office of Water Enforcement, U. S. Environmental Protection Agency,
Washington, DC 20460. (MCD-62).
USEPA. 1979d. Good laboratory practice standards for health effects.
Paragraph 772.110-1, Part 772 - Standards for development of test data.
Fed. Reg. 44:27362-27375, May 9, 1979.
USEPA. 1980a. Appendix B - Guidelines for Deriving Water Quality Criteria
for the Protection of Aquatic Life and Its Uses. Federal Register, Vol;
45, No. 231, Friday, November 28, 1980.
USEPA. 1980b. Proposed good laboratory practice guidelines for toxicity
testing. Paragraph 163.60-6. Fed. Reg. 45:26377-26382, April 18, 1980.
USEPA. 1980c. Physical, chemical, persistence, and ecological effects
testing; good laboratory practice standards (proposed rule). 40 CFR
772, Fed. Reg. 45:77353-77365, November 21, 1980.
USEPA. 1981 a. Acute toxicity test standard using mysid shrimp in static and
flow-through systems. Toxic Substances Control Act, Section 4. U. S.
Environmental Protection Agency, Office of Toxic Substances, Health and
Environmental Review Division, Washington, DC 20460.
USEPA. 1981b. Chronic toxicity test standard using mysid shrimp in a flow-
through system. Toxic Substances Control Act, Section 4. U. S.
Environmental Protection Agency, Office of Toxic Substances, Health and
Environmental Review Division, Washington, DC 20460.
USEPA. 1981c. Technical support document for using mysid shrimp in acute and
chronic toxicity tests. Toxic Substances Control Act, Section 4. U. S.
Environmental Protection Agency, Office of Toxic Substances, Health and
Environmental Review Division, Washington, DC 20460.
USEPA. 1981d. Effluent toxicity screening test using Daphnia and mysid
shrimp. Weber, C. I., and W. H. Peltier. Environmental Monitoring and
Support Laboratory, U. S. Environmental Protection Agency, Cincinnati,
Ohio 45268.
388
-------
USEPA. 1981e. In situ acute/chronic toxicological monitoring of industrial
effluents for the NPDES biomonitoring program using fish and amphibian
embryo/larval stages as test organisms. Birge, W. J., and J. A. Black.
Office of Water Enforcement and Permits, U. S. Environmental Protection
Agency, Washington, DC. OWEP-82-001.
USEPA. 1982. Methods for organic chemical analysis of municipal and
industrial wastewater. Environmental Monitoring and Support Laboratory,
U. S. Environmental Protection Agency, Cincinnati, Ohio 45268,
EPA/600/4-82/057.
USEPA. 1983. Guidelines and format for EMSL-Cincinnati methods. Kopp, J.F.
Environmental Monitoring and Support Laboratory, U. S. Environmental
Protection Agency, Cincinnati, Ohio 45268, EPA/600/8-83/020.
USEPA. 1984a. Development of water quality-based permit limitations for
toxic pollutants: National Policy. Fed. Reg. 49:9016-9019.
USEPA. 1984b. Effluent and ambient toxicity testing and instream community
response on the Ottawa River, Lima, Ohio. Mount, D. I., N. A. Thomas,
T. J. Norberg, M. T. Barbour, T. H. Roush, and W. F. Brandes.
Environmental Research Laboratory, U. S. Environmental Protection
Agency, Duluth, Minnesota, EPA/600/3-84/080.
USEPA. 1985a. Technical Support Document for Water Quality-based Control.
Office of Water, U. S. Environmental Protection Agency, Washington, DC
20460.
USEPA. 1985b. Ambient water quality criteria for ammonia 1984. Office of
Water Regulations and Standards, Office of Water, U. S. Environmental
Protection Agency, Washington, DC 20460, EPA/440/5-85/001.
USEPA. 1985c. Ambient water quality criateria for ammonia 1984. Office of
Water Regulations and Standards, Office of Water, U. S. Environmental
Protection Agency, Washington, DC 20460, EPA/440/5-85/001.
USEPA. 1985d. Methods for measuring the acute toxicity of effluents to
freshwater and marine organisms. Third Edition. Peltier, W., and C. I.
Weber, eds. Environmental Monitoring and Support Laboratory, U. S.
Environmental Protection Agency, Cincinnati, Ohio 45268,
EPA/600/4-85/013.
USEPA. 1985e. Short-term methods for estimating the chronic toxicity of
effluents and receiving waters to freshwater organisms. Horning, W.
B., and C. I. Weber. Environmental Monitoring and Support Laboratory,
U. S. Environmental Protection Agency, Cincinnati, Ohio 45268,
EPA/600/4-85/014.
USEPA. 1985f. Validity of effluent and ambient toxicity tests for predicting
biological impact, Scippo Creek, Circleville, Ohio. Mount, D. I., and
T. J. Norberg-King, eds. Environmental Research Laboratory, U. S.
Environmental Protection Agency, Duluth, Minnesota, EPA/600/3-85/044.
389
-------
USEPA. 1985g. Validity of effluent and ambient toxicity testing for
predicting biological impact on Five Mile Creek, Birmingham, Alabama.
Mount, D. I., A. E. Steen, and T. J. Norberg-King, eds. Environmental
Research Laboratory, U. S. Environmental Protection Agency, Duluth,
Minnesota. EPA/600/8-85/015.
USEPA. 1985h. Validity of effluent and ambient toxicity tests for predicting
biological impact, Ohio River, near Wheeling, West Virginia. Mount, D.
I., A. E. Steen, and T. J. Norberg-King, eds. Environmental
Research Laboratory, U. S. Environmental Protection Agency, Duluth,
Minnesota, EPA/600/3-85/071.
USEPA. 1985i. Hazard evaluation division: Standard evaluation procedure:
Acute toxicity test for estuarine and marine organisms (shrimp 96-hour
acute toxicity test). Rieder, D., Office of Pesticides Programs, U. S.
Environmental Protection Agency, Washington, DC 20460, EPA/540/9-85/010.
USEPA. 1986a. Occupational health and safety manual. Office of
Administration, U. S. Environmental Protection Agency, Washington, DC.
USEPA. 1986b. Validity of effluent and ambient toxicity tests for predicting
biological impact, Skeleton Creek, Enid, Oklahoma. Norberg, T. J., and
D. I. Mount, eds. Environmental Research Laboratory, U. S.
Environmental Protection Agency, Duluth, Minnesota, EPA/600/8-86/002.
USEPA. 1986c. Validity of effluent and ambient toxicity tests for predicting
biological impact, Kanawha River, Charleston, West Virginia. Mount, D.
I., and T. Norberg-King. Environmental Research Laboratory, U. S.
Environmental Protection Agency, Duluth, Minnesota, EPA/600/3-86/006.
USEPA. 1986d. Validity of effluent and ambient toxicity tests for predicting
biological impact, Back River, Baltimore Harbor, Maryland. Mount, D.
I., A. E. Steen, and T. Norberg-King. Environmental Research
Laboratory, U. S. Environmental Protection Agency, Duluth, Minnesota,
EPA/600/8-86/001.
USEPA. 1986e. Validity of effluent and ambient toxicity tests for predicting
biological impact, Naugatuck River, Connecticut. Mount, D. I., T.
Norberg-King, and A. E. Steen. Environmental Research Laboratory, U. S.
Environmental Protection Agency, Duluth, Minnesota, EPA/600/8-86/005.
USEPA. 1987a. Methods for spawning, culturing and conducting toxicity-tests
with early life stages of four antherinid fishes: the inland silverside,
Menidia beryl Una, Atlantic silverside, M. mem'dia, tidewater
silverside, M. peninsulas, and California grunion, Leuresthes tenuis.
Middaugh, D. P., M. J. Hemmer, and L. R. Goodman. Office of Research
and Development, U. S. Environmental Protection Agency, Washington, DC
20460, EPA/600/8-87/004.
USEPA. 1987b. Guidance manual for conducting complex effluent and receiving
water larval fish growth-survival studies with the sheepshead minnow
(Cyprinodon variegatus). Contribution No. X104. Hughes, M. M., M. A.
390
-------
Heber, S. C. Schimmel, and W. J. Berry. In: Schimmel, S. C., ed.
Users guide to the conduct and interpretation of complex effluent
toxicity tests at estuarine/marine sites. Environmental Research
Laboratory, U. S. Environmental Protection Agency, Narragansett, Rhode
Island. Contribution No. 796., 265 pp.
USEPA. 1987c. Guidance manual for rapid chronic toxicity tests on
effluents and receiving waters with larval inland silversides (Mem'dia
beryllina). Contribution No. 792. Heber, M. A., M. M. Hughes, S. C.
Schimmel, and D. A. Bengtson. In: Schimmel, S. C. ed., Users guide to
the conduct and interpretation of complex effluent toxicity tests at
estuarine/marine sites. Environmental Research Laboratory, U. S.
Environmental Protection Agency, Narragansett, Rhode Island.
Contribution No. 796., 265 pp.
USEPA. 1987d. Users guide to the conduct and interpretation of complex
effluent toxicity tests at estuarine/marine sites. Schimmel, S. C., ed.
Environmental Research Laboratory, U. S. Environmental Protection
Agency, Narragansett, Rhode Island. Contribution No. 796., 265 pp.
USEPA. 1987e. Guidance manual for conducting sexual reproduction test with
the marine macroalga Champia parvula for use in testing complex
effluents. Contribution No. X103. Thursby, G. B., and R. L. Steele.
In: Schimmel, S. C., ed. Users guide to the conduct and interpretation
of complex effluent toxicity tests at estuarine/marine sites.
Environmental Research Laboratory, U. S. Environmental Protection
Agency, Narragansett, Rhode Island. Contribution No. 796., 265 pp.
USEPA. 1987f. Guidance manual for conducting seven day mysid
survival/growth/reproduction study using the estuarine mysid, Mysidopsis
bahia. Contribution No. X106. Lussier, S. M., A. Kuhn, and J. Sewall.
In: Schimmel, S. C., ed. Users guide to the conduct and interpretation
of complex effluent toxicity tests at estuarine/marine sites.
Environmental Research Laboratory, U. S. Environmental Protection
Agency, Narragansett, Rhode Island. Contribution No. 796., 265 pp.
USEPA. 1987g. Guidance manual for conducting sperm cell tests with the sea
urchin, Arbacia punctulata, for use in testing complex effluents.
Nacci, D., R. Walsh, and E. Jackim. Contribution No. X105. In:
Schimmel, S.C., ed. Users guide to the conduct and interpretation of
complex effluent toxicity tests at estuarine/marine sites.
Environmental Research Laboratory, U. S. Environmental Protection
Agency, Narragansett, Rhode island. Contribution No. 796., 265 pp.
USEPA. 1988a. NPDES compliance inspection manual. Office of Water
Enforcement and Permits (EN-338), U. S. Environmental Protection
Agency, Washington, DC 20460.
USEPA. 1988b. 40 CFR Part 160 - Good laboratory practice standards, pages
153-164.
391
-------
USEPA. 1988c. Methods for aquatic toxicity identification evaluations:
Phase II, toxicity identification procedures. D. I. Mount, and L.
Anderson-Carnahan. Environmental Research Laboratory, U. S.
Environmental Protection Agency, Duluth, Minnesota, EPA-600/3-88/035.
USEPA. 1988d. Methods for aquatic toxicity identification evaluations:
Phase III, toxicity confirmation procedures. D. I. Mount. Environmental
Research Laboratory, U. S. Environmental Protection Agency, Duluth,
Minnesota, EPA-600/3-88/036.
USEPA. 1988e. An interpolation estimate for chronic toxicity: The ICp
approach. Norberg-King, T. J. Technical Report 05-88, National
Effluent Toxicity Assessment Center, Environmental Research Laboratory,
U. S. Environmental Protection Agency, Duluth, Minnesota.
USEPA. 1989a. Short-term methods for estimating the chronic toxicity of
effluents and receiving waters to marine and estuarine organisms.
Weber, C. I., W. B. Horning, II, D. J. Klemm, T. W. Neiheisel, P. A.
Lewis, E. L. Robinson, J. Menkedick, and F. Kessler (eds.).
Environmental Monitoring and Support Laboratory, U. S. Environmental
Protection Agency, Cincinnati, Ohio 45268, EPA/600/4-87/028.
USEPA. 1989b. Short-term methods for estimating the chronic toxicity of
effluents and receiving waters to freshwater organisms. Second
Edition. Weber, C. I., W. H. Peltier, T. J. Norbgerg-King, W. B.
Horning, II, F. A. Kessler, J. R. Menkedick, T. W. Neidheisel, P. A.
Lewis, D. J. Klemm, Q. H. Pickering, E. L. Robinson, J. M. Lazorchak, L.
J. Wymer, and R. W. Freyberg (eds.). Environmental Monitoring Systems
Laboratory, U. S. Environmental Protection Agency, Cincinnati, Ohio
45268, EPA/600/4-89/001.
USEPA. 1989c. Toxicity reduction evaluation protocol for municipal
wastewater treatment plants. J. A. Botts, J. W. Braswell, J. Zyman, W.
L. Goodfellow, and S. B. Moore. Risk Reduction Engineering Laboratory,
U. S. Environmental Protection Agency, Cincinnati, Ohio 45268,
EPA/600/2-88/062.
USEPA. 1989d. Generalized methodology for conducting industrial toxicity
reduction evaluations (TREs). J. A. Fava, D. Lindsay, W. H. Clement, R.
Clark, G. M. DeGraeve, J. D. Cooney, S. Hansen, W. Rue, S. Moore, and P.
Lankford. Risk Reduction Engineering Laboratory, U. S. Environmental
Protection Agency, Cincinnati, Ohio 45268, EPA/600/2-88/070.
USEPA. 1990a. Macroinvertebrate field and laboratory methods for evaluating
the biological integrity of surface waters. Klemm, D. J., P. A. Lewis,
F. Fulk, and J. M. Lazorchak. Environmental Monitoring Systems
Laboratory, U. S. Environmental Protection Agency, Cincinnati, Ohio
45268, EPA/600/4-90/030.
USEPA. 1990b. Supplemental methods and status reports for short-term
saltwater toxicity tests. G. Morrison, and G. Chapman. ERL Contrib.
392
-------
No. 1199. Environmental Research Laboratory, U. S. Environmental
Protection Agency, Narragansett, Rhode Island, 127 pp.
USEPA. 1991a. Technical support document for water quality-based toxic
controls. Office of Water Enforcement and Permits and Office of Water
Regulations and Standards, U. S. Environmental Protection Agency,
Washington, DC 20460, EPA/505/2-90-001.
USEPA. 1991b. Manual for the evaluation of laboratories performing aquatic
toxicity tests. Klemm, D. J., L. B. Lobring, and W. H. Horning.
Environmental Monitoring Systems Laboratory, U. S. Environmental
Protection Agency, Cincinnati, Ohio 45268, EPA/600/4-90/031.
USEPA. 1991c. Methods for measuring the acute toxicity of effluents and
receiving waters to freshwater and marine organisms. Fourth Edition.
Weber, C. I. (ed.). Environmental Monitoring Systems Laboratory, U. S.
Environmental Protection Agency, Cincinnati, Ohio 45268, EPA/600/4-
90/027.
USEPA. 1991d. Methods for aquatic toxicity identification evaluations:
Phase I, toxicity characterization procedures. T. Norberg-King, D. I.
Mount, E. Durhan, G. Ankley, L. Burkhard, J. Amato, M. Lukasewycz, M.
Schubauer-Berigan, and L. Anderson-Carnahan. Environmental Research
Laboratory, U. S. Environmental Protection Agency, Duluth, Minnesota,
EPA/600/6-91/003.
USEPA. 1991e. Toxicity identification evaluation: Characterization of
chronically toxic effluents, Phase 1. T. J. Norberg-King, D. I. Mount,
Jr. Amato, D. A. Jensen, J. A. Thompson. Environmental Research
Laboratory, U. S. Environmental Protection Agency, Duluth, Minnesota,
EPA-600/6-91/005.
USEPA. 1992. Short-term methods for estimating the chronic toxicity of
effluents and surface waters to freshwater organisms. Third edition.
Lewis, P. A., D. J. Klemm, and J. M. Lazorchak (eds.). Environmental
Monitoring Systems Laboratory, U. S. Environmental Protection Agency,
Cincinnati, Ohio 45268, EPA/600/4-91/022.
Usher, R.R., and D.A. Bengtson. 1981. Survival and growth of sheepshead
minnow larvae and juveniles on diet of Artemia nauplii. Prog. Fish-Cult.
43(2):102-105.
Vanhaecke, P. and Sorgeloos, P. 1980. "International Study on Artemia. IV.
The biometrics of Artemia strains from different geographical origin."
In: G. Persoone, P. Sorgeloos, 0. Roels and E. Jaspers (eds.), The
Brine Shrimp Artemia, Vol. 3, Ecology, Culturing Use in Aquaculture,
Universa Press, Wetteren, Belgium, 1980, pp. 393-405.
Vanhaecke, P., Steyaert, H. and Sorgeloos, P- 1980. "International Study on
Artemia. III. The use of Coulter Counter equipment for the biometrical
analysis of Artemia cysts. Methodology and Mathematics," In: G.
Persoone, P. Sorgeloos, 0. Roels, and E. Jaspers (eds.), The Brine
393
-------
Shrimp Artemia, Vol. 1, Morphology, Genetics, Radiobiology, Toxicology.
Universa Press, Wettersen, Belgium, 1980, pp. 107-115.
Vanhaecke, P., and P. Sorgeloos. 1983. International study on Artemia.
XIX. Hatching data for ten commercial sources of brine shrimp cysts and
re-evaluation of the "hatching efficiency" concept. Aquacult. 30:43-52.
Walsh, G. E., L. H. Bahner, and W. B. Horning. 1980. Toxicity of textile
mill effluents to freshwater and estuarine algae, crustaceans, and
fishes. Environ. Pollut. (Ser. A.) 21:169-179.
Walters, D. B., and C. W. Jameson. 1984. Health and safety for toxicity
testing. Butterworth Publ., Woburn, Massachusetts.
Ward, G. S., and P. R. Parrish. 1980. Evaluation of early life stage
toxicity tests with embryos and juveniles of sheepshead minnows. In:
J. G. Eaton, P. R. Parrish and A. C. Hendricks, Aquatic Toxicology, ASTM
STP 707, American Society for Testing and Materials, Philadelphia,
Pennsylvania, p. 243-247.
Ward, S. H. 1984. A system for laboratory rearing of the mysid, Mysidopsis
bahia Molenock. Progr. Fish-Cult. 46(3):170-175.
Ward, S. H. 1987. Feeding response of the mysid Mysidopsis bahia reared on
Artemia. Prog. Fish-Cult. 49:29-33.
Ward, S. H. 1990. Procedures for toxicant testing and culture of the marine
macroalga, Champia parvula. Technical Report, U. S. Environmental
Protection Agency, Region 2, New York, NY 20278.
Welch, P.S. 1948. Limnological methods. McGraw-Hill Book Company, New York,
NY.
Weltering, D. M. 1984. The growth response in fish chronic and early life
stage toxicity tests: A critical review. Aquat. Toxicol. 5:1-21.
Zaroogian, G. E., G. Pesch, and G. Morrison. 1969. Formulation of an
artificial sea water media suitable for oyster larvae development.
Amer. Zool. 9:1141.
Zillioux, E. J., H. R. Foulk, J. C. Prager, and J. A. Cardin. 1973. Using
Artemia to assay oil dispersant toxicities. JWPCF 45:2389-2396.
394
-------
APPENDICES
A. Independence, Randomization, and Outliers 397
1. Statistical Independence 397
2. Randomization 397
3. Outliers 402
B. Validating Normality and Homogeneity of Variance
Assumptions 404
1. Introduction 404
2. Test for Normal Distribution of Data 404
3. Test for Homogeneity of Variance 412
4. Transformations of Data 413
C. Dunnett's Procedure 416
1. Manual Calculations 416
2. Computer Calculations 423
D. T-test with Bonferroni's Adjustment 463
E. Steel's Many-one Rank Test 468
F. Wilcoxon Rank Sum Test 473
G. Probit Analysis 479
H. Single Concentration Toxicity Test - Comparison of Control with
100% Effluent or Receiving Water 497
I. Linear Interpolation Method 501
1. General Procedure 501
2. Data Summary and Plots 501
3. Monotonicity 501
4. Linear Interpolation Method 502
5. Confidence Intervals 503
395
-------
APPENDICES (CONTINUED)
6. Manual Calculations 503
7. Computer Calculations 507
396
-------
APPENDIX A
INDEPENDENCE, RANDOMIZATION, AND OUTLIERS1
1. STATISTICAL INDEPENDENCE
1.1 Dunnett's Procedure and the t-test with Bonferroni's adjustment are
parametric procedures based on the assumptions that (1) the observations
within treatments are independent and normally distributed, and (2) that the
variance of the observations is homogeneous across all toxicant concentrations
and the control. Of the three possible departures from the assumptions,
non-normality, heterogeneity of variance, and lack of independence, those
caused by lack of independence are the most difficult to deal with (see
Scheffe, 1959). For toxicity data, statistical independence means that given
knowledge of the true mean for a given concentration or control, knowledge of
the error in any one actual observation would provide no information about the
error in any other observation. Lack of independence is difficult to assess
and difficult to test for statistically. It may also have serious effects on
the true alpha or beta level. Therefore, it is of utmost importance to be
aware of the need for statistical independence between observations and to be
constantly vigilant in avoiding any patterned experimental procedure that
might compromise independence. One of the best ways to help insure
independence is to follow proper randomization procedures throughout the test.
2. RANDOMIZATION
2.1 Randomization of the distribution of test organisms among test chambers,
and the arrangement of treatments and replicate chambers is an important part
of conducting a valid test. The purpose of randomization is to avoid
situations where test organisms are placed serially into test chambers, or
where all replicates for a test concentration are located adjacent to one
another, which could introduce bias into the test results.
2.2 An example of randomization of the distribution of test organisms among
test chambers, and an example of randomization of arrangement of treatments
and replicate chambers are described using the Sheepshead Minnow Larval
Survival and Growth test. For the purpose of the example, the test design is
as follows: Five effluent concentrations are tested in addition to the
control. The effluent concentrations are as follows: 6.25%, 12.5%, 25.0%,
50.0%, and 100.0%. There are four replicate chambers per treatment. Each
replicate chamber contains ten fish.
2.3 Randomization of Fish to Replicate Chambers Example
Prepared by Florence Fulk, Environmental Monitoring Systems Laboratory, U.S.
Environmental Protection Agency; Laura Gast and Cathy Poore, Computer
Sciences Corporation, 26 VI. Martin Luther King Drive, Cincinnati, Ohio 45268;
Phone (513) 569-7968.
397
-------
2.3.1 Consider first the random assignment of the fish to the replicate
chambers. The first step is to label each of the replicate chambers with the
control or effluent concentration and the replicate number. The next step is
to assign each replicate chamber four double digit numbers. An example of
this assignment is provided in Table A.I. Note that the double digits 00 and
97 through 99 were not used.
TABLE A.I. RANDOM ASSIGNMENT OF FISH TO REPLICATE CHAMBERS EXAMPLE
ASSIGNED NUMBERS FOR EACH REPLICATE CHAMBER
Assigned Numbers
Replicate Chamber
01,
02,
03,
04,
05,
06,
07,
08,
09,
10,
11,
12,
13,
14,
15,
16,
17,
18,
19,
20,
21,
22,
23,
24,
25,
26,
27,
28,
29,
30,
31,
32,
33,
34,
35,
36,
37,
38,
39,
40,
41,
42,
43,
44,
45,
46,
47,
48,
49,
50,
51,
52,
53,
54,
55,
56,
57,
58,
59,
60,
61,
62,
63,
64,
65,
66,
67,
68,
69,
70,
71,
72,
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
Control ,
Control ,
Control ,
Control ,
6.25% effluent,
6.25% effluent,
6.25% effluent,
6.25% effluent,
12.5% effluent,
12.5% effluent,
12.5% effluent,
12.5% effluent,
25.0% effluent,
25.0% effluent,
25.0% effluent,
25.0% effluent,
50.0% effluent,
50.0% effluent,
50.0% effluent,
50.0% effluent,
100.0% effluent,
100.0% effluent,
100.0% effluent,
100.0% effluent,
repl
repl
repl
repl
repl
repl
repl
repl
repl
repl
repl
repl
repl
repl
repl
repl
repl
repl
repl
repl
repl
repl
repl
repl
icate
icate
icate
icate
icate
icate
icate
icate
icate
icate
icate
icate
icate
icate
icate
icate
icate
icate
icate
icate
icate
icate
icate
icate
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
2.3.2 The random numbers used to carry out the random assignment of fish to
replicate chambers are provided in Table A.2. The third step is to choose a
starting position in Table A.2, and read the first double digit number. The
first number read identifies the replicate chamber for the first fish taken
from the tank. For the example, the first entry in row 2 was chosen as the
starting position. The first number in this row is 37. According to Table
A.I, this number corresponds to replicate chamber 1 of the 25.0% effluent
concentration. Thus, the first fish taken from the tank is to be placed in
replicate chamber 1 of the 25.0% effluent concentration.
2.3.3 The next step is to read the double digit number to the right of the
first one. The second number identifies the replicate chamber for the second
398
-------
TABLE A.2. TABLE OF RANDOM NUMBERS1
10 09 73 26 33
37 54 20 48 06
08 42 26 89 53
99 01 90 25 29
12 80 79 99 70
06 06 57 47 17
31 06 01 08 05
85 26 97 76 02
63 57 33 21 36
73 79 64 67 53
98 52 01 77 67
11 80 50 54 31
83 46 29 96 34
88 68 64 02 00
99 69 46 73 48
86 48 11 76 74
80 12 43 66 35
74 35 09 98 17
69 91 62 68 03
09 88 32 05 05
91 49 91 45 23
80 33 69 45 98
44 10 48 19 49
12 55 07 37 42
83 60 64 93 29
61 19 69 04 46
15 47 44 52 66
94 56 72 85 73
42 48 11 62 13
23 52 37 83 17
04 49 36 24 94
00 64 99 76 64
35 96 31 63 07
59 80 80 83 91
46 05 88 52 36
32 17 90 05 97
69 23 46 14 06
19 56 64 14 30
45 15 51 49 38
94 86 43 19 94
98 08 82 48 28
33 18 61 62 S3
80 96 10 04 08
79 75 24 91 40
18 63 33 25 37
74 03 94 39 02
M 17 84 56 11
11 88 44 98 83
48 33 47 79 38
89 07 49 41 38
76 52 01 35 86
64 89 47 42 96
19 64 50 93 03
09 37 67 07 15
80 16 73 61 47
34 07 27 68 50
45 57 18 24 06
02 06 16 66 92
06 32 54 70 48
03 52 96 47 78
14 90 56 86 07
39 80 82 77 32
06 28 89 80 83
86 50 75 84 01
87 51 76 49 69
17 46 85 09 50
17 72 70 80 15
77 40 27 72 14
66 25 22 91 48
14 22 56 85 14
68 47 92 76 86
26 94 03 68 68
85 15 74 79 54
11 10 00 20 40
16 50 53 44 84
26 45 74 77 74
95 27 07 99 53
67 89 75 43 87
97 34 40 87 21
73 20 88 98 37
75 24 63 38 24
64 05 18 81 59
26 89 80 93 54
45 42 72 68 42
01 39 00 22 86
87 37 92 62 41
20 11 74 52 04
01 75 87 53 79
19 47 60 72 46
38 16 81 08 51
46 24 02 84 04
41 94 15 09 49
96 38 27 07 74
71 96 12 82 96
98 14 50 65 71
77 56 73 22 70
80 99 33 71 43
52 07 98 48 27
31 24 96 47 10
87 63 79 19 76
34 67 35 48 76
24 80 52 40 37
23 20 90 25 60
38 31 13 11 65
64 03 23 66 53
36 69 73 61 70
35 30 34 26 14
68 66 67 48 18
90 65 35 75 48
35 80 83 42 82
22 10 94 05 68
50 72 56 82 48
13 74 67 00 78
36 76 66 79 51
91 82 60 89 28
58 04 77 69 74
45 31 82 23 74
43 23 60 02 10
36 93 68 72 03
46 42 75 67 88
46 16 28 35 64
70 29 73 41 35
32 97 92 65 75
12 86 07 46 97
40 21 95 25 63
51 92 43 37 29
59 36 78 38 48
54 62 24 44 31
16 86 84 87 67
68 93 59 14 16
45 86 25 10 25
96 11 96 38 96
33 35 13 64 62
83 60 94 97 00
77 28 14 40 77
05 56 70 70 07
15 95 66 00 00
40 41 92 15 85
43 66 79 48 43
34 88 88 15 63
44 99 90 88 96
89 43 64 85 81
20 15 12 33 87
69 86 10 26 91
31 01 02 46 74
97 79 01 71 19
06 33 51 29 69
59 38 17 16 39
02 29 63 68 70
35 58 40 44 01
80 95 90 91 17
20 63 61 04 02
15 95 33 47 64
88 67 67 43 97
98 95 11 68 77
65 81 33 98 85
86 79 90 74 39
73 05 38 52 47
28 46 82 87 09
60 93 52 03 44
60 97 09 34 33
29 40 52 42 01
18 47 64 06 10
90 36 47 64 93
93 78 56 13 68
73 03 95 71 86
21 11 57 82 53
45 52 16 42 37
76 62 11 39 90
96 29 77 88 22
94 75 08 99 23
53 14 03 33 40
57 60 04 08 81
96 64 48 94 39
43 65 17 70 82
65 39 45 95 93
82 39 61 01 18
91 19 04 25 92
03 07 11 20 59
26 25 22 96 63
61 96 27 93 35
54 69 28 23 91
77 97 45 00 24
13 02 12 48 92
93 91 08 36 47
86 74 31 71 57
18 74 39 24 23
66 67 43 68 06
69 04 79 00 33
01 54 03 54 56
39 09 47 34 07
88 69 64 19 94
26 01 62 62 98
74 85 22 05 39
06 45 56 14 27
52 52 75 80 21
66 12 71 92 :<5
09 97 33 34 40
32 30 75 75 46
10 51 82 16 15
39 29 27 49 45
00 82 29 16 65
35 08 03 36 06
04 43 62 76 59
12 17 17 68 33
11 19 92 91 70
23 40 30 97 32
18 62 38 85 79
83 49 12 56 24
35 27 38 84 35
50 50 07 39 98
52 77 56 78 51
68 71 17 78 17
29 60 91 10 62
23 47 83 41 13
40 21 81 65 44
14 38 55 37 63
96 28 60 26 55
94 40 05 64 18
54 38 21 45 98
37 08 92 00 48
42 05 08 23 41
22 22 20 64 13
28 70 72 58 15
07 20 73 17 90
42 58 26 05 27
33 21 15 94 66
92 92 74 59 73
25 70 14 66 70
05 52 28 25 62
65 33 71 24 72
23 28 72 95 29
90 10 33 93 33
78 56 52 01 06
70 61 74 29 41
85 39 41 18 38
97 11 89 63 38
84 96 28 52 07
20 82 66 95 41
05 01 45 11 76
35 44 13 18 80
37 54 87 30 43
94 82 46 11 71
00 38 75 95 79
77 93 89 19 38
80 81 45 17 48
36 04 09 03 24
88 46 12 33 56
15 02 00 99 94
01 84 87 69 38
1From Dixon and Massey (1983)
399
-------
fish taken from the tank. Continuing the example, the second number read in
row 2 of Table A.2 is 54. According to Table A.I, this number corresponds to
replicate chamber 2 of the 6.25% effluent concentration. Thus, the second
fish taken from the tank is to be placed in replicate chamber 2 of the 6.25%
effluent concentration.
2.3.4 Continue in this fashion until all the fish have been randomly assigned
to a replicate chamber. In order to fill each replicate chamber with ten
fish, the assigned numbers will be used more than once. If a number is read
from the table that was not assigned to a replicate chamber, then ignore it
and continue to the next number. If a replicate chamber becomes filled and a
number is read from the table that corresponds to it, then ignore that value
and continue to the next number. The first ten random assignments of fish to
replicate chambers for the example are summarized in Table A.3.
TABLE A.3. RANDOM ASSIGNMENT OF FISH TO REPLICATE CHAMBERS EXAMPLE
ASSIGNMENT OF FIRST TEN FISH
Fish Assignment
First
Second
Third
Fourth
Fifth
Sixth
Seventh
Eighth
Ninth
Tenth
fi
sh
fish
fi
fi
fi
fi
fi
fi
fi
fi
sh
sh
sh
sh
sh
sh
sh
sh
taken
taken
taken
taken
taken
taken
taken
taken
taken
taken
from
from
from
from
from
from
from
from
from
from
tank
tank
tank
tank
tank
tank
tank
tank
tank
tank
25.0% effluent,
6.25% effluent,
50.0% effluent
100.0% effluent,
6.25% effluent,
25.0% effluent,
50.0% effluent,
100.0% effluent,
50.0% effluent,
100.0% effluent,
repl
repl
repl
repl
repl
repl
repl
repl
repl
repl
icate
icate
icate
icate
icate
icate
icate
icate
icate
icate
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
chamber
1
2
4
4
1
4
1
3
2
4
2.3.5 Four double digit numbers were assigned to each replicate chamber
(instead of one, two, or three double digit numbers) in order to make
efficient use of the random number table (Table A.2). To illustrate, consider
the assignment of only one double digit number to each replicate chamber: the
first column of assigned numbers in Table A.I. Whenever the numbers 00 and 25
through 99 are read from Table A.2, they will be disregarded and the next
number will be read.
2.4 Randomization of Replicate Chambers to Positions Example
2.4.1 Next consider the random assignment of the 24 replicate chambers to
positions within the water bath (or equivalent). Assume that the replicate
chambers are to be positioned in a four row by six column rectangular array.
The first step is to label the positions in the water bath. Table A.4
provides an example layout.
400
-------
TABLE A.4. RANDOM ASSIGNMENT OF REPLICATE CHAMBERS TO POSITIONS EXAMPLE
LABELLING THE POSITIONS WITHIN THE WATER BATH
1 2
7 8
13 14
19 20
3
9
15
21
4
10
16
22
5
11
17
23
6
12
18
24
2.4.2 The second step is to assign each of the 24 positions four double digit
numbers. An example of this assignment is provided in Table A.5. Note that
the double digits 00 and 97 through 99 were not used.
TABLE A.5. RANDOM ASSIGNMENT OF REPLICATE CHAMBERS TO POSITIONS EXAMPLE
ASSIGNED NUMBERS FOR EACH POSITION
Assigned Numbers
Position
01,
02,
03,
04,
05,
06,
07,
08,
09,
10,
11,
12,
13,
14,
15,
16,
17,
18,
19,
20,
21,
22,
23,
24,
25,
26,
27,
28,
29,
30,
31,
32,
33,
34,
35,
36,
37,
38,
39,
40,
41,
42,
43,
44,
45,
46,
47,
48,
49,
50,
51,
52,
53,
54,
55,
56,
57,
58,
59,
60,
61,
62,
63,
64,
65,
66,
67,
68,
69,
70,
71,
72,
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
401
-------
2.4.3 The random numbers used to carry out the random assignment of replicate
chambers to positions are provided in Table A.2. The third step is to choose
a starting position in Table A.2, and read the first double digit number. The
first number read identifies the position for the first replicate chamber of
the control. For the example, the first entry in row 10 of Table A.2 was
chosen as the starting position. The first number in this row was 73.
According to Table A.5, this number corresponds to position 1. Thus, the
first replicate chamber for the control will be placed in position 1.
2.4.4 The next step is to read the double digit number to the right of the
first one. The second number identifies the position for the second replicate
chamber of the control. Continuing the example, the second number read in row
10 of Table A.2 is 79. According to Table A.5, this number corresponds to
position 7. Thus, the second replicate chamber for the control will be placed
in position 7.
2.4.5 Continue in this fashion until all the replicate chambers have been
assigned to a position. The first four numbers read will identify the
positions for the control replicate chambers, the second four numbers read
will identify the positions for the lowest effluent concentration replicate
chambers, and so on. If a number is read from the table that was not assigned
to a position, then ignore that value and continue to the next number. If a
number is repeated in Table A.2, then ignore the repeats and continue to the
next number. The complete randomization of replicate chambers to positions
for the example is displayed in Table A.6.
TABLE A.6. EXAMPLE OF RANDOM ASSIGNMENT OF REPLICATE CHAMBERS TO
POSITIONS: ASSIGNMENT OF ALL 24 POSITIONS
Control
Control
100.0%
50.0%
100.0%
12.5%
50.0%
50.0%
6.25%
Control
100.0%
25.0%
6.25%
25.0%
Control
50.0%
6.25%
12.5%
100.0%
12.5%
12.5%
25.0%
25.0%
6.25%
2.4.6 Four double digit numbers were assigned to each position (instead of
one, two, or three) in order to make efficient use of the random number table
(Table A.2). To illustrate, consider the assignment of only one double digit
number to each position: the first column of assigned numbers in Table A.5.
Whenever the numbers 00 and 25 through 99 are read from Table A.2, they will
be disregarded and the next number will be read.
3. OUTLIERS
3.1 An outlier is an inconsistent or questionable data point that appears
unrepresentative of the general trend exhibited by the majority of the data.
Outliers may be detected by tabulation of the data, plotting, and by an
402
-------
analysis of the residuals. An explanation should be sought for any
questionable data points. Without an explanation, data points should be
discarded only with extreme caution. If there is no explanation, the analysis
should be performed both with and without the outlier, and the results of both
analyses should be reported.
3.2 Gentleman-Milk's A statistic gives a test for the condition that the
extreme observation may be considered an outlier. For a discussion of this,
and other techniques for evaluating outliers, see Draper and John (1981).
403
-------
APPENDIX B
VALIDATING NORMALITY AND HONOGENEITY OF VARIANCE ASSUMPTIONS1
1. INTRODUCTION
1.1 Dunnett's Procedure and the t-test with Bonferroni's adjustment are
parametric procedures based on the assumptions that the observations within
treatments are independent and normally distributed, and that the variance of
the observations is homogeneous across all toxicant concentrations and the
control. These assumptions should be checked prior to using these tests, to
determine if they have been met. Tests for validating the assumptions are
provided in the following discussion. If the tests fail (if the data do not
meet the assumptions), a nonparametric procedure such as Steel's Many-one Rank
Test may be more appropriate. However, the decision on whether to use
parametric or nonparametric tests may be a judgement call, and a statistician
should be consulted in selecting the analysis.
2. TEST FOR NORMAL DISTRIBUTION OF DATA
2.1 A formal test for normality is the Shapiro-Wilk's Test (Conover, 1980).
The test statistic is obtained by dividing the square of an appropriate linear
combination of the sample order statistics by the usual symmetric estimate of
variance. The calculated W must be greater than zero and less than or equal
to one. This test is recommended for a sample size of 50 or less. If the
sample size is greater than 50, the Kolmogorov "D" statistic (Stephens, 1974)
is recommended. An example of the Shapiro-Wilk's test is provided below.
2.2 The example uses growth data from the Sheepshead Minnow Larval Survival
and Growth Test. The same data are used in the discussion of the homogeneity
of variance determination in Paragraph 3 and Dunnett's Procedure in Appendix
C. The data and the mean and standard deviation of the observations at each
concentration, including the control, are listed in Table B.I.
2.3 The first step of the test for normality is to center the observations by
subtracting the mean of all observations within a concentration from each
observation in that concentration. The centered observations are listed in
Table B.2.
2.4 Calculate the denominator, D, of the test statistic:
n
D = E (X,- - X)2
Where: X,- = the centered observations and~X is the overall mean of
Prepared by Laura Gast, Cathy Poore, Ron Freyberg, Florence Kessler, John
Menkedick and Larry Wymer, Computer Sciences Corporation, 26 W. Martin Luther
King Drive, Cincinnati, Ohio 45268; Phone (513) 569-7968.
404
-------
the centered observations. For this set of data, X = 0,
and D = 0.1589.
2.5 Order the centered observations from smallest to largest,
X(1) X<2> - X(n)
and let X(1) denote the ith order statistic. The ordered observations are
listed in Table B.3.
2.6 From Table B.4, for the number of observations, n, obtain the
coefficients a.,, a2, ..., ak, where k is n/2 if n is even, and (n-l)/2 if n is
odd. For the data in this example, n = 15, k = 7, and the a,, values are
listed in Table B.5.
2.7 Compute the test statistic, W, as follows:
k
W = 1 [ S a, (X(n-i+1) - X(i))]2
D i=l
The differences, x(n'i+1) X(i), are listed in Table B.5.
2.8 The decision rule for this test is to compare the critical value from
Table B.6 to the computed W. If the computed value is less than the critical
value, conclude that the data are not normally distributed. For this example,
the critical value is 0.835. The calculated value, 0.9516, is not less than
the critical value. Thus, the conclusion of the test is that the data are
normally distributed.
2.9 In general, if the data fail the test for normality, a transformation
such as to log values may normalize the data. After transforming the data,
repeat the Shapiro-Wilk's Test for normality.
405
-------
TABLE B.I. SHEEPSHEAD MINNOW, CYPRINODON VARIEGATUS, LARVAL GROWTH
DATA (WEIGHT IN MG) FOR THE SHAPIRO-WILK'S TEST
Effluent Concentration (%)
Replicate
1
2
3
Mean
s;
i
Control
1.017
0.745
0.862
0.875
0.019
1
6.25
1.157
0.914
0.992
1.021
0.015
2
12.5
0.998
0.793
1.021
0.937
0.016
3
25.0
0.837
0.935
0.839
0.882
0.0031
4
50.0
0.715
0.907
1.044
0.889
0.027
5
TABLE B.2. EXAMPLE OF SHAPIRO-WILK'S TEST CENTERED OBSERVATIONS
Effluent Concentration (%)
Replicate Control
6.25
12.5
25.0
50.0
1 0.142
2 - 0.13
3 - 0.013
0.136
- 0.107
- 0.029
0.061
- 0.144
0.084
0.009
0.053
- 0.043
- 0.174
0.018
0.155
406
-------
TABLE B.3. EXAMPLE OF THE SHAPIRO-MILK'S TEST: ORDERED OBSERVATIONS
1 - 0.174
2 - 0.144
3 - 0.130
4 - 0.107
5 - 0.043
6 - 0.029
7 - 0.013
8 - 0.009
9 0.018
10 0.053
11 0.061
12 0.084
13 0.136
14 0.142
15 0.155
407
-------
TABLE B.4. COEFFICIENTS FOR THE SHAPIRO-tflLK'S TEST1
1
2
3
4
5
2
\
0.7071
3
0.7071
0.0000
4
0.6872
0.1667
5
0.6646
0.2413
0.0000
e
0.6431
0.2806
0.0875
0.
0.
0.
0.
7
6233
3031
1401
0000
8
0.6052
0.3164
0.1743
0.0561
9
0.5888
0.3244
0.1976
0.0947
0.0000
10
0.5739
0.3291
0.2141
0.1224
0.0399
\
i\
1
2
3
4
5
6
7
8
9
10
11
\
0.5601
0.3315
0.2260
0.1429
0.0695
0.0000
12
0.5475
0.3325
0.2347
0.1586
0.0922
0.0303
13
0.5359
0.3325
0.2412
0.1707
0.1099
0.0539
0.0000
14
0.5251
0.3318
0.2460
0.1802
0.1240
0.0727
0.0240
15
0.5150
0.3306
0.2495
0.1878
0.1353
0.0880
0.0433
0.0000
0.
0.
0.
0.
0.
0.
0.
0.
16
5056
3290
,2521
1939
.1447
1005
0593
0196
17
0.4968
0.3273
0.2540
0.1988
0.1524
0.1109
0.0725
0.0359
0.0000
18
0.4886
0.3253
0.2553
0.2027
0.1587
0.1197
0.0837
0.0496
0.0163
19
0.4808
0.3232
0.2561
0.2059
0.1641
0.1271
0.0932
0.0612
0.0303
0.0000
20
0.4734
0.3211
0.2565
0.2085
0.1686
0.1334
0.1013
0.0711
0.0422
0.0140
\
1
2
3
4
5
6
7
g
9
10
11
12
13
14
15
21
\
0.4643
0.3185
0.2578
0.2119
0.1736
0.1399
0.1092
0.0804
0.0530
0.0263
0.0000
22
0.4590
0.3156
0.2571
0.2131
0.1764
0.1443
0.1150
0.0878
0.0618
0.0368
0.0122
23
0.4542
0.3126
0.2563
0.2139
0.1787
0.1480
0.1201
0.0941
0.0696
0.0459
0.0228
0.0000
24
0.4493
0.3098
0.2554
0.2145
0.1807
0.1512
0.1245
0.0997
0.0764
0.0539
0.0321
0.0107
25
0.4450
0.3069
0.2543
0.2148
0.1822
0.1539
0.1283
0.1046
0.0823
0.0610
0.0403
0.0200
0.0000
0
0
0
26
.4407
.3043
.2533
0.2151
0
0
0
0
0
0
0
0
0
.1836
.1563
.1316
.1089
.0876
.0672
.0476
.0284
.0094
27
0.4366
0.3018
0.2522
0.2152
0.1848
0.1584
0.1346
0.1128
0.0923
0.0728
0.0540
0.0358
0.0178
0.0000
28
0.4328
0.2992
0.2510
0.2151
0.1857
0.1601
0.1372
0.1162
0.0965
0.0778
0.0598
0.0424
0.0253
0.0084
29
0.4291
0.2968
0.2499
0.2150
0.1864
0.1616
0.1395
0.1192
0.1002
0.0822
0.0650
0.0483
0.0320
0.0159
0.0000
30
0.4254
0.2944
0.2487
0.2148
0.1870
0.1630
0.1415
0.1219
0.1036
0.0862
0.0697
0.0537
0.0381
0.0227
0.0076
From Conover (1980)
408
-------
TABLE B.4. COEFFICIENTS FOR THE SHAPIRO-WILK'S TEST (Continued)
\n
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
31
\
0.4220
0.2921
0.2475
0.2145
0.1874
0.1641
0.1433
0.1243
0.1066
0.0899
0.0739
0.0585
0.0435
0.0289
0.0144
0.0000
32
0.4188
0.2898
0.2462
0.2141
0.1878
0.1651
0.1449
0.1265
0.1093
0.0931
0.0777
0.0629
0.0485
0.0344
0.0206
0.0068
33
0.4156
0.2876
0.2451
0.2137
0.1880
0.1660
0.1463
0.1284
0.1118
0.0961
0.0812
0.0669
0.0530
0.0395
0.0262
0.0131
0.0000
34
0.4127
0.2854
0.2439
0.2132
0.1882
0.1667
0.1475
0.1301
0.1140
0.0988
0.0844
0.0706
0.0572
0.0441
0.0314
0.0187
0.0062
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
35
4096
2834
2427
2127
1883
1673
1487
1317
1160
1013
0873
0739
0610
0484
0.0361
0.0239
0.
0.
0119
0000
_
36
0.4068
0.2813
0.2415
0.2121
0.1883
0.1678
0.1496
0.1331
0.1179
0.1036
0.0900
0.0770
0.0645
0.0523
0.0404
0.0287
0.0172
0.0057
37
0.4040
0.2794
0.2403
0.2116
0.1883
0.1683
0.1505
0.1344
0.1196
0.1056
0.0924
0.0798
0.0677
0.0559
0.0444
0.0331
0.0220
0.0110
0.0000
38
0.4015
0.2774
0.2391
0.2110
0.1881
0.1686
0.1513
0.1356
0.1211
0.1075
0.0947
0.0824
0.0706
0.0592
0.0481
0.0372
0.0264
0.0158
0.0053
39
0.3989
0.2755
0.2380
0.2104
0.1880
0.1689
0.1520
0.1366
0.1225
0.1092
0.0967
0.0848
0.0733
0.0622
0.0515
0.0409
0.0305
0.0203
0.0101
0.0000
40
0.3964
0.2737
0.2368
0.2098
0.1878
0.1691
0.1526
0.1376
0.1237
0.1108
0.0986
0.0870
0.0759
0.0651
0.0546
0.0444
0.0343
0.0244
0.0146
0.0049
\ n
i\
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
41
\
0.3940
0.2719
0.2357
0.2091
0.1876
0.1693
0.1531
0.1384
0.1249
0.1123
0.1004
0.0891
0.0782
0.0677
0.0575
0.0476
0.0379
0.0283
0.0188
0.0094
0.0000
42
0.3917
0.2701
0.2345
0.2085
0.1874
0.1694
0.1535
0.1392
0.1259
0.1136
0.1020
0.0909
0.0804
0.0701
0.0602
0.0506
0.0411
0.0318
0.0227
0.0136
0.0045
43
0.3894
0.2684
0.2334
0.2078
0.1871
0.1695
0.1539
0.1398
0.1269
0.1149
0.1035
0.0927
0.0824
0.0724
0.0628
0.0534
0.0442
0.0352
0.0263
0.0175
0.0087
0.0000
44
0.3872
0".2667
0.2323
0.2072
0.1868
0.1695
0.1542
0.1405
0.1278
0.1160
0.1049
0.0943
0.0842
0.0745
0.0651
0.0560
0.0471
0.0383
0.0296
0.0211
0.0126
0.0042
0
0
0
0
0
0
0
0
0
0
0
45
.3850
.2651
.2313
.2065
.1865
.1695
.1545
.1410
.1286
.1170
.1062
0.0959
0.0860
0
0
0
0
0
0
0
0
0
0
.0765
.0673
.0584
.0497
.0412
.0328
.0245
.0163
.0081
.0000
46
0.3830
0.2635
0.2302
0.2058
0.1862
0.1695
0.1548
0.1415
0.1293
0.1180
0.1073
0.0972
0.0876
0.0783
0.0694
0.0607
0.0522
0.0439
0.0357
0.0277
0.0197
0.0118
0.0039
47
0.3808
0.2620
0.2291
0.2052
0.1859
0.1695
0.1550
0.1420
0.1300
0.1189
0.1085
0.0986
0.0892
0.0801
0.0713
0.0628
0.0546
0.0465
0.0385
0.0307
0.0229
0.0153
0.0076
0.0000
48
0.3789
0.2604
0.2281
0.2045
0.1855
0.1693
0.1551
0.1423
0.1306
0.1197
0.1095
0.0998
0.0906
0.0817
0.0731
0.0648
0.0568
0.0489
0.0411
0.0335
0.0259
0.0185
0.0111
0.0037
49
0.3770
0.2589
0.2271
0.2038
0.1851
0.1692
0.1553
0.1427
0.1312
0.1205
0.1105
0.1010
0.0919
0.0832
0.0748
0.0667
0.0588
0.0511
0.0436
0.0361
0.0288
0.0215
0.0143
0.0071
0.0000
50
0.3751
0.2574
0.2260
0.2032
0.1847
0.1691
6.1554
0.1430
0.1317
0.1212
0.1113
0.1020
0.0932
0.0846
0.0764 .
0.0685
0.0608
0.0532
0.0459
0.0386
0.0314
0.0244
0.0174
0.0104
0.0035
409
-------
TABLE B.5. EXAMPLE OF THE SHAPIRO-MILK'S TEST:
TABLE OF COEFFICIENTS AND DIFFERENCES
a,-
1 0.5150 0.329
2 0.3306 0.286
3 0.2495 0.266
4 0.1878 0.191
5 0.1353 0.104
6 0.0880 0.082
7 0.0433 0.031
410
-------
TABLE B.6. QUANTILES OF THE SHAPIRO-WILK'S TEST STATISTIC1
n
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
2*
29
34
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
0.01
0.753
0.687
0.686
0.713
0.730
0.749
0.764
0.781
0.792
0.805
0.814
0.825
0.835
0.844
0.851
0.858
0.863
0.868
0.873
0.878
0.881
0.884
0.888
0.891
0.894
0.896
0.898
0.900
0.902
0.904
0.906
0.908
0.910
0.912
0.914
0.916
0.917
0.919
0.920
0.922
0.923
0.924
0.926
0.927
0.928
0.02
0.756
0.707
0.715
0.743
0.760
0.778
0.791
0.806
0.817
0.828
0.837
0.846
0.855
0.863
0.869
0.874
0.879
0.884
0.888
0.892
0.895
0.898
0.901
0.904
0.906
0.908
0.910
0.912
0.914
0.915
0.917
0.919
0.920
0.922
0.924
0.925
0.927
0.928
0.929
0.930
0.932
0.933
0.934
0.935
0.936
0.05
0.767
0.748
0.762
0.788
0.803
0.818
0.829
0.842
0.850
0.859
0.866
0.874
0.881
0.887
0.892
0.897
0.901
0.905
0.908
0.911
0.914
0.916
0.918
0.920
0.923
0.924
0.926
0.927
0.929
0.930
0.931
0.933
0.934
0.935
0.936
0.938
0.939
0.940
0.941
0.942
0.943
0.944
0.945
0.945
0.946
0.10
0.789
0.792
0.806
0.826
0.838
0.851
0.859
0.869
0.876
0.883
0.889
0.895
0.901
0.906
0.910
0.914
0.917
0.920
0.923
0.926
0.928
0.930
0.931
0.933
0.935
0.936
0.937
0.939
0.940
0.941
0.942
0.943
0.944
0.945
0.946
0.947
0.948
0.949
0.950
0.951
0.951
0.952
0.953
0.953
0.954
0.50
0.959
0.935
0.927
0.927
0.928
0.932
0.935
0.938
0.940
0.943
0.945
0.947
0.950
0.952
0.954
0.956
0.957
0.959
0.960
0.961
0.962
0.963
0.964
0.965
0.965
0.966
0.966
0.967
0.967
0.968
0.968
0.969
0.969
0.970
0.970
0.971
0.971
0.972
0.972
0.972
0.973
0.973
0.973
0.974
0.974
0.90
0.998
0.987
0.979
0.974
0.972
0.972
0.972
0.972
0.973
0.973
0.974
0.975
0.975
0.976
0.977
0.978
0.978
0.979
0.980
0.980
0.981
0.981
0.981
0.982
0.982
0.982
0.982
0.983
0.983
0.983
0.983
0.983
0.984
0.984
0.984
0.984
0.984
0.985
0.985
0.985
0.985
0.985
0.985
0.985
0.985
0.95
0.999
0.992
0.986
0.981
0.979
0.978
0.978
0.978
0.979
0.979
0.979
0.980
0.980
0.981
0.981
0.982
0.982
0.983
0.983
0.984
0.984
0.984
0.985
0.985
0.985
0.985
0.985
0.985
0.986
0.986
0.986
0.986
0.986
0.986
0.987
0.987
0.987
0.987
0.987
0.987
0.987
0.987
0.988
0.988
0.988
0.98
1.000
0.996
0.991
0.986
0.985
0.984
0.984
0.983
0.984
0.984
0.984
0.984
0.984
0.985
0.985
0.986
0.986
0.986
0.987
0.987
0.987
0.987
0.988
0.988
0.988
0.988
0.988
0.988
0.988
0.988
0.989
0.989
0.989
0.989
0.989
0.989
0.989
0.989
0.989
0.989
0.990
0.990
0.990
0.990
0.990
0.99
1.000
0.997
0.993
0.989
0.988
0.987
0.986
0.986
0.986
0.986
0.986
0.986
0.987
0.987
0.987
0.988
0.988
0.988
0.989
0.989
0.989
0.989
0.989
0.989
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.991
0.991
0.991
0.991
0.991
0.991
0.991
0.991
0.991
1From Conover (1980).
411
-------
3. TEST FOR HOMOGENEITY OF VARIANCE
3.1 For Dunnett's Procedure and the t-test with Bonferroni's adjustment, the
variances of the data obtained from each toxicant concentration and the
control are assumed to be equal. Bartlett's Test is a formal test of this
assumption. In using this test, it is assumed that the data are normally
distributed.
3.2 The data used in this example are growth data from a Sheepshead Minnow
Larval Survival and Growth Test, and are the same data used in Appendices C
and D. These data are listed in Table B.7, together with the calculated
standard deviation for the control and each toxicant concentration.
3.3 The test statistic for Bartlett's Test (Snedecor and Cochran, 1980) is
as follows:
P -2 P 2
[(S V,. In S ) - S V,- In S,T
i=l i=l
B = _
Where:
= degrees of freedom for each effluent concen-
p =
tration and control, (v^
n,-
1)
number of levels of toxicant concentration
including the control
s2 [iE-iV
-------
3.5 For the data in this example,
The calculated B value is:
-2
= 2, p = 5, S = 0.0158, and C = 1.2,
2 [5(ln 0.0158) - E In
B = i
1.2
2[5( 4.1477) ( 22.1247)]
1.2
= 2.3103
3.6 Since B is approximately distributed as chi-square with p 1 degrees of
freedom when the variances are equal, the appropriate critical value for the
test is 13.3 for a significance level of 0.01. Since B is less than 13.3, the
conclusion is that the variances are equal.
TABLE B.7. SHEEPSHEAD MINNOW, CYPRINODON VARIEGATUS, LARVAL GROWTH DATA
(WEIGHT IN MG) USED FOR BARTLETT'S TEST FOR HOMOGENEITY OF
VARIANCE
Replicate
Control
Effluent Concentration (%)
6.25
12.5
25.0
50.0
1
2
3
Mean
s;
i
1.017
0.745
0.862
0.875
0.019
1
1.157
0.914
0.992
1.021
0.015
2
0.998
0.793
1.021
0.937
0.016
3
0.837
0.935
0.839
0.882
0.0031
4
0.715
0.907
1.044
0.889
0.027
5
4. TRANSFORMATIONS OF THE DATA
4.1 When the assumptions of normality and/or homogeneity of variance are not
met, transformations of the data may remedy the problem, so that the data can
be analyzed by parametric procedures, rather than a nonparametric technique
such as Steel's Many-one Rank Test or Wilcoxon's Rank Sum Test. Examples of
transformations include log, square root, arc sine square root, and recipro-
cals. After the data have been transformed, the Shapiro-Wilk's and Bartlett's
tests should be performed on the transformed observations to determine whether
the assumptions of normality and/or homogeneity of variance are met.
413
-------
4.2 Arc Sine Square Root Transformation1
4.2.1 For data consisting of proportions from a binomial (response/no
response; live/dead) response variable, the variance within the i-th treatment
is proportional to Pj (1 P,), where Pf is the expected proportion for the
treatment. This clearly violates the homogeneity of variance assumption
required by parametric procedures such as Dunnett's or a t-test since the
existence of a treatment effect implies different values of P,- for different
treatments, i. Also, when the observed proportions are based on small
samples, or when P,- is close to zero or one, the_normal ity assumption may be
invalid. The arc sine square root (arc sine / P ) transformation is commonly
used for such data to stabilize the variance and satisfy the normality
requirement.
4.2.2 Arc sine transformation consists of determining the angle (in radians)
represented by a sine value. In the case of arc sine square root
transformation of mortality data, the proportion of dead (or affected)
organisms is taken as the sine value, the square root of the sine value is
determined, and the angle (in radians) for the square root of the sine value
is determined. Whenever the proportion dead is 0 or 1, a special modification
of the arc sine square root transformation must be used (Bartlett, 1937). An
explanation of the arc sine square root transformation and the modification is
provided below.
4.2.3 Calculate the response proportion (RP) at each effluent concentration,
where:
RP = (number of dead or "affected" organisms)/(number exposed).
Example: If 8 of 20 animals in a given treatment die:
RP = 8/20
= 0.40
4.2.4 Transform each RP to arc sine, as follows:
4.2.4.1 For RPs greater than zero or less than one:
Angle (radians) = arc sine (RP)0'5.
Example: If RP = 0.40:
Angle = arc sine (0.40)0'5
= arc sine 0.6325
= 0.6847 radians
'From USEPA (1985d).
414
-------
4.2.4.2 Modification of the arc sine when RP = 0.
Angle (in radians) = arc sine (1/4N)0'5
Where: N = Number of animals/treatment
Example: If 20 animals are used:
Angle = arc sine (1/80)0'5
= arc sine 0.1118
= 0.1120 radians
4.2.4.3 Modification of the arc sine when RP = 1.0.
Angle = 1.5708 radians - (radians for RP = 0)
Example: Using above value:
Angle = 1.5708 - 0.1120
= 1.4588 radians
415
-------
APPENDIX C
DUNNETT'S PROCEDURE
1. MANUAL CALCULATIONS1
1.1 Dunnett's Procedure is used to compare each concentration mean with the
control mean to decide if any of the concentrations differ from the control.
This test has an overall error rate of alpha, which accounts for the multiple
comparisons with the control. It is based on the assumptions that the
observations are independent and normally distributed and that the variance of
the observations is homogeneous across all concentrations and control. (See
Appendix B for a discussion on validating the assumptions). Dunnett's
Procedure uses a pooled estimate of the variance, which is equal to the error
value calculated in an analysis of variance. Dunnett's Procedure can only be
used when the same number of replicate test vessels have been used at each
concentration and the control. When this condition is not met, the t-test
with Bonferroni's adjustment is used (see Appendix D).
1.2 The data used in this example are growth data from a Sheepshead Minnow
Larval Survival and Growth Test, and are the same data used in Appendices B
and D. These data are listed in Table C.I. One way to obtain an estimate of
the pooled variance is to construct an ANOVA table including all sums of
squares, using the following formulas:
TABLE C.I. SHEEPSHEAD MINNOW, CYPRINODON VARIEGATUS, LARVAL GROWTH
DATA (WEIGHT IN MG) USED FOR DUNNETT'S PROCEDURE
Effluent
Cone (%)
Replicate Test Vessel
1 2 3
Total
Mean
Control
6.25
12.5
25.0
50.0
1
2
3
4
5
1.017
1.157
0.998
0.873
0.715
0.745
0.914
0.793
0.935
0.907
0.862
0.992
1.021
0.839
1.044
2.624
3.063
2.812
2.647
2.666
0.875
1.021
0.937
0.882
0.889
Prepared by Laura Cast, Cathy Poore, Ron Freyberg, Florence Kessler, John
Menkedick and Larry Wymer, Computer Sciences Corporation, 26 W. Martin Luther
King Drive, Cincinnati, Ohio 45268; Phone (513) 569-7968.
416
-------
1.3 One way to obtain an estimate of the pooled variance is to construct an
ANOVA table including all sums of squares, using the following formulas:
Total Sum of Squares: SST = s Y?j - G2/N
Between Sum of Squares: SSB = 2 T/n,- - G2/N
i
Within Sum of Squares: SSW = SST - SSB
Where: G = The grand total of all sample observations; G = 2 T,
i
N = The total sample size; N = 2 n,-
i
n; = The number of replicates for concentration "i".
TJ = The total of the replicate measurements for concentration "i".
YJJ = The jth observation for concentration "i".
1.4 Calculations:
Total Sum of Squares: SST = 2 Y2j G2/N
ij
= 12.922 13.8122
15
= 0.204
Between Sum of Squares: SSB = 2 T/n, - G2/N
= 12.763 - (13.812)2/15
= 0.045
Within Sum of Squares: SSW = SST SSB
= 0.204 - 0.045
= 0.159
417
-------
1.5 Prepare the ANOVA table as follows:
TABLE C.2. GENERALIZED ANOVA TABLE
Source DF Sum of
Squares (SS)
Mean Square (MS)
(SS/DF)
Between p* - 1
SSB
= SSB/(p-l)
Within N - p
SSW
SSW/(N-p)
Total
N - 1
SST
*p = Number of different concentrations, including the control.
1.6 The completed ANOVA table for this data is provided below:
TABLE C.3. COMPLETED ANOVA TABLE FOR DUNNETT'S PROCEDURE
Source
DF
SS
Mean Square
Between 5-1=4 0.045
Within 15 - 5 = 10 0.159
0.011
0.016
Total
14
0.204
418
-------
1.7 To perform the individual comparisons, calculate the t statistic for each
concentration and control combination, as follows:
(Y1 Y,
L
SH / (l/n,) + (1/n,-)
Where: Y,- = Mean for each concentration
Y,, = Mean for the control
Sw = Square root of the within mean square
n., = Number of replicates in the control.
n( = Number of replicates for concentration "i".
1.8 Table C.4 includes the calculated t values for each concentration and
control combination.
TABLE C.4. CALCULATED T VALUES
Effluent
Concentration
(*)
6.25
12.5
25.0
50.0
i
2
3
4
5
t,
1.414
- 0.600
- 0.068
- 0.136
419
-------
1.9 Since the purpose of the test is only to detect a decrease in growth from
the control, a one-sided test is appropriate. The critical value for the
one-sided comparison (2.47), with an overall alpha level of 0.05, 10 degrees
of freedom and four concentrations excluding the control is read from the
table of Dunnett's "T" values (Table C.5; this table assumes an equal number
of replicates in all treatment concentrations and the control). Comparing
each of the calculated t values in Table C.4 with the critical value, no
decreases in growth from the control were detected. Thus the NOEC is 50.0%.
1.10 To quantify the sensitivity of the test, the minimum significant
difference (MSD) may be calculated. The formula is as follows:
MSD = d SH /(1/n, ) + (1/n)
Where: d = Critical value for the Dunnett's Procedure
Sw = The square root of the within mean square
n = The number of replicates at each concentration,
assuming an equal number of replicates at all
treatment concentrations
n1 = Number of replicates in the control
For example:
MSD = 2.47 (0.126)[ /(1/3) + (1/3) ] = 2.47 (0.126) /2/3
= 2.47 (0.126)(0.816)
= 0.254
1.11 For this set of data, the minimum difference between the control mean
and a concentration mean that can be detected as statistically significant is
0.254 mg. This represents a decrease in growth of 29% from the control.
1.11.1 If the data have not been transformed, the MSD (and the percent
decrease from the control mean that it represents) can be reported as is.
1.11.2 In the case where the data have been transformed, the MSD would be in
transformed units. In this case carry out the following conversion to
determine the MSD in untransformed units.
420
-------
1.11.2.1 Subtract the MSD from the transformed control mean. Call this
difference D. Next, obtain untransformed values for the control mean and the
difference, D.
MSDU = Controlu - Du
Where:
MSDU = The minimum significant difference for untransformed data
Controlu = The untransformed control mean
Du = The untransformed difference
1.11.2.2 Calculate the percent reduction from the control that MSDU
represents as:
MSDU
Percent Reduction = X 100
Controlu
1.11.3 An example of a conversion of the MSD to untransformed units, when the
arc sine square root transformation was used on the data, follows.
Step 1. Subtract the MSD from the transformed control mean. As an
example, assume the data in Table C.I were transformed by
the arc sine square root transformation. Thus:
0.875 - 0.254 = 0.621
Step 2. Obtain untransformed values for the control mean (0.875) and
the difference (0.621) obtained in Step 1, above.
[Sin(0.875)]2 = 0.589
[Sin(0.621)]2 = 0.339
Step 3. The untransformed MSD (MSD ) is determined by subtracting
the untransformed values obtained in Step 2.
MSDU = 0.589 - 0.339 = 0.250
In this case, the MSD would represent a 42% decrease in survival from
the control [(0.250/0.589)(100)].
421
-------
TABLE C.5. DUNNETT'S "T" VALUES1
(One-tailed) d k ,
X
5
6
7
8
»
10
11
12
13
H
15
16
17
18
IB
20
24
30
40
go
120
1
2.02
1.94
1.89
1.86
1.83
1.81
1.80
1.78
1.77
1.7«
1.75
1.75
1.74
1.73
1.73
1.72
1.71
1.70
1.68
1.67
1.66
1.64
2
2.44
2.34
2.27
2.22
2.18
2.15
2.13
2.11
2.09
2.08
2.07
2.06
2.05
2.04
2.03
2.03
2.01
1.99
1.97
1.95
1.93
1.92
3
2.68
2.56
2.48,
2.42
2.37
2.34
2.31
2.29
2.27
2.25
2.24
2.23
2.22
2.21
2.20
2.19
2.17
2.15
2.13
2.10
2.08
2.06
4
2.85
2.71
2.62
2.SS
2. SO
2.47
2.44
2.41
2.39
2.37
2.36
2.34
2.33
2.32
2.31
2.30
2.28
2.25
2.23
2.21
2.18
2.16
a . .05
5
2.98
2.83
2.73
2.66
2.60
2.56
2.53
2.50
2.48
2.46
2.44
2.43
2.42
.41.
.40
.39
.36
.33
.31
2.28
2.26
2.23
6
3.08
2.92
2.82
2.74
2.68
2.64
2.60
2.58
2.55
2.53
2.51
2.50
.49
.48
.47
. 4«
.43
2.40
2.37
2.35
2.32
2.29
7
3. 16
3.00
2.89
2.81
2.75
2.70
2.87
2.64
2.61
2.59
2.57
2. 5«
2.54
2.53
2.52
2.51
2.48
2.45
2.42
2.19
2.37
2.34
8
3.24
3.07
2.95
2.87
2.81
2.76
2.72
2.69
2.66
2.64
2.62
2.61
2.59
2.58
2.57
2.56
2.53
2.50
2.47
2.44
2.41
2.38
9
3'.30
3.12
3.01
2.92
2.86
2.81
2.77
2.74
2.71
2.69
2.67
2.65
2.64
2.62
2.61
2.60
2.57
2.54
2.51
2.48
2.45
2.42
1
3.37
3.14
3.00
2.90
2.82
2.76
2.72
2.68
2.65
2.62
2.60
2.58
2.57
2.55
2.54
2.53
2.49
2.46
2.42
2.39
2.36
2.33
2
.90
.61
.42
.29
.19
.11
.06
.01
.91
.94
.91
.88
.86
.84
.8}
.81
.77
.72
.68
.64
2.60
2.56
3
4.21
3.88
3.66
3.51
3.40
3.31
3.25
3.19
3.15
Ml
3.08
3.05
3.03
3.01
2.99
2.97
2.92
2.87
2.82
2.7S
2.73
2.68
4
4.43
4.07
3.83
3.67
3.55
3.45
3.38
3.32
3.27
3.23
3.20
3.17
3.14
3.12
3.10
3.08
3.03
2.97
2.92
2.87
2.82
2. 71
a- .01
4.60
4.21
.96
.79
.66
.56
.48
.42
.37
.32
.29
.26
.23
.21
.IS
.17
.11
.05
.99
.94
.89
.84
4.73
4.33
4.07
3.88
3.75
3.64
3.56
3.50
3.44
3.40
3.36
3.33
3.30
3.21
3.25
3.23
3.11
3.11
3.05
3.00
2.94
2.89
4.85
4.43
.15
.96
.82
.71
.63
.56
.51
.48
.42
.39
.36
.33
.31
.29
.22
.16
.10
.04
.99
.93
.94
.51
.23
.03
.89
.71
.69
.62
.56
.51
.47
.44
.41
.38
.36
.34
.21
.21
.14
.08
3.03
2.91
5.03
4.59
4.30
4.09
3.94
3. S3
3.74
3.67
3.61
3.M
3.52
3.48
3.45
3.42
3.40
3.38
3.31
3.24
3.18
3.12
3.06
3.00
Vrom Miller (1981)
422
-------
2. COMPUTER CALCULATIONS
2.1 This computer program incorporates two analyses: an analysis of variance
(ANOVA), and a multiple comparison of treatment means with the control mean
(Dunnett's Procedure). The ANOVA is used to obtain the error value.
Dunnett's Procedure indicates which toxicant concentration means (if any) are
statistically different from the control mean at the 5% level of significance.
The program also provides the minimum difference between the control and
treatment means that could be detected as statistically significant, and tests
the validity of the homogeneity of variance assumption by Bartlett's Test.
The multiple comparison is based on Dunnett, C. W., 1955, "Multiple Comparison
Procedure for Comparing Several Treatments with a Control," J. Amer. Statist.
Assoc. 50:1096-1121.
2.2 The source code for the Dunnett's program is structured into a series of
subroutines, controlled by a driver routine. Each subroutine has a specific
function in the Dunnett's Procedure, such as data input, transforming the
data, testing for equality of variances, computing P values, and calculating
the one-way analysis of variance.
2.3 The program compares up to seven toxicant concentrations against the
control, and can accommodate up to 50 replicates per concentration.
2.4 If the number of replicates at each toxicant concentration and control
are not equal, a t-test with the Bonferroni adjustment is performed instead of
Dunnett's Procedure (see Appendix D).
2.5 The program was written in IBM-PC FORTRAN (XT and AT) by D. L. Weiner,
Computer Sciences Corporation, 26 W. Martin Luther King Drive, Cincinnati,
Ohio 45268. A compiled version of the program can be obtained from the
Environmental Monitoring Systems Laboratory (EMSL-Cincinnati) by sending a
diskette with a written request.
2.6 Data Input and Output
2.6.1 Data on the proportion of surviving mysids, Mysidopsis bahia, from a
survival, growth and fecundity test, listed in Table C.6, below, are used to
illustrate the data input and output for this program.
2.6.2 Data Input
2.6.2.1 When the program is entered, the user has the following options:
1. Create a data file
2. Edit a data file
3. Perform ANOVA (analysis) on existing data set
4. Exit the program
423
-------
TABLE C.6. SAMPLE DATA FOR DUNNETT'S PROGRAM.
MYSIDS, HYSIDOPSIS BAHIA
PROPORTION OF SURVIVING
Replicate
Treatment
1 (Control)
2
3
4
5
6
0
1
1
0
0
0
.50
.00
.00
.50
.33
.00
0
1
0
0
0
0
.75
.00
.50
.00
.50
.50
0.67
1.00
0.67
0.75
1.00
0.33
0.
1.
1.
1.
1.
0.
67
00
00
00
00
50
0.50
1.00
0.50
1.00
0.00
0.00
1.00
0.50
1.00
1.00
0.33
0.50
1.00
0.50
1.00
0.67
0.50
0.50
0.00
0.50
0.50
0.67
0.00
0.50
2.6.2.2 When Option 1 (Create a data file) is selected, the program prompts
the user for the following information:
1. Number of groups, including control
2. For each group:
- Number of observations
- Data for each observation
2.6.2.3 After the data have been entered, the user may save the file on a
disk, and the program returns to the main menu (see below).
2.6.2.4 Sample data input is shown below.
424
-------
MAIN MENU AND DATA INPUT
l) Create a data file
2) Edit a data file
3) Perform ANOVA on existing data set
4) Stop
Your choice ? 1
Number of observations for group 1 ? 8
Biter the data for group 1 one observation at a time.
NO. 1? 0.80
NO. 2? 0.80
NO. 3? 1.00
NO. 4? 1.00
NO. 5? 1.00
NO. 6? 1.00
NO. 7? 1.00
NO. 8? 0.80
Number of observations for group 2 ? 8
Do you wish to save the data on disk ?y
Disk file for output ? sample
425
-------
2.6.3 Program Output
2.6.3.1 When Option 3 (Perform ANOVA on existing data set) is selected from
the main menu, the user is asked to select the transformation desired, and
indicate whether they expect the means of the test groups to be less or
greater than the mean for the control group (see below).
1) Create a data file
2) Edit a data file
3) Perform ANOVA on existing data set
4) Stop
Your choice ? 3
File name ? sample
Available Transformations
1) no transform
2) square root
3) loglO
4) arcsine square root
Your choice ? 4
Dunnett's test as implemented in this program is
a one-sided test. You must specify the direction
the test is to be run; that is, do you expect the
means for the test groups to be less than or
greater than the mean for the control group mean.
Direction for Dunnetts test : L=less than, G=greater than
426
-------
2.6.3.2 Summary statistics for the raw and transformed data, if applicable,
the ANOVA table, results of Bartlett's Test, the results of the multiple
comparison procedure and the minimum detectable difference are included in the
program output.
Smmary Statistics for Raw Data
Group
Mean
s.d.
cv%
1
= control
2
3
4
5
8
8
8
8
8
.9250
.9000
.8500
.7250
.1000
.1035
.1069
.1773
.2375
.1512
11.2
11.9
20.9
32.8
151.2
Summary Statistics and ANCVA
Transformation = Arcsine Square Foot
Group
Mean
s.d.
cv%
1
= control
2
3
4*
5*
8
8
8
8
8
1.3969
1.3390
1.2837
1.0570
.2015
.2400
.2478
.3181
.3066
.2863
17.2
18.5
24.8
29.0
142.1
*) the mean for this group is significantly less than
the control mean at alpha = 0.05 (1-sided) by Dunnett's test
Minumum detectable difference for Dunnett's test = -.316663
Tnis corresponds to a difference of -.192003 in original units
Bus difference corresponds to -19.79 percent of control
Between groups sum of squares = 7.826581 with 4 degrees of freedom.
Error mean square = .079230 with 35 degrees of freedom.
Bartlett's test p-value for equality of variances = .934
427
-------
2.7 Listing of Computer Program for Dunnett's Procedure.
$storage:2
c
c.
c EPA Dunnett's test program. : Version 1.1 by D.L.Weiner, 9/10/87
c Written for IBM PC and full compatibles. May require modifications
c for other systems. This version compiled using Microsoft FORTRAN
c compiled, V3.3.
c
c Driver program for D.exe
c
c This program does the following :
c 1) creation and/or editing of ascii files in the following
c format (data are for example purposes only):
c
c 1, 23.4 The first column denotes the group and
c 1, 17.6 the second column the data value. Note
c 1, 59.0 that it is assumed that group 1 is the
c 2, 44.1 control. Delimiters can be commas or
c 2, 50 blanks. Groups can have unequal sample
c 2, 51.7 sizes.
c 3, 49.8
c 3, 39.0
c 3, 56.2
c 4, 73.4
c 4, 64.9
c
c 2) temporary transformation of the data for analysis purposes -
c the transformed data are not permanently saved
c 3) a one-way ANOVA
c 4) Bartlett's test for equality of variances
c 5) Dunnett's test to compare the comtrol mean vs. each of the
c test group means if the sample sizes are equal; otherwise
c simple t-tests (using a pooled error term) are done with
c Bonferroni adjustment of p-value.
c
C PROGRAM RESTRICTIONS
c * number of groups must be between 2 and 8 (inclusive)
c * number of observations per group must be between 1 and 50
c
c
c Version 1.1 : D.L.Weiner, July,1987
c
program d
implicit real*8 (a-h,o-z)
real*8 mdd
character*79 title
character*! ans
dimension wk(2000), n(20), y(400), ip(20)
dimension wkt(2000), nt(20), yt(400)
data n/20*0/, nt/20*0/
c
c variables : ng=number of groups
c n(g)=array of untransformed sample sizes
c ntot-sum of n(i), i»l,g
c y(ntot)=array of response data
428
-------
2.7 Listing of Computer Program for Dunnett's Procedure (Continued).
c iunit=unit number for output
c nt(g)=array of sample sizes for transformed data
c yt(ntotal)=array of transformed data
c ntott=sum of nt(i), i=l,g
c title=user specified title for output
c ssb,ssw=between and within group sum of squares
c sst=total corrected sum of squares
c ems=error mean square with idf degrees of freedom
c wk()=work array - contains means and variances
c wkt()=work array - contains transformed means and va
c work arrays are also used for other purposes als
c mdd=minimum detectable difference
c
c see source code for individual subroutines for additional documenta
c
1 call input(ng,n,ntot,y,iunit,title)
call trans(ng,n,ntot,y,nt,yt,ntott,inum)
c
c summarize raw data if transformation requested in addition
c to summarizing raw data
c
idf = ntott - ng
if(ng.lt.2.or.idf.It.5) then
write(*,'(/a/)') ' Not enough data values for analysis."
goto 10
endif
if(ng.gt.S) then
writef*,'(/a/)') ' Too many groups for analysis.'
goto 10
endif
if(inum.ne.1) call oneway(ng,n,ntot,y,sst,ssb,ssw,wk)
call oneway(ng,nt,ntott,yt,sst,ssb,ssw,wkt)
ems = ssw / idf
call eqvar(ng,nt,wkt(ng-H),p)
c
c p is p value for Bartlett's test : if p > 1 then the test couldn't
c be run as one or more of the variances are zero
c
call dunnet(ng,idf,ems,wkt(l),nt,wkt(2*ng+l),iside,ip,mdd)
c
c call summary to summarize raw data if transformation requested
c in addition to summarizing raw data : next to last arg = 0
c means no ANOVA summary - n, mean, sd only
c
c summary is called twice - once for screen output and once for
c printer or disk output
c
call els
c
c summarize raw data here (if analysis is on transformed data)
c note that ssb,ssw,p,ip,isid« ar« not used in this call (dummy)
c
429
-------
2.7 Listing of Computer Program for Dunnett's Procedure (Continued)
if (inum.ne.l) then
call summary (ng,n,ntot,wk(l) ,wk(ng+l) ,ssb,ssw,p, ip,iside,
& title, 0,0, mdd)
pause ' '
call summary (ng,n,ntot,wk(l) ,wk(ng+l) ,ssb,ssw,p, ip, iside,
& title, iunit,0, mdd)
end if
c
c summarize transformed data here
c
call summary (ng,nt,ntott,wkt(l) ,wkt(ng+l) , ssb,ssw,p, ip, iside,
& title, 0, inum, mdd)
call summary (ng,nt,ntott,wkt(l) ,wkt(ng+l) ,ssb,ssw,p, ip, iside,
& title, iunit, inum, mdd)
write(iunit, ' (lx,al) ') char(12)
close (iunit)
10 write (*,' (/a\) ') ' Do you wish to restart the program ? '
read(*. ' (al) ') ans
if (ans.eq. 'V .or.ans.eq. "y1 ) goto 1
if (ans.ne. 'N1 .and. ans.ne. 'n' ) goto 10
stop 'Normal ending. '
end
430
-------
2.7 Listing of Computer Program for Dunnett's Procedures.
$storage:2
c
c
c
c
c
c
c
c
c
c
c
c
input.for
on output
data input routine
variable type description
g
ntot
n(g)
y(ntot)
iunit
title
i2
i2
12
r8
12
a79
number og groups
total I of obs.
# obs. per group
data values
unit for output
title
&
&
&
&
&
&
subroutine input(g,n,ntot,y,iunit,title)
implicit real*8 (a-h,o-z)
dimension y(l),n(l)
integer*2 g
character*64 fname
character*79 title
character*! ans
logical iochk
call els
call gotorc(8,l)
write(*,'(a/,a/,a/,a/,a)')
P A Dunnett''s
Version 1.1
Program
els
2
write(*,'(/a\)') ' Title ? '
read(*,'(a79)') title
write(*,'(/a\)') ' Output to printer or disk file ? '
read(*,'(al)') ans
if (ans.ne.'p'.and.ans.ne.'P'.and.ans.ne.*d'.and.ans.ne.'D') then
write(*,*) ' Please respond with a p or a d '
pause
call
goto
endif
if(ans.eq.'p'.or.ans.eq.'P1) then
iunit - 20
open(iunit,flie-'prn1)
goto 5
endif
iunit - 10
write(*,'(/a\)') Disk file for output ? '
read(*,'(a64)') fname
inquire(file-fname,exist-iochk)
if(iochk) then
write (*,'(/a\)«) ' File already exists, overwrite it ?
read(*,'(al)') ans
if(ans.eq.'N1.or.ans.eq."n") goto 3
if(ans.ne.'Y1.and.ans.ne.'y1) then
431
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
write(*,*) ' Please answer yes or no.
goto 4
endif
endif
open(iunit,file-fname,access-'sequential',status-'new')
c
5 call els
write(* *) ' 1) Create a data file1
write(*
write(*
write(*
write(*
*) '2) Edit a data file1
*) '3) Perform ANOVA on existing data set1
*) ' 4) Stop1
'(/a\)') ' Your choice ? '
read(*,'(al)') ans
if(ans.ne.'I1.and.ans.ne.'2'.and.ans.ne.'3'.
& and.ans.ne.'4') then
write(*,'(/a/)') ' Please enter a number 1,2 or 3'
goto 6
endif
if(ans.eq.'4') stop 'Normal Ending.1
if(ans.eq.'3') goto 30
if(ans.eq.'2') goto 100
c
c input from keyboard here
c
15 call inkb(g,n,ntot,y)
call els
16 write(*,'(/a\)') ' Do you wish to save the data on disk ?'
read(*.'(al)') ans
if(ans.eq.'N1.or.ans.eq."n1) then
write(*,'(/a,a\)') ' Are you sure ? Data will be lost ! ',
& ' (Y or N) ? '
read(*.'(al)') ans
if(ans.eq.'Y'.or.ans.eq.'y') goto 5
goto 16
endif
if(ans.ne.'y1.and.ans.ne.'Y') then
write(*,'(/a)') ' Please respond yes or no.1
goto 16
endif
write(*,'(/a\)') ' Disk file for output ? '
read(*,'(a64)') fname
inquire(file-fname,exist-iochk)
if(iochk) then
write (*,'(/a\)') ' File already exists, overwrite it ? '
read(*,'(al)') ans
if(ans.eq.'N'.or.ans.eq.'n') goto 16
endif
iunitl - 11
open(iunitl,file-fname,access-'sequential,status-'new')
1-0
do 20 j-l,g
do 22 k-l,n(j)
1-1 + 1
22 write(iunitl,'(Ix,i4,2x,a,2x,fl8.6)'
20 continue
432
-------
9
ntot
n(g)
y(ntot)
i2
i2
i2
r8
number og groups
total I of obs.
# obs. per group
data values
2.7 Listing of Computer Program for Dunnett's Procedures (Continued)
close(iunitl)
25 goto 5
c
c input from disk file here
c
30 continue
call readf(g,n,ntot,y,iflag,fname)
if(iflag.eq.O) return
goto 5
100 continue
call readf(g,n,ntot,y,iflag,fname)
if(iflag.eg.0) call edit(g,n,ntot,y,fname)
goto 5
end
c
c inkb.for - keyboard input
c
c on output : variable type description
c
c
c
c
c
c
subroutine inkb(g,n,ntot,y)
implicit real*8(a-h,o-z)
dimension y(l),n(l)
integer g t
call els
5 write(*,'(/a\)') ' Number of groups, including control ? '
read(*,*,err-10) g
if(g.le.l.or.g.gt.S) then
write(*,*) ' The number of groups must be 2 to 8 (inclusive)'
goto 5
endif
goto 20
10 write(*,'(/a\)') ' Invalid number of groups, please reenter.'
goto 5
20 continue
1-0
do 200 i-l,g
22 write(*,'(/a,i2,a,\)') ' Number of observations for group ',
& i,1 ? '
read(*,*,err-25) n(i)
if(n(i).le.O.or.n(i).gt.50) then
write(*,*) ' The number of observations per group must be ',
& '1 to 50'
goto 22
ndif
goto 28
25 writ«(*,'(/a/)') ' Invalid number. Please reenter.1
goto 22
28 write(*,'(/a,i2,a)') ' Enter the data for group ',i,
& ' one observation at a time.'
do 100 j-l,n(i)
433
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued)
g
ntot
n(g)
y(ntot)
iflag
fname
i2
i2
i2
r8
i2
a64
number og groups
total f of obs.
1 obs. per group
data values
0 - ok read, 1 = not ok
file that was read in
30 write(*,'(/a,i2,a,\)') ' NO. ',j,'?
read(*,*,err=35) ynum
1-1+1
y(l) - ynum
goto 100
35 write(*.'(/a/)') ' Invalid number. Please reenter.'
goto 30
100 continue
200 continue
ntot - 1
return
end
c
c readf.for - read a data file
c
c on output : variable type description
c
c
c
c
c
c
c
c
subroutine readf(g,n,ntot,y,iflag,fname)
implicit real*8 (a-h,o-z)
integer g
character*64 fname
character*! ans
logical iochk
dimension y(l),n(l)
call els
iflag - 0
30 write(*,'(/a\)') ' File name ?
read(*,'(a64)') fname
if(fname.eq. ') goto 31
inquire(file-fname, exist-iochk)
if(iochk) goto 50
31 write(*,*) ' The file you specified does not exist, or you need"
write(*,*) ' to specify a different disk as part of the name.1
write(*,'(/a\)') 'Do you wish to reenter the name ?
read(*,'(al)') ans
if (ans.eq.'Y1.or.ans.eq.'y1) goto 30
iflag - 1
return
50 continue
c begin file read
iunit - 30
open(iunit,file-fname,access-'sequential',status-'old')
read(iunit,*,«nd-l00) igrp, x
c i is obsf, ig is groupl, nn counts fobs per group
i - 1
ig - 1
nn - 1
ilag - ig
434
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
60
100
150
c
c
c
c
c
c
c
c
c
c
c
1
5
iglag = igrp
read(iunit,*,end-100,err=150) igrp, x
if(iglag.eq.igrp) nn - nn + 1
if(iglag.ne.igrp) then
n(ig) - nn
ig - ig + 1
iglag = igrp
nn = 1
endif
i - i + 1
y(i) - x
if(nn.gt.SO) then
write(*,'(\a,i2,a\)')
' Max « 50'
iflag - 1
pause
return
endif
goto 60
continue
n(ig) - nn
ntot - i
g - ig
close(iunit)
return
continue
write(*,'(a,i2,a,a)') ' There is an
'your data file. Please correct
pause
iflag - 1
return
end
Too many observations for group ',ig,
error
it.'
on line ',i,' of
edit.for
on input
file edit routine
variable type
g i2
ntot i2
n(g) i2
y(ntot) r8
fname a64
description
number og groups
total # of obs.
# obs. per group
data values
file to be edited
subroutine edit(g,n,ntot,y,fname)
implicit real*8(a-h,o-z)
dimension y(l),n(l)
integer g
character*64 fname
call els
fuzz l.e-20
iunit - 30
open(iunit,file»fname,access-'sequential'
write(*,'(/a\)') ' Edit values for which
read(*,*,err-10) ig
,status-
group ?
old')
435
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
if(ig.ge.l.and.ig.le.g) goto 100
10 write(*,'(//a,i2,/)') ' Please respond 1 - ",g
pause
goto 1
100 continue
call els
call gotorc(0,0)
write(*,'(a,i2,//)') ' The following values are for group ',ig
call gotorc(5,0)
loff «= 0
if (ig.eq.l) goto 105
c compute offset
do 102 i=l,ig-l
102 loff - loff + n(i)
c
105 'Jo 110 j=l,n(ig)
irow = 5 + (j-1) / 4
k - mod(j,4)
if(k.eq.l) icol = 0
if(k.eq.2) icol - 20
if(k.eq.3) icol - 40
if(k.eq.O) icol - 60
call gotorc(irow,icol)
110 write(*,'(f!8.6,\)') y(loff + j)
120 call gotorc(23,0)
write(*,'(/a\)')
& ' Modify a value, Edit a different Group or quit (M,E,Q) ? '
read(*,'(al)') ans
c
if(ans.eq.'q1.or.ans.eq.'Q') goto 200
c
if(ans.eq.'e'.or.ans.eq.'E1) then
call els
goto 1
end if
c
if(ans.ne.'m1.and.ans.ne.'M') then
write(*,'(/a\)') ' Please respond with a M, E or Q1
pause
goto 100
endif
c
write(*,'(/a\)') ' Enter old value, new value: '
read(*,*,err-140) yold, ynew
do 130 i=l,n(ig)
130 if(dabs(y(loff + i)-yold).le.fuzz) goto 150
write(*,*) ' The old value which you specified does not exist1
pause
goto 100
140 write(*,*) ' one or both of the numbers which you have entered1
write(*,*) ' are invalid. Please reenter the numbers.1
pause
goto 100
150 y(loff + i) « ynew
goto 100
436
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued)
200 continue
write(*,'(/a\)') ' Do you wish to save the changes ? '
read(*,'(al)') ans
if(ans.eq.'n'.or.ans.eq.'N1) then
close(iunit)
return
endif
if(ans.ne.'y1.and.ans.ne.'Y') goto 200
close(iunit)
open(iunit,file»fname,access"'sequential',status-'new')
1-0
do 250 j-l,g
do 252 k-l,n(j)
1 = 1 + 1
252 write(iunit,'(Ix,i4,2x,a,2x,fl8.6)') j, ','
250 continue
close(iunitl)
return
end
437
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued)
$storage:2
c 0 ... 79
c cursor positioning, goto (irow,icol) :
c 24
subroutine gotorc (irow,icol)
character*! dummy
if (irow.lt.0.or.irow.gt.24) irow = 0
if (icol.lt.o.or.icol.gt.79) icol = 0
read (dummy,'(Ix)')
call locate (0,0,ier)
write (*,'(\)')
call locate (irow,icol,ier)
read (dummy,'(Ix)')
return
end
438
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
$storage
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
:2
trans . for
on input :
on output
subroutine
- computes transformed data
variable type description
g i2 number of groups
n() i2 n values for each group
ntot i2 total 1 of obs.
y() r8 data values
: variable type description
nt() i2 n values after transformati
yt() r8 data values after transform
ntott r8 ntot after transformations
inum i2 transformation number
trans (g,n, ntot, y,nt,yt, ntott, inum)
implicit real*8 (a-h,o-z)
c
5
integer g
dimension n
iflag - 0
call els
write (* *)
write (* *)
write (* *)
write (* *)
write(* *)
(1) , y(l) , nt(l) , yt(l)
Available Transformations'
1) no transform1
2) square root1
3) loglO'
4) arcsine square root'
write(* '(/a\)') ' Your choice ? '
read(*,*,err-50) inum
if(inum.It.1.or.inum.gt.4) goto 50
goto 60
50 write(*,*) ' Please answer 1-4 '
pause ' '
goto 5
60 k - 0
kk - 0
ntott - 0
do 52 i-l,g
nt(i) - 0
do 54 j=l,n(i)
k - k + 1
goto(70,72,74,76),inum
70 temp - y(k)
goto 53
72 if (y(k).lt.O.dO) then
iflag - 1
goto 54
endif
temp - dsqrt(y(k))
goto 53
74 if (y(k).le.O.dO) then
iflag - 1
439
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
goto 54
end if
temp - dloglO(y(k))
goto 53
76 if (y(k).It.O.dO.or.y(k).gt.l.dO) then
iflag - 1
goto 54
end if
temp = dasin(dsqrt(y(k)))
53 kk - kk + 1
nt(i) « nt(i) + 1
ntott - ntott + 1
yt(kk) » temp
54 continue
52 continue
c
if(iflag.gt.O) then
write(*,'(/a)')
& ' One or more data values could not be transformed. These '
write(*,'(a/)') ' values will not be included in the analyses.1
pause ' '
endif
c
c check to see if each group has at least 1 observation
c
do 100 i-1, g
100 if(nt(i).le.O) goto 110
return
110 call els
write(*,'(a,a,i2)') ' After transformation, all of the values',
& ' are missing for group ', i
write(*,'(/a)') ' ANOVA cannot be performed. Program ending.1
stop ' '
end
440
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
$storage:2
c
c oneway . for
c
c on input :
c
c
c
c
c
c
c on output :
c
c
c
c
c
c
computes
variable
g
n()
ntot
y()
wk()
variable
wk(l)-wk(g)
wk(g+l)-wk(g+g)
sst
ssb
ssw
oneway
type
i2
i2
i2
r8
r8
type
r8
r8
r8
r8
r8
anova
description
number of groups
n values for each group
total sample size
data values (possible trans
work array (dimensioned 2*g
description
group means
group variances
corrected total sum of s
between group sum of squ
within sum of squares
subroutine oneway(g,n,ntot,y,sst,ssb,ssw,wk)
implicit real*8 (a-h,o-z)
integer*2 g
dimension n(l), y(l) , wk(l)
c
sy - 0.0
syy - 0.0
sst - 0.0
ssb = 0.0
ssw - 0.0
k «= 0
temp - 0.0
c
do 10 i»l, 2*g
10 wk(i)=0.0
c
do 20 i-1, g
do 30 j-l,n(i)
k - k + 1
sy - sy + y(k)
syy - syy + y(k) * y(k)
wk(i) - wk(i) + y(k)
30 wk(g+i) - wk(g+i) + y(k)*y(k)
c
c compute and store ith group variance in wk(g+i)
c
wk(g+i) - wk(g-t-i) - wk(i) * wk(i) / n(i)
if(n(i).gt.l) wk(g+i) - wk(g+i) / (n(i) - 1)
temp - temp + wk(i) * wk(i) / n(i)
c
c compute and store ith mean in wk(i)
c
20 wk(i) - wk(i) / n(i)
c
sst - syy - sy * sy / ntot
ssb - temp - sy * sy / ntot
ssw - sst - ssb
return
end
441
-------
g
n()
var(g)
y()
variable
P
i2
i2
12
r8
type
r8
number of groups
n values for each group
array of variances
data values (possible trans
description
Bartlett's test p-value
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
$storage:2
c eqvar.for - computes tests for equality of variances
c on input : variable type description
c
c
c
c
c
c
c on output :
c
c
c
subroutine egyar(g,n,var,p)
implicit real*8(a-h,o-z)
integer g
dimension var(l), n(l)
c
vmin - var(l)
vmax - var(l)
imin - 1
imax - 1
vsum - var(l)
sse - (n(l)-l) * var(l)
idf - n(l) - 1
cl - O.dO
c2 - O.dO
if(var(l).le.l.d-10) goto 100
cl = (n(l)-l) * dloglO(var(l))
if ((n(i)-l).gt.O) c2 - l.dO / (n(l)-l)
do 10 i=2,g
vsum - vsum + var(i)
sse - sse + (n(i)-l) * var(i)
idf = idf + (n(i) - 1)
if(var(i).le.l.d-10) goto 100
cl - cl + (n(i)-l) * dloglO(var(i))
if((n(i)-l).gt.O) c2 - c2 + l.dO / (n(i)-l)
if (var(i).le.vmin) then
vmin - var(i)
imin - i
end if
if (var(i).gt.vmax) then
vmax « var(i)
imax - i
endif
10 continue
f - vmax / vmin
idfl - n(imax) - 1
idf2 - n(imin) - 1
call pvalue (3,2,f,idfl,idf2,pl)
c
c note pi - p-value for F max test
c
442
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued)
c = l.dO + (l.dO / (3*(g-l))) * (c2 - l.dO/idf)
chi - 2.303 * (idf * dloglO(sse/idf) - cl)
chi = chi / c
idfl - g-1
idf2 - 0
call pvalue (4,2,chi,idfl,idf2,p)
return
100 continue
c set p to 2 (a flag) if 1 or more variances - 0
p - 2.do
return
end
443
-------
itype
112
value
idfl
idf2
variable
P
i2
12
r8
i2
i2
type
r8
l=t, 2-z, 3=f, 4=chi square
1=1 sided, 2=2 sided
value of test statistic
1st degrees of freedom t,z,
2nd degrees of freedom f
description
p-value
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
$storage:2
c
c subroutine pvalue.for - compute p values
c
c
c on input : variable type description
c
c
c
c
c
c
c
c on output :
c
c
c
subroutine pvalue(itype,i!2,value,idf1,idf2,p)
implicit real*8 (a-h,o-z)
goto (10,11,12,13) itype
10 t - value
idf - idfl
aa-idf/2.do
bb-0.5dO
xx-l.dO/(l.dO+(t**2)/idf)
bta-beta(xx,aa,bb)
if(i!2.eq.l.and.t.gt.0.dO) p-bta/2.dO
if(i!2.eq.l.and.t.lt.0.dO) p-(l.dO-bta/2.dO)
if(i!2.eq.2) p - bta
return
11 z - value
aa-l.dlO/2.dO
bb-0.5dO
xx-l.dlO/(l.dlO+(z**2))
bta-beta(xx,aa,bb)
if(i12.eq.1.and.z.gt.0.do) p-bta/2.do
if(i!2.eq.1.and.z.It.0.do) p-(1.dO-bta/2.dO)
if(i!2.eq.2) z - bta
return
12 f - value
aa-idf2/2.dO
bb-idfl/2.dO
xx=l.dO/(l.dO+(idfl*f)/idf2)
p-beta(xx,aa,bb)
return
13 chi - value
idf - idfl
aa-l.dlO/2.dO
bb-idf/2.dO
xx-l.dlO/(l.dlO+chi)
p-beta(xx,aa,bb)
return
stop
end
444
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued)
double precision function b«ta(xx,aa,bb)
implicit real*8(a-i,k-z)
integer*2 j
ier - 0
BETA=1.0
IF (XX - 1.0) 1,30,30
1 BETA-0.0
IF (XX) 30,30,4
4 A-AA
B-BB
x-xx
LO-DLOG(X)
Ll-DLOG(l.O-X)
M-DLGAMA(A) + DLGAMA(B) - DLGAMA(A+B)
5 IF (A-1.5) 6,7,7
6 BETA-BETA+DEXP(A*LO+B*L1-M)/A
M-M+DLOG(A)-DLOG(A+B)
A-A+1.0
GO TO 5
7 IF (B-1.5) 8,9,9
8 BETA-BETA-DEXP(A*LO+B*L1-M)/B
M-M+DLOG(B)-DLOG(A+B)
B-B-H.O
GO TO 7
9 Y-1.28155156553942DO
L-(Y*Y-3.0)/6.0
H-(2.0*A-1.0)*(2.0*B-1.0)/(A+B-1.0)
Z-(L+(5.0*H-4.0)/(6.0*H))*(A-B)/(A+B-1.0)*2.0/H
L-DSQRT(H+L)/H
CUTY*L-Z
CUT-A/(A+B*DEXP(2.0*CUT))
IF (CUT-0.5) 12,12,10
10 CUT-Y*L-Z
CUT-A/(A+B*DEXP(2.0*CUT))
IF (CUT-0.5) 11,12,12
11 CUT-0.5
12 H-1.0
IF (X-CUT) 14,14,13
13 H-LO
LO-L1
Ll-H
X-l.O-X
H-A
A-B
B-H
BETA-BETA+1.0
H1.0
14 EPS-l.OD-16
M-DEXP(A*LO+(B-l.0)*L1-M) /A
X-X/(1.0-X)
L-0.0
Y-1.0
Z-1.0
DO 18 J-1,100
I-J
445
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
L0=(I-B)*(A+I-l. 0)/(A+2.0*1-2.0)*X/(A+2.0*1-1.0)
Ll=(A+B+I-1.0)*!/(A+2.0*1)*X/(A+2.0*1-1.0)
L-L*LO+M
Z-Z*LO+Y
M=M*L1+L
Y-Y*L1+Z
IF (Z) 15,18,15
15 S«=L/Z
T-M/Y
IF (T) 16,20,16
20 IF (S) 18,19,18
16 IF (DABS(S/T-1.0)-EPS) 19,19,18
18 CONTINUE
19 BETA-BETA+H*T
30 RETURN
RETURN
END
DOUBLE PRECISION FUNCTION DLGAMA(XX)
DOUBLE PRECISION XX,ZZ,TERM,RZ2,DLNG
IER=0
zz-xx
IF (XX-1.D10) 2,2,1
1 IF (XX-1.D35) 8,9,9
2 IF (XX-l.D-9) 3,3,4
3 IER1
DLNO-'1.D38
GO TO 10
4 TERM-l.DO
5 IF (ZZ-18.DO) 6,6,7
6 TERM-TERM*ZZ
ZZ-ZZ+l.DO
GO TO 5
7 RZ2«1.DO/ZZ**2
DLNG »(ZZ-0.5DO)*DLOG(ZZ)-ZZ +0.9189385332046727 -DLOG(TERM)+
l(l.DO/ZZ)*(.8333333333333333D-l-(RZ2*(.2777777777777777D-2+(RZ2*
2(.7936507936507936D-3-(RZ2*(.5952380952380952D-3)))))))
GO TO 10
8 DLNG-ZZ*(DLOG(ZZ)-1.DO)
GO TO 10
9 IER-+1
DLNG-1.D38
10 DLGAMA-DLNG
RETURN
END
446
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
$storage:2
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
10
20
c
c
c
22
subroutine dunnet.for - compute p values
on input
variable
type
description
on output
ng
idf
ems
mean(ng)
n(ng)
t(8,50)
iside
variable
ip(ng)
12
i2
r8
r8
12
r8
12
type
12
number of groups 2<-ng<=8
degrees of freedom for erro
error mean square
array of means
n per each group
work array
0-trts lower, 1-trts higher
description
0-NS, 1-sig @ alpha=0.05
HDD
r8
ip(l) - 0 -> Dunnetts te
ip(l) - 1 => Bonferroni
min. detectable diff. in
original units
note : the calling program must check to see that 2<-ng<=8 and
that idf is >- 5
subroutine dunnet(ng, idf ,ems,mean,n,t, iside, ip,mdd)
implicit real*8 (a-h,o-z)
real*8 mean(l) , mdd
character*! ans
logical iochk
dimension t (7,49) , ip(l) ,
read in Dunnett's t values
n(l)
inquire ( f lie- ' dunnet . f 11 ' , exist-iochk)
if (iochk) goto 10
write (*,*) ' The file containing Dunnett t values is not on the'
write (*,*) ' default drive. The file name is DUNKET.FIL '
write (*,*) ' Please copy it over to the default drive and rerun'
write (*,*) ' this program. '
stop ' '
open ( 97, file-1 dunnet. fil1 , status-1 old ')
do 20 j-1,49
r«ad(97, ' (7f5.2) ') (t(i,j) ,1-1,7)
close (97)
read in direction for the test
call cl«
write(*,*)
write(*,*)
write(*,*)
write(*,*)
write(*,*)
Dunnett''s test as implemented in this program is',
a one-sided test. You must specify the direction ',
the test is to be run; that is, do you expect the1
means for the test groups to be less than or '
greater than the mean for the control group mean.'
447
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
write (*,'(/a\)')
& ' Direction for Dunnetts test : L-less than, G-greater than ?
read(*. ' (al) ') ans
if (ans.ne. '!' .and. ans. ne. 'L1 .and. ans. ne. 'g1 .
& and . ans . ne . ' G ' ) then
write(*,*) ' Please respond L or G. '
goto 22
endif
if (ans.eq. '!' .or.ans.eq. 'L') iside=0
if (ans. eg. 'g' .or. ans. eg. 'G') iside-1
c
cmean - mean(l)
c check to see if sample sizes are equal : if not, do Bonferroni
c
do 30 i-2,ng
if(n(i)-n(i-l)) 100, 30, 100
3 0 continue
xn - l.dO * n(l)
c
denom - dsqrt(2.dO * ens / xn)
c
c recover Dunnett's t value
c
icol - ng - 1
if (idf.le.50) dunt - t(icol , idf - 4)
if (idf. gt.50.and.idf.lt. 60) dunt - t(icol , 46)
if (idf. ge.60.and.idf.lt. 120) dunt - t(icol , 47)
if (idf. ge.120.and.idf.lt. 150) dunt - t(icol , 48)
if (idf. ge. 150) dunt - t(icol , 49)
c
ipd) 0
do 50 i-2,ng
if (iside.eq.O) diff - cmean - mean(i)
if (iside.eq.l) diff - mean(i) - cmean
stat - diff / denom
50 if ( stat. gt. dunt) ip(i)-l
c
c compute HDD
c
mdd » dunt * denom
if (iside.eq.O) mdd - -l.dO * mdd
c fixup if transformed is done by summary
return
100 continue
c Bonferroni adjustment here
ipd) - 1
alpha - O.OSdO / (ng - 1)
pctile - l.dO - alpha
tval - tinv (pctile, idf)
nl - n(l)
do 150 i-2,ng
ni -
448
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued)
denom - dsqrt(ems * (l.dO/nl + l.dO/ni) )
if (iside.eq.O) diff - ernean - mean(i)
if (iside.eq.l) diff - mean(i) - cmean
stat - diff / denom
150 if (stat.gt.tval) ip(i)=l
c
c compute HDD
c
denom - dsqrt(2.dO * ems / nl)
mdd - tval * denom
if(iside.eq.O) mdd - -l.dO * mdd
c fixup if transformed is done by summary
return
end
449
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
$storage:2
c
summary.for
on input
variable
computes oneway anova
type
description
g i2 number of groups
n() 12 n values for each group
ntot i2 total sample size
mean() r8 data values (possible trans
var() r8 work array (dimensioned 2*g
ssb r8 between group sum of sgu
ssw r8 within sum of squares
p r8 Bartlett's test p-value
ip() i2 flag for Dunnett's test res
0-ns, 1-sig
ip(l) - 0 > Dunnetts test
ip(l) - 1 -> Bonferroni t-t
iside i2 direction of Dunnett's test
0-lower, 1-upper
title a79 title
iunit i2 unit | for output
inum i2 0 means summarize raw data,
1-no trans, 2-sqrt,
3-loglO, 4-arcsine
HDD r8 min. detectable diff. for
Dunnett'a test
subroutine summary(g,n,ntot,mean,var,ssb,ssw,p,ip,iaide,title,
iunit,inum,mdd)
implicit real*8 (a-h,o-z)
real*8 mean(l), mdd, mddl
integer*2 g
character*79 title
character*20 tran(4)
character*12 direct
character*! pstar
dimension n(l), var(l), ip(l)
if(inum.eq.
tran(l)-'
tran(2)-'
tran(3)'
0)
goto 2
None '
Square Root '
Logio '
tran(4)-'Arcsine Square Root'
convert mdd to original units
if (inum.eq.l) then
tempc - mean(l)
tempt » mean(l) + mdd
mddl tempc - tempt
endif
if (inum.eq.2)
endif
then
tempc - mean(l) ** 2
tempt (mean(l) + mdd)
ddl - tempc - tempt
** 2
450
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued)
if (inum.eq.3) then
tempo = 10.0 ** (mean(l))
tempt - 10.0 ** (mean(l) + mdd)
mddl - tempc - tempt
endif
if (inum.eq.4) then
tempc - (dsin(mean(l)))**2
tern - (mean(l) + mdd)
tempt - (dsin(tem))**2
mddl - tempc - tempt
endif
mddl - (dabs(mdd)/mdd) * dabs(mddl)
pctrl - 100.do * mddl / dabs(tempo)
c
c check to see if all sample sizes are equal
c
2 ieqn - 0
do 3 i - 2, g
if(n(i) - n(i-l)) 4,3,4
3 continue
goto 5
4 ieqn - 1
c
5 if(iunit.gt.O) write(iunit,'(lx,al)') char(12)
write(iunit,'(/,lx,a/)') title
if(inum.eq.O) write(iunit,'(/,15x,a//)')
& ' Summary Statistics for Raw Data1
if(inum.ge.l) then
write(iunit,'(/,17x,a)')
S, * summary Statistics and ANOVA1
write(iunit,'(/,13x,a,a/)') Transformation- ',tran(inum)
endif
write(iunit,*)
& ' Group n Mean a.d. cv%'
write(iunit,*)
& '
if(var(l).gt.O.dO) then
sd - dsqrt(var(l))
else
sd - O.dO
endif
if(mean(l).ne.O.dO) then
cv - dabs(100.dO * sd / mean(l))
else cv - O.dO
ndif
write (iunit,30) n(l),mean(l),sd,cv
30 format (' 1 - control1,lx,i2,3x,f12.4,lx,f12.4,9x,f6.1)
&
do 10 i-2,g
if(var(i).gt.O.dO) then
d dsqrt(var(i))
else
d - O.dO
endif
if(mean(i).ne.O.dO) then
451
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
cv - dabs(100.dO * sd / mean(i))
else
cv » O.dO
endif
pstar-' '
if(ip(i).eq.l) pstar-1*'
if(inum.ne.O)
& write (iunit,(4x,i2,al,6x,i2,3x,fl2.4,lx,fl2.4,9x,f6.1)')
& i,pstar,n(i),mean(i),sd,cv
c
c don't write out pstar if inum - 0 - summarize raw data
c
if(inum.eg.0)
& write (iunit,'(4x,i2,7x,i2,3x,fl2.4,lx,fl2.4,9x,f6.1)')
& i,n(i),mean(i),sd,cv
10 continue
write(iunit,*)
& '. '
c
if(inum.eq.O) return
c
c
if (iside.eq.O) then
direct-'less than1
else
direct-'greater than'
endif
c
if(ip(l).lt.l) then
write(iunit,'(/a,a,/,a,a/)')
S ' *) the mean for this group is significantly ',direct,
& ' the control mean at alpha - 0.05 (1-sided) by Dunnetf's',
& ' test1
else
write(iunit,'(/a,a,/,a,/,a/)')
& ' *) the mean for this group is significantly ',direct,
& ' the control mean at alpha - 0.05 (1-sided) by a t - test',
& ' with Bonferroni adjustment of alpha level'
endif
c
if(iunit.eq.O) pause ' '
idfl - g - 1
idf2 - ntot - g
f - (ssb/idfl) / (ssw/idf2)
call pvalue(3,2,f,g-1,ntot-g,pval)
if(pval.lt.0.001) pval - 0.001
c
write(iunit,'(//)')
if(ip(l).lt.l) then
write(iunit,'(a,a,f!5.6)') ' Minimum detectable difference for',
& ' Dunnetf's test - ',mdd
else
write(iunit,'(a,/,a,f!5.6)') ' Minumum detectable difference for',
t ' t-tests with Bonferroni adjustment - ',»dd
endif
452
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
if(inum.gt.l) write(iunit,(a,f!5.6,a) ')
1 This corresponds to a difference of 'jmddl,1 in original units'
write(iunit,'(a,f8.2,a//)') ' This difference corresponds to ',
pctrl, ' percent of control1
if(ieqn.eq.1) then
write(iunit,*) "A***********************************************
write(iunit,*) '* *
write(iunit,*) '* Note - the above value for the minimum *
write(iunit,*) '* detectable difference is approximate as *
write(iunit,*) '* the sample sizes are not the same for all of *
write(iunit,*) '* the groups. *
write(iunit,*) '* *
write(iunit,*) 'A***********************************************
write(iunit,'(//)')
endif
100
',ssb,' with ',idfl,
pval
1 with ',idf2,
write(iunit,(a,f!6.6,a,i2,a/)')
& ' Between groups sum of squares -
& ' degrees of freedom.'
write(iunit,'(a,f6.3,/)') ' p <- ',
write(iunit,(a,f!6.6,a,i2,a/)')
& ' Error mean square - I,ssw/idf2,
& ' degrees of freedom.'
if(p.gt.l.dO) goto 100
if(p.It.0.001) then
p - 0.001
write(iunit,'(a,f6.3,/)')
& ' Bartletf's test p-value for equality of variances <- ',p
Ise
write(iunit,(a,f6.3,/)')
& ' Bartletf's test p-value for equality of variances - ',p
endif
if(p.gt.0.01) return
i************************************************
* *
'* Warning - the test for equality of variances
'* is significant (p less than 0.01). The
i* results of this analysis should be inter-
i* preted with caution.
write(iunit,*)
write(iunit,*)
write(iunit,*)
write(iunit,*)
write(iunit,*)
write(iunit,*)
write(iunit,*)
write(iunit,*)
return
write(iunit,*)
writ*(iunit,*)
write(iunit,*)
write(iunit,*)
write(iunit,*)
writ*(iunit,*)
write(iunit,*)
return
end
I************************************************
I************************************************
'* Warning - the test for equality of variances
'* could not be computed as 1 or more of the
'* variances is zero.
«*
*************************************************
453
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
$storage:2
REAL*8 FUNCTION TINV(P,NDF)
IMPLICIT REAL*8(A-H,0-Z)
DF=NDF*1.DO
Z=GAUINV(P)
T-Z
T=T+(Z**3+Z)/(4.DO* DF)
T=T+(5.DO*Z**5+16.DO*Z**3+3.DO*Z)/(96.DO* DF**2)
T«T+(3.DO*Z**7+19.DO*Z**5+17.DO*Z**3-15.DO*Z)/(384.DO* DF**3)
T-T+(79.DO*Z**9+776.DO*Z**7+1482.DO*Z**5-1920.DO*Z**3-945.DO*Z)/
$(92160.00* DF**4)
TINV-T
RETURN
END
REAL*8 FUNCTION GAUINV(P)
IMPLICIT REAL*8(A-H,0-Z)
D=P
IF(D.GT..5DO) D-l.-D
T2-DLOG(1./(D*D))
T-DSQRT(T2)
GAUINV-T-(2.515517DO+0.802853DO*T+0.010328DO*T2)/
* (1.0DO+1.432788DO*T+0.189269DO*T2+.001308DO*T*T2)
IF(P.LT..5DO) GAUINVGAUINV
RETURN
END
454
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
$include:'speedy.fil1
C 0 ... 79
c cursor positioning, goto (irow,icol) :
C 24
subroutine gotorc (irow,icol)
character*! dummy
if (irow.lt.0.or.irow.gt.24) irow - 0
if (icol.lt.0.or.icol.gt.79) icol - 0
read (dummy,'(Ix)')
call locate (0,0,ier)
write (*,'(\)')
call locate (irow,icol,ier)
read (dummy,'(Ix)')
return
end
c clears the screen, returns to text mode if not already in text mode,
c and puts cursor in upper left corner
subroutine clrscr
call qrmode (im,i)
if (im.ne.2) call qsmode(2)
call qclear (0,7)
call gotorc (0,0)
write (*,'(\)')
return
end
c clears to end of line from current cursor position, does not move
c cursor
subroutine clreol
call qcpos (icol,irow)
if (icol.eq.79) then
if (irow.ne.24) write (*,'(al)') ' '
else
call qstext (32,0,7,(79-icol))
end if
return
end
c gotorc + clreol
subroutine goclrc (irow,icol)
call gotorc (irow,icol)
call clreol
return
end
455
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued)
c converts a character (ascii value) to upper case (if lower case)
character*! function upchar (ch)
character*! c,ch
equivalence (ich,c)
c = ch
if (ich.ge.97.and.ich.le.122) ich - ich - 32
upchar - c
return
end
c converts a fortran string of length len to uppercase
subroutine upstr (str,len)
character*! str(l)
character*! upchar
do 10 i=l,len
10 str(i) - upchar(str(i))
return
end
c back up (move left) one text position on the screen
subroutine backup
call qcpos (icol,irow)
call qcmov ((icol-1),irow)
return
end
c pauses for idelay seconds
subroutine delay (idelay)
call qtime (ihr,imin,isec,ihun)
10 call qtime (ihr,imin,nsec,ihun)
if (isec.gt.nsec) then
iexp - 60 - isec + nsec
else
iexp - nsec - isec
endif
if (iexp.gt.idelay) return
goto 10
return
end
c get a string from the keyboard, echo it to the screen, but handle
456
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued)
c each character separately, so there is no scrolling, etc.
c if user hits escape, str(l) «« ESC (char(27)), then return.
c if string is of length zero, str(l) - char(13).
c maximum length of string accepted is len.
subroutine getstr (str,len)
character*1 str(1),gstr(80)
equivalence (gstr(l),ifirst)
ilen - 0
10 call qinkey (iext,key)
if ((iext.eq.l).and.((key.eq.13).or.(key.eq.27))) goto 100
if (iext.eq.1.and.key.ge.32.and.key.le.126.and.ilen.It.len) then
write (*,'(al,\)') key
ilen - ilen + 1
gstr(ilen) - key
endif
if (iext.eq. Land.key.eq.8.and. ilen.gt.O) then
call backup
write (*,'(al,\)') ' '
call backup
ilen - ilen - 1
endif
goto 10
100 if (key.eq.27) then
ifirst - 27
ilen - 1
else
do 105 k-ilen+l,len
105 gstr(k) - ' '
endif
if (ilen.eq.O) ifirst - 13
do 110 k-l,len
110 str(k) - gstr(k)
return
end
c this routine gets a password for a protected data set.
c it works about the same as getstr, but does not echo the input
c characters to the screen, maximum length is 10 characters.
subroutine gpwrd (gstr)
character*! gstr(10)
ilen - 0
10 call qinkey (iext,key)
if (iext.eq.O) goto 10
if (key.eq.27) goto 27
if (key.eq.13) goto 13
if (key.ge.32.and.key.le.126.and.ilen.lt.10) then
write (*,'(a\)') ' '
ilen - ilen + 1
457
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
13
27
gstr(ilen) - key
end if
if (key.eq.S.and.ilen.gt.O)
call backup
ilen = ilen - 1
endif
goto 10
if (ilen.eq.O) goto 27
call pad (gstr,ilen,10)
return
gstr(l) - char(key)
return
end
then
c this routine decodes an encrypted password stored in the header
c file and returns the original ascii string.
subroutine dpwrd (pw)
character*! pw(10), n(10)
call copy (pw,
pw(l)
pw(2)
pw(3)
pw(4)
pw(5)
pw(6)
pw(7)
pw(8)
pw(9)
pw(10)
return
end
char
char
char
char
char
char
char
char
char
char
n,10)
(ichar(n(6)) / 2 + 6)
(ichar(n(3)) / 2 + 3)
(ichar(n(8)) / 2 + 8)
(ichar(n(4)) / 2 + 5)
(ichar(n(10)) / 2 + 9)
(ichar(n(7)) / 2 + 4)
(ichar(n(5)) / 2 + 2)
(ichar(n(2)) / 2 + 1)
(ichar(n(9)) / 2 + 7)
(ichar(n(l)J / 2 + 0)
c finds length of a string (str) of max length mien (position of last
c non-blank character)
function length (str,mien)
character*! str(mien)
k - mien + 1
10 k - k - 1
if (k.eq.O) goto 20
if (str(k).eq.1 ') goto 10
20 length - k
return
end
c centers a string (str) in a field of len characters, padded on both
c sides by blanks, strips away leading blanks, then pads it back out.
458
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued)
subroutine center (str,len)
character*! str(l)
if (length(str,len).eg.O) return
kl - 0
5 kl - kl + 1
if (str(kl).eg.' ') goto 5
k2 - length (str,len)
ilen - k2 - kl + 1
do 10 k-kl,k2
10 str(k-kl-H) - str(k)
call pad (str,ilen,len)
imov - (len-ilen)/2
do 30 k-ilen,l,-l
30 str(k+imov) - str(k)
do 40 k-l,imov
40 str(k) - ' '
return
end
c left justifies a string, pads with blanks to the right
subroutine leftj (str,len)
character*! str(l)
if (length(str,len).eg.O) return
kl - 0
S kl - kl + 1
if (str(kl).eg.' ') goto 5
k2 - length (str,len)
ilen - k2 - kl + 1
do 10 k-kl,k2
10 str(k-kl+l) - str(k)
call pad (str,ilen,len)
return
end
c right justifies a string of length len, pads with blanks at left
subroutine rightj (str,len)
character*! str(l)
if (length(str,len).eg.O) return
kl - 0
5 kl - kl + 1
if (str(kl).eq.' ') goto 5
k2 length (str,len)
ilen - k2 - kl + 1
do 10 k-k2,kl,-l
10 str(k-k2+l«n) - str(k)
call pad (str,0,len-ilen)
return
end
459
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued).
c strips the dollar sign (col 1) from a fortran string of length len,
c pads with a blank at the right.
subroutine stripp (str,len)
character*! str(l)
if (str(l).eq.'$') then
do 10 k-l,len-l
10 str(k) - str(k+l)
str(len) - ' '
endif
return
end
c copies LEH characters from fortran string FROM to fortran string TO
subroutine copy (from,to,len)
character*! from(len),to(len)
do 5 k-l,len
5 to(k) - from(k)
return
end
o pads fortran string of length LEN to length NEWLEN with blanks
subroutine pad (str,len,newlen)
character*! str(newlen)
do 5 k-len+l,newlen
5 str(k) - ' '
return
end
c forces user to press a function key between Fl and Fn.
c ESC is considered equivalent to Fl.
function ifnkey (n)
10 call qinkey (i,k)
if (i.eq.o.and.k.ge.59.and.k.le.(58+n)) goto 20
if (i.eq.l.and.k.eq.27) then
ifnkey - 1
return
ndif
goto 10
20 ifnkey - k - 58
return
end
460
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued)
c tests for upper-case equality of strings - capitalizes, then tests.
function uequal (a,b,len)
character*! a(len),b(len),al(80),a2(80)
logical uegual,equal
call copy (a,al,len)
call copy (b,a2,len)
call upstr (al,len)
call upstr (a2,len)
if (equal(al,a2,len)) then
uequal - .true.
else
uequal « .false.
endif
return
end
c prompts user to press any key to continue, message is centered
c at the bottom of the screen.
subroutine retcon
call goclrc (23,27)
write (*,'(a\)') 'press any key to continue1
call gotorc (23,27)
call qinkey (ji,ji)
return
end
461
-------
2.7 Listing of Computer Program for Dunnett's Procedures (Continued)
c tests whether the first len character* of two strings are
c identical.
logical function equal (sl,s2,len)
character*! sl(l),s2(l)
equal ** .true.
do 5 k-1,1en
5 if (sl(fc).ne.s2(k)) goto 10
return
10 equal - .false.
return -^
end
c tests for "same" col name - upcase, stripp
function same (a,b,len)
character*! a(len),b(len),al(80),a2(80)
logical same,equal
call copy (a,al,len)
call copy (b,a2,len)
call upstr (al,len)
call upstr (a2,len)
call stripp (al,len)
call stripp (a2,len)
if (equal(al,a2,len)) then
same - .true.
else
same - .false.
endif
return
end
462
-------
APPENDIX D
T-TEST WITH BONFERRONI'S ADJUSTMENT
1. The t-test with Bonferroni's adjustment is used as an alternative to
Dunnett's Procedure when the number of replicates is not the same for all
concentrations. This test sets an upper bound of alpha on the overall error
rate, in contrast to Dunnett's Procedure, for which the overall error rate is
fixed at alpha. Thus, Dunnett's Procedure is a more powerful test.
2. The t-test with Bonferroni's adjustment is based on the same assumptions
of normality of distribution and homogeneity of variance as Dunnett's
Procedure (See Appendix B for testing these assumptions), and, like Dunnett's
Procedure, uses a pooled estimate of the variance, which is equal to the error
value calculated in an analysis of variance.
3. An example of the use of the t-test with Bonferroni's adjustment is
provided below. The data used in the example are the same as in Appendix C,
except that the third replicate from the 50% effluent treatment is presumed to
have been lost. Thus, Dunnett's Procedure cannot be used. The weight data
are presented in Table D.I.
TABLE D.I. SHEEPSHEAD MINNOW, CYPRINODON VARIEGATUS, LARVAL GROWTH
DATA (WEIGHT IN MG) USED FOR THE T-TEST WITH BONFERRONI'S
ADJUSTMENT
Effluent i
Cone (%)
Replicate Test Vessel
1 2 3
Total
Mean
Control
6.25
12.5
25.0
50.0
1
2
3
4
5
1.017
1.157
0.998
0.873
0.715
0.745
0.914
0.793
0.935
0.907
0.862
0.992
1.021
0.839
(Data lost)
2.624
3.063
2.812
2.647
1.622
0.875
1.021
0.937
0.882
0.811
Prepared by Laura Gast, Cathy Poore, Ron Freyberg, Florence Kessler, John
Menkedick and Larry Wymer, Computer Sciences Corporation, 26 W. Martin Luther
King Drive, Cincinnati, Ohio 45268; Phone (513) 569-7968.
463
-------
3.1 One way to obtain an estimate of the pooled variance is to construct an
ANOVA table including all sums of squares, using the following formulas:
Total Sum of Squares: SST = X Y2j - G 2/N
ij
Between Sum of Squares: SSB = X Tj/n,- - G /N
i
Within Sum of Squares: SSW = SST - SSB
Where: G = The grand total of all sample observations; G = 2 T,-
i
N = The total sample size; N = S n{
i
n( = The number of replicates for concentration "i".
TJ = The total of the replicate measurements for concentration "i".
Y,.j = The jth observation for concentration "i".
3.2 Calculations:
Total Sum of Squares: SST = X Y?j - G /N
ij
= 11.832 - 12.7682
14
= 0.188
Between Sum of Squares: SSB = Z T^/n,- - G2/N
i
= 11.709 - (12.768)2/14
= 0.064
Within Sum of Squares: SSW = SST SSB
- 0.188 - 0.064
= 0.124
464
-------
3.3 Prepare the ANOVA table as follows:
TABLE D.2. GENERALIZED ANOVA TABLE
Source
OF Sum of
Squares (SS)
Mean Square (MS)
(SS/DF)
Between p* - 1
SSB
SB = SSB/(p-l)
Within N p
SSW
SSW/(N-p)
Total
N 1
SST
*p = Number of different concentrations, including the control
3.4 The completed ANOVA table for this data is provided below:
TABLE D.3. COMPLETED ANOVA TABLE FOR THE T-TEST WITH BONFERRONI'S
ADJUSTMENT
Source
DF
SS
Mean Square
Between 5-1=4 0.064
Within 14 - 5 = 9 0.124
0.016
0.014
Total
13
0.188
465
-------
3.5 To perform the individual comparisons, calculate the t statistic for
each concentration and control combination, as follows:
tf
(Y, - Yf)
[Su /
d/n,)
Where:
Y,
Y, =
n, =
Mean for concentration i
Mean for the control
Square root of the within mean square
Number of replicates in the control.
Number of replicates for concentration "i"
3.6 Table D.4 includes the calculated t values for each concentration and
control combination.
TABLE D.4. CALCULATED T VALUES.
Effluent
Concentration
6.25
12.5
25.0
50.0
2
3
4
5
- 1.511
- 0.642
- 0.072
- 0.592
3.7 Since the purpose of the test is only to detect a decrease in growth from
the control, a one-sided test is appropriate. The critical value for the
one-sided comparison (2.686), with an overall alpha level of 0.05, nine
degrees of freedom and four concentrations excluding the control, was obtained
from Table D.5. Comparing each of the calculated t values in Table D.4 with
the critical value, no decreases in growth from the control were detected.
Thus the NOEC is 50.0%.
466
-------
TABLE D.5. CRITICAL VALUES FOR THE T-TEST WITH BONFERRONI'S
ADJUSTMENT
P = 0.05 CRITICAL LEVEL, ONE TAILED
D.F. K - 1 K
1 <
2
3
* ,
5 4
6
9
10
12
13
1*
15
16
17
18
19
20
21
22
23
2*
25
26
27
28
29
30
31
33
3*
35
36
37
38
39
*0
50
60
70
80
90
100
110
= 2
i.31* 12.707
>'.920 *.303
J.35* 3.183
'.132 2.777
>.016 2.571
..9** 2.**7
.895 2.365
.860 2.307
.83*
.813
. 796
.783
.77 1
. 762
.75*
.7*6
. 7*0
. 735
.730
. 725
. 721
.718
. 71*
.711
.709
. 706
. 70*
.702
. 700
6 V 8
.696
.69*
.693
.691
.690
.609
.638
.686
.685
.68*
.676
.671
.667
.665
.662
.661
.659
120 1.658
INF. 1.6*5
D.F. = Degrees
K = Number
2.229
2 .201
2. 179
2.161
2. 1*5
2.132
2.120
2. 1 10
2. 101
2.09*
2.086
2.030
2.07*
2.069
2.06*
2.060
2.056
2.052
2.0*9
2 .0*6
2.0*3
2.0*0
2 .037
2.035
2.033
2 .031
2.029
2.027
2.025
2.023
2.022
2.009
2.001
1.995
1.991
1.987
1.93*
1 .982
1.980
1.960
of
K = 3
19.002
5.3*0
3. 7*1
3. 107
2.912
2. 750
2.6*2
2.567
2. 510
2.*C6
2.*32
2.*0*
2.300
2.360
2.3*3
2.329
2.316
2.305
2.295
2.206
2.278
2.271
2.26*
2.253
2.253
2.2*8
2.2*3
2.239
2.235
2.231
2.228
2 .22*
2.221
2.219
2.216
2.213
2.211
2.209
2.207
2.205
2.189
2. 179
2.171
2.166
2. 162
2.158
2.156
2.153
2.129
freedom
K = *
25.*52
6.206
*. 177
3.*96
3.16*
2. 969
2. 8*2
2.752
2. 686
2.63*
2.59*
2.561
2.533
2.510
2.*90
2.*73
2.*59
2.**6
2.*3*
2.*2*
2.*l*
2.*06
2.398
2.391
2.385
2.379
2.37*
2.369
2.36*
2.360
2. 356
2.352
2.3*9
2.3*6
2.3*2
2. 3*0
2.337
2.33*
2.332
2.329
2.311
2.300
2.291
2.285
2.280
2.276
2.273
2.270
2.2*2
for Mean
of concentrations to
K = 5
31.821
6 .965
*.5*1
3.7*7
3.365
3. 1*3
2.998
2.897
2.822
2. 76*
2.719
2.681
2.651
2.625
2.603
2.58*
2.567
2.553
2.5*0
2.528
2.518
2.509
2.500
2.*93
2.*86
2.*79
2.*73
2.*66
2.*63
2.*58
2. *53
2.4*9
2.**5
2.**2
2.*38
2.*35
2.*32
2.*29
2.*26
2.*2*
2.*0*
2.391
2. 381
2.37*
2.369
2.365
2. 361
2.358
2.327
Square
K = 6
38. 189
7.6*9
*.857
3.961
3.535
3.288
3.123
3.016
2 .93*
2.871
2.821
2. 780
2. 7*6
2.718
2.69*
2.67*
2.655
2.6*0
2.626
2.613
2.602
2.592
2.583
2.57*
2.566
2.559
2.553
2.5*7
2.5*1
2 .536
2.531
2.527
2.523
2.519
2.515
2.512
2.508
2.505
2.502
2.*99
2.*78
2.*63
2.*53
2.**6
2 .**0
2.*35
2.*32
2.*29
2.39*
Error
be compared
K = 7
**. 556
8.277
5.130
*. 1*0
3.681
3.*12
3.239
3. 118
3.029
2. 961
2.907
2 .363
2. 327
2.797
2.771
2. 7*9
2. 729
2.712
2. 697
2.68*
2.672
2. 661
2.651
2 .6*2
2.63*
2.627
2 .620
2.613
2.607
2.602
2.597
2.592
2.587
2.583
2.579
2.575
2.572
2.560
2.565
2.562
2.539
2.52*
2.513
2.505
2 .*99
2 .*9*
2.*90
2.*87
2.*50
(MSE)
to the
K « 8
50.92*
8.861
5.392
*.315
3.811
3.522
3.336
3. 206
3.111
3.039
2.981
2 .935
2. 897
2.86*
2 .83 7
2.81*
2. 793
2. 775
2. 759
2. 7*5
2.732
2. 721
2. 710
2.701
2.692
2. 68*
2.677
2.670
2. 66*
2.658
2.652
2.6*7
2.6*3
2.638
2.63*
2.630
2 .626
2.623
2.619
2.616
2. 592
2.576
2.56*
2. 556
2.5*9
2.5**
2. 5*0
2.536
2 .*98
K = 9
57.290
9.*08
5.626
*.*66
3.927
3.619
3.*22
3. 285
3.185
3.108
3.0*7
2.998
2.950
2.92*
2.095
2.871
2.8*9
2.830
2.013
2. 798
2.785
2.773
2. 762
2. 752
2.7*3
2. 73*
2.727
2.720
2.713
2. 707
2.701
2.696
2.691
2.686
2. 682
2. 678
2.67*
2.670
2.667
2.663
2.638
2.621
2.609
2. 600
2. 593
2.588
2. 583
2.580
2.5*0
K = 10
63.
9.
5.
* .
*.
3.
3'.
3!
3.
3 '.
2.
2 "
2.
2.
2 .
2.
2.
2.'
2.
z'.
2.
2.
2.
2l
2.
2 *
2 J
2 "
2 .
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
657
925
8*1
605
033
708
500
356
250
170
106
055
013
977
9*7
921
899
879
86 1
8*6
832
319
308
797
788
779
771
76*
757
750
7*5
739
73*
729
72*
720
716
712
708
705
678
661
6*3
639
632
626
622
618
576
from ANOVA.
control .
467
-------
APPENDIX E
STEEL'S MANY-ONE RANK TEST1
1. Steel's Many-one Rank Test is a nonparametric test for comparing
treatments with a control. This test is an alternative to the Dunnett's
Procedure, and may be applied to the data when the normality assumption has
not been met. Steel's Test requires equal variances across the treatments and
the control, but it is thought to be fairly insensitive to deviations from
this condition (Steel, 1959). The tables for Steel's Test require an equal
number of replicates at each concentration. If this is not the case, use
Wilcoxon's Rank Sum Test, with Bonferroni's adjustment (See Appendix F).
2. For an analysis using Steel's Test, for each control and concentration
combination, combine the data and arrange the observations in order of size
from smallest to largest. Assign the ranks to the ordered observations (1 to
the smallest, 2 to the next smallest, etc.). If ties occur in the ranking,
assign the average rank to the observation. (Extensive ties would invalidate
this procedure). The sum of the ranks within each concentration and within
the control is then calculated. To determine if the response in a
concentration is significantly different from the response in the control, the
minimum rank sum for each concentration and control combination is compared to
the critical value in Table E.5. In this table, k equals the number of
treatments excluding the control and n equals the number of replicates for
each concentration and the control.
3. An example of the use of this test is provided below. The test employs
survival data from a mysid 7-day, chronic test. The data are listed in Table
E.I. Throughout the test, the control data are taken from the site water
control. Since there is 0% survival for all eight replicates for the 50%
concentration, it is not included in this analysis and is considered a
qualitative mortality effect.
4. For each control and concentration combination, combine the data and
arrange the observations in order of size from smallest to largest. Assign
the ranks (1, 2, 3, ..., 16) to the ordered observations (1 to the smallest, 2
to the next smallest, etc.). If ties occur in the ranking, assign the average
rank to the observation.
5. An example of assigning ranks to the combined data for the control and
3.12% effluent concentration is given in Table E.2. This ranking procedure is
repeated for each control and concentration combination. The complete set of
rankings is listed in Table E.3. The ranks are then summed for each effluent
concentration, as shown in Table E.4.
Prepared by Laura Gast, Cathy Poore, Ron Freyberg, Florence Kessler, John
Menkedick and Larry Wymer, Computer Sciences Corporation, 26 W. Martin Luther
King Drive, Cincinnati, Ohio 45268; Phone (513) 569-7968.
468
-------
6. For this set of data, we wish to determine if the survival in any of the
effluent concentrations is significantly lower than the survival of the
control organisms. If this occurs, the rank sum at that concentration would
be significantly lower than the rank sum of the control. Thus, we are only
concerned with comparing the rank sums for the survival at each of the various
effluent concentrations with some "minimum" or critical rank sum, at or below
which the survival would be considered to be significantly lower than the
control. At a probability level of 0.05, the critical rank sum in a test with
four concentrations and eight replicates per concentration, is 47 (see Table
F.4).
7. Of the rank sums in Table E.4, none are less than 47- Therefore, due to
the qualitative effect at the 50% effluent concentration, the NOEC is 25%
effluent and the LOEC is 50% effluent.
469
-------
TABLE E.I. EXAMPLE OF STEEL'S MANY-ONE RANK TEST: SURVIVAL
DATA FOR MYSID, MYSIDOPSIS BAHIA, 7-DAY CHRONIC TEST
Effluent Replicate
Concentration Chamber
Control
(Site Water)
Control
(Brine &
Dilution Water)
3.12*
6.25X
12. 5X
25. OX
50. OX
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
Number of
Myslds at
Start of Test
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
Number of
Live Myslds
at End of Test
4
4
5
4 -
5
4
4
5
3
5
3
3
4
4
3
3
4
4
4
5
4
4
5
3
3
4
5
4
4
4
5
5
5
4
5
3
5
4
4
3
5
5
5
5
3
5
4
4
0
0
0
0
0
0
0
0
470
-------
TABLE E.2. EXAMPLE OF STEEL'S MANY-ONE RANK TEST: ASSIGNING
RANKS TO THE CONTROL AND 3.12% EFFLUENT CONCENTRATIONS
Rank
1
6.5
6.5
6.5
6.5
6.5
6.5
6.5
6.5
6.5
6.5
14
14
14
14
14
Number of Live
Mysids, Mysidopsis bahia
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
Control or 3.12% Effluent
3.12%
Control
Control
Control
Control
Control
3.12%
3.12%
3.12%
3.12%
3.12%
Control
Control
Control
3.12%
3.12%
TABLE E.3. TABLE OF RANKS
Replicate
Chamber
1
2
3
4
5
6
7
8
Effluent Concentration (%)
Control1
4 (6.5,6,6.5,5)
4 (6.5,6,6.5,5)
5 (14,13.5,13.5,12.5)
4 (6.5,6,6.5,5)
5 (14,13.5,13.5,12.5)
4 (6.5,6,6.5,5)
4 (6.5,6,6.5,5)
5 (14,13.5,13.5,12.5)
3.12
4 (6.5)
4 (6.5)
4 (6.5)
5 (14)
4 (6.5)
4 (6.5)
5 (14)
3 (1)
6.25
3 (1)
4 (6)
5 (13.5)
4 (6)
4 (6)
4 (6)
5 (13.5)
5 (13.5)
12.5
5 (13.5)
4 (6.5)
5 (13.5)
3 (1.5)
5 (13.5)
4 (6.5)
4 (6.5)
3 (1.5)
25.0
5 (12.5)
5 (12.5)
5 (12.5)
5 (12.5)
3 (1)
5 (12.5)
4 (5)
4 (5)
1Control ranks are given in the order of the concentration with which they
were ranked.
471
-------
TABLE E.4. RANK SUMS
Effluent
Concentration
Rank Sum
3.12
6.25
12.50'
25.00
61.5
65.5
63.0
73.5
TABLE E.5. SIGNIFICANT VALUES OF RANK SUMS: JOINT CONFIDENCE
COEFFICIENTS OF 0.95 (UPPER) and 0.99 (LOWER) FOR
ONE-SIDED ALTERNATIVES1
n
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
k =
2
11
18
15
27
23
37
32
49
43
63
56
79
71
97
87
116
105
138
125
161
147
186
170
213
196
241
223
272
252
304
282
339
315
number
3
10
17
-
26
22
36
31
48
42
62
55
77
69
95
85
114
103
135
123
158
144
182
167
209
192
237
219
267
248
299
278
333
310
of treatments
4 5
10
17
-
25
21
35
30
47
41
61
54
76
68
93
84
112
102
133
121
155
142
180
165
206
190
234
217
264
245
296
275
330
307
10
16
-
25
21
35
30
46
40
60
53
75
67
92
83
111
100
132
120
154
141
178
164
204
188
232
215
262
243
294
273
327
305
(excluding
6 7
10
16
-
24
-
34
29
46
40
59
52
74
66
91
82
110
99
130
119
153
140
177
162
203
187
231
213
260
241
292
271
325
303
-
16
-
24
-
34
29
45
40
59
52
74
66
90
81
109
99
129
118
152
139
176
161
201
186
229
212
259
240
290
270
323
301
control)
8 9
-
16
-
24
-
33
29
45
39
58
51
73
65
90
81
108
98
129
117
151
138
175
160
200
185
228
211
257
239
288
268
322
300
-
15
-
23
-
33
29
44
39
58
51
72
65
89
80
108
98
128
117
150
137
174
160
199
184
227
210
256
238
287
267
320
299
1From Steel (1959).
472
-------
APPENDIX F
WILCOXON RANK SUM TEST1
1. Wilcoxon's Rank Sum Test is a nonparametric test, to be used as an
alternative to Steel's Many-one Rank Test when the number of replicates are
not the same at each concentration. A Bonferroni's adjustment of the pairwise
error rate for comparison of each concentration versus the control is used to
set an upper bound of alpha on the overall error rate, in contrast to Steel's
Many-one Rank Test, for which the overall error rate is fixed at alpha. Thus,
Steel's Test is a more powerful test.
2. The use of this test may be illustrated with reproduction data from the
mysid test in Table F.I. The site water control and the 12.5% effluent
concentration each have seven replicates for the proportion of females bearing
eggs, while there are eight replicates for each of the remaining three
concentrations.
3. For each concentration and control combination, combine the data and
arrange the values in order of size, from smallest to largest. Assign ranks
to the ordered observations (a rank of 1 to the smallest, 2 to the next
smallest, etc.). If ties in rank occur, assign the average rank to the
observation.
4. An example of assigning ranks to the combined data for the control and
effluent concentration 3.12% is given in Table F.2. This ranking procedure is
repeated for each of the three remaining control versus test concentration
combinations. The complete set of ranks is listed in Table F.3. The ranks
are then summed for each effluent concentration, as shown in Table F.4.
5. For this set of data, we wish to determine if the fertility in any of the
test concentrations is significantly lower than in the control. When this is
the case, the rank sum for that concentration compared to the control will be
significantly lower than the rank sum given by the average rank over both
concentrations times the number of replicates at that test concentration.
Thus, we are concerned with comparing the rank sums for fertility at each of
the effluent concentrations with some "minimum" or critical rank sum, at or
below which the fertility would be considered to be significantly lower than
the control. At a probability level of 0.05, the critical rank in a test with
four concentrations and seven replicates in the control is 44 for those
concentrations with eight replicates, and 34 for those concentrations with
seven replicates (see Table F.5, for R = 4).
6. Of the rank sums in Table F.4, only the 25% effluent concentration does
not exceed its critical value of 44. Therefore, the LOEC for the test on
fertility is 25% effluent, and the NOEC is 12.5% effluent.
Prepared by Laura Cast, Cathy Poore, Ron Freyberg, Florence Kessler, John
Menkedick and Larry Wymer, Computer Sciences Corporation, 26 W. Martin Luther
King Drive, Cincinnati, Ohio 45268; Phone (513) 569-7968.
473
-------
TABLE F.I. EXAMPLE OF WILCOXON'S RANK SUM TEST: FECUNDITY
DATA FOR MYSID, MYSIDOPSIS BAHIA, 7-DAY CHRONIC
TEST
Effluent Replicate
Concentration Chamber
Control
(Site Water)
Control
(Brine &
Dilution Water)
3.12X
6.25X
12. 5X
25. OX
50. OX
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
a
1
2
3
4
5
6
7
8
Number of
Mysids at
Start of Test
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
Number of
Live Mysids
at End of Test
4
4
5
4
5
4
4
5
3
5
3
3
4
4
3
3
4
4
4
5
4
4
5
3
3
4
5
4
4
4
5
5
5
4
5
3
5
4
4
3
5
5
5
5
3
5
4
4
0
0
0
0
0
0
0
0
Proportion
of Females
with Eggs
0.50
--
0.75
0.67
0.67
0.50
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.50
0.50
0.50
1.00
0.50
0.67
1.00
0.50
1.00
1.00
0.00
0.50
0.00
0.75
1.00
1.00
1.00
0.67
0.67
0.33
0.50
1.00
--
1.00
0.00
0.33
0.50
0.00
0.50
0.33
0.00
0.50
0.00
0.50
0.50
__
--
--
--
--
--
--
--
474
-------
TABLE F.2. EXAMPLE OF WILCOXON'S RANK SUM TEST: ASSIGNING
RANKS TO THE CONTROL AND 3.12% EFFLUENT CONCENTRATIONS
Rank Proportion of Site Water Control
Females W/Eggs or 3.12% Effluent
1
3.5
3.5
3.5
3.5
7
7
7
9
12.5
12.5
12.5
12.5
12.5
12.5
0.00
0.50
0.50
0.50
0.50
0.67
0.67
0.67
0.75
1.00
1.00
1.00
1.00
1.00
1.00
3.12%
Control
Control
3.12%
3.12%
Control
Control
3.12%
Control
Control
Control
3.12%
3.12%
3.12%
3.12%
TABLE F.3. TABLE OF RANKS1
Rep Proper- Site Water Effluent Concentration (%)
tion Control Rank 3.12 6.25 12.5 25.0
1
2
3
4
5
6
7
8
0.50
0.75
0.67
0.67
0.50
1.00
1.00
(3.5,3,5.5,7.5)
-
(9,9.5,10,13)
(7,6.5,8.5,11.5)
(7,6.5,8.5,11.5)
(3.5,3,5.5,7.5)
(12.5,13,12.5,14.
(12.5,13,12.5,12.
5)
5)
1.00
0.50
0.67
1.00
0.50
1.00
1.00
0.00
(12.5)
(3.5)
(7)
(12.5)
(3.5)
(12.5)
(12.5)
(1)
0.50
0.00
0.75
1.00
1.00
1.00
0.67
0.67
(3)
(1)
(9.5)
(13)
(13)
(13)
(6.5)
(6.5)
0.33
0.50
1.00
--
1.00
0.00
0.33
0.50
(2.5)
(5.5)
(12.5)
(12.5)
(1)
(2.5)
(5.5)
0.00
0.50
0.33
0.00
0.50
0.00
0.50
0.50
(2)
(7.5)
(4)
(2)
(7.5)
(2)
(7.5)
(7.5)
1Control ranks are given in the order of the concentration with which they
were ranked.
475
-------
TABLE F.4. RANK SUMS
Effluent
Concentration
3.12
6.25
12.50
25.00
Rank Sum
65
65.5
42
40
No. of
Replicates
8
8
7
8
Critical
Rank Sum
44
44
34
44
TABLE F.5. CRITICAL VALUES FOR WILCOXON'S RANK SUM TEST WITH
BONFERRONI'S ADJUSTMENT OF ERROR RATE FOR COMPARISON
OF "K" TREATMENTS VERSUS A CONTROL FIVE PERCENT CRITICAL
LEVEL (ONE-SIDED ALTERNATIVE: TREATMENT CONTROL)
No. Replicates No. of Replicates:Effluent Concentration
in Control
3456789 10
1 3
4
5
6
7
8
9
10
6
6
7
8
8
9
10
10
10
11
12
13
14
15
16
17
16
17
19
20
21
23
24
26
23
24
26
28
29
31
33
35
30
32
34
36
39
41
43
45
39
41
44
46
49
51
54
56
49
51
54
57
60
63
66
69
59
62
66
69
72
72
79
82
3 -- -- 15 22 29 38 47 58
4 -- 10 16 23 31 40 49 60
5 6 11 17 24 33 42 52 63
6 7 12 18 26 34 44 55 66
7 7 13 20 27 36 46 57 69
8 8 14 21 29 38 49 60 72
9 8 14 22 31 40 51 62 75
10 9 15 23 32 42 53 65 78
476
-------
TABLE F.5. CRITICAL VALUES FOR WILCOXON'S RANK SUM TEST WITH
BONFERRONI'S ADJUSTMENT OF ERROR RATE FOR COMPARISON
OF "K" TREATMENTS VERSUS A CONTROL FIVE PERCENT CRITICAL
LEVEL (ONE-SIDED ALTERNATIVE: TREATMENT CONTROL)(CONTINUED)
K No. Replicates No. of Replicates:Effluent Concentration
in Control
3456789 10
3 3
4
5
6
7
8
9
10
4 3
4
5
6
7
8
9
10
5 3
4
5
6
7
8
9
10
6 3
4
5
6
7
8
9
10
6
7
7
7
8
--
6
6
7
7
7
..
6
6
7
7
6
6
6
7
10
11
11
12
13
13
14
10
11
12
12
13
14
..
--
10
11
11
12
13
13
..
10
11
11
12
12
13
16
17
18
19
20
21
22
15
16
17
18
19
20
21
..
15
16
17
18
19
20
21
15
16
16
17
18
19
20
21
22
24
25
26
28
29
31
21
22
23
24
26
27
28
30
..
22
23
24
25
27
28
29
..
21
22
24
25
26
27
29
29
30
32
33
35
37
39
41
28
30
31
33
34
36
38
40
28
29
31
32
34
35
37
39
28
29
30
32
33
35
37
38
37
39
41
43
45
47
49
51
37
38
40
42
44
46
48
50
36
38
40
42
43
45
47
49
36
38
39
41
43
45
47
49
46
48
51
53
56
58
61
63
46
48
50
52
55
57
60
62
46
48
50
52
54
56
59
61
45
47
49
51
54
56
58
60
57
59
62
65
68
70
73
76
56
59
61
64
67
69
72
75
56
58
61
63
66
68
71
74
56
58
60
63
65
68
70
73
477
-------
TABLE F.5. CRITICAL VALUES FOR WILCOXON'S RANK SUM TEST WITH
BONFERRONI'S ADJUSTMENT OF ERROR RATE FOR COMPARISON
OF "K" TREATMENTS VERSUS A CONTROL FIVE PERCENT CRITICAL
LEVEL (ONE-SIDED ALTERNATIVE: TREATMENT CONTROL)(CONTINUED)
No. Replicates No. of Replicates:Effluent Concentration
in Control
3456789 10
7 3
4
5
6
7
8
9
10
8 3
4
5
6
7
8
9
10
9 3
4
5
6
7
8
9
10
10 3
4
5
6
7
8
9
10
--
--
--
--
6
6
7
6
6
6
--
--
6
6
--
--
--
--
--
6
6
--
--
10
11
11
12
13
10
11
11
12
12
--
--
10
10
11
11
12
--
--
10
10
11
11
12
--
15
16
17
18
19
20
15
16
17
18
19
19
--
15
16
17
18
18
19
15
16
16
17
18
19
21
22
23
25
26
27
28
21
22
23
24
25
27
28
21
22
23
24
25
26
28
21
22
23
24
25
26
27
29
30
32
33
35
36
38
29
30
31
33
34
36
37
28
30
31
33
34
35
37
28
29
31
32
34
35
37
36
37
39
41
43
44
46
48
36
37
39
40
42
44
46
48
37
39
40
42
44
46
47
37
38
40
42
43
45
47
45
47
49
51
53
55
58
60
45
47
49
51
53
55
57
59
45
46
48
50
52
55
57
59
45
46
48
50
52
54
56
58
56
58
60
62
65
67
70
72
55
57
59
62
64
67
69
72
55
57
59
62
64
66
69
71
55
57
59
61
64
66
68
71
478
-------
APPENDIX G
PROBIT ANALYSIS
1. This program calculates the EC1 and EC50 (or LCI and LC50), and the
associated 95% confidence intervals.
2. The program is written in IBM PC Basic for the IBM compatible PC by D.L.
Weiner, Computer Sciences Corporation, 26 W. Martin Luther King Drive,
Cincinnati, Ohio 45268. A compiled version of the program can be obtained
from the Environmental Monitoring Systems Laboratory (EMSL Cincinnati) by
sending a diskette with a written request. The current version (vl.4) of the
Probit program makes no distinction between EC and LC endpoints, and labels
all results as EC values. When the response in question is mortality, the EC
values output by the program should be treated as the corresponding LC values.
2.1 Data input is illustrated in Table G.I by a set of mortality data from a
sheepshead minnow embryo-larval survival and teratogenicity test. The program
begins with a request for the following information:
1. Output designation (P = printer, D = disk file).
2. Title for the output.
3. A selection of model fitting options (see sample
output for a detailed description of options).
If option 2 is selected, the theoretical lower
threshold needs to be entered. If option 3 is
selected, the program requests that the number
of animals responding in the control group and
the total number of orginal animals in the control
group be entered.
4. Toxicant concentration data.
2.2 The program output includes the following:
1. A table of the observed proportion mortality, the adjusted observed
proportion mortality, and the predicted proportion mortality for each
toxicant concentration (see Table G.2).
2. The calculated chi-squared statistic for heterogeneity. This test is
one indicator of how well the data fit the model. The program will
issue a warning when the test indicates that the data do not fit the
model (see Table G.2).
3. Estimates of the mean (mu) and the standard deviation (sigma) of the
underlying Iog10 tolerance distribution (see Table G.2).
4. A table of the estimated EC values and associated 95% confidence
intervals (see Table G.2).
5. A plot of the fitted regression line with observed data overlaid on
the plot (see Figure G.I).
479
-------
TABLE G.I. SAMPLE DATA INPUT FOR USEPA PROBIT ANALYSIS PROGRAM,
VERSION 1.4
I
I
I
I
EPA PROBIT ANALYSIS PROGRAM
USED FOR CALCULATING EC VALUES
Version 1.4
I
I
I
I
Output to printer or disk file (P / D)? P
Title ? Example of Probit Analysis
Model Fitting Options Which Are Available
1) Fit a model which includes two parameters: an intercept and
a slope. This model assumes that the spontaneous response
(in controls) is zero. No control data are entered if this
option is specified.
2) Fit a model which includes three parameters: an intercept,
a slope and a theoretical lower threshold which represents
the level of spontaneous response (in controls). This
option requires the user to input the theoretical lower
threshold (the value must be between 0.0 and 0.99). No
control data is entered if this option is specified.
3) Fit a model which includes three parameters, an intercept,
a slope and a lower threshold. The lower threshold is
estimated based on control data which are input by the user.
If the number responding in the control group is zero, then
this option is indentical to option two (above).
Your choice (1, 2, or 3)? 3
Number of responders in the control group = ? 2
Number of animals exposed in the concurrent control group
Number of administered concentrations ? 5
? 20
480
-------
TABLE G.I. SAMPLE DATA INPUT FOR USEPA PROBIT ANALYSIS PROGRAM,
VERSION 1.4 (CONTINUED)
Input data starting with the lowest concentration
Concentration - ? 0.5
Number responding -72
Number exposed - ? 20
Concentration - 7 1.0
Number responding = ? 1
Number exposed - ? 20
Concentration = 7 2.0
Number responding = 74
Number exposed - ? 20
Concentration - ? 4.0
Number responding » 7 16
Number exposed - 7 20
Concentration - ? 8.0
Number responding = 7 20
Number exposed * 7 20
Number
I
2
3
4
5
Cone.
0.5000
1.0000
2.0000
4.0000
8.0000
Number
Resp.
2
1
4
16
20
Number
Exposed
20
20
20
20
20
Do you wish to modify your data 7 n
The number of control animals which responded
The number of control animals exposed = 20
Do you wish to modify these values ? n
481
-------
TABLE G.2. SAMPLE DATA OUTPUT FOR USEPA PROBIT ANALYSIS PROGRAM
VERSION 1.4
EPA PROBIT ANALYSIS PROGRAM
USED FOR CALCULATING EC VALUES
Version 1.4
Example for Probit Analysis
Cone.
Control
0.5000
1.0000
2.0000
4.0000
8.0000
Number
Exposed
20
20
20
20
20
20
Number
Resp.
2
2
1
4
16
20
Observed
Proportion
Responding
0.1000
0.1000
0.0500
0.2000
0.8000
1.0000
Adjusted
Proportion
Responding
0.0000
0.0174
-.0372
0.1265
0.7816
1.0000
Predicted
Proportion
Responding
0.0841
0.0000
0.0007
0.1179
0.7914
0.9975
Chi - Square Heterogeneity
Mu
Sigma
Parameter
0.479736
0.150766
Estimate
0.441
Std. Err.
95* Confidence Limits
Intercept
Slope
1.818003
6.632814
0.976915
1.804695
( -0.096749,
( 3.095611,
3.732756)
10.170017)
Spontaneous
Response Rate
0.084104
0.036007
0.013529,
0.154678)
Estimated EC Values and Confidence Limits
.00
.00
Point
EC 1.
EC 5.
EC10.00
EC15.00
EC50.00
EC85.00
EC90.00
EC95.00
EC99.00
Cone.
1.3459
1.7051
1.9343
2.1061
3.0181
4.3250
4.7093
5.3423
6.7680
Lower Upper
95% Confidence Limits
0.4533
0.7439
0.9654
1.1484
2.2676
3.5656
3.8443
4.2566
5.0712
1.9222
2.2689
2.4871
2.6523
3.6717
6.3827
7.5099
9.6486
15.6871
482
-------
Figure G.I. Plot of Adjusted Probits and Predicted Regression Line.
Example for Probit Analysis
PLOT OF ADJUSTED PROBITS AND PREDICTED REGRESSION LINE
Probit
10+
9+
8 +
7+
6+
. .o.
5+
4+
.o
1+
0+0
_+ + + + + + +_
EC01 EC10 EC25 EC50 EC75 EC90 EC99
483
-------
2.3 Listing of Computer Program for Probit Analysis.
10
20
30
40
50
60
70
80
90
1 00
1 1 0
1 20
130
1 40
1 50
160
1 70
180
1 90
200
210
220
230
240
250
260
270
280
290
300
310
320
330
340
350
360
370
380
390
400
410
420
430
440
45'.
460
470
480
490
500
510
520
530
540
output is probit
pvaIues
'EPA PROBIT analysis program : Version 1.4 by D.L.Weiner, 3/24/88
'Written for IBM PC and full compatibles. May require minor
'modifications for other systems.
DIM NDOSE(20),NRESP(20),DOSE(20),LDOSE(20),CHISQ(18),T05(18),OP<14),
OY( 14 )
DIM WTC20), YPROB(20), WKY(20), XPRIME(20), Y(20)
DIM P1 (4) ,Q1 (4),P2(8) ,02(8),P3(5 ) ,Q3(5) , PLTT$(51 , 71 )
BIGC=0! : PI 3.1415926535*
GOSUB 720 ' input data
GOSUB 320 ' check for valid data file type
CIS : LOCATE 12,35 :PRINT "Working ..."
GOSU8 1970 ' read in t and chisq values
' sub 1950 is used to compute probits : input is nd,r
' sub 2270 is used to compute area under normal curve
FOR I 1 TO N
UT(I } = 1! ' initialize weights to 0 or 1 for first iteration
IF NRESP(I) = 0 OR NRESP( I ) = NDOSE( I ) THEN WT(I) = 0!
IF BIGC >= NRESP( I )/NDOSE( I ) THEN UT(I) = 0!
XPR I ME(I ) 0!
NEXT I
ITER 0 ' begin zeroth iteration
GOSUB 3180 ' model fitting
GOSUB 3500 ' compute predicted probits (y), working probits (wky)
' and weights (wt)
OK = 0
GOSUB 3640 ' check for convergence
IF 0 K = 1 THEN GOTO 3950 ' convergence achieved, finish up
ITER ITER + 1 ' begin next iteration
GOTO 210
REM subroutine to force form feeds
PRINT #2, CHRSC12)
RETURN
REM check for valid data files , wt is used as a work array here
N1 = 1 : N2 = 1 : N3 = 0 : UT(1 ) =OOSE(1 )
FOR I - 2 TO N
FOR J - 1 TO N2
IF DOSE(I) WT(J) THEN GOTO 390
NEXT J
N1=N1 + 1 : WT(N1)=DOSE(1 ) : N2 = N1
NEXT I
N3=0
FOR 1=1
TO N
IF NRESP(I)>0 AND N R E S P ( I ) < N D 0 S E ( I ) THEN N3=N3+1
NEXT I
IF N3<=1 THEN GOTO 620
N2 = N3
FOR 1=1 TO N3 1
FOR J=I+1 TO N3
IF WT( I )»UT< J) THEN N2 = N2-1
NEXT J
NEXT I
N3 = N2
IF N1 >= 3 AND N3 >= 2 THEN RETURN
IF N1 >= 3 THEN GOTO 620
C L S -.PRINT " " : P R I N T " "
UT < N 3 ) =0 OS E ( I )
484
-------
2.3 Listing of Computer Program for Probit Analysis (Continued).
550
560
570
580
590
500
61 0
620
630
640
650
660
670
680
690
700
710
720
730
740
750
760
770
780
790
800
810
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
I F N3
CLS :
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
GOTO
REH
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
PRINT
INPUT
II
II
II
"
II
II
> =
PR
M
(i
n
M
||
II
II
II
t
t
t
t
[
t
2 THEN
INT "
C
t
t
t
t
[
C
" : P R I
t t t C t C [
Note
[ t t C C [ t
: data
[ t t
f i
unique
t C t t t t t
GOTO 5
" : P R I N T
t t t t t t (
Note
C t t C t t t
550
M II
t t C t [ C C
: data
C t [
t t [
f i
concent
respond
[ [ [ t [ C [
NT " "
[ C [ C C [ C
t C C
[ C t
I e
t [ t C
must
concent
t t t
t [ t
I e
ra t
i ng
t t t
[[ft
t [ C [
must
ions
i s
t [ [ C
t t [ m t t [ t t
contain at
ration leve
t t C [ t [ [ [ t [ [
[ t t [ t t t C [ [ [
contain at
for which
between OX
t t t C [ C C t t [ [
C C t t [ t t [ [ t C [ C [
least three
I s .
t [ [ [ t t [ t t t C t C t
C [ C t [ t t [ t t t t t t
least two
the percent
and 100X
[ t C C [ [ [ C C C C [ C C
C t C t [[[[[[[
[
[
c
[
t C C [ t t t t t t t
C C [ C C t t [ C [ C
[
[
t
[
[
[[[ C [[[[[[[
5550
ii
»
n
n
n
n
n
n
IF ANSS
Data
C
C
[
C
t
c
" -. P R I
Output
Input subrouti
C t C C C [ t
[[[[[[[
ne
t C C
i I I
EPA
USED
[[[[[[[
NT " "
to pr i
="" THEN GOTO
t C t t [ [ t
n t e r or
800
t t C
di
t t [
sk
[[[C
t t C [ [ t t t [ t t
PROBIT ANAL YS 1
FOR
t t [ t
file
CALCULATING
Version 1.4
t t t [ t t t t t t C
[ (t( t m [ C t tt[
S PROGRAM
EC VALUES
C C C C C [ C C [ t C t t [
[ t [ [ t [ t [ [ [ [
t
[
[
[
1 1 [ [ [ [ t [ 1 1 1
( P / D ) " ; A N S S
EPA PROBIT ANALYSIS PROGRAM"
USED FOR CALCULATING EC VALUES"
V e r s i o n 1 . 4 "
PRINT #2," " : P R I N T #2," ": CLS
1) Fit a
Model Fitting Options Which Are Available ":PRINT " "
model which includes two parameters: an intercept and
820 ANS$ = LEFTS(ANSS , 1 ) : I F ASC(ANSS)>96 THEN A N S S = C H R S ( A S C ( A N S $ ) - 3 2 )
830 IF ANSS<>"P" AN" ANSSo"D" GOTO 800
840 IF ANS$="P" THEN FlLE$="lpt1:"
850 IF ANSS="D" THEN INPUT "File name for output"; FILES
860 INPUT "Title ";TITLES
870 OPEN FILES FOR OUTPUT AS #2
880 PRINT #2," "
890 PRINT #2,"
900 PRINT #2,"
910 PRINT # 2 , "
920 PRINT #2,"
930 PRINT "
940 PRINT "
a"
950 PRINT "
960 PRINT »
970 PRINT "
980 PRINT "
a"
990 PRINT «
the"
1000 PRINT "
1010 PRINT "
1020 PRINT "
I s"
1030 PRINT "
1040 PRINT M
e"
slope. This model assumes that the spontaneous response"
(in controls) is lero. No control data are entered if this"
option is s p e c i f i e d.":P RIN T " "
2) Fit a model which includes three parameters: an intercept,
slope and a theoretical lower threshold which represents
level of spontaneous response (in controls). This option"
requires the user to input the theoretical lower threshold"
(the value must be between 0.0 and 0.99). No control data
entered if this option is specified.":PRINT " "
3) Fit a model which includes three parameters, an intercept,
485
-------
2.3 Listing of Computer Program for Probit Analysis (Continued).
1050 PRINT "
estimated
1060 PRINT "
number"
1070 PRINT "
i s. "
PRINT "
INPUT "
slope and a lower threshold. The lower threshold is
based on control data which are input by the user. If the
responding in the control group is zero, then this option
1080 PRINT " indent i c a I to option two (above).":PRINT " "
1090 INPUT " Your choice (1, 2, or 3)";CHO$
1100 IF CHO$ <> "1" AND CHO$ <> "2" AND CHO$ <> "3" THEN CLS : GOTO 930
1110 IF CHOI - "1" THEN CLS GOTO 1240
1120 IF CHO$ "3" THEN CLS : GOTO 1160
1130 CLS : INPUT "Spontaneous response rate ";BIGC
1140 IF BIGO1! OR BIGC<0! THEN PRINT "Value must be between 0 and 1 ":GOTO
1130
1150 GOTO 1240
1160 INPUT "Number of responders in the control group ";NRCTRL
1170 IF NRCTRL<=0 THEN PRINT "Number must be greater than 0, please reenter":
GOTO 1160
1180 INPUT "Number of animals exposed in the concurrent control group - ";
NCTRL
1190 IF NCTRL < 0 THEN PRINT "Invalid number, please reenter":GOT 0 1180
1200 IF NRCTRL>NCTRL THEN PRINT "The number of responders must be no greater"
1210 IF NRCTRL>NCTRL THEN PRINT "than the number exposed, please reenter":
GOTO 1160
1220 C NRCTRL / NCTRL ' empirical control resp
1230 BIGC - C ' initial predicted control resp
1240 INPUT "Number of administered concentrations ";N
1250 N1 = N : IF NCTRL > 0 THEN N1 N1 + 1
1260 IF N>=3 THEN GOTO 1290
1270 CLS
1280 PRINT "Not enough concentration levels to fit a model":GOTO 5550
1290 IF N1>20 THEN PRINT "Maximum number of concentration levels exceeded":
GOTO 5550
1300 CLS
1310 PRINT:PRINT"Input data starting with the lowest concentration":PRINT
1320 DOSE(O) = 0 ' dummy value
1330 FOR J 1 TO N
1340 INPUT "Concentration - »;DOSE(J)
1350 IF DOSE(J)< = 0 THEN PRINT "Invalid concentration please reenter" : GOT 0
1340
1360 IF DOSE(J) > DOSE(J-I) THEN GOTO 1390
1370 PRINT :PRINT "Concentrations must be entered in ascending order low to
high."
1380 PRINT "Please reenter the data.":GOTO 1310
1390 LDOSE(J)=LOG(DOSE(J))/2.3025851* 'Iog10 dose
1400 INPUT "Number responding - ";NRESP(J)
1410 IF NRESP(J)<0 THEN PRINT "Invalid number - please reenter":GOTO 1400
1420 INPUT "Number exposed - ";NDOSE(J)
1430 IF NRESP( J)>NDOSE ( J) THEN PRINT"Number responding can't be greater than
number exposed please reenter": GOTO 1340
1440 PRINT
1450 NEXT J
1460 CLS
1470 PRINT " Number Number "
1480 PRINT" Number Cone. Resp. Exposed "
486
-------
2.3 Listing of Computer Program for Probit Analysis (Continued).
1490 PRINT " it
1500 FOR I = 1 TO N
1510 PRINT USING " MM MMMMit.MMMM MMMtt #### " ; I ; D0SE ( I ) ,
NRESP( I ) , NDOSE( I)
1520 NEXT I
1530 PRI'NT: INPUT "Do you wish to modify your data ";ANS$
1540 IF ANS$="" THEN GOTO 1530
1550 ANSS=LEFTS(ANSS , 1 ) : I F ASC(ANS$)>96 THEN ANS$ = CHR$(ASC(ANS$) - 32)
1560 IF ANSS<>"N" AND ANSS<>"Y" THEN GOTO 1530
1570 IF A N S S = " N " THEN GOTO 1680
1580 INPUT "Observation Number to be modified ";IOB
1590 IF IOB<=0 OR IOB>N THEN PRINT" Invalid Number":GOTO 1530
1600 INPUT "Concentration ";DOSE(IOB)
1610 IF DOSE(IOB)<=0 THEN PRINT "Invalid concentration please reenter":GOTO
1600
1620 LOOSE( I 08 ) = LOG(DOSE( I OB ) 1/2 .3025851M 'Iog10 dose
1630 INPUT "Number responding ";NRESP(IOB)
1640 IF NRESP(IOB)<0 THEN PRINT "Invalid number please reenter":GOTO 1630
1650 INPUT "Number exposed ";NOOSENDOSE(I08) THEN PRINT"Number responding can't be greater
than number exposed please reenter": GOTO 1600
1670 GOTO 1460
1680 CLS
1690 IF CHOS - "3" THEN GOTO 1790
1700 PRINT "The theoretical spontaneous rate * ";BIGC
1710 INPUT "Do you wish to modify it"; ANSS
1720 IF ANSS*"" THEN GOTO 1710
1730 ANSS=LEFTS(ANSS,1):IF ASC(ANSS)>96 THEN ANS$=CHRS(ASC(ANSS)32)
1740 IF ANS$<>"N" AND ANS$<>"Y" THEN GOTO 1710
1750 IF A N S $ = " N " THEN GOTO 1950
1760 INPUT "Spontaneous response rate ";BIGC
1770 IF BIGC>11 OR BIGC96 THEN ANSS = CHR$(ASC(ANSS) - 32)
1840 IF ANS$<>"N" AND ANSS<>"Y" THEN GOTO 1680
1850 IF ANS$="N" THEN GOTO 1950
1860 INPUT "Number of responders in the control group - ";NRCTRL
1870 IF NRCTRL<=0 THEN PRINT "Number must be greater than 0, please reenter"
GOTO 1860
1880 INPUT "Number of animals exposed in the concurrent control group ";
NCTRL
1890 IF NCTRL < 0 THEN PRINT "Invalid number, please reenter":GOTO 1880
1900 IF NRCTRL>NCTRL THEN PRINT "The number of responders must be no greater"
1910 IF NRCTRL>NCTRL THEN PRINT "than the number exposed, please reenter":
GOTO 1860
1920 C = NRCTRL / NCTRL ' empirical control resp
1930 BIGC = C ' initial predicted control resp
1940 GOTO 1680
1950 CLS
1960 RETURN
487
-------
2.3 Listing of Computer Program for Probit Analysis (Continued).
1970
1 980
1 990
2000
2010
2020'
2030
2040
2050
2060
2070
2080
2090
2100
2110
2120
2130
2140
2150
2160
2170
2180
2190
2200
2210
2220
2230
2240
2250
2260
2270
2280
2290
2300
2310
2320
2330
2340
2350
2360
2370
2380
2390
2400
2410
2420
2430
2440
2450
2460
2470
2480
2490
2500
2510
1 subroutine to load in op, oy, t05, chisq
»
FOR I 1 TO 9
READ OP(I) ' op values are percentiles for predicted curve
NEXT I
DATA .01, .05,.!,. 15, .5, .85, .9, .95, .99
FOR I 1 TO 9
READ OY(I) ' oy values are probits corresponding to op values
NEXT I
DATA 2.6737,3.3551,3.7184,3. 9636, 5.,6. 0364
DATA 6.2816,6.6449,7.3263
1 T05 values are 97.5 percentiles (alpha= 0.05 two-sided)
1 for a t-distribution . I denotes the degrees of freedom (df)
1 and runs from 1df to 1Sdf
FOR I 1 TO 18
READ T05( I )
NEXT I
DATA 12.706,4.303,3.182,2.776,2.571,2.447,2.365,2.306
DATA 2.262,2.228,2.201,2.179,2.16,2.145,2.131,2.12,2. 11,2.101
1 CHISQ values are 95 percentiles (alpha =0.05 two sided)
1 for a chi-square distribution. I denotes the degrees of freedom
1 (df) and runs from 1df to 18df.
FOR I 1 TO 18
READ CHISQ( I )
NEXT I
DATA 3.841,5.991,7.815,9.488,11.07.12.592,14.067,15.507
DATA 16.919,18.307.19.675,21.026,22.362,23.685,24.996,26.296
DATA 27.587,28.869
RETURN
1 subroutine to comput probits
1 input is nd,r where nd=#exposed and r=#responding
1 output - probit
1 precision equals that of BASIC
PP = (P
BIGC )/( 1 I
THEN XP=-105 : GOTO 2570
THEN XP= 105 : GOTO 2570
GOTO 2570
P - R/NO
PPP=PP
IF R=0 AND ITER'O
IF R=NO AND ITER=0
IF BIGC >= P THEN XP=-105 :
P0=-.322232431088*
P1»-1I : P2=-.342242088547*
P4=- .0000453642210148*
00».099348462606* : 01 = .58858 1 570495 *
02 » .531103462366* : 03». 1 0353775285 *
04= .0038560700634*
IF PPP> .5 THEN PP = 1PP
IF PP».5 THEN GOTO 2570
Y = SQR(LOG(1/(PP*PP) ) )
T1 = ( Y*P4 + P3) * Y
(T1 + P2) * Y
(T2 + P1) * Y
(T3 + PO)
BIGC) ' Abbot t s f ormu I a
infinity if r
infinity if r
1 -infinity if p < c
P3=-.0204231210245*
T2
T3
T4
U1 = (Y*04 » 03)
488
-------
2.3 Listing of Computer Program for Probit Analysis (Continued).
2520
2530
2540
2550
2560
2570
2580
2590
2600
2610
2620
2630
2640
2650
2660
2670
2680
2690
2700
2710
2720
2730
2740
2750
2760
2770
2780
2790
2800
2810
2820
2830
2840
2850
2860
2870
2880
2890
2900
2910
2920
2930
2940
2950
2960
2970
2980
2990
3000
3010
3020
3030
3040
3050
3060
U2 - (U1 + 02 ) * Y
U3 = (U2 + 01 ) * Y
U4 = 3 0 THEN Z = Z/SOR(2
= 242 . 66795523053 1 8*
=6.996383488619135*
3215. 058875869861 2*
= 15. 08279763040779*
= 300 .459261 0201616*
= 339 .32081 67343437*
= 43 . 16222722205673*
=.564195517478974* :
=300.4592609569833*
3931 . 3540948506096*
=277.5854447439876*
=12 . 78272731962942*
=-2.9961070770354220
= . 2269565935396869*
= 2 .231924597341847D
= 1 .0620923052846790 -
= 1 . 05 1675 1 07067932*
= 1 *
=.3989422804014327*
> 4* GOTO 3010
>= .46875* GOTO 2920
1(1) : 80T3Q1 ( 1 )
= 2 TO 4
2 * ( I 1 ) : ZZ = Z A I
TOP + PHI) * ZZ
BOT + 01(1) * ZZ
I
TOP / BOT : ERFX Z
)
: P1 (2
: P1 (4
01(2
: 01(4
: P2(2
P2(4
: P2(6
P2(8)
: 02(2
: 02(4
: 02(6
I of
) = 21
) = -3
)=91
) = 1*
)=45
normal distribution
.97926161 894152*
. 560984370181 5390-02
. 1649054045 149*
1 .91895371 18729*
) = 1 52 .9892850469404*
)=7.
= \ .
21 1 758250883094*
3686485738271670-07
)=790. 950925327898*
)=638. 980264465631 2*
) = 77
.00015293522947*
: 02(8)=1#
-03 :
: P3(
02
P3(2
4> = -
02 : 03(2)
: 03(4
1
* R1
) = 1 .
)= 4. 94 73 091 062325070-02
.2786613086096478*
=. 1913089261078298*
98733201817^353*
: GOTO 3110
TOP=P2( 1 ) : BOT=02{ 1 )
FOR I
I 1 =
TOP »
BOT =
NEXT
R2
I F Z
ERFX
Z2 =
FOR I
I 1 *
= 2 TO 8
I - 1 : ZZ = Z * I 1
TOP + P2( I ) ZZ
BOT + 02(1) ZZ
I
TOP / BOT : ERFCX *
< 13.038* THEM ERFCX
* 1# - ERFCX : GOTO
0*
= EXP
3110
1# / (Z*Z) : TOP * P3( 1 ) :
» 2 TO 5
2 * ( I - 1 ) : ZZ = Z2
TOP=TOP + P3(I)*ZZ : BOT=
NEXT
R3 *
I
TOP / BOT
* 11
BOT +
( -Z*
BOT
Q3( I
Z) * R2
= 03(1 )
)*ZZ
489
-------
2.3 Listing of Computer Program for Probit Analysis (Continued).
3070
3080
3090
3100
3110
3120
3130
3140
3150
3160
3 1 70
3180
3190
3200
3210
3220
3230
3240
3250
3260
3270
3280
3290
3300
3310
3320
3330
3340
3350
3360
3370
3380
3390
3400
3410
3420
3430
3440
3450
3460
3470
3480
3490
3500
3510
3520
3530
3540
3550
3560
3570
3580
3590
3600
361 0
CONST
ERFCX
IF Z <
ERFX
ERFDU =
I'F Z1
IF
Z
IF
IF
S Q R ( 2 #)
E1
CONST
CONST
0#
13 . 038# THEN ERFCX
1# ERFCX
R FX
0 THEN PVALUE= (1
Z1 >= 0 THEN PVALUE= (
Z 1 ' recover z
RVALUE <=0 THEN RVALUE
RVALUE >=1 THEN RVALUE
RETURN
1 subroutine to fit probit model
SNW 0! : SNUX=0! : SNUY=OI : SNUXY=OI
SNUXP 0! : SNUXPP 0! : SNUXPY - 0! :
FOR I 1 TO N
ND NDOSE(I) : U UT(I) : X - LDOSE(I)
X2
R3
(EXP(-Z*Z)/Z) * E1
- ER FDU ) / 2!
1 + ERFDW)/2!
.000001
.999999
SNUXX=0I :
SNUXXP 01
SNUYY = 0!
NRESP( I )
NU ND*U : XPR
IF ITER <=0 THEN
IF ITER > 0 THEN
SNUY = SNUY « NU
SNU - SNU * NU :
SNUXP SNUXP +
SNUXPY SNUXPY
NEXT I
XBAR - SNUX/SNU :
SYY - SNUYY SNUY
SXY SNUXY SNUX
- XPR I ME ( I )
GOSUB 2270 : YPROS(I ) = PROB I T : '
Y - UKY(I) ' working probit if
'Y : SNUYY - SMUYY + NU*Y*Y : SNUXY
SNUX = SNUX + NU*X : SNUXX SNUXX +
NU'XPR : SNUXPP = SNUXPP * NU*XPR*XPR
+ NU*XPR*Y : SNUXXP SNUXXP « N U * X
= YPROB( I }
not 0 iteration
SNUXY
NU*X*X
XPR
NU*X*Y
YBAR - SNUY/SNU
* SNUY / SNU :
* SNUY / SNU :
: XPBAR - SNUXP / SNU
SXX * SNUXX SNUX * SNUX / SNU
SXPXP SNUXPP SNUXP * SNUXP /
SNU
SXXP -
SXPY -
I F SXX
SLOPE
INTCPT
SNUXXP
SNUXPY
= 0 THEN
SNUX * SNUXP
SNUXP * SNUY
GOTO 5610
SNU
SNU
SXY /
- YBAR
IF NCTRL <= 0
SXPXP SXPXP +
SXPY SXPY +
NUM - (SXY/SXX)
SXX
SLOPE * XBAR
THEN RETURN > otherwise adjust
+ NCTRL * (11 BIGO/B I GC
for natural response rate
NCTRL * (C
((SXXP *
11 ((SXXP
NUH / DEN
SXY / SXX
(II BIGC) *
SI GC + DELC
= YBAR SLOPE
* SXXP)
- BI GO/B IGC
SXPY) / (SXPXP * SXX))
/ (SXPXP * SXX))
((SXPY
XBAR
TEHP * SXXP) / (SXPXP
(DELC
XPBAR / (1 I
SXXP*SXXP)/SXX)
BIGC))
DEN =
SLOPE
TEMP =
DELC -
BIGC -
INTCPT
RETURN
1 subroutine to compute y, wky, w, zv : input is n, Idose, intcpt, slope
FOR K - 1 T.Q N
Y(K) » INTCPT + LDOSE(K) * SLOPE ' predicted probit
Z = Y(K) 51 : GOSUB 2590 ' get p value
ZV = (1 I/(SQR(2!«PI ) ) )*EXP( - .5*Z*Z)
IF ZV< = 0 THEN ZV=.000001
P - NRESP(K)/NDOSE(K) : P=(P - BIGC)
UKY(K)» Y(K) + (P - PVALUE) / ZV '
XPRIME(K) -(II - PVALUE) / ZV
ADJ = BIGC / (1 I BIGC)
UT(K) = ZV * ZV / ( (PVALUE + ADJ)*(1 I
NU - NDOSE(K)*UT(K)
/ (11 - BIGC)
working probit
PVALUE) )
compute weight
490
-------
2.3 Listing of Computer Program for Probit Analysis (Continued).
3620
3630
3640
3650
3660
3670
3680
3690
3700
3710
3720
3730
3740
3750
3760
3770
3780
3790
3800
3810
3820
3830
3840
3850
3860
3870
3880
3890
3900
3910
3920
3930
3940
3950
3960
3970
3980
3990
4000
4010
4020
4030
4040
4050
NEXT K
RETURN
' subroutine to check for convergence
IF ITER > 25 THEN GOTO 3750 ' modify here if more iterations required
IF ITER > 0 THEN GOTO 3690
OLDINT INTCPT : OLOSLO SLOPE
GOTO 3940
IF COLDINT=0) OR (OLDSLO=0) THEN GOTO 3730
CHK1 ABSCOLDINT INTCPT) / OLOINT
CHK2 ABSCOLDSLO SLOPE) / OLDSLO
IF (CHK1 < .005) AND (CHK2 < .0001) THEN 0 K = 1 ' this is cvgt check
OLDINT INTCPT : OLOSLO SLOPE
GOTO 3940
CLS
PRINT " ":PRINT " "
PRINT #2," **«****«******««.*««*.««.«.*.....«..*...
Note : No convergence in 25 iterations. This usually
means that a probit model is not appropriate
for your concentration response data.
PRINT #2 , "
* II
PRINT « 2 , "
* II
PRINT #2 , "
* II
PRINT #2 , "
PRINT # 2,"
* ii
PRINT #2,"
PRINT #2, TITLES : PRINT #2, " "
PRINT #2," ":PRINT #2," "
PRINT " [[[C CC C t C C[[[[ C C C C t [ C [[[ C C C t t [[[[([ C t [ t [[[[[[[ t t [ C [C t CC C [[[[[
P R I N T " C [
PRINT " [ Note : No convergence in 25 iterations. This usually [
PRINT " C means that a probit model is not appropriate [
PRINT" C foryourconcentrationresponsedata. [
P R I N T " C t
PRINT » t C t C t t C C t C [ t [ t t t [ [ C C [ [C CI[C C[C tIC t[IC t t C t[C C C t t C C[[[C C C C t t C C C
PRINT " ":PRINT " ": GOTO 5550
RETURN
1 goodness of fit - printed output
PRINT #2. TITLES : PRINT #2, " "
PRINT #2,
Predicted '
PRINT #2, "
Proper t i on"
PRINT #2, "
Responding"
PRINT #2. » "
IF NCTRL = 0 THEM GOTO 4040
RATIO = NRCTRL / NCTRL : ARATIO=0
PRINT #2.USING " Control #### *#*# *.###* *.*#*#
#.####";NCTRL,NRCTRL,RATIO,ARATIO.BIGC
FOR L = 1 TO N
Z = Y(L) 5 : GOSUB 2590
Number
Cone. Exposed
Observed
Number Proportion
Resp. Respond!ng
Adjusted
P report i on
Respondi ng
491
-------
2.3 Listing of Computer Program for Probit Analysis (Continued).
4060 RATIO = NRESP(L) / NOOSE(L) :ARATIO ( R A T I 0 - B I G C ) / ( 1 ! - B I G C )
4070 PRINT #2,USING "#####.#### ##*# #### #.#### #.###*
#.####";DOSE(L),NDOSE(L),NRESP(L>,RATIO,ARATIO,RVALUE
4080 NEXT L
4090 CHIHET SYY SLOPE * SXY
4100 I -F NCTRL>0 THEN CHIHET CHIHET DELC * SXPY / (1! 8 I G C )
4110 PRINT # 2 , " ": PRINT #2, " ":PRINT #2, "Chi Square Heterogeneity
4120 PRINT #2, USING "*##.###";CHIHET : PRINT #2, " "
4130NDF N 2 T 1.96 : MET 1!
4140 IF CHIHET <= CHISQ(NDF) THEN GOTO 4240
4150 T T05(NDF) : HET - CHIHET / (N 2)
4160 PRINT #2, ii ***»************«******«******«*******««***««««.*,...«,.,
* it
4170 PRINT #2, "* WARNING
* it
4180 PRINT * 2 , "*
* it
4190 PRINT #2, " * Significant heterogeneity exists. The results reported
4200 PRINT #2, "* for this data set may not be valid. The results should
* ii
4210 PRINT #2, " be interpreted with appropriate caution.
* ii
4220 PRINT #2, 'I**************************************.*******-..*.***.*..*..
* II
4230 PRINT #2, " "
4240 SEB (1 ! / SXX )
4250 IF NCTRL > 0 THEN SE8 1! / (SXX (SXXP*SXXP)/SXPXP)
4260 G= HET * T*T * SEB / (SL0PE*SL0PE)
4270 IF G < 1 I THEN GOTO 4350
4280
4290
4300
4310
4320
4330
PR
PR
PR
* ii
PR
* ii
PR
* ii
PR
* ii
I NT
I NT
I NT
I NT
I NT
I N T
# 2 , " * NOTE
# 2 , " *
#2, " * Slope not significantly different from zero.
#2, "* EC fiducial limits cannot be computed.
4340 PRINT #2, " "
4350 IF NCTRL > 0 THEN SEC - <1! B1GCK2 / (SXPXP (SXXP SXXP)/SXX)
4360 SEI (11/SNW) + (X8AR * XBAR SE8)
4370 IF NCTRL > 0 THEN SEI = SEI + (XPBAR»XPBAR*SEC/<(1 I -BIGC)A2) )
4380 PRINT #2,"Mu = ";:PRINT #2, USING "####.######";
(5! INTCPT)/SLOPE
4390 PRINT #2,"Sigma = " ; : P R I N T #2, USING "####.#####*»;( 1 I/SLOPE>
4400 PRINT #2, " "
4410 PRINT #2,"Parameter Estimate Std. Err. 9 5 X Confidence
Limits"
4420 PRINT #2,
4430 XL = INTCPT - T * SQR : XU - INTCPT + T * SQR(SEI*HET)
492
-------
2.3 Listing of Computer Program for Probit Analysis (Continued).
4440
4450
4460
4470
4480
4490
4500
45 10
4520
4530
4540
4550
4560
4570
4580
4590
4600
4610
4620
4630
4640
4650
4660
4670
4680
4690
4700
4710
4720
4730
4740
4750
4760
4770
4780
4790
4800
4810
4815
4820
4830
4840
4850
4860
4870
4880
4890
4900
4910
4920
PRINT *2, USING "Intercept «###.*#««## ###«.*#«##* (###*#.######,
#####.'######>";INTCPT,SQR(SEI*HET),XL,XU
XL SLOPE T * SQR(SEB*HET) : XU SLOPE + T * SQR(SEB*HET)
PRINT #2, USING "Slope ####.#####* *###.##«### (#*###.*#####,
#####.##*###)";SLOPE,SQR(SEB*HET),XL,XU
PRINT #2, " "
IF NCTRL <=0 THEN GOTO 4530
XL - BIGC T * SOR(SEC'HET) : XU - BIGC + T SQR(SEC*HET)
PRINT #2, USING "Spontaneous ##*#.###### #*##.#*#**# (»####.;
#####.######)";BIGC,SQR(SEC'HET),XL,XU
PRINT #2, "Response Rate"
GOTO 4550
PRINT #2 ,
PRINT
GOSU8
PRINT
PRINT
PRINT
PRINT
PRINT
FOR I
M - (
IF G> =
"Theoretical Spontaneous Response Rate
USING "#.####";BIGC
#2 ,
290
#2, " " :
#2, "
it 2 , "
# 2 , "Point
# 2 , " "
1 TO 9
OY( I )
1 I THEN
PRINT #2 ,
Estimated
TITLES : PRINT #2. " "
EC Values and Confidence Limits'
Lower
: PRINT
Upper"
Cone.
95X Confidence Limits"
SLOPE
0!
(G
( 1 I
T '
+ T '
THEN
THEN
SE
IF NCTRL >
TEMP M +
SE SOR (
XL - TEMP
XU - TEMP
MSG = 0
IF XL<- 10
IF XU> 20
GOTO 4920
1 fixup formulas
C11 = SEB
TEMPC - ((11 BIGC)
C22 = SEC / TEMPC
I NTCPT ) /
GOTO 4900
THEN GOTO 4730
/ (11 G ) )*(H
G)/SNW + (CM'
* SE / (SLOPE *
/ (SLOPE *
XBAR)
X B A R ) " 2 )
S E B )
SE
SE
SQR(HET)
(1!
(1!
G))
G))
XL=
XU=
10
20
MSG = 1
if a threshold (spontaneous response) was estimated
(11
BIGC) )
C12
R1
R2
R3
R4
1 there
R5 « G
TEMP '
11
/ (SXXP - (SXX * SXPXP)/SXXP)
M + (G / (1*-G)) ((M XBAR) XPBAR * C12 / C11)
T / (SLOPE * (11 G))
11/SNU 2I*(M - XBAR) * XPBAR * C12
XPBAR * XPBAR * C22 + C11 ((M XBAR)*2)
is some disagreement whether the following * sign is correct
(11 /SNU
SOR((R3 *
+ XPBAR
R4 R5)
' XPBAR
HET)
(C22
C12*C12/C1 1 ) )
R1
R1
XL «
XU »
MSG'O
IF XL<- 1 0
IF XU> 20
R2
R2
TEMP
TEMP
THEN XL«- 10
THEN XU= 20 :
IF G<1I THEN GOTO 4920
PRINT #2,USING "EC##.##
GOTO 4930
PRINT # 2.USING "EC**.**
MSG = 1
***#****.****"; 100*OP(I); 10'M
#*#«#**#*#.****";
4930 NEXT I
1000P( I ) ;
#*******.****
M; 10 * X L; 10
********.****
'XU
493
-------
2.3 Listing of Computer Program for Probit Analysis (Continued).
4940
4950
4960
4970
4980
4990
5000
5010
5020
5030
5040
5050
5060
5070
5030
5090
51 00
5110
5120
5130
5140
5150
5160
51 70
5 180
5 190
5200
5210
5220
5230
5240
5250
5260
5270
5280
5290
5300
5310
5320
5330
5340
5350
5360
5370
5380
5390
5400
5410
5420
5430
5440
IF MSG = 1 THEN PRINT #2. " " : PRINT #2," NOTE Upper limits greater than
or eqUal to 1.E20 are really infinite"
GOSUB 290
subroutine to do the probit plot
#2,
#2 ,
TITLES
PRINT #2, " "
PLOT OF ADJUSTED
PRINT #2,
, "Probit"
: RADJ
PRINT
PR I NT
LINE" :
PRINT #2
DROW=10!
LLD01 (2.6732 - INTCPT )/SLOPE
ADJ (LLD99 LL001)/ 68
DCOL LLD01 : CADJ ADJ
FOR J=1 TO N
DROU YPROBC J )
RDIFF/RADJ + 1
LOOSE(J) DCOL
CD I FF/CADJ + 1
0 THEN IROWX=51
10 THEN IROUX=1
PROBITS AND PREDICTED REGRESSION
LLD99
(7.327- INTCPT)/SLOPE
RD I F F
I ROUX
CD I F F
I COLX -
IF YPROB(J)
IF YPROB(J )
I COLX*1
I COLX = 71
1 change plotting symbol here
IF LDOSE(J)LLD99 THEN
PLTT$(IROUX,ICOLX)="o"
NEXT J
ND = 100 ' number of points on the predicted curve
FOR 1=1 TO 99
if desired
p is negative
IF I/NO <= BIGC THEN GOTO 5300
GOSUB 2270 ' compute probit
LOGD=(PROBIT INTCPT)/SLOPE 'EC value
RDIFF DROW - PROBIT
IROWX RDIFF/RADJ » 1
CDIFF - LOGD DCOL
ICOLX = CDIF F/CADJ » 1
IF PROBIT < 0 THEN IROUX=51
IF PROBIT > 10 THEN IROUX=1
IF LOGD < LLD01 THEN ICOLX=1
IF LOGD > LLD99 THEN ICOLX=71
IF PLTT$( IROWX, ICOLX) <>"o" THEN PLTT*{IROVX, I COLX)
plotting symbols here if desired
NEXT I
NUM=10 : LAG = 0
FOR 1=1 TO 51
11=1-1
JJX=(II-2)/5 : JJX*II - <5«JJX)
IF JJX*0 THEN PRINT #2, USING "
IF JJX<>0 THEN PRINT *2 , " -";
FOR J«1 TO 71
IF (PLTTSCI.J) <> "o" AND PLTT$(I,J) <>
plotting symbols here if desired
JJ=J-1-LAG : IF JJ <» 0 THEN GOTO 5430
FOR K= 1 TO JJ
PRINT # 2, " " ;
NEXT K
PR INT #2, PLTTSCI ,J ) ;
LAG = J
change
#*";NUM;:PRINT #2,"+";:NUH=NUN-1
." ) THEN GOTO 5450 ' change
print blanks out to next symbol
494
-------
2.3 Listing of Computer Program for Probit Analysis (Continued).
5450 NEXT J
5460 PRINT #2,
5470 LAG=0
5480 NEXT I
5500 PRINT #2,"
5510 BOT$="EC01
11 ; B 0 T $
EC10 EC25
- - - - * * -*
EC50 EC75 EC90
5520
5530
5540
5550
5560
5570
5580
5590
5600
5610
5620
5630
5640
5650
5660
5670
5675
5680
5690
5700
5710
5720
5730
5740
5750
5760
5770
5780
5790
5800
5810
5820
5830
5840
5850
EC99"
PRINT #2," " ; B 0 T t
GOSUB 290
IF FILE$<>"Ipt1 : " THEN LOCATE 12,30 : PRINT "Output stored in " ; f I LE$
LOCATE 15,1 rINPUT "Fit another dataset ";ANS$
IF A N S J = " " THEN GOTO 5550
ANS$ = LEFT$(ANSt, 1 ) : I F ASC(ANS$)>96 THEN ANSS = CHR$(ASC(ANSI}- 32)
IF ANS$o"N" AND ANS*<>"Y" THEN GOTO 5550
IF A N S S = " Y " THEN CLEAR : CLOSE #2 : CLS : GOTO 40
GOTO 5910
CLS
PRINT M ":PRINT " "
PRINT t m 11 c m c c m m m m m m m m m m m m m m m m t'
P R I N T " C ['
PRINT " [ Note : Iterations are not converging. This usually ['
PRINT" [ means that only one concentration is on the C'
PRINT " t linear portion of the concentration response C1
PRINT" [ curve. ['
P R I N T " [ ['
PRINT » m m m c m m m m m m m 11 m m m m m m c m m t [ [
PR I NT :PR I NT
IF NCTRL - 0 THEN GOTO 5810
PRINT:PRINT "It may be possible to fit the data assuming the spontaneous"
PRINT "control rate is zero.":PRINT
INPUT "Would you like to try";ANS$
IF ANS$="" THEN GOTO 5740
ANS$=LEFT$96 THEN ANS$*CHR$(ASC(ANS$)- 32)
IF ANS$o"N" AND A N S $ <>" Y » THEN GOTO 5740
IF A N S $ = " N " THEN GOTO 5810
NCTRL«0: NRCTRL»0: BIGC=0
CLS : LOCATE 12,35 :PRINT "Working ...": GOTO 140
PRINT #2, TITLES : PRINT #2, " »
PRINT #2," «:PRINT #2," "
PRINT #2," ****«*********«
PRINT #2, "
* n
PRINT #2,"
PRINT #2,"
* H
PRINT *2,"
* H
PRINT #2,"
5860
5870
5875
5880 PRINT #2,"
Note : Iterations are not converging. This usually
means that only one concentration is on the
linear portion of the concentration response
curve.
495
-------
2.3 Listing of Computer Program for Probit Analysis (Continued).
5890 PRINT #2,"
5900 PRINT " ":PRINT " ": GOTO 5550
5910 END
496
-------
APPENDIX H
SINGLE CONCENTRATION TOXICITY TEST - COMPARISON OF CONTROL
WITH 100% EFFLUENT OR RECEIVING WATER
1. To statistically compare a control with one concentration, such as
100% effluent or the instream waste concentration, a t-test is the
recommended analysis. The t-test is based on the assumptions that the
observations are independent and normally distributed and that the
variances of the observations are equal between the two groups.
2. Shapiro-Wilk's test may be used to test the normality assumption (See
Appendix B for details). If the data do not meet the normality
assumption, the nonparametric test, Wilcoxon's Rank Sum Test, may be
used to analyze the data. An example of this test is given in
Appendix F. Since a control and one concentration are being compared,
the K = 1 section of Table F.5 contains the needed critical values.
3. The F test for equality of variances is used to test the homogeneity
of variance assumption. When conducting the F test, the alternative
hypothesis of interest is that the variances are not equal.
4. To make the two-tailed F test at the 0.01 level of significance, put
the larger of the two variances in the numerator of F.
F =
where S1 > S2
5. Compare F with the 0.005 level of a tabled F value with n,, - 1 and
n2 - 1 degrees of freedom, where n1 and n2 are the number of replicates
for each of the two groups.
6. A set of mysid growth data from an effluent (single concentration) test
will be used to illustrate the F test. The raw data, mean and variance for
the control and 100% effluent are given in Table H.I.
TABLE H.I. MYSID, MYSIDOPSIS BAHIA, GROWTH DATA FROM AN
EFFLUENT (SINGLE CONCENTRATION) TEST
Replicate
1
8
Control
100% Effluent
0.183 0.148 0.216 0.199 0.176 0.243 0.213 0.180 0.195 0.000861
0.153 0.117 0.085 0.153 0.086 0.193 0.137 0.129 0.132 0.00131
497
-------
7. Since the variability of the 100% effluent is greater than the
variability of the control, S2 for the 100% effluent concentration is
placed in the numerator of the F statistic and S2 for the control is
placed in the denominator.
0.00131
F = = 1.52
0.000861
8. There are 8 replicates for the effluent concentration and 8 replicates for
the control. Thus, both numerator and denominator degrees of freedom are
equal to 7. For a two-tailed test at the 0.01 level of significance, the
critical F value is obtained from a table of the F distribution (Snedecor and
Cochran, 1980). The critical F value for this test is 8.89. Since 1.52 is
not greater than 8.89, the conclusion is that the variances of the control and
100% effluent are homogeneous.
9. Equal Variance T-Test.
9.1 To perform the t-test, calculate the following test statistic:
t =
I/ 1 + 1
ni n2
Where:
Y, = Mean for the control
Y2 = Mean for the effluent concentration
_ (n, - 1JS,2 + (n2 - 1) S22
P A< n, + n2 - 2
S = Estimate of the variance for the control
S = Estimate of the variance for the effluent
2 concentration
n., = Number of replicates for the control
n2 = Number of replicates for the effluent
concentration
498
-------
9.2 Since we are usually concerned with a decreased response from the
control, such as a decrease in survival or a decrease in reproduction, a
one-tailed test is appropriate. Thus, you would compare the calculated t
with a critical t, where the critical t is at the 5% level of significance
with n1 + n2 2 degrees of freedom. If the calculated t exceeds the critical
t, the mean responses are declared different.
9.3 Using the data from Table H.I to illustrate the t-test, the
calculation of t is as follows:
0.195 - 0.132
t = = 3.83
(8 1) 0.000861 + (8 1) 0.00131
Where: SD = / = 0.0329
" 8 + 8-2
9.4 For an 0.05 level of significance test with 14 degrees of freedom,
the critical t is 1.762 (Note: Table D.5 for K = 1 includes the critical
t values for comparing two groups). Since 3.83 is greater than 1.762,
the conclusion is that the growth for the 100% effluent concentration is
significantly lower than growth for the control.
10. Unequal Variance T-Test.
10.1 If the F test for equality of variance fails, the t-test is still a
valid test. However, the denominator of the t statistic is adjusted as
follows:
t =
Where:
Y, = Mean for the control
Y2 = Mean for the effluent concentration
S1 = Estimate of the variance for the control
499
-------
S2 = Estimate of the variance for the effluent
concentration
n, = Number of replicates for the control
n2 = Number of replicates for the effluent
concentration
10.2 Additionally, the degrees of freedom for the test are adjusted
using the following formula:
(n, D(n2 - 1)
df =
(nz 1)C2 + (1 C)2 (n, - 1)
Where:
C =
10.3 The modified degrees of freedom is usually not an integer. Common
practice is to round down to the nearest integer.
10.4 The t-test is then conducted as the equal variance t-test. The
calculated t is compared to the critical t at the 0.05 significance level
with the modified degrees of freedom. If the calculated t exceeds the
critical t, the mean responses are found to be statistically different.
500
-------
APPENDIX I
LINEAR INTERPOLATION METHOD
1. GENERAL PROCEDURE
1.1 The Linear Interpolation Method is used to calculate a point estimate of
the effluent or other toxicant concentration that causes a given percent
reduction (e.g., 25%, 50%, etc.) in the reproduction or growth of the test
organisms (Inhibition Concentration, or 1C). The procedure was designed for
general applicability in the analysis of data from short-term chronic toxicity
tests, and the generation of an endpoint from a continuous model that allows a
traditional quantitative assessment of the precision of the endpoint, such as
confidence limits for the endpoint of a single test, and a mean and
coefficient of variation for the endpoints of multiple tests.
1.2 The Linear Interpolation Method assumes that the responses (1) are
monotonically non-increasing, where the mean response for each higher
concentration is less than or equal to the mean response for the previous
concentration, (2) follow a piecewise linear response function, and (3) are
from a random, independent, and representative sample of test data. If the
data are not monotonically non-increasing, they are adjusted by smoothing
(averaging). In cases where the responses at the low toxicant concentrations
are much higher than in the controls, the smoothing process may result in a
large upward adjustment in the control mean. Also, no assumption is made
about the distribution of the data except that the data within a group being
resampled are independent and identically distributed.
2. DATA SUMMARY AND PLOTS
2.1 Calculate the mean responses for the control and each toxicant
concentration, construct a summary table, and plot the data.
3. MONOTONICITY
3.1 If the assumption of monotonicity of test results is met, the observed
response means (Y,-) should stay the same or decrease as the toxicant
concentration increases. If the means do not decrease monotonically, the
responses are "smoothed" by averaging (pooling) adjacent means.
3.2 Observed means at each concentration are considered in order of
increasing concentration, starting with the control mean_(Y1). If the mean
observed response at the lowest_toxicant concentration (Y2) is equal to or
smaller than the control mean (Y,), it is used as the response. If it is
larger than the control mean, it is averaged with the control, and this
average is used for both the control response (M,) and the lowest toxicant
concentration response (M2). This mean is then compared to the mean
observed response for the next higher toxicant concentration (Y3). Again,
if the mean observed response for the next higher toxicant concentration is
smaller than the mean of the control and the lowest toxicant concentration, it
501
-------
is used as the response. If it is higher than the mean of the first two,' it
is averaged with the first two, and the mean is used as the response for the
control and two lowest concentrations of toxicant. This process is continued
for data from the remaining toxicant concentrations. A numerical example of
smoothing the data is provided below. (Note: Unusual patterns in the
deviations from monotonicity may require an additional step of smoothing).
Where Y,- decrease monotonically, the Y,- become M,- without smoothing.
4. LINEAR INTERPOLATION METHOD
4.1 The method assumes a linear response from one concentration to the next.
Thus, the 1C is estimated by linear interpolation between two concentrations
whose responses bracket the response of interest, the (p) percent reduction
from the control .
J+1
4.2 To obtain the estimate, determine the concentrations Cj and C
which bracket the response M,,(l p/100), where M., is the smoothed control
mean response and p is the percent reduction in response relative to the
control response. The linear interpolation estimate is calculated as follows:
ICp = ^
p/100) - Mj](CJ+1
M
J+I
(1)
M
Where:
J+1
M1
Mj
MJ+1
p
ICp
The tested concentration whose observed
mean response is greater than M^l - p/100).
The tested concentration whose observed
mean response is less than M^l - p/100).
Smoothed mean response for the control .
Smoothed mean response for concentration J.
Smoothed mean response for concentration J+l.
Percent reduction in response relative to
the control response.
The estimated concentration at which there is
p% reduction from the smoothed mean control response.
This ICp is reported for the test, together with
the 95% confidence interval from the Bootstrap Method,
as described below.
4.3 If C, is the highest concentration tested, then the ICp is specified as
"greater than C/1.
502
-------
5, CONFIDENCE INTERVALS
5.1 Due to the use of a linear interpolation technique to calculate an
estimate of the ICp, standard statistical methods for calculating confidence
intervals are not applicable for the ICp. This limitation is avoided by
using the Bootstrap Method proposed by Efron (1982) for deriving point
estimates and confidence intervals.
5.2 In the Linear Interpolation Method, the smoothed response means are
used to obtain the ICp estimate reported for the test. The Bootstrap Method
is then used to determine the 95% confidence interval for the true mean. In
the Bootstrap Method, the test data, Y-,-, are randomly resampled with
replacement, to produce a new set of data, YjM*, that is statistically
equivalent to the orginal data, but which produces a new and somewhat
different estimate of the ICp (ICp*). This process is repeated 80 times
in the procedure developed by Battelle (Marcus and Holtzman, 1988),
resulting in multiple "data" sets, each with an associated ICp* estimate.
The distribution of the ICp* estimates derived from the sets of resampled
data approximates the sampling distribution of the ICp estimate. The
standard error of the ICp is estimated by the standard deviation of the
individual ICp* estimates. Empirical confidence intervals are derived
from the quantiles of the ICp* empirical distribution. For example, if
the test data are resampled a minimum of 80 times, the empirical 2.5% and
97.5% confidence limits are approximately the second smallest and second
largest ICp* estimates (Battelle, 1988).
5.3 The width of the confidence intervals calculated by the Bootstrap
Method is related to the variability in the data. When the intervals are
wide, the reliability of the 1C estimate is in question. However, narrow
intervals do not necessarily indicate that the estimate is highly reliable,
because of undetected violations of assumptions and the fact that confidence
limits based on the empirical quantiles of a Bootstrap distribution of 80
resamples may be unstable.
5.4 The Bootstrap Method is computationally intensive, and not amenable to
hand calculations. For this reason, all of the calculations associated with
the Linear Interpolation Method have been incorporated into a single
computer program, BOOTSTRP, described in Paragraph 7 below.
6. MANUAL CALCULATIONS
6.1 Data Summary and Plots
6.1.1 The data used in this example are the mysid growth data used in the
example in Section 14. Table I.I includes the raw data and the mean growth
for each concentration. A plot of the data is provided in Figure I.I.
503
-------
TABLE I.I. MYSID, MYSIDOPSIS BAHIA, GROWTH DATA
Toxicant Concentration (ppb)
Replicate Control 50.0 100.0 210.0 450.0
1
2
3
4
5
6
7
8
Mean (7,)
i
0.183
0.148
0.216
0.199
0.176
0.243
0.213
0.180
0.195
1
0.192
0.193
0.237
0.237
0.256
0.191
0.152
0.177
0.204
2
0.190
0.172
0.160
0.199
0.165
0.241
0.259
0.186
0.197
3
0.153
0.117
0.085
0.153
0.086
0.193
0.137
0.129
0.132
4
_
0.060
0.009
-
0.203
0.091
5
504
-------
en
O
en
0.30 -
0.28 -_
0.26 :
0.24 -_
0.22 -
0.20
0.18
0.16 :
0.14 -
0.12
0.10 -
0.08 :
0.06
0.04 -
0.02
0.00
* *
INDIVIDUAL REPLICATE MEAN WEIGHT
CONNECTS THE OBSERVED MEAN VALUE
CONNECTS THE SMOOTHED MEAN VALUE
*
*
50
I
100
210
*
~T
450
Figure I.I
TOXICANT CONCENTRATION (PPB)
Plot of observed and smoothed means for mysid, Hysidopsis bahia, growth data.
-------
6.2 Monotonicity
6.2.1 As can be seen from the plot, Figure I.I, the observed means are not
monotonically non-increasing with respect to concentration. Therefore, the
means must be smoothed prior to calculating the 1C.
6.2.2_ Starting with the control mean, Y1 = 0.195 and Y2 = 0.204, we see that
smoothed means:
M, = M2 = (^ + Y2)/2 = 0.200
Y1 < Y2. Calculate the smoothed means:
6.2.3 Since Y5 = 0.091 < Y, = 0.132 < Y3 = 0.197 < M2, set M3 = 0.197 and M,
0.132, and M5 = 0.091. Table 1.2 contains the smoothed means and Figure I.I
gives a plot of the smoothed response curve.
TABLE 1.2. MYSID, MYSIDOPSIS BAHIA, MEAN GROWTH
RESPONSE AFTER SMOOTHING
Toxicant
Cone.
(ppb)
Control
50.0
100.0
210.0
450.0
i
1
2
3
4
5
MI
(mg)
0.200
0.200
0.197
0.132
0.091
6.3 Linear Interpolation
6.3.1 Estimates of the IC25 and IC50 can be calculated using the Linear
Interpolation Method. A 25% reduction in mean weight, compared to the
controls, would result in a mean weight of 0.150 mg, where M^l-p/100) =
0.200(1-25/100). A 50% reduction in mean weight, compared to the controls,
would result in a mean weight of 0.100 mg. Examining the means and their
associated concentrations (Table 1.2), the response, 0.150 mg, is bracketed by
C, = 100 ppb and C, = 210 ppb. The response, 0.100 mg, is bracketed by C4 =
210 ppb and C5 = 450 ppb.
6.3.2 Using Equation 1 from 4.2, the estimate of the IC25 is calculated as
follows:
ICp = Cd + [M^l p/100) - MJ (CJ+1 -
(MJ+1
506
-------
IC25 = 100 + [0.200(1 - 25/100) - 0.197] (210 - 100)
(0.132 - 0.197)
= 179 ppb.
6.3.3 Using Equation 1 from 4.2, the estimate of the IC50 is calculated as
follows:
ICp = Cj + [M,(l - p/100) - MJ (CJ+1 -
(M
J+1
IC50 = 210 + [0.200(1 50/100) - 0.132] (450 - 210)
(0.091 - 0.132)
= 397 ppb.
6.4 Confidence Intervals
6.4.1 Confidence intervals for the ICp are derived using the Bootstrap
Method. As described above, this method involves randomly resampling the
individual observations and recalculating the ICp at least 80 times, and
determining the mean ICp, standard deviation, and empirical 95% confidence
interval. For this reason it is not practical to
manually. However, a computer program, BOOTSTRP,
perform these calculations
described below, is
available to carry out all of the calculations, including the Bootstrap
Method, required for the Linear Interpolation Method.
7. COMPUTER CALCULATIONS
7.1 The computer program, BOOTSTRP, prepared for the Linear Interpolation
Method, was written in Fortran for IBM compatible PCs. The program was
developed by the Battelle Laboratories, Columbus, Ohio, with funding from the
Environmental Research Laboratory, U. S. Environmental Protection Agency,
Duluth, Minnesota (USEPA, 1988e). For information concerning the computer
program and program documentation, contact the Environmental Research
Laboratory - Duluth.
7.2 BOOTSTRP performs the following functions: (1) calculates the observed
response means (Y,-), (2) checks the responses for monotonicity, (3)
calculates smoothed means (M,-) if necessary, (4) uses the means, M,-, to
calculate the initial ICp of choice by linear interpolation, (5) performs a
user-specified number of Bootstrap Method resamples, (6) calculates the mean
and standard deviation of the Bootstrap ICp estimates, and (7) provides an
empirical 95% confidence interval to be used with the initial ICp.
7.3 A maximum number of 8 concentrations and 10 replicates per concentration
507
-------
are allowed by the program. Equal replication across all concentrations is
not required. The value of p can range from 1% to 99%.
7.4 Data Input.
7.4.1 Data may be entered on the screen or read from an external file.
Instructions for creating external data files, such as with the use of Word
Perfect, are included with the documentation.
7.4.2 When data are entered on the screen, the program prompts the user for
the following information:
1. Concentration group number
2. Concentration amount (i.e., percent effluent)
3. Response (i.e., weight, number of young)
7.4.2.1 After data values have been entered, the program displays them on the
screen and prompts the user to verify them. An example of sample data input
on the screen is shown in Figure 1.2.
7.4.3 When data are entered from an existing file, the program prompts the
user for the file name. An example of sample data input using an existing
file is shown in Figure 1.3.
508
-------
ENTER "S" IF DATA WILL BE INPUT TO THE PROGRAM ON THE SCREEN.
ENTER "F" IF DATA WILL BE INPUT THROUGH AN INPUT FILE.
ENTER "S" OR "F
WHEN PROMPTED FOR THE CONCENTRATION GROUP NUMBER, THE NUMBER SHOULD
BE AN INTEGER FROM 1 TO 8. GROUP NUMBER 1 CORRESPONDS TO THE CONTROL
GROUP, GROUP NUMBER 2 CORRESPONDS TO THE LOWEST CONCENTRATION GROUP,
CONTINUING TO THE HIGHEST ASSIGNED GROUP NUMBER WHICH CORRESPONDS TO
THE HIGHEST CONCENTRATION GROUP.
ENTER THE CONCENTRATION GROUP NUMBER (1,2,...)
(ENTER "X" TO EXIT THE INPUT PROCEDURE): 1
ENTER THE CONCENTRATION (IN MG/L, % EFFLUENT, ETC.): -000
ENTER THE RESPONSE VALUE: .183
YOU HAVE ENTERED THE FOLLOWING VALUES:
CONCENTRATION ID = 1
CONCENTRATION = .000
RESPONSE = .183
PRESS RETURN IF THESE VALUES ARE CORRECT.
PRESS "X", THEN RETURN, IF ANY OF THE VALUES ARE INCORRECT:
THESE VALUES HAVE BEEN SUCCESSFULLY INPUT TO THE PROGRAM.
THE NEXT SET OF VALUES CAN NOW BE ENTERED.
Figure 1.2. Example of BOOTSTRP program data input on the screen.
509
-------
ENTER "S" IF DATA WILL BE INPUT TO THE PROGRAM ON THE SCREEN.
ENTER "F" IF DATA WILL BE INPUT THROUGH AN INPUT FILE.
II r II .
ENTER "S" OR "F
ENTER THE INPUT FILE NAME (SPECIFYING THE DRIVE AND
SUBDIRECTORY IF NECESSARY): mysid.dat
ENTER THE VALUE P, THE DESIRED PERCENT REDUCTION IN RESPONSE
RELATIVE TO THE CONTROL GROUP (P = 50 IS THE DEFAULT): 25
THE VALUE OF P IS 25.0
ENTER THE NUMBER OF BOOTSTRAP RESAMPLES TO BE TAKEN
(80 IS THE RECOMMENDED NUMBER): 80
Figure 1.3. Example of BOOTSTRP program data input from an existing data
file.
510
-------
7.4.4 After all of the data have been entered, the user is asked to enter the
ICp estimate desired (e.g., IC25 or IC50) and the number of Bootstrap Method
resamples that are to be taken. The program has the capability of performing
any number of resamples from 2 - 200. However, Marcus and Holtzman (1988)
recommend that a minimum of 80 Bootstrap Method resamples be used (see Figure
1.3 for example).
7.5 Data Output
7.5.1 BOOTSTRP program output includes the following:
1. A table of the test concentrations, observed response means (Y(),
and smoothed (pooled) means (M,.).
2. The linear interpolation estimate of the ICp using the means, M,.
(This ICp is reported for the test.)
3. The mean ICp and standard deviation from the Bootstrap Method
resampling.
4. The empirical 95% confidence interval calculated by the Bootstrap
Method for the ICp. (This confidence interval is used for ICp
obtained in Item 2, above).
7.6 BOOTSTRP program output for the analysis of the mysid growth data in
Table I.I is provided in Figures 1.4 and 1.5.
7.6.1 When the Bootstrap program was used to analyze this set of data,
requesting 80 resamples, the mean estimate of the IC25 was 183.0960 ppb, with
a standard deviation of 22.0142 ppb (coefficient of variation = 12.0%). The
empirical 95.0% confidence interval for the true mean was (149.2220 ppb,
239.4104 ppb).
7.6.2 When the Bootstrap program was used to analyze this set of data for the
IC50, requesting 80 resamples, the mean estimate of the IC50 was 336.9059 ppb,
with a standard deviation of 51.5538 ppb (coefficient of variation = 15.3%).
The empirical 95.7% confidence interval for the true mean was (259.7452 ppb,
424.8817 ppb).
511
-------
THE NUMBER OF RESAMPLES IS 80
*** LISTING OF GROUP CONCENTRATIONS (% EFF-) AND RESPONSE MEANS ***
CONC. (%EFF) RESPONSE MEAN MEAN AFTER POOLING
0.000 0.195 0.200
50.000 0.204 0.200
100.000 0.196 0.196
210.000 0.132 0.132
450.000 0.091 0.091
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
FROM THE INPUT SAMPLE IS 179.4003.
BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY *
OF THE ESTIMATED ICp *
THE MEAN OF THE BOOTSTRAP ESTIMATES IS 183.0960.
THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS 22.0142.
AN EMPIRICAL 95.0% CONFIDENCE INTERVAL FOR THE
BOOTSTRAP ESTIMATE IS (149.2220,239.4104).
Figure 1.4. Example of BOOTSTRP program output for the IC25.
512
-------
THE NUMBER OF RESAMPLES IS 80
*** LISTING OF GROUP CONCENTRATIONS (% EFF.) AND RESPONSE MEANS ***
CONC. (%EFF) RESPONSE MEAN MEAN AFTER POOLING
0.000 0.195 0.200
50.000 0.204 0.200
100.000 0.196 0.196
210.000 0.132 0.132
450.000 0.091 0.091
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
FROM THE INPUT SAMPLE IS 396.5921.
BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY *
OF THE ESTIMATED ICp *
THE MEAN OF THE BOOTSTRAP ESTIMATES IS 336.9059.
THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS 51.5538.
AN EMPIRICAL 95.7% CONFIDENCE INTERVAL FOR THE
BOOTSTRAP ESTIMATE IS (259.7452,424.8817).
*** NOTE: THE ABOVE BOOTSTRAP CALCULATIONS WERE BASED ON 47
INSTEAD OF 80 RESAMPLINGS. THOSE RESAMPLES NOT
USED HAD ESTIMATES ABOVE THE HIGHEST CONCENTRATION / % EFF.
Figure 1.5. Example of BOOTSTRP program output for the IC50.
«U.S GOVERNMENT PRINTING OFFICE: 1 9 9 2 . 6 k 8 - 0 0 3". 1 s 0 7 513
------- |