&EPA
          United States
          Environmental Protection
          Agency
            Office of Research and
            Development
            Washington, DC 20460
   EPA/600/4-91/022
   February 1992
Short-Term
Methods for
Estimating the
Chronic Toxicity of
Effluents and
Receiving  Waters to
Freshwater Organisms
Final Draft

-------

-------
                                  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.
                                    n

-------
                                 FOREWORD

    Environmental measurements are required to determine the quality of
ambient waters and the character of waste effluents.  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 response 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 solid 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 a third edition of the freshwater chronic toxicity test manual for
effluents, first published (EPA/600/4-85/014) by EMSL-Cincinnati in December,
1985 and revised (EPA/600/4-89/001) in March, 1989.  It provides updated
methods for estimating the chronic toxicity of effluents and receiving waters
to freshwater organisms for use by the U. S. Environmental Protection Agency
(USEPA) regional and state programs, and National Pollutant Discharge
Elimination System (NPDES) permittees.
                                    Thomas A. Clark
                                    Director
                                    Environmental Monitoring Systems
                                    Laboratory - Cincinnati

-------
                                 PREFACE

      This manual represents the third edition of the Agency's methods manual
for estimating the chronic toxicity of effluents and receiving waters to
freshwater organisms initially published by EMSL - Cincinnati in December,
1985.  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,  Environmental Services Division,  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 and 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,  Human Health and Ecological  Criteria Division, Office
      of Science and Technology, Office of Water
    Daniel Rieder, Hazard Evaluation Division, Office of Pesticides 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
                          James M. Lazorchak, Ph.D.
                          Chief, Bioassessment and Ecotoxicology Branch
                          Ecological Monitoring Research Division
                          Environmental Monitoring Systems Laboratory  -
                          Cincinnati

-------
                                 ABSTRACT

    This manual describes four short-term (four to seven days)  methods for
estimating the chronic toxicity of effluents and receiving waters to three
freshwater species:  the fathead minnow,  Pimephales promelas,  a daphnid,
Cen'odaphm'a dubia, and a green alga,  Selenastrum capn'cornutum.   The methods
include single and multiple concentration static renewal and 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.

-------
                                 CONTENTS

Foreword	   iii
Preface	    iv
Abstract	    vi
Figures  	     x
Tables	xiii
Acknowledgments  	 xviii

Section Number                                                      Page

      1. Introduction   	     1
      2. Short-Term Methods for Estimating Chronic Toxicity  ...     4
            Introduction 	     4
            Types of Tests	     6
            Static Tests 	     7
            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 Short-Term Chronic Toxicity Test  .  .    15
            Analytical  Methods 	    15
            Calibration and Standardization  	    15
            Replication and Test Sensitivity 	    15
            Variability in Toxicity Test Results 	    16
            Test Precision	    16
            Demonstrating Acceptable Laboratory Performance  ...    17
            Documenting Ongoing Laboratory Performance 	    18
            Reference Toxicants  	    20
            Record Keeping 	    20
            Video Tapes of USEPA Culture and Toxicity
              Test Methods	    20
            Supplemental Reports for Training Video Tapes   ....    21
      5. Facilities, Equipment, and Supplies   	    22
            General Requirements 	    22

                                      vii

-------
                        CONTENTS (Continued)

       Test Chambers	    22
       Cleaning Test Chambers and Laboratory Apparatus  ...    23
       Apparatus and Equipment for Culturing and
         Toxicity Tests 	    24
       Reagents and Consumable Materials  	    25
       Test Organisms	    25
       Supplies	    25
 6. Test Organisms	    27
       Test Species	    27
       Sources of Test Organisms	    28
       Life Stage	    29
       Laboratory Culturing 	    29
       Holding and Handling Test Organisms  	    29
       Transportation to the Test Site	    30
       Test Organism Disposal 	    31
 7. Dilution Water	    32
       Types of Dilution Water	    32
       Standard, Synthetic Dilution Water 	    32
       Use of Receiving Water as Dilution Water 	    34
       Use of Tap Water as Dilution Water	    35
       Dilution Water Holding 	    35
 8. Effluent and Receiving Water Sampling, Sample Handling,
       and Sample Preparation for Toxicity Tests  	    36
       Effluent Sampling  	    36
       Effluent Sample Types  	    36
       Effluent Sampling Recommendations  .  .  . .	    37
       Receiving Water Sampling 	    38
       Effluent and Receiving Water Sample Handling,
         Preservation, and Shipping 	    39
       Sample Receiving 	    40
       Persistence of Effluent Toxicity During Sample
         Shipping and Holding ..... 	    40
       Preparation of Effluent and Receiving Water Samples
         for Toxicity Tests	    40
       Preliminary Toxicity Range-Finding Tests 	   .  .    42
       Multi-concentration (Definitive) Effluent
         Toxicity Tests 	    43
       Receiving Water Tests  	    43
 9. Chronic Toxicity Test Endpoints and Data Analysis ....    45
       Endpoints	    45
       Relationship Between Endpoints Determined by
         Hypothesis Testing and Point Estimation Techniques  .    46
       Precision	    48
       Data Analysis	    48
       Choice of Analysis 	    50
       Hypothesis Tests 	    51
       Point Estimation Techniques  	    54
10. Report Preparation  	    57
       Introduction 	    57

                                viii

-------
                             CONTENTS  (Continued)

            Plant Operations	    57
            Sources of Effluent, Receiving Water, and
              Dilution Water 	    57
            Test Methods	    58
            Test Organisms	    58
            Quality Assurance   	    58
            Results	    58
            Conclusions and Recommendations  	    59
     11. Fathead Minnow, Pimephales promelas, Larval Survival
           and Growth Test	    60
     12. Fathead Minnow, Pimephales promelas, Embryo-larval
           Survival and Teratogenicity Test  	   114
     13. Daphnid, Ceriodaphnia dubia,  Survival  and Reproduction
           Test	   146
     14. Green Alga, Selenastrum capricornutum, Growth Test  .  .   .   197

Selected References  	   231
Appendices	   246
      A. Independence, Randomization,  and Outliers 	   248
      B. Validating Normality and Homogeneity of Variance
           Assumptions	   255
      C. Dunnett's Procedure 	   276
      D. T-test with Bonferroni's Adjustment 	   288
      E. Steel's Many-one Rank Test	   295
      F. Wilcoxon Rank Sum Test	   300
      G. Fisher's Exact Test	   306
      H. Single Concentration Toxicity Test - Comparison
           of Control with 100% Effluent or Receiving Water  ...   315
      I. Probit Analysis	   319
      J. Linear Interpolation Method 	   324

-------
                                  FIGURES


SECTIONS 1-10

 Number                                                              Page

   1. Control (cusum) charts 	    19

   2. Flowchart for statistical analysis of test data	    52

 SECTION 11

 Number                                                              Page

   1. Data form for the fathead minnow, Pimephales promelas,
      larval survival and growth test.  Routine chemical and
      physical determinations  	    78

   2. Mortality data for the fathead minnow, Pimephales promelas,
      larval survival and growth test	    80

   3. Weight data for the fathead minnow, Pimephales promelas,
      larval survival and growth test	    81

   4. Summary data for the fathead minnow, Pimephales promelas,
      larval survival and growth test	    82

   5. Flowchart for statistical analysis of fathead minnow,
      Pimephales promelas, larval survival data  	    85

   6. Plot of mean survival proportion data in Table 3	    87

   7. Plot of adjusted probits and predicted regression
      line from EPA Probit Program	    96

   8. Flowchart for statistical analysis of fathead
      minnow, Pimephales promelas, larval growth data   	    98

   9. Plot of mean weight data from fathead minnow, Pimephales
      promelas, larval survival and growth test  	    99

  10. Plot of raw data, observed means, and smoothed means for
      the fathead minnow, Pimephales promelas, growth data
      in Tables 2 and 19	   107

  11. BOOTSTRP program output for the IC25	   109

  12. BOOTSTRP program output for the IC50	   110

-------
                              FIGURES  (Continued)

 SECTION 12

 Number                                                              Page

   1.  Data form for the fathead minnow,  Pimephales promelas,
      embryo-larval survival and teratogenicity test.   Routine
      chemical  and physical determinations 	    125

   2.  Data form for the fathead minnow,  Pimephales promelas,
      embryo-larval survival and teratogenicity test.   Survival
      and terata data	    127

   3.  Summary data for the fathead minnow, Pimephales promelas,
      embryo-larval survival and teratogenicity test 	    129

   4.  Flowchart for statistical analysis of fathead minnow,
      Pimephales promelas, embryo-larval data  	    131

   5.  Plot of fathead minnow, Pimephales promelas, total  mortality
      data from the embryo-larval test	    134

   6.  Plot of adjusted probits and predicted regression line  ...    141

SECTION 13

 Number                                                              Page

   1.  Examples of a test board and randomizing template	    161

   2.  Data form for the daphnid, Ceriodaphm'a dubia, survival
      and reproduction test.  Routine chemical and physical
      determinations 	    168

   3.  Data form for the daphnid, Ceriodaphm'a dubia, survival
      and reproduction test.  Daily summary of data  	    170

   4.  Flowchart for statistical analysis of the daphnid,
      Ceriodaphm'a dubia, survival data	    173

   5.  Flowchart for statistical analysis of the daphnid,
      Ceriodaphm'a dubia, reproduction data   	    177

   6.  Plot of number of young per adult female from the daphnid,
      Ceriodaphm'a dubia, survival and reproduction test 	    178

   7.  Plot of raw data, observed means,  and smoothed means for
      the daphnid, Ceriodaphm'a dubia, reproductive data 	    189

   8.  Example of BOOTSTRP program output for  IC25	    191

-------
                              FIGURES (Continued)
SECTION 13 (Continued)
 Number                                                              Page
   9. Example of BOOTSTRP program output for IC50	   192
SECTION 14
 Number                                                              Page
   1. Data form for the green alga, Selenastrum capricornutum,
      growth test.  Routine chemical and physical
      determinations  	   213
   2. Data form for the green alga, Selenastrum capricornutum,
      growth test.  Cell density determinations  	  .   213
   3. Flowchart for statistical analysis of the green alga,
      Selenastrum capricornutum, growth response data  	   214
   4. Plot of Iog10  transformed  cell  count data  from  the  green
      alga, Selenastrum capricornutum, growth response test. .  .  .   217
   5. Plot of raw data, observed means, and smoothed means
      for the green alga, Selenastrum capricornutum,  growth
      data	   225
   6. BOOTSTRP program output for the IC25	   228
   7. BOOTSTRP program output for the IC50	   229
                                     xii

-------
                                  TABLES

SECTIONS 1-10

Number                                                              Page

   1. National interlaboratory study of chronic toxicity
      test precision, 1991:  Summary of responses using
      a reference toxicant 	    17

   2. Commercial sources of brine shrimp, Artemia, cysts 	    26

   3. Preparation of synthetic freshwater using reagent
      grade chemicals	    33

   4. Preparation of synthetic freshwater using
      mineral water	    34

   5. Percent un-ionized NH3 in aqueous ammonia solution:
      Temperatures 15 - 26°C and pH 6.0 - 8.9	    42

SECTION 11

Number                                                               Page

   1. Summary of test conditions and test acceptability
      criteria for fathead minnow, Pimephales promelas,
      larval survival and growth toxicity tests with
      effluents and receiving waters 	    76

   2. Summary of survival and growth data for fathead minnow,
      Pimephales promelas, larvae exposed to a reference
      toxicant for seven days	    83

   3. Fathead minnow, Pimephales promelas, survival  data 	    86

   4. Centered observations for Shapiro-Wilk's example 	    88

   5. Ordered centered observations for the
      Shapiro-Wilk's example 	    88

   6. Coefficients and differences for Shapiro-Wilk's example   .  .    89

   7. ANOVA table	    91

   8. ANOVA table for Dunnett's Procedure example  	    92

   9. Calculated t values	    93

  10. Data for Probit Analysis	    94
                                     xm

-------
                              TABLES (Continued)

SECTION 11 (Continued)

 Number                                                              Page
  11. Output from EPA Probit Analysis program for
      EC50 values, version 1.4	     95

  12. Fathead minnow, Pimephales promelas,  growth data 	     97

  13. Centered observations for Shapiro-Milk's example 	    100

  14. Ordered centered observations for Shapiro-Milk's example .  .    101

  15. Coefficients and differences for Shapiro-Milk's example  .  .    101

  16. ANOVA table	    103

  17. ANOVA table for Dunnett's Procedure example  	    104

  18. Calculated t values	    105

  19. Fathead minnow, Pimephales promelas,  mean growth response
      after smoothing	    108

  20. Precision of the fathead minnow, Pimephales promelas,
      larval survival and growth test, using NaPCP as a
      reference toxicant 	    112

  21. Combined frequency distribution for survival NOECs
      for all laboratories	    112

  22. Combined frequency distribution for weight NOECs
      for all laboratories	    113

SECTION 12

Number                                                               Page

   1. Summary of test conditions and test acceptability
      criteria for fathead minnow, Pimephales promelas,
      embryo-larval survival and teratogenicity toxicity
      tests with effluents and receiving waters	    123

   2. Data from fathead minnow, Pimephales promelas, embryo-larval
      toxicity test with trickling filter waste   	    132

   3. Fathead minnow, Pimephales promelas, embryo-larval total
      mortality data	    133

   4. ANOVA table	    135
                                      xw

-------
                              TABLES (Continued)
SECTION 12 (Continued)
 Number                                                              Page
   5. ANOVA table for Dunnett's Procedure example  	   137
   6. Calculated t values	   138
   7. Data for Probit Analysis	   139
   8. Output for EPA Probit Analysis program, version 1.4  ....   140
   9. Precision of the fathead minnow, Pimephales promelas,
      embryo-larval survival  and teratogenicity test, using
      cadmium as a reference toxicant  	   143
  10. Precision of the fathead minnow, Pimephales promelas,
      embryo-larval, survival and teratogenicity test, using
      diquat as a reference toxicant 	   144
  11. Precision of the fathead minnow, Pimephales promelas,
      embryo-larval survival  and teratogenicity static-renewal
      test conducted with trickling filter effluent  	   145
SECTION 13
 Number                                                              Page
   1. Nutrient stock solutions for maintaining algal  stock
      cultures and test control cultures 	   152
   2. Final concentration macronutrients and micronutrients
      in the culture medium	   153
   3. Summary of test conditions and test acceptability
      criteria for daphnid, Ceriodaphnia dubia, survival
      and reproduction toxicity tests with effluents
      and receiving waters 	   166
   4. Summary of survival and reproduction data for the
      daphnid, Ceriodaphnia dubia, exposed to an effluent
      for seven days	   174
   5. Format of the 2X2 contingency table	   174
   6. 2X2 contingency table fbr control and 25% effluent 	   175
   7. Data for Probit Analysis	   176
   8. The daphnid, Ceriodaphnia dubia, reproduction data 	   179
                                      xv

-------
                              TABLES (Continued)

SECTION 13 (Continued)

 Number                                                             Page

   9. Centered observations for Shapiro-Milk's example 	   180

  10. Ordered centered  observations for Shapiro-Milk's example .  .   181

  11. Coefficients and  differences for Shapiro-Milk's example. .  .   182

  12. ANOVA table	   184

  13. ANOVA table for Dunnett's Procedure example	   185

  14. Calculated t values	   186

  15. Daphnid, Cen'odaphm'a dubia, reproduction mean
      response after smoothing 	   188

  16. Single laboratory precision of the daphnid,  Cen'odaphm'a
      dubia, survival and reproduction test,  using NaPCP as a
      reference toxicant 	   194

  17. The daphnid, Cen'odaphm'a dubia, seven-day survival  and
      reproduction test precision for a single laboratory using
      NaPCP as the reference toxicant	   194

  18. Interlaboratory precision of the daphnid, Cen'odaphm'a
      dubia, survival and reproduction test with copper
      spiked effluent  	   195

  19. Interlaboratory precision data for the daphnid,
      Cen'odaphm'a dubia, summarized for eight reference
      toxicants and effluents	   196

SECTION 14

 Number                                                              Page

   1. Nutrient stock solutions for maintaining algal stock
      cultures and test control cultures 	   202

   2. Final concentration of macronutrients and micronutrients
      in the culture medium	   203

   3. Summary of test conditions and test acceptability
      criteria for green alga, Selenastrum capn'cornutum,
      growth toxicity tests with effluents and receiving
      waters	   211

                                      xv i

-------
                              TABLES (Continued)
SECTION 14 (Continued)
 Number                                                             Page
   4. Green alga, SeTenastrum capn'cornutum, growth
      response data	    215
   5. Centered observations for Shapiro-Milk's example 	    216
   6. Ordered centered observations for Shapiro-Wilk's example .  .    218
   7. Coefficients and differences for Shapiro-Wilk's example. .  .    219
   8. ANOVA table	    221
   9. ANOVA table for Dunnett's Procedure example  	    222
  10. Calculated t values	    223
  11. Algal growth means 	    226
  12. Single laboratory precision of the green alga,
      SeTenastrum capn'cornutum, 96-h toxicity tests,
      using the reference toxicant cadmium chloride	    230
                                     xvi i

-------
                             ACKNOWLEDGMENTS

      The principal  authors of this document are Philip A. Lewis, Donald J.
Klemm, Cornelius I.  Weber,  James M. Lazorchak, and Florence Fulk,
Environmental  Monitoring Systems Laboratory   Cincinnati; William H. Peltier,
Environmental  Services Division, Region 4, Athens, GA; Teresa Norberg-King,
Environmental  Research Laboratory,  Duluth, MM; and Cathy Poore, Computer
Science Corporation, Cincinnati, OH.   Contributors to specific sections of
this manual  are listed below.

1.  Sections 1-10;  General  Guidelines

    Teresa Norberg-King, ERL   Duluth
    Margarete Heber, OST, Office of Water
    Philip A.  Lewis, EMSL   Cincinnati
    Donald J.  Klemm, EMSL - Cincinnati
    Cornelius I. Weber, EMSL   Cincinnati
    William H. Peltier, ESD, Region 4

2.  Sections 11-13

    Teresa Norberg-King, ERL - Duluth
    Donald Mount, ERL - Duluth
    Philip A.  Lewis, EMSL - Cincinnati
    Donald J.  Klemm, EMSL   Cincinnati
    Quentin Pickering, EMSL - Cincinnati
    James M. Lazorchak, EMSL   Cincinnati

3.  Data Analysis (Sections 9, 11-13 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:

Celeste P. Barr, Environmental Services Division, Biology Section, U.S.
  Environmental Protection Agency,  Region 1, Lexington, MA.
Michael W. Tucker,  Bioassay Lab. Environmental Services Division, U.S.
  Environmental Protection Agency,  Region 7, Kansas City, MO.
Michael G. Morton,  Environmental Services Division, U.S. Environmental
  Protection Agency, Region 6, Dallas, TX.
Jerry Smrchek, Environmental Effects Branch, Health and Environmental Review
  Division,  U.S. Environmental Protection Agency, Washington, DC.
Robert Donaghy, Environmental Services Division, U.S. Environmental  Protection
  Agency, Wheeling,  WV.
Philip A. Crocker,  Water Quality Management Branch, U.S. Environmental
  Protection Agency, Dallas, TX.
Chick Steiner, Central Regional Laboratory, U.S. Environmental Protection
  Agency, Region 5,  Chicago, IL.
Tom Waller,  Institute of Applied Sciences, University of North Texas, Denton,
  TX.
                                     xvm

-------
                             ACKNOWLEDGMENTS (Continued)

      Many, very useful public comments on the second edition of the
freshwater toxicity test methods (EPA/600/4-89/001) were received in response
to the proposed rule, published in the Federal Register, December 4, 1989
[FR 54(231):50216-50224], regarding the Agency's intent to include the short-
term chronic toxicity tests in Table la, 40 CFR Part 136.  These comments were
considered in the preparation of this third edition of the freshwater test
methods manual (EPA/600/4-91/022), and are included in the Public Docket for
rulemaking, located in room 2904, Waterside Mall, EPA Headquarters,
Washington, DC.

      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, MN, 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, OH, 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, DC, (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, OH,
EPA-600/4-85/013  (USEPA, I985d); USEPA, 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, OH, EPA/600/4-89/001 (1989a); A Seven-day
Life-cycle Cladoceran Test,  Environ. Toxicol. Chem. 3:425-434 (Mount, D. I.
and T. J. Norberg, 1984); A New Fathead Minnow (Pimephales promelas)
Subchronic Toxicity Test, Environ. Toxicol. Chem. 4:711-718 (Norberg, T. J.,
and D. I. Mount, 1985a); The Selenastrum capricornutum Printz Algal Assay
Bottle Test, Environmental Research Laboratory, U. S. Environmental Protection
Agency, Corvallis, OR, EPA-600/9-78-018 (USEPA, 1978b); Short-term Methods for
Estimating the Chronic Toxicity of Effluents and Receiving Waters to Marine
and Estuarine Organisms, Environmental Monitoring and Support Laboratory
Cincinnati, U. S. Environmental Protection Agency, Cincinnati, OH,
EPA-600/4-87/028  (USEPA, 1988c); Short-term Methods for Estimating the Chronic
Toxicity of Effluents and Receiving Waters to Marine and Estuarine Organisms,
Environmental Monitoring Systems Laboratory   Cincinnati, U.S. Environmental
Protection Agency, Cincinnati, OH, EPA/600/4-91/021 (USEPA, 1992); Methods for
Measuring the Acute Toxicity of Effluents and Receiving Waters to Freshwater
and Marine Organisms  (Fourth Edition), Environmental Monitoring Systems
Laboratory - Cincinnati, U.S. Environmental Protection Agency, Cincinnati, OH,
EPA/600/4-90/027  (USEPA, 1991b); and Technical Support Document for Water
Quality-Based Toxics Control, Office of Water Enforcement and Permits, Office
of Water Regulations and Standards, U.S. Environmental Protection Agency,
Washington, DC, EPA/505/2-90-001 (USEPA, 1991c).

      The fathead minnow, Pimephales promelas, larval survival and growth test
method and the daphnid, Ceriodaphnia dubia, survival and reproduction test
method in this manual were adapted from methods developed at the Environmental

                                      xix

-------
                        ACKNOWLEDGMENTS (Continued)

Research Laboratory   Duluth by Donald Mount and Teresa Norberg-King.  The
fathead minnow, Pimephales promelas, embryo-larval survival and teratogenicity
test method was developed by Wesley Birge and Jeffrey Black, University of
Kentucky, Lexington, under a cooperative agreement with the Environmental
Monitoring and Support Laboratory   Cincinnati.  The algal growth test method
was adapted from the green alga, Selenastrum capricornutum, algal assay bottle
test developed at the Environmental Research Laboratory   Corvallis by William
E. Miller, Joseph C. Greene, and Tamotsu Shiroyama.

      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.
                                    xx

-------
                                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, IC50 or IC25 (see Section 9, Chronic Toxicity 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 were 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, 1988a; USEPA, 1988d; USEPA, 1989b; USEPA, 1989c;
USEPA, 1991a; USEPA, 1991c and USEPA, 1991d).

1.4  This methods manual serves as a companion to the acute toxicity test
methods for freshwater and marine organisms (USEPA, 1991b), the short-term
chronic toxicity test methods for marine and estuarine organisms (USEPA,
1992), and the manual for evaluation of laboratories performing aquatic
toxicity tests (USEPA, 1991a).

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, 1991c).

1.6  These freshwater short-term toxicity tests are similar to those developed
for marine and estuarine organisms to evaluate the toxicity of effluents
discharged to marine and estuarine waters under the NPDES permit program.
Methods are presented in this manual for three species of freshwater organisms
from three phylogenetic groups.  The methods are all  static renewal type
seven-day tests except the green alga, Selenastrum capn'cornutum, test which
lasts four days.

-------
1.7  The three species for which test methods are provided are the fathead
minnow, Pimephales promelas-,  the daphnid,  Ceriodaphm'a dubiz; and the green
alga, Selenastrum capricornutum.

1.7.1  The tests included in  this document are based on the following methods:

    1. "A new fathead minnow  (Pimephales promelas) subchronic toxicity test,"
       by Teresa J. Norberg and Donald I.  Mount,  1985, Environmental
       Toxicology and Chemistry (Norberg and Mount,  1985a).

    2. "In-situ acute/chronic toxicological  monitoring of industrial effluents
       for the NPDES biomonitoring program using  fish and amphibian
       embryo/larval stages as a test organism,"  by Wesley J. Birge and
       Jeffrey A. Black, 1981, OWEP-82-001,  Office of Water Enforcement and
       Permits, U.S. Environmental Protection Agency, Washington, DC (USEPA,
       1981).

    3. "A seven-day life-cycle cladoceran test,", by Donald I. Mount and
       Teresa Norberg, 1984,  Environmental Toxicology and Chemistry (Mount and
       Norberg, 1984).

    4. "The Selenastrum capricornutum Printz algal assay bottle test," by
       William E. Miller, Joseph C. Greene and Tamotsu Shiroyama, 1978,
       Environmental Research Laboratory,  U.S. Environmental Protection
       Agency, Con/all is, OR.  EPA/600/9-78/018 (USEPA, 1978b).

1.7.2  Two of the methods incorporate the chronic endpoint of growth in
addition to lethality and one incorporates reproduction.  The fathead minnow,
Pimephales promelas, embryo-larval survival  and teratogenicity test
incorporates teratogenic effects in addition to lethality.  The green alga,
Selenastrum capricornutum, growth test has the advantage of a relatively short
exposure period (96 h).

1.8  The validity of the freshwater chronic  methods in predicting adverse
ecological impacts of toxic discharges was demonstrated in field studies
(USEPA, 1985e; USEPA, 1985f;  USEPA, 1985g; USEPA, 1986b; USEPA,  1986c; USEPA,
1986d; USEPA, 1986e; Birge et al., 1989; and Eagleson et al., 1990).

1.9  These methods are restricted to use by, or under the supervision of,
analysts experienced in the use or conduct of aquatic toxicity tests and the
interpretation of data from aquatic toxicity testing.  Each analyst must
demonstrate the ability to generate acceptable test results with the methods
using the procedures described in this methods manual.

1.10  This manual was prepared in the established EMSL-Cincinnati format
(USEPA, 1983).

-------
                                  SECTION 2

             SHORT-TERM 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
promeTas, 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 ot 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 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 consensus that the ELS  test data usually would be adequate for
estimating chronically safe concentrations, there was a rapid shift by aquatic
toxicologists to 30-90 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, Weltering (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, 1981; Birge et al., 1985),
and a seven-day larval survival  and growth test (Norberg and Mount, 1985a).

-------
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, 1984),  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  (1985a) 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  Mount and Norberg  (1984) developed a seven-day cladoceran partial
life-cycle test and experimented with a number of diets for use in culturing
and testing the daphnid, Ceriodaphnia reticulata (Norberg and Mount, 1985b).
As different laboratories began to use this cladoceran test, it was discovered
that apparently more than one species was involved in the tests conducted by
the same laboratory.  Berner (1986) studied the problem and determined that
perhaps as many as three variant forms were involved and it was decided to
recommend the use of the more common Ceriodaphnia dubia rather than the
originally reported Ceriodaphnia reticulata.  The method was adopted for use
in the first edition of the freshwater short-term chronic methods (USEPA,
1985c).

2.1.15  The green alga, Selenastrum capricornutum, bottle test was developed,
after extensive design, evaluation, and application, for the National
Eutrophication Research Program (USEPA, 1971).  The test was later modified
for use in the assessment of receiving waters and the effects of wastes
originating from industrial, municipal, and agricultural point and non-point
sources (USEPA, 1978b).

2.1.16  The use of short-term toxicity tests including subchronic and chronic
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 21-day daphnid, Daphnia magna, 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 (four to  seven days).  The
results of the tests are expressed  in terms of the highest  concentration that
has no statistically significant observed effect on those responses when
compared to 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
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 where applicable.

2.3  STATIC TESTS

2.3.1  Static non-renewal 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,
            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  non-renewal  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:

       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.
                                      8

-------
                                 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.
This program should include (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 each of 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 lack of oxygen or presence of noxious gases.

3.1.3  Prior to sample collection and laboratory work, personnel  will
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 should 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  or other
means 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
rules pertaining to the handling of hazardous materials (see Safety
Adapted from USEPA (1985c)  and USEPA (1991b).

                                       9

-------
Manuals, Paragraph 3.5).  It is recommended that personnel collecting
samples and performing toxicity tests 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 are to 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 industrial safety manuals, including USEPA (1986f) 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, the toxicity testing activities.  Local fire officials
should be notified of any potentially hazardous conditions.

                                      10

-------
                                  SECTION 4

                             QUALITY ASSURANCE1
4.1  INTRODUCTION
4.1.1  Development and maintenance of a toxicity test laboratory quality
assurance (QA) program (USEPA, 1991a) 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 a written description 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 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 test 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
1991a; DeWoskin, 1984; and Taylor, 1987.

4.1.4  Guidance for the evaluation of laboratories performing toxicity tests
and laboratory evaluation criteria may be found in USEPA (1991a).

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 culturing or testing areas, and from testing 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
contact with the effluent (see Section 5, Facilities, Equipment and Supplies,
Adapted from USEPA (1978a),  USEPA (1985d),  and USEPA (1991b).

                                      11

-------
and the specific toxicity test method).

4.3  TEST ORGANISMS

4.3.1  The test organisms used in the procedures described in this manual are
the fathead minnow, Pimephales promelas, the daphnid, Ceriodaphnia dubia,
and the .green alga, Selenastrum capricornutum.  The fish and invertebrates
should appear healthy, behave normally, feed well, and have low mortality in
the cultures, during holding, and in test controls.  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 in part on the objectives of the study and
logistical constraints, as discussed in detail in Section 7, Dilution Water.
For tests performed to meet NPDES objectives, synthetic, moderately hard water
should be used.  The dilution water used for internal quality assurance tests
with organisms, food, and reference toxicants should be the water routinely
used with success in the laboratory.  Types of water are discussed in Section
5, Facilities, Equipment and Supplies.  Water used for culturing and test
dilution should be analyzed at least quarterly for toxic metals and organics.
The concentration of the metals Al, As, Cr, Co, Cu, Fe, Pb, Ni, and Zn,
expressed as total metal, should not exceed 1 /ug/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 concentration 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, Sample Handling, and Sample
Preparation for Toxicity Tests.

4.6  TEST CONDITIONS

4.6.1  Water temperature 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 test vessel for the duration of each test.  DO
concentration and pH should be checked at the beginning of each test and daily
throughout the test period.
                                      12

-------
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
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 Subsections 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.  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 cysts, 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 a 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 M9/9 wet weight, or the total concentration of organochlorine
pesticides plus PCBs exceeds 0.30 ng/g wet weight, or toxic metal exceed 20
/xg/g wet weight, the food should not be used (for analytical methods see
USEPA, 1979b and 1982b).
                                      13

-------
4.9  ACCEPTABILITY OF SHORT-TERM CHRONIC TOXICITY TESTS

4.9.1  To be acceptable, control survival in fathead minnow, Pimephales
promelas, and the daphnid, Ceriodaphm'a dubia, tests must be at least 80%.  At
the end of the test, the average dry weight of seven-day-old fathead minnows
in the controls must equal or exceed 0.30 mg.  In the Cen'odaphnia dubia
controls., at least 80% of animals must have a brood, at least 60% of the
animals should have produced their third brood in 7 days, and the number of
young per surviving adult must be 20 or greater.   In algal toxicity tests, the
mean cell density in the controls after 96 h must equal or exceed 2 X 10
cells/ml and not vary more than 20% among replicates.

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 condition
summaries).   The acceptability of the test would 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 the 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 and USEPA, 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, and conductivity, must be calibrated
and standardized according to instrument manufacturer's 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 EPA methods (see EPA 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

                                      14

-------
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; and (5) 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 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 data from chronic toxicity tests with two
species using the reference toxicants potassium chloride and copper sulfate
are shown 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
Technical Support Document for Water Quality-Based Control (see pp. 2-4, and
11-15 in USEPA, 1991c).

4.14.5  In cases where the test data are used in Probit Analysis (see  Section
9) 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.  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 ± (NOEC-LOEC).

4.14.6  It should be noted here that the dilution factor selected for a test

                                      15

-------
TABLE 1.   NATIONAL INTERLABORATORY STUDY OF CHRONIC TOXICITY TEST PRECISION,
          1991:   SUMMARY OF RESPONSES USING A REFERENCE TOXICANT1
 Organism
Endpoint
No. Labs    % Effluent'
                                                              SD
CV(%)
Pimephales
promelas


Ceriodaphm'a
dubia


Survival, NOEC
Growth, IC25
Growth, IC50
Growth, NOEC
Survival, NOEC
Reproduction, IC25
Reproduction, IC50
Reproduction, NOEC
146
124
117
142
162
155
150
156
NA
4.67
6.36
NA
NA
2.69
3.99
NA
NA
1.87
2.04
NA
NA
1.96
2.35
NA
NA
40.0
32.1
NA
NA
72.9
58.9
NA
   1From a national study of inter!aboratory precision of toxicity test data
   performed in 1991 by the Environmental  Monitoring Systems Laboratory -
   Cincinnati, U.S. Environmental Protection Agency, Cincinnati, OH 45268.
   Participants included Federal, state, and private laboratories engaged in
   NPDES permit compliance monitoring.

   Expressed as % effluent; in reality it was a reference toxicant (KC1) but
   was not known by the persons conducting the tests.
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 variability as high as
± 300%.  With a dilution factor of 0.5, the NOEC could be considered to have a
relative variability of ± 100%.  Therefore, USEPA recommends a dilution factor
of 0.5 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 the 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 the same data analysis

                                      16

-------
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
toxicity test method 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 for each reference-toxicant-
organism combination, and successive toxicity endpoints (NOECs, LC50s, IC25s,
ICBOs, etc.) 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 (XJ  from successive tests with a given
reference toxicant.  The types of control charts illustrated (see USEPA,
1979a) are used to evaluate the cumulative trend of results from a series of
samples.  For Probit Analysis results (such as LCSOs), the mean (X) and upper
and lower control limits (+ 2S) are recalculated with each successive point,
until the statistics stabilize.  Precision may vary with the test species,
reference toxicant, and type of test.  Five or six tests are adequate for
establishing the control charts.

4.16.3  Outliers, which are values that fall 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
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.
                                      17

-------
  o
  UJ
  o
            UPPER CONTROL LIMIT



                CENTRALTENDENCY
           LOWER CONTROL LIMIT



       I !  I I  I !  I  I  I I  I I  I I  I  I !!  I I
                     10       15       20
           UPPER CONTROL LIMIT(X+2S)       B

  o     "
                CENTRALTENDENCY
           LOWER CONTROL LI MIT (X - 2S)


      !  I I  I I  I  I  I I  I I  I I  I I I  I I  i  1
     0       5       10       15       20

    TOXICITY TEST WITH REFERENCE TOXICANTS
           _
               n
                          "  \2
                            r,)/fi
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).

                           18

-------
4.17  REFERENCE TOXICANTS

4.17.1  Reference toxicants such as Nad, KC1, cadmium, copper, SDS, and
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 should be
maintained for each individual toxicity test or group of tests on closely
related samples.  This file should 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
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 Administration, Customer Services Section, 8700 Edgeworth Dr.,
Capitol Heights, MD 20743-3701.  Charge to VISA or Master Card by calling
toll-free (800) 683-1300 or FAX your purchase order by dialing (301) 763-6025.
For other information call (301) 763-1896.

4.19.2  The complete EPA Test Methods for Freshwater Effluent Toxicity Tests
video package includes a 23-minute VMS tape and a 26-page report on the
Daphnid, Ceriodaphnia dubia,  Survival and Reproduction Toxicity Tests, a 15-
minute VMS tape and a 18-page report of Fathead Minnow Larval Survival and
Growth Toxicity Tests, and a  249-page USEPA Manual titled "Short-term Methods
for Estimating the Chronic Toxicity of Effluents and Receiving Waters to
Freshwater Organisms."  The package may be purchased for $45.00.


                                      19

-------
4.20  SUPPLEMENTAL REPORTS FOR TRAINING VIDEO TAPES

4.20.1  Additional materials that may be ordered from the National Archives
Trust Fund Board at the above address include the following:

   1. "Culturing of Ceriodaphm'a dubia", a video training tape and
      supplemental report (EPA/505/8-89-002a).
   2. "Culturing of Fathead Minnows,  Pimephales promelas", a video training
      tape and supplemental  report (EPA/505/89-002b).
   3. "Ceriodaphm'a dubia Survival and Reproduction Tests" - a video training
      tape and supplemental  report (EPA/505/8-89-001a).
   4. "Fathead Minnow Larval Growth and Survival  Toxicity Tests" - a video
      training tape and supplemental  report (EPA/505/8-89-001b).
                                     20

-------
                                 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 holding and acclimating
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 ground water, receiving water, dechlorinated tap water, or synthetic
water.  Dechlorination can be accomplished by carbon filtration, or the use of
sodium thiosulfate.  Use of 3.6 mg (anhydrous) sodium thiosulfate/L will
reduce 1.0 mg chlorine/L.  After dechlorination, total residual chlorine
should be nondetectable.  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 BAISTON" 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 from fumes.  Laboratory
ventilation systems should be checked to ensure that return air from chemistry
laboratories and/or sample holding 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 test 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., that come in
contact with the effluent and dilution water should be carefully chosen.
Tempered glass and perfluorocarbon plastics (TEFLON")  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 to ship, store and transfer effluents and receiving waters,
but they should not be reused unless absolutely necessary, because they could
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 for test chambers.  The
use of large (z 20 L) glass carboys is discouraged for safety reasons.
Adapted from USEPA (1985d)  and USEPA (1991b).

                                      21

-------
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 and
stainless steel, 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.  It is usually sufficient to
rinse them twice with deionized water and once with sample dilution water
before use.   New glassware should be soaked overnight in 10% acid (see below)
and rinsed well in deionized water and dilution water.

5.3.2  All nondisposable sample containers, test vessels, tanks, and other
equipment that have come in contact with effluent must be washed after use to
remove contaminants as  described below.

    1. Soak  15 min 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) hydrochloric or
       nitric acid to remove scale, metals and bases.  To prepare a 10%
       solution of acid, 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.
                                      22

-------
5.3.3  Special requirements for cleaning glassware used in the green alga,
Selenastrum capricornutum, toxicity tests (Section 14).  Prepare all graduated
cylinders, test flasks, bottles, volumetric flasks, centrifuge tubes and vials
used in algal bioassays as follows:

    1.  Wash with nonphosphate detergent solution, preferably heated to
        50°C or hotter.  Brush the inside of flasks with a stiff-bristle brush
        to loosen any  attached material.  The use of a commercial laboratory
        glassware washer or heavy-duty kitchen dishwasher (under-counter type)
        is highly recommended.

    2.  Rinse with tap water.

    3.  Test flasks should be thoroughly rinsed with acetone and a 10%
        solution (by volume) of reagent grade hydrochloric acid  (HC1).  It may
        be advantageous to soak the flasks in 10% HC1 for several days.  Fill
        vials and centrifuge tubes with the 10% HC1 solution and allow to
        stand a few minutes; fill all larger containers to about one-tenth
        capacity with  HC1 solution and swirl so that the entire  surface is
        bathed.

    4.  Rinse twice with MILLI-QR, or equivalent,  water.

    5.  New test flasks, and all flasks which through use may become
        contaminated with toxic organic substances, must be rinsed with
        pesticide-grade acetone or heat-treated before use.  To  thermally
        degrade organics, place glassware in a high temperature  oven at 400°C
        for 30 min.  After cooling, go to 7.  If acetone is used, go to 6.

    6.  Rinse thoroughly with MILLI-QR,  or equivalent,  water,  and dry in an
        105°C oven.  All glassware should be autoclaved before use and between
        uses.

    7.  Cover the mouth of each chamber with aluminum foil or other closure,
        as appropriate, before storing.

5.3.4  The use of sterile, disposable pipets will eliminate the  need for pipet
washing and minimize the possibility of contaminating the cultures with toxic
substances.

5.3.5  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 CULTURIN6 AND TOXICITY TESTS

5.4.1  Apparatus and equipment requirements for culturing and testing are
specified in each toxicity test method.  Also, see USEPA, 1991b.

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  in sufficient capacity

                                      23

-------
for laboratory needs.  Deionized water may be obtained from MILLIPORER, Milli-
QR,  or equivalent systems.   If large quantities of high quality deionized
water are needed, it may be advisable to supply the laboratory grade 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-QR 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

      1.  Brine shrimp, Artemia sp., cysts -- A list of commercial
          sources is listed in Table 2.

      2.  Frozen adult brine shrimp -- Available from most pet supply  shops or
          from San Francisco Bay Brand,  8239 Enterprise Dr.,  Newark, CA 94560
          (415-792-7200).

      3.  Flake fish food -- TETRAMINR and BIORILR  are  available from most pet
          shops.

      4.  Trout chow -- Available from Zeigler Bros.,  P.O. Box 95, Gardners,
          PA 17324 (717-677-6181 or 800-841-6800); Glencoe Mills,  1011 Elliott
          St., Glencoe, MN 55336 (612-864-3181); or Murray Elevators,  118 West
          4800 South, Murray, UT 84107 (800-521-9092).

      5.  CEROPHYLL" -- Available from Ward's  Natural  Science Establishment,
          Inc., P.O. Box 92912, Rochester, NY 14692-9012 (716-359-2502) or as
          cereal leaves from Sigma Chemical  Company, P.O.  Box 14508, St.
          Louis, MO 63178 (800-325-3010).

      6.  Yeast - Available from Lake States Yeast, Rynland,  WI.

      7.  Alfalfa Rabbit Pellets -- Available from feed stores as Purina
          rabbit chow.

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).

5.5.2  Reagents and consumable materials are specified in each toxicity test
method section (see Section 4, Quality Assurance).

5.6  TEST ORGANISMS

5.6.1  Test organisms should be obtained from inhouse cultures or from
commercial  suppliers (see specific test  method; Section 4, Quality Assurance;
and Section 6,  Test Organisms).
                                      24

-------
5.7  SUPPLIES
5.7.1  See test methods (Sections 11-14) for specific supplies.
                                       25

-------
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, other]

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
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
(Columbia)
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  -
IH 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
Telephone 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)
  From  D.  A.  Bengston,  University of Rhode Island,  Narragansett,  RI.

                                      26

-------
                                    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 the toxicity tests must be identified to species.  If there is any doubt as
to the identity of the test organism, representative specimens should be sent
to a taxonomic expert to confirm the identification.

6.1.2  Toxicity test conditions and culture methods for the species listed in
Subsection 6.1.3 are provided in this manual  and in USEPA, 1991b.

6.1.3  The organisms used in the short-term chronic toxicity tests described
in this manual are the fathead minnow, Pimephales promelas, the daphnid,
Cen'odaphm'a dubia (Berner, 1986), and the green alga,  Selenastrum
capricornutum.

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 this manual.  However, USEPA allows the  use of indigenous
species only where state regulations require their use  or  prohibit importation
of the recommended species.  Where state regulations prohibit importation of
non-native fishes or the use of recommended test species,  permission must be
requested from the appropriate state agency prior to their use.

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 the one or more 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,
1991c).

      1.  Where the salinity of the receiving water is  < l°/oo, freshwater
          organisms are used regardless of the salinity of the effluent.


                                      27

-------
      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  SOURCE OF TEST ORGANISMS

6.2.1  The test organisms recommended in this manual  can be cultured in the
laboratory using culturing and handling methods described in the respective
test sections.  The fathead minnow, Pimephales promelas, culture method is
given in  Section 11 and  not repeated in Section 12.  Also,  see USEPA, 1991b.

6.2.2  Inhouse cultures  should be established wherever it is cost effective.
If inhouse cultures cannot be maintained or it is not cost  effective, test
organisms or starter cultures should be purchased from experienced commercial
suppliers (see USEPA, 1991b).

6.2.3  Starter cultures  of the green algae, Selenastrum capricornutum, S.
minutum,  and Chlamydomonas reinhardti are available from the following
sources:

      1.   American Type  Culture Collection (Culture No.  ATCC 22662), 12301
          Parklawn Drive, Rockville, MD 10852.

      2.   Culture Collection of Algae,  Botany Department, University of Texas,
          Austin, TX 78712.

6.2.4  Because the daphnid, Ceriodaphm'a dubia, must  be cultured individually
in the laboratory for at least seven days before the  test begins, it will  be
necessary to obtain a starter culture from a commercial  source at least three
weeks before the test is to begin if they are not being cultured inhouse.

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 test methods.

6.2.6  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 a logical  approach.  However,  it is generally
impractical and not recommended 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 collecting 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.  The fathead minnow, Pimephales  promelas, the
          daphnid, Ceriodaphm'a dubia,  and the green  alga,  Selenastrum
          capricornutum, are easily cultured in the laboratory or readily

                                      28

-------
          available commercially.
      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 the test organism be
          known  to species level,  it would be necessary to examine each
          organism caught in the wild to confirm its identity.   This 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 caught by electroshocking must not be
          used in toxicity testing.

6.2.6.2  Guidelines for collecting naturally occurring organisms are provided
in USEPA (1973 and 1990).

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 can be obtained from
commercial stock 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 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 the respective test methods (also, see USEPA,
1991b).

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 2 units of pH in any 24-h period.

6.5.2  The 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 a dry surface or are
injured during handling must be discarded.  Wide-bore, smooth glass tubes (4
to 8 mm inside diameter) with rubber bulbs or pipettors (such as PROPIPETTER)
should be used for transferring the organisms.


                                      29

-------
6.5.3  Holding tanks for fish are supplied with good quality water (see
Section 5) with flow-through rate of at least two tank volumes per day.
Otherwise use a recirculation system where 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 must be avoided because it will stress the organisms and lower
the DO concentrations  to unacceptable levels.  The DO must be maintained at a
minimum of 4.0 mg/L.  Aerate gently if necessary.

6.5.5  The organisms should be observed carefully each day for signs of
disease, stress,  physical damage, or mortality.  Dead and abnormal organisms
should be removed as soon as observed.  It is not uncommon for some fish
mortality (5-10%) to occur during the first 48 h in a holding tank because of
individuals that  refuse to feed on artificial food and die of starvation.
Organisms in the  holding tanks should generally be fed as in the cultures (see
culturing methods in the respective methods).

6.5.6  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 TETRAMINR or BIORILR
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.

6.5.7  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 or by use of
an airstone supplied by a portable pump.  The DO concentration must not fall
below 4.0 mg/L.

6.6.2  Upon arrival at the test site, the 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-minute period with dilution
water.  If receiving water is to be used as the 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 may be transferred
to the dilution water.
                                      30

-------
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 the receiving water is used as
dilution 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.

6.7  TEST ORGANISM DISPOSAL

6.7.1  When the toxicity test is concluded, all test organisms (including
controls) should be humanely destroyed and disposed of in an  appropriate
manner.
                                      31

-------
                                  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 the primary objective of NPDES permit-related toxicity
testing,  a synthetic (standard)  dilution water (moderately hard 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 either upstream and outside the influence of the
outfall,  or with other uncontaminated natural  water or standard dilution water
having approximately the same  characteristics (hardness, alkalinity, and
conductivity) as the receiving water.  Seasonal variations in the quality of
receiving waters may affect effluent toxicity.   Therefore, the pH, alkalinity,
hardness, and conductivity of  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 immediately upstream or outside the influence  off 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 or mineral  water (Tables 3 and 4).  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-QR,  or equivalent,
system.  It is advisable to provide a preconditioned (deionized) feed water by
using a CulliganR,  Continental", or equivalent  system  in  front of  the  MILLI-Q
System to extend the life of the MILLI-QR 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,

                                      32

-------
(3) carbon, and (4) organic cleanup (such as ui\uAiitA-q", or equivalent)
followed by a final bacteria filter.

7.2.3  Standard, Synthetic Freshwater

7.2.3.1  To prepare 20 L of synthetic, moderately hard, reconstituted water,
use the reagent grade chemicals in Table 3 as follows:
    1.

    2.
    3.
    4.
Place 19 L of MILLI-QR,  or equivalent,  water in a properly cleaned
plastic carboy.
Add 1.20 g of MgS04,  1.92 g NaHC03,  and  O.OSOg  KC1  to  the  carboy.
Aerate overnight.
Add 1.20 g of CaS04-2H20  to 1  L  of MILU-QR or  equivalent deionized
water in a separate flask.  Stir on magnetic stirrer until calcium
sulfate is dissolved, add to the 19 L above, and mix well.
For Ceriodaphnia dubia culturing and testing,  add sufficient sodium
selenate (Na2Se04)  to  provide  2  /ig selenium  per liter  of final  dilution
water.
Aerate the combined solution vigorously for an additional  24 h to
dissolve the added chemicals and stabilize the medium.
The measured pH, hardness, etc., should be as  listed in Table 3.
TABLE 3.  PREPARATION OF SYNTHETIC FRESHWATER USING REAGENT GRADE CHEMICALS3
Reagent Added (mq/L)D
Final Water Quality
Water
Type
Very
Soft

soft

Moderately
Hard
Very

hard
NaHCO3
12.0
48.0
Hard 96.0
192.0
384.0
CaS04-
7
30
60
120
240
2H20
.5
.0
.0
.0
.0
MgS04
7.5
30.0
60.0
120.0
240.0
KC1
0.
2.
4.
8.
16.

5
0
0
0
0
PHC
6.4-6
7.2-7
7.4-7
7.6-8
8.0-8

.8
.6
.8
.0
.4
Hardnessd
10-13
40-48
80-100
160-180
280-320
Alka-
linity6
10-13
30-35
60-70
110-120
225-245
aTaken in part from Marking and Dawson (1973).
bAdd reagent grade chemicals to deionized water.
Approximate equilibrium pH after 24 h of aeration.
dExpressed as mg CaCOj/L .
7.2.3.2  If large volumes of synthetic reconstituted water will be needed, it
may be advisable to mix 1 L portions of concentrated stock solutions of
NaHC03,  MgS04,  and  KC1  for  use  in  preparation  of  the  reconstituted  water.

7.2.3.3  To prepare 20 L of standard, synthetic,  moderately hard,
reconstituted water, using mineral water such as PERRIERR Water, or equivalent
(Table 4),  follow the instructions below.

                                      33

-------
    1. Place 16 L of MILLI-QR or equivalent water in a properly cleaned
       plastic carboy.
    2. Add 4 L of PERRIER* Water,  or equivalent.
    3. Aerate vigorously for 24 h to stabilize the medium.
    4. The measured pH, hardness and alkalinity of the aerated water will be
       as indicated in Table 4.
    5. This synthetic water is referred to as diluted mineral water (DMW) in
       the toxicity test methods.
      TABLE 4.   PREPARATION OF  SYNTHETIC  FRESHWATER USING  MINERAL WATER2
Water
Type
Volume of
Mineral Water
Added (mL/L)b
Proportion
of Mineral
Water (%)
Final Water Oual itv
Alka-
pHc Hardnessd linityd
Very soft           50
Soft               100
Moderately Hard    200
Hard               400
Very hard6         —
 2.5
10.0
20.0
40.0
7.2-8.1
7.9-8.3
7.9-8.3
7.9-8.3
 10-13
 40-48
 80-100
160-180
 10-13
 30-35
 60-70
110-120
      Mount et al.  (1987),  and  data  provided  by  Philip  Lewis,  EMSL-Cincinnati.
 Add mineral  water  to Milli-QR water, or equivalent, to prepare Diluted
 Mineral Water (DMW).
Approximate  equilibrium pH after  24 h  of  aeration.
 Expressed as mg, CaCOj/L.

Dilutions of PERRIER" Water form  a  precipitate when concentrations equivalent
 to "very hard water" are aerated.
7.3  USE OF RECEIVING WATER AS DILUTION WATER

7.3.1  If the objectives of the test require the use of uncontaminated surface
water as dilution water, and the receiving water is uncontaminated, it may be
possible to collect a sample of the receiving water upstream of, or close to,
but outside of the zone influenced by the outfall.   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,
the sample should be chilled to 4 C during  or  immediately following
collection, and maintained at that temperature prior to use  in the test.
                                      34

-------
7.3.3  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 \im mesh openings prior to use.

7.3.4  Where toxicity-free dilution water is required in a test, the water is
considered acceptable if test organisms show the required survival, growth,
and reproduction in the controls during the test.

7.3.5  The regulatory authority may require that the hardness of the dilution
water be comparable to the receiving water at the discharge site.  This
requirement can be satisfied by collecting an uncontaminated receiving water
with a suitable hardness, or adjusting the hardness of an otherwise suitable
receiving water by addition of reagents as indicated in Table 3.

7.4  USE OF TAP WATER AS DILUTION WATER

7.4.1  The use of tap water as dilution water is discouraged unless it is
dechlorinated and passed through a deionizer and carbon filter.  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, p. 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.

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.
                                      35

-------
                                   SECTION 8

          EFFLUENT AND RECEIVING WATER SAMPLING,  SAMPLE HANDLING,
                AND SAMPLE PREPARATION FOR TOXICITY TESTS1
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,
1991c).

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
1Adapted  from  USEPA  (1989a)  and  USEPA  (1991b).

                                      36

-------
       not masked by dilution.

    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 a spike 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 a single
       grab sample 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, except for the green alga, Selenastrum
capricornutum, test which is not renewed.

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 volume 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 a continuously discharged effluent is less than 14 days and the
       variability of the waste 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.


                                      37

-------
    2.  If the calculated retention time of a continuously discharged effluent
       is greater than 14 days,  or if it can be demonstrated that the
       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 single 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, at the end the shift, clean up  activities result in the
       discharge of a slug of toxic wastes (one sample is collected).

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, the  test is performed using a single grab sample
of water consisting of receiving water collected daily upstream from or (in
lakes)  away from the influence of the outfall.

8.4.2  The sampling point is determined by the objectives of the test.  In
rivers, samples should be collected from mid-stream and at mid-depth, if
accessible.  In lakes the samples are collected at mid-depth.

8.4.3  To determine the extent of the zone of toxicity in the receiving water
downstream from the outfall,  receiving water samples are collected at several
distances downstream from the discharge.  The time required for the effluent-
receiving-water mixture to travel to sampling points downstream from the
outfall,  and the rate and degree of mixing,  may be difficult to ascertain.
Therefore, it may not be possible to correlate downstream toxicity with
effluent toxicity at the discharge point unless a dye study is performed.   The
toxicity of receiving water samples from five stations downstream from the
discharge point can be evaluated using the same number of test vessels and


                                      38

-------
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, they should be chilled and maintained 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.

8.5.3  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 must 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 regulatory 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.  Insulating material must not be placed
between the ice and the sample in the shipping container.

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

                                      39

-------
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 multiconcentration
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 should be held under the same conditions as during
shipment and holding.

8.8  PREPARATION OF EFFLUENT AND RECEIVING WATER SAMPLES FOR TOXICITY TESTS

8.8.1  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  is used),
or by using an appropriate discharge valve (spigot).

8.8.2   With the daphnid, Ceriodaphnia dubia, and fathead minnow, Pimephales
promelas, tests, effluents and receiving waters must be filtered through a
60-jim plankton net to remove indigenous organisms that may attack or be
confused with the test organisms (see  the daphnid, Ceriodaphnia dubia, test
method for details).  Receiving waters used in green alga, Selenastrum
capricornutum, toxicity tests must be  filtered through a 0.45-nm pore diameter
filter before use.  It may be necessary to first coarse-filter the dilution
and/or wastewater through a nylon sieve having 2-to 4-mm holes to remove
debris and/or break up large floating  or suspended solids.  Because filtration
may increase the DO in the effluent, the DO should be checked both before and
after filtering.  Caution: filtration  may remove some toxicity.

8.8.3  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
with a DO probe after reaching test temperature and, if the DO is greater than

                                      40

-------
100% saturation or lower than 4.0 mg/l, the solutions are aerated moderately
(approximately 500 mL/min) for a few minutes, using an airstone, until the DO
is within the prescribed range (>4.0 mg/L).  Caution:  avoid excessive
aeration.

8.8.4  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.

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 pH.  However, the DO
in the test solutions must not be allowed to fall below 4.0 mg/L.

8.8.4.2  In static tests (renewal or non-renewal), low DOs may commonly occur
in the higher concentrations of wastewater.  Aeration is accomplished by
bubbling air through a pipet at a 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, 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, receiving 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 (see Subsection 7.4.1).  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  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

                                      41

-------
      TABLE 5.   PERCENT UN-IONIZED  NH,  IN  AQUEOUS AMMONIA SOLUTION:
                TEMPERATURES  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
6.3
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
8.3
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
0.0427
0.0537
0.0676
0.0851
0.107
0.135
0.170
0.214
0.269
0.338
0.425
0.535
0.672
0.845
1.061
1.33
1.67
2.10
2.62
3.28
4.10
5.10
6.34
7.85
9.69
11.90
14.5
17.6
21.2
25.3
22 23
0.0459 0.0493
0.0578 0.0621
0.0727 0.0781
0.0915 0.0983
0.115 0.124
0.145 0.156
0.182 0.196
0.230 0.247
0.289 0.310
0.363 0.390
0.457 0.491
0.575 0.617
0.722 0.776
0.908 0.975
1.140 1.224
1.43 1.54
1.80 1.93
2.25 2.41
2.82 3.02
3.52 3.77
4.39 4.70
5.46 5.85
6.78 7.25
8.39 8.96
10.3 11.0
12.7 13.5
15.5 16.4
18.7 19.8
22.5 23.7
26.7 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.
8.32
10.9
13.3
16.2
19.5
23.4
27.8
32.6

'Table provided by Teresa Norberg-King, Environmental Research Laboratory,
 Duluth, MN. See also Emerson et al. (1975), Thurston et al. (1974),  and USEPA (1985b).


(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 by
adding IN NaOH  or IN HN03 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.
                                       42

-------
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,
IC50, or LC50, or (2) a no-observed-adverse-effect concentration (NOEC)
defined in terms of mortality, growth, reproduction, and/or teratogenicity and
obtained by hypothesis testing.  The tests may be static renewal or static
nonrenewal.

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%.

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 hardness of the control should be comparable
to the receiving water.

8.12.2  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
acceptability criteria in the specific test method).


                                      43

-------
8.12.3  In cases where the objective of the test is to estimate the degree of
toxicity of the receiving water,  a multiconcentration test is performed by
preparing dilutions of the receiving water, using a * 0.5 dilution series,
with a suitable control  water.
                                     44

-------
                                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 subchronic 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 quantal, "all
or nothing," response (such as death, immobilization, or serious
incapacitation) in a given percent of the 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

                                      45

-------
the observed 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 non-quantal 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, EC, 1C, 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

                                      46

-------
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 effluents
and reference toxicant data sets analyzed.  The data sets included short-term
chronic toxicity tests for the fathead minnow, Pimephales promelas, and the
daphnid, Cen'odaphnia dubia.  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 NOECs for several organisms.

                                      47

-------
Similarly,  Norberg-King (USEPA,  1988)  reported that the IC25s were comparable
to the NOECs for a set of daphnid,  Cen'odaphnia 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
(twofold  difference between adjacent  concentrations), would be fourfold.

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

                                      48

-------
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 estimation 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 of the different toxicity test data sets for which they are recommended,
(2) powerful statistical tests, (3) hopefully "easily" understood by
non-statisticians, 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.
Test sensitivity generally increases as the number of replicates is increased,
but the point of diminishing returns in sensitivity may be reached rather

                                      49

-------
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 test 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 the 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.
                                      50

-------
9.5.3  Analysis of Growth and Reproduction Data

9.5.3.1  Growth data from the fathead minnow, Pimephales promelas, larval
survival and growth test are analyzed using hypothesis testing or point
estimation techniques 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.  Data from effluent concentrations that have
tested significantly different from the control for survival are excluded from
further hypothesis tests concerning growth effects.  When analyzing the data
using point estimation techniques, data from all concentrations are included
in the analysis.

9.5.3.2  Reproduction data from the daphnid, Ceriodaphm'a dubia, survival and
reproduction test are analyzed using hypothesis testing or point estimation
techniques according to the flowchart in Figure 2.  In hypothesis testing,
data from effluent concentrations that have significantly lower survival than
the control, as determined by Fisher's Exact test, are not included in the
hypothesis tests for growth effects.  Data from all concentrations are
included when using point estimation techniques.

9.5.4  Analysis of Algal Growth Response Data

9.5.4.1  The growth response data from the green alga, Selenastrum
capricornutum, toxicity test, after an appropriate transformation, if
necessary, to meet the assumptions of normality and homogeneity of variance,
may be analyzed by hypothesis testing according to the flowchart in Figure 2.
Point estimates, such as the IC25 and IC50, would also be appropriate in
analyzing algal growth data.

9.5.5  Analysis of Mortality Data

9.5.5.1  Mortality data from the fathead minnow, Pimephales promelas, larval
survival and growth test and the fathead minnow, Pimephales promelas,
embryo-larval survival and teratogenicity 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
Appendix B), according to the flowchart in Figure 2.

9.5.5.2  Mortality data from the daphnid, Ceriodaphm'a dubia, survival and
reproduction test are analyzed by Fisher's Exact Test (Appendix G) prior to
the analysis of the reproduction data.  The mortality data may also be
analyzed by Probit Analysis, if appropriate.

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

                                      51

-------
                 DATA (SURVIVAL, GROWTH, REPRODUCTION, ETC.)
                         HYPOTHESIS TESTING
                                T
                           TRANSFORMATION?
  ENDPOINT ESTIMATE
      LC, EC, 1C
T
                                      TEST
              NORMAL DISTRIBUTION
               NON-NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
                             BARTLETT'S TEST
                                              HETEROGENEOUS
                                                 VARIANCE
                     "
NO STATISTICAL ANALYSIS
RECOMMENDED

NU
YE
4 OR MORE
REPLICATES?
:s

             t
EQUAL NUMBER OF
REPLICATES?
YES
1
'
EQUAL NUMBER OF
REPLICATES?
1
YES
!

ADJUSTMENT



DUNNETT'S
TEST

STEEL'S MANY-ONE
RANK TEST




WILCOXON RANK SUM
TEST WITH
BONFERRONI ADJUSTMENT


                                  T
                            ENDPOINT ESTIMATES
                                NOEC.LOEC
   Figure  2.   Flowchart  for statistical  analysis of  test data.

                                  52

-------
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
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  A 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.

9.6.2.2  The assumptions upon which the use of the t-test with Bonferroni's
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
the t-test with Bonferroni's 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
for 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

                                      53

-------
efficient (Hodges and Lehmann,  1956).

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
treatments 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
concentration), 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 calculate the 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  (see USEPA, 1985d; 1991b).  If the test results in 100% survival and
100% mortality in adjacent treatments (all or nothing effect), a LC50 may be
estimated using the Graphical method.

                                      54

-------
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 I.  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 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 nonincreasing (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.
The assumption for piecewise 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 piecewise 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

                                      55

-------
with caution when the results of a toxicity test approach an "all  or nothing1
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 J for a more detailed discussion of the use of this method and a
computer program available for the calculations.
                                     56

-------
                               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 tests are 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. Collection dates and times
         c. Sample collection method
         d. Physical and chemical data
         e. Mean daily discharge on sample collection date
         f. Lapsed time from sample collection to delivery
         g. Sample temperature when received at the laboratory

      2. Receiving Water Samples
         a. Sampling point
         b. Collection dates and times
         c. Sample collection method
         d. Physical and chemical data
         e. Streamflow (at 7Q10 and at time of sampling)
         f. Sample temperature when received at the laboratory
         g, Lapsed time from sample collection to delivery
Adapted  in  part  from USEPA (1985d)  and  USEPA (1991b).

                                      57

-------
    3.  Dilution Water Samples
         a.  Source
         b.  Collection date(s)  and time(s)
         c.  Pretreatment
         d.  Physical  and chemical  characteristics

10.4  TEST METHODS

      1.  Toxicity test method used (title,  number,  source)
      2.  Endpoint(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 and volume of test chambers
      7.  Volume of solution used per chamber
      8.  Number of organisms per test chamber
      9.  Number of replicate test chambers  per treatment
     10.  Acclimation  of test organisms (temperature mean and range)
     11.  Test temperature (mean and range)
     12.  Specify if aeration was needed
     13.  Feeding frequency, and amount and  type of food

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  and 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 LCBOs, NOECs, IC25, IC50, etc.
      3.  Indicate statistical methods used  to calculate endpoints
      4.  Provide summary table  of physical  and chemical data
      5.  Tabulate QA  data
                                      58

-------
10.8  CONCLUSIONS AND RECOMMENDATIONS

      1.  Relationship between test endpoints and permit limits
      2.  Actions to be taken
                                      59

-------
                                 SECTION 11

                                 TEST METHOD

     FATHEAD MINNOW,  PIMEPHALES PROHELAS, LARVAL SURVIVAL AND GROWTH TEST
                                METHOD 1000.0
1.  SCOPE AND APPLICATION

1.1  This method estimates the chronic toxicity of effluents and
receiving water to the fathead minnow, Pimephales promelas,  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 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 LC50s).

1.3  Detection limits of the toxicity of an effluent or pure substance are
organism dependent.

1.4  Brief 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
degradable or highly volatile toxicants present in the  source may not be
detected in  the test.

1.5  This test 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  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 (final dry weight) of the larvae.

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
concentrations of suspended and/or dissolved  solids, and extremes of pH,
alkalinity,  or hardness, may mask the presence of toxic substances.

3.3  Improper effluent sampling and sample handling may adversely affect test
results (see Section 8, Effluent and Receiving Water Sampling, Sample

                                      60

-------
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.  Daily renewal of solutions, however,
will reduce the probability of reduction of toxicity caused by feeding.

4.  SAFETY

4.1  See Section 3, Health and Safety.

5.  APPARATUS AND EQUIPMENT

5.1  Fathead minnow and brine shrimp culture units -- see USEPA, 1985d and
USEPA, 1991b.  This test requires 180-360 larvae.  It is preferable to obtain
larvae from an in-house fathead minnow culture unit.  If it is not feasible to
culture fish in-house, embryos or newly hatched larvae can be shipped in well
oxygenated water in insulated containers.

5.2  Samplers -- automatic sampler, preferably with sample cooling
capability, that can collect a 24-h composite sample of 2 L or more.

5.3  Sample containers -- for sample shipment and storage (see Section 8,
Effluent and Receiving Water Sampling, Sample Handling,  and Sample Preparation
for Toxicity Tests).

5.4  Environmental chamber or equivalent facility with temperature control
(25 ± 1°C).

5.5  Water purification system -- MILLIPORE MILLI-QR deionized water or
equivalent (see Section 5, Facilities, Equipment, and Supplies).

5.6  Balance -- analytical, capable of accurately weighing larvae to
0.00001 g.

5.7  Reference weights, Class S -- for checking performance of balance.
Weights should bracket the expected weights of the weighing pans and the
expected weights of the pans plus fish.

5.8  Test chambers -- four (minimum of three) borosilicate glass or nontoxic
disposable plastic test chambers are required for each concentration and
control.  Test chambers may be 1L, 500 mL or 250 ml beakers, 500 mL plastic
cups, or fabricated rectangular (0.3 cm thick) glass chambers, 15 cm x 7.5 cm
x 7.5 cm.  To avoid potential contamination from the air and excessive
evaporation of test solutions during the test, the chambers should be covered
with safety glass plates or sheet plastic (6 mm thick).

5.9  Volumetric flasks and graduated cylinders -- Class A, borosilicate glass
or nontoxic plastic labware, 10-1000 mL for making test solutions.


                                      61

-------
5.10  Volumetric pipets-- Class A, 1-100 ml.
5.11  Serological pipets-- 1-10 ml, graduated.
5.12  Pipet bulbs and fillers -- PROPIPETR,  or equivalent.
5.13  Droppers, and glass tubing with fire polished edges,  4 mm ID -- for
transferring larvae.
5.14  Wash bottles -- for rinsing small  glassware and instrument electrodes
and probes.
5.15  Glass or electronic thermometers -- for measuring water temperatures.
5.16  Bulb-thermograph or electronic-chart type thermometers -- for
continuously recording temperature.
5.17  National Bureau of Standards Certified thermometer (see USEPA Method
170.1, USEPA 1979b).
5.18  Meters, pH, DO, and specific conductivity -- for routine physical and
chemical measurements.
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 recording data.
6.3  Vials, marked -- for preserving larvae (Optional).
6.4  Weighing boats, aluminum -- for weighing larvae.
6.5  Tape, colored -- for labeling test chambers.
6.6  Markers, water-proof -- for marking containers, etc.
6.7  Reagents for hardness and alkalinity tests (see USEPA Methods 130.2 and
310.1, USEPA 1979b).
6.8  Standard pH buffers 4, 7, and 10 (or as per instructions of instrument
manufacturer) for instrument calibration (see USEPA Method 150.1, USEPA
1979b).
6.9  Specific conductivity standards (see USEPA Method 120.1, USEPA 1979b).
6.10  Membranes and filling solutions for DO probe (see Method 360.1, USEPA,
1979b), or reagents for modified Winkler analysis.
                                      62

-------
6.11  Laboratory quality control samples and standards --  for calibration of
the above methods.

6.12  Reference toxicant solutions (see Section 4, Quality Assurance).

6.13  Ethanol (70%)  or formalin (4%) -- for use as a preservative for the
fish larvae.

6.14  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).

6.15  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  BRINE SHRIMP, ARTEMIA, NAUPLII (see USEPA, 1991b).

6.16.1  Newly-hatched Artemia nauplii are used as food for fathead minnow,
Pimephales promelas, larvae in toxicity tests and frozen brine shrimp and
flake food are used 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 (see Section 4, Quality Assurance, 4.8, Food Quality).

6.16.2  Each new batch of brine shrimp, Artemia, cysts must be evaluated for
size  (Vanhaecke and Sorgeloos, 1980, and Vanhaecke et al., 1980) and
nutritional suitability (see Leger et al., 1985; Leger et al., 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 Branch, Environmental Monitoring
Systems Laboratory, Cincinnati, Ohio.  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 ^g/g wet weight or the total concentration of
organochlorine pesticides plus PCBs exceeds 0.30 \ig/g wet weight.  (For
analytical methods see USEPA, 1982b).

6.16.3  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,
        1991b and ASTM designation E1203-87, for details on Artemia culture
        and quality control).

                                      63

-------
    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 jum Nitex
        or stainless steel  screen,  and rinse with seawater, or equivalent,
        before use.

6.16.4  Testing Artemia nauplii as  food for toxicity test organisms.

6.16.4.1  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 fathead minnow larvae (see Subsection 12.  ACCEPTABILITY OF
TEST RESULTS).  The fish 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.

6.16.4.2  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.

6.16.4.3  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.

6.16.5  Limited quantities  of reference Artemia cysts, information on
commercial sources of good  quality  brine shrimp, Artemia, cysts, and
procedures for determining  cyst suitability are available from the Quality
Assurance Research Division,  Environmental Monitoring Systems Laboratory,
U. S. Environmental Protection Agency, Cincinnati,  Ohio, 45268.

6.17  TEST ORGANISMS,  FATHEAD MINNOWS, PIMEPHALES PROMELAS

6.17.1  Newly hatched fish  less than 24 h  old should be used for the test.  If
organisms must be shipped to the testing site,  fish up to 48 h old may be
used, all hatched within a  24-h window.

6.17.2  If the fish are kept in a holding  tank  or container, most of the water
should be siphoned off to concentrate the  fish.  The fish are then transferred
one at a time randomly to the test  chambers until each chamber contains ten
fish.  Alternately, fish may be placed one or two at a time into small beakers
or plastic containers until they each contain five fish.  Three (minimum of
two) of these beakers/plastic containers are then assigned to randomly-
arranged control and exposure chambers.
                                      64

-------
6.17.2.1  The fish are transferred directly to the test vessels or
intermediate beakers/plastic containers, using a large-bore, fire-polished
glass tube (6 mm to 9 mm I.D. X 30 cm long) equipped with a rubber bulb, or a
large volumetric pipet with tip removed and fitted with a safety type bulb
filler.  The glass or plastic containers should only contain a small volume of
dilution water.

6.17.2.2  It is important to note that larvae should not be handled with a
dip net.  Dipping small fish with a net may result in damage to the fish
and cause mortality.

6.17.3  The test is conducted with four (minimum of three) test chambers at
each toxicant concentration and control.  Fifteen (minimum of ten) embryos are
placed in each replicate test chamber.  Thus 60 (minimum of 30) fish are
exposed at each test concentration.

6.17.4  Sources of organisms

6.17.4.1  Fathead minnows, Pimephales promelas, may be obtained from
commercial biological supply houses.  Fish obtained from outside sources for
use as brood stock or in toxicity tests may not always be of suitable age and
quality.  Fish provided by supply houses should be guaranteed to be of (1) the
correct species, (2) disease free, (3) in the requested age range, and (4) in
good condition.  This can be done by providing the record of the date on which
the eggs were laid and hatched, and information on the sensitivity of
contemporary fish to reference toxicants.

6.17.5  Inhouse Sources of Fathead Minnows, Pimephales promelas^

6.17.5.1  Problems in obtaining suitable fish from outside laboratories can be
avoided by developing an inhouse laboratory culture facility.  Fathead
minnows, Pimephales promelas, can be easily cultured in the laboratory from
eggs to adults in static, recirculating, or flow-through systems.   The larvae,
juveniles, and adult fish should be kept in 60 L (15 gal) or 76 L (20 gal)
rearing tanks supplied with reconstituted water, dechlorinated tap water, or
natural water.  The water should be analyzed for toxic metals and organics
quarterly (see Section 4, Quality Assurance).

6.17.5.1.1  If a static or recirculating system is used, it is necessary to
equip each tank with an outside activated carbon filter system, similar to
those sold for tropical fish hobbyists (or one large activated carbon filter
system for a series of tanks) to prevent the accumulation of toxic metabolic
wastes (principally nitrite and ammonia) in the water.

6.17.5.2  Flow-through systems require large volumes of water and may not be
feasible in some laboratories.  The culture tanks should be shielded from
extraneous disturbances using opaque curtains, and should be isolated from
toxicity testing activities to prevent contamination.
'Adapted  from USEPA (1991b).

                                      65

-------
6.17.5.3  To avoid the possibility of inbreeding of the inhouse brood stock,
fish from an outside source should be introduced yearly into the culture unit.

6.17.5.4  Dissolved oxygen -- The DO concentration in the culture tanks should
be maintained near saturation,  using gentle aeration with 15 cm air stones if
necessary.  Brungs (1971), in a carefully controlled long-term study, found
that the. growth of fathead minnows was reduced significantly at all dissolved
oxygen concentrations below 7.9 mg/L.  Soderberg (1982) presented an
analytical approach to the reaeration of flowing water for culture systems.

6.17.5.5  Culture Maintenance

6.17.5.5.1  Adequate procedures for culture maintenance must be followed to
avoid poor water quality in the culture system.   The spawning and brood stock
culture tanks should be kept free of debris (excess food, detritus, waste,
etc.) by siphoning the accumulated materials (such as dead brine shrimp
nauplii or cysts) from the bottom of the tanks daily with a glass siphon tube
attached to a plastic hose leading to the floor drain.  The tanks are more
thoroughly cleaned as required.  Algae, mostly diatoms and green algae,
growing on the glass of the spawning tanks are left in place, except for the
front of the tank, which is kept clean for observation.  To avoid excessive
build up of algal growth,  the walls of the tanks are periodically scraped.
The larval culture tanks are cleaned once or twice a week to reduce the mass
of fungus growing on the bottom of the tank.

6.17.5.5.2  Activated charcoal  and floss in the tank filtration systems should
be changed weekly, or more often if needed.  Culture water may be maintained
by preparation of reconstituted water or use of dechlorinated tap water.
Distilled or deionized water is added as needed to compensate for evaporation.

6.17.5.5.3  Before new fish are placed in tanks, salt deposits are removed by
scraping or with 5% acid solution, the tanks are washed with detergent,
sterilized with a hypochlorite solution, and rinsed well with hot tap water
and then with laboratory water.

6.17.5.6  Obtaining Embryos for Toxicity Tests

6.17.5.6.1  Embryos can be shipped to the laboratory from an outside source or
obtained from adults held in the laboratory as described below.

6.17.5.6.2  For breeding tanks, it is convenient to use 60 L (15 gal) or 76 L
(20 gal) aquaria.  The spawning unit is designed to simulate conditions in
nature conducive to spawning, such as water temperature and photoperiod.
Spawning tanks must be held at a temperature of 25 ± 2°C.   Each aquarium is
equipped with a heater, if necessary, a continuous filtering unit, and
spawning substrates.  The photoperiod for the culture system should be
maintained at 16 h light and 8 h darkness.  For the spawning tanks, this
photoperiod must be rigidly controlled.  A convenient photoperiod is 5:00 AM
to 9:00 PM.  Fluorescent lights should be suspended about 60 cm above the
surface of the water in the brood and larval tanks.  Both DURATESTR and cool-
white fluorescent lamps have been used, and produce similar results.  An
illumination level of 50 to 100 ft-c is adequate.

                                      66

-------
6.17.5.6.3  To simulate the natural spawning environment, it is necessary to
provide substrates (nesting territories) upon which the eggs can be deposited
and fertilized, and which are defended and cared for by the males.  The
recommended spawning substrates consist of inverted half-cylinders, 7.6 cm X
7.6 cm (3 in. X 3 in.) of schedule 40, PVC pipe.  The substrates should be
placed equidistant from each other on the bottom of the tanks.

6.17.5.6.4  To establish a breeding unit, 15-20 prespawning adults six to
eight months old are taken from a "holding" or culture tank and placed in a
76-L spawning tank.  At this point, it is not possible to distinguish the
sexes.  However, after less than a week in the spawning tank, the breeding
males will develop their distinct coloration and territorial behavior, and
spawning will begin.  As the breeding males are identified, all but two are
removed, providing a final ratio of 5-6 females per male.  The excess spawning
substrates are used as shelter by the females.

6.17.5.6.5  Sexing of the fish to ensure a correct female/male ratio in each
tank can be a problem.  However, the task usually becomes easier as experience
is gained (Flickinger, 1966).  Sexually mature females usually have large
bellies and a tapered snout.  The sexually mature males are usually
distinguished by their larger overall size, dark vertical color bands, and the
spongy nuptial tubercles on the snout.  Unless the males exhibit these
secondary breeding characteristics, no reliable method has been found to
distinguish them from females.  However, using the coloration of the males and
the presence of enlarged urogenital structures and other characteristics of
the females, the correct selection of the sexes can usually be achieved by
trial and error.

6.17.5.6.6  Sexually immature males are usually recognized by their aggressive
behavior and partial banding.  These undeveloped males must be removed from
the spawning tanks because they will eat the eggs and constantly harass the
mature males, tiring them and reducing the fecundity of the breeding unit.
Therefore, the fish in the spawning tanks must be carefully checked
periodically for extra males.

6.17.5.6.7  A breeding unit should remain in their spawning tank about four
months.  Thus, each brood tank or unit is stocked with new spawners about
three times a year.  However, the restocking process is rotated so that at any
one time the spawning tanks contain different age groups of brood fish.

6.17.5.6.8  Fathead minnows spawn mostly in the early morning hours.  They
should not be disturbed except for a morning feeding (8:00 AM) and daily
examination of substrates for eggs in late morning or early afternoon.  In
nature, the male protects, cleans, and aerates the eggs until they hatch.  In
the laboratory, however, it is necessary to remove the eggs from the tanks to
prevent them from being eaten by the adults, for ease of handling for purposes
of recording embryo count and hatchability, and for the use of the newly
hatched young fish for toxicity tests.

6.17.5.6.9  Daily, beginning six to eight hours after the lights are turned  on
(11:OOAM   1:OOPM), the substrates in the spawning tanks are each lifted
carefully and inspected for embryos.  Substrates without embryos are

                                      67

-------
immediately returned to the spawning tank.   Those with embryos are immersed in
clean water in a collecting tray,  and replaced with a clean substrate.  A
daily record is maintained of each spawning site and the estimated number of
embryos on the substrate.

6.17.5.6.10  Three different methods are described for embryo incubation.

      1.  Incubation of Embryos on  the Substrates:  Several  (2-4) substrates
      are placed on end in a circular pattern (with the embryos on the inner
      side) in 10 cm of water in a tray.  The tray is then  placed in a
      constant temperature water bath,  and  the embryos are  aerated with a 2.5
      cm airstone placed in the center of the circle.  The  embryos are
      examined daily,  and  the dead and fungused embryos are counted,
      recorded, and removed with forceps.   At an incubation temperature of
      25°C,  50% hatch  occurs  in  five days.   At  22°C embryos incubated  on
      aerated tiles require 7 days for 50% hatch.

      2.  Incubation of Embryos in  a Separatorv Funnel:  The embryos are
      removed from the substrates  with a rolling action of  the index finger
      ("rolled off")(Gast  and Brungs, 1973), their total volume is measured,
      and the number of embryos is calculated using a conversion factor of
      approximately 430 embryos/mL.  The embryos are incubated in about 1.5 L
      of water in a 2 L separatory funnel  maintained in a water bath.   The
      embryos are stirred  in the separatory funnel by bubbling air from the
      tip of a plastic micro-pipette placed at the bottom,  inside the
      separatory funnel.  During the first  two days,  the embryos are taken
      from the funnel  daily,  those that are dead and fungused are removed, and
      those that are alive are returned to  the separatory funnel in clean
      water.  The embryos  hatch in four days at a temperature of 25°C.
      However, usually on  day three the eyed embryos are removed from  the
      separatory funnel and placed in water in a plastic tray and gently
      aerated with an air  stone.  Using this method,  the embryos hatch in five
      days.  Hatching time is greatly influenced by the amount of agitation of
      the embryos and the  incubation temperature.  If on day  three the embryos
      are transferred from the separatory funnel to a static, unaerated
      container, a 50% hatch will  occur in  six days (instead  of five)  and a
      100% hatch will  occur in seven days.   If the culture  system is operated
      at 22°C,  embryos incubated on aerated tiles require seven days for 50%
      hatch.

      3.  Incubation in Embryo Incubation Cups:   The embryos are "rolled off"
      the substrates,  and  the total number  is estimated by  determining the
      volume.  The embryos are then placed  in incubation cups attached to a
      rocker arm assembly  (Mount,  1968).  Both flow-through and static renewal
      incubation have been used.  On day one, the embryos are removed  from the
      cups and those that  are dead and fungused are removed.   After day one
      only dead embryos are removed from the cups.  During  the incubation
      period, the eggs are examined daily for viability and fungal growth,
      until they hatch.  Unfertilized eggs, and eggs that have become  infected
      by fungus, should be removed with forceps using a table top magnifier-
      illuminator.  Nonviable eggs become milky and opaque, and are easily
      recognized.  The nonviable eggs are very susceptible  to fungal

                                      68

-------
      infection, which may then spread throughout the egg mass.  Removal of
      fungus should be done quickly, and the substrates should be returned
      to the incubation tanks as rapidly as possible so that the good eggs are
      not damaged by desiccation.  Hatching takes 4-5 days at an optimal
      temperature of 25°C.   Hatching can be delayed several  (two to four) days
      by incubating at lower temperatures.  A large plastic tank receiving
      recirculating water from a temperature control unit, can be used as a
      water bath for incubation of embryos.

6.17.5.6.11  Newly-hatched larvae are transferred daily from the egg
incubation apparatus to small rearing tanks, using a large bore pipette, until
the hatch is complete.  New rearing tanks are set up on a daily basis to
separate fish by age group.  Approximately 1500 newly hatched larvae are
placed in a 60-L (15 gal) or 76-L (20 gal) all-glass aquarium for 30 days.  A
density of 150 fry per liter is suitable for the first four weeks.  The water
temperature in the rearing tanks is allowed to follow ambient laboratory
temperatures of 20-25°C,  but sudden, extreme variations in temperature must be
avoided.

6.17.5.7  Food and Feeding

6.17.5.7.1  The amount of food and feeding schedule affects both growth and
egg production.  The spawning fish and pre-spawners in holding tanks usually
are fed all the adult frozen brine shrimp and tropical fish flake food or dry
commercial fish food (No. 1 or No. 2 granules) that they can eat (ad libitum)
at the beginning of the work day and in the late afternoon (8:00 AM and 4:00
PM).  The fish are fed twice a day  (twice a day with dry food and once a day
with adult shrimp) during the week and once a day on weekends.

6.17.5.7.2  Fathead minnow larvae are fed freshly-hatched brine shrimp
(Artemia) nauplii twice daily until they are four weeks old.  Utilization of
older (larger) brine shrimp nauplii may result in starvation of the young fish
because they are unable to ingest the larger food organisms (see Subsection
6.16 or USEPA, 1991b Appendix A.5 for instructions on the preparation of brine
shrimp nauplii).

6.17.5.7.3  Fish older than four weeks are fed frozen brine shrimp and
commercial fish starter (#1 and #2), which is ground fish meal enriched with
vitamins.  As the fish grow, larger pellet sizes are used, as appropriate.
(Starter, No. 1 and N. 2 granules, U.S. Fish and Wildlife Service
Formulation Specification Diet SD9-30, can be obtained from Zeigler Bros.,
Inc., P.O. Box 90, Gardners, PA 17324).  Newly hatched brine shrimp nauplii,
and frozen adult brine shrimp (San Francisco Bay Brand) are fed to the fish
cultures in volumes based on age, size, and number of fish in the tanks.

6.17.5.7.4  Fish in the larval tanks (from hatch to 30 days old) are fed
commercial starter fish food at the beginning and end of the work day, and
newly hatched brine shrimp nauplii (from the brine shrimp culture unit) once a
day, usually midmorning and midafternoon.

6.17.5.7.5  Attempts should be made to avoid introducing Artemia cysts and
empty shells when the brine shrimp nauplii are fed to the fish larvae.  Some

                                      69

-------
of the mortality of the larval  fish observed in cultures could be caused from
the ingestion of these materials.

6.17.5.8  Disease Control

6.17.5.8.1  Fish are observed daily for abnormal  appearance or behavior.
Bacterial  or fungal infections  are the most common diseases encountered.
However, if normal  precautions  are taken,  disease outbreaks will rarely, if
ever,  occur.  Hoffman and  Mitchell (1980)  have put together a list of some
chemicals  that have been used commonly for fish diseases and pests.

6.17.5.8.2  In aquatic culture  systems where filtration is utilized, the
application of certain antibacterial  agents should be used with caution.  A
treatment  with a single dose of antibacterial  drugs can interrupt nitrate
reduction  and stop nitrification for various periods of time, resulting in
changes in pH, and in ammonia,  nitrite and nitrate concentrations (Collins
et al., 1976).  These changes could cause  the death of the culture
organisms.

6.17.5.8.3  Do not transfer equipment from one tank to another without first
disinfecting tanks and nets.  If an outbreak of disease occurs, any
equipment, such as nets, airlines, tanks,  etc., which has been exposed to
diseased fish should be disinfected with sodium hypochlorite.  Also to
avoid  the  contamination of cultures or spread of disease, each time nets are
used to remove live or dead fish from tanks, they are first sterilized with
sodium hypochlorite or formalin, and rinsed in hot tap water.  Before a new
lot of fish is transferred to culture tanks, the tanks are cleaned and
sterilized as described above.

6.17.5.9  It is recommended that chronic toxicity tests be performed monthly
with a reference toxicant.  Newly hatched  fathead minnow larvae less than 24 h
old are used to monitor the chronic toxicity of the reference toxicant to the
test fish  produced by the  culture unit.  See Section 4, Quality Assurance,
Subsection 4.17, Reference Toxicants.

6.17.5.9  Record Keeping

6.17.5.9.1  Records, kept  in a  bound notebook, include: (1) type of food and
time of feeding for all fish tanks; (2) time of examination of the tiles
for embryos, the estimated number of embryos on the tile, and the tile
position number; (3) estimated  number of dead embryos and embryos with
fungus observed during the embryonic development stages; (4) source of all
fish;  (5)  daily observation of the condition and behavior of the fish; and (6)
dates  and  results of reference  toxicant tests performed (see Section 4,
Quality Assurance).

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.
                                      70

-------
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.
Receiving water toxicity is determined with samples used directly as collected
or after samples are passed through a 60 /urn NITEXR filter and compared without
dilution, against a control.  Using four replicate chambers per test, each
containing 250 ml, and 400 mL for chemical analyses, would require
approximately 1.5 L or more of sample per test per day.

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 the dilution factor is increased beyond 0.5, and
declines rapidly if a smaller dilution factor is 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%).  If a high rate of mortality is observed during the first
1  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 250 ml of test solution, is approximately
2.5 L.  Sufficient test solution (approximately 1500 ml) is prepared at each
effluent concentration to provide 400 mL additional volume for chemical
analyses at the high, medium, and low test concentrations.  If the sample is
used for more than one daily renewal of test solutions, the volume must be
increased proportionately.

10.1.2.4  Tests 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 for the first time in a test more than 72 h after sample
collection.

                                      71

-------
10.1.2.5  Just prior to test initiation (approximately one h) the temperature
of sufficient quantity of the sample to make the test solutions should be
adjusted to the test temperature and maintained at that temperature during the
addition of dilution water.

10.1.2.6  The DO of the test solutions should be checked prior to the test
initiation.  If any of the solutions are supersaturated with oxygen or any
solution has a DO concentration below 4.0 mg/L, all  of the solutions and the
control must be gently aerated.

10.1.3  Dilution Water

10.1.3.1  Dilution water may be uncontaminated receiving water, a standard
synthetic (reconstituted) water, or some other uncontaminated natural water
(see Section 7, Dilution Water).

10.2  START Cr THE TEST

10.2.1  Label the test chambers with a marking pen.   Use of color-coded tape
to identify each treatment and replicate is helpful.   A minimum of five
effluent concentrations and a control are used for each effluent test.  Each
treatment (including the control) should have four (minimum of three)
replicates.

10.2.2  Tests performed in laboratories that have in-house fathead minnow
breeding cultures should use larvae less than 24 h old.  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 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.

10.2.3  Randomize the position of 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.

10.2.4  The larvae are pooled and placed one or two at a time into each
randomly arranged test chamber or intermediate container in sequential order,
until each chamber contains 15 (minimum of 10) larvae, for a total of 60
larvae (minimum of 30) for each concentration (see Appendix A).  The test
organisms should come from a pool of larvae consisting of at least three
separate spawnings.  The amount of water added to the chambers when
transferring the larvae to the compartments should be kept to a minimum to
avoid unnecessary dilution of the test concentrations.

10.2.4.1  The chambers may be placed on a light table to facilitate counting
the larvae.

10.3  LIGHT, PHOTOPERIOD, AND TEMPERATURE

10.3.1  The light quality and intensity should be at  ambient laboratory
levels, which is approximately 10-20 /iE/m/s,  or 50 to 100 foot candles

                                      72

-------
(ft-c), with a photoperiod of 16 h of light and 8 h of darkness.  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 satisfactory DO concentrations.  The DO
concentrations should be measured in the new solutions at the start of the
test (Day 0) and before daily renewal of the test solutions on subsequent
days.  The DO concentrations 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 concentrations 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 equivalent.   Care should be
taken to ensure that turbulence resulting from aeration does not cause undue
physical stress to the fish.

10.5  FEEDING

10.5.1  The fish in each test chamber are fed 0.1 g (approximately 700 to
1000) of a concentrated suspension of newly hatched (less than 24-h old) brine
shrimp nauplii three times daily at 4-h intervals or, as a minimum, 0.15 g are
fed twice daily at an interval of 6 h.  Equal amounts of nauplii must be added
to each replicate chamber to reduce variability in larval weight.  Sufficient
numbers of nauplii should be provided to assure that some remain alive in the
test chambers at the next feeding, but not in excessive amounts which will
result in depletion of DO below acceptable levels (below 4.0 mg/L).

10.5.2  The feeding schedule will depend on when the test solutions are
renewed.  If the test is initiated after 12:00 PM, the larvae may be fed only
once the first day.  On following days, the larvae normally would be fed at
the beginning of the work day, at least 2 h before test solution renewal, and
at the end of the work day, after test solution renewal.  However, if the test
solutions are changed at the beginning of the work day, the first feeding
would be after test solution renewal in the morning, and the remaining
feeding(s) would be at the appropriate intervals.  The larvae are not fed
during the final 12 h of the test.

10.5.3  The nauplii should be rinsed with freshwater to remove salinity  before
use (see USEPA, 1991b).  At feeding time pipette about 5 mL (5 g) of
concentrated newly hatched brine shrimp nauplii into a 120 mesh nylon net or
plastic cup with nylon mesh bottom.  Slowly run freshwater through the net or
rinse by immersing the cup in a container of fresh water several times.
Resuspend the brine shrimp in 10 mL of fresh water in a 30 mL beaker or  simply
set the cup of washed brine shrimp in 1/4 inch of fresh water so that the cup
contains about 10 mL of water.  Allow the container to set for a minute  or two
to allow dead nauplii and empty cysts to settle or float to the surface  before
collecting the brine shrimp from just below the surface in a pipette for
feeding.  Distribute 2 drops (0.1 g) of the brine shrimp to each test chamber.
If the survival rate in any test chamber falls below 50%, reduce the feeding
in that chamber to 1 drop of brine shrimp at each subsequent feeding.

                                      73

-------
10.6  OBSERVATIONS DURING THE TEST

10.6.1  Routine Chemical  and Physical  Observations

10.6.1.1  DO is measured  at the beginning and end of each 24-h exposure period
in at least one test chamber at each test concentration and in the control.

10.6.1.2  Temperature and pH are measured at the end of each 24-h exposure
period in at least one test chamber at each test concentration and in the
control.  Temperature should also be monitored continuously or observed and
recorded daily for at least two locations in the environmental control system
or the samples.  Temperature should be measured in a sufficient number of test
vessels at least at the end of the test to determine the temperature variation
in the environmental chamber.

10.6.1.3  The pH is measured in the effluent sample each day before new test
solutions are made.

10.6.1.4  Conductivity, alkalinity and hardness are measured in each new
sample (100% effluent or  receiving water) and in the control.

10.6.1.5  Record all the  measurements  on the data sheet (Figure 1).

10.6.2  Routine Biological  Observations

10.6.2.1  The number of live larvae in each test chamber are recorded on the
data sheets daily (Figure 2), and the  dead larvae are discarded.

10.6.2.2  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 these operations.

10.7  DAILY CLEANING OF TEST CHAMBERS

10.7.1  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  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 or pipet  when cleaning
the test 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 retrieved and returned
to the chambers.  Any incidence of removal of live larvae from the test
chambers during cleaning, and subsequent return to the chambers,  should be
noted in the records.

10.8  TEST SOLUTION RENEWAL

10.8.1  Freshly prepared  solutions are used to renew the tests daily

                                      74

-------
immediately after cleaning the test chambers.  For on-site toxicity studies,
fresh effluent or receiving water samples should be collected daily, and no
more than 24 h should elapse between collection of the samples and their use
in the tests (see Section 8, Effluent and Receiving Water Sampling, Sample
Holding, and Sample Preparation for Toxicity Tests).  For off-site tests, a
minimum of three samples are collected, preferably on days one, three, and
five.  Maintain the samples in the refrigerator at 4°C until  used.

10.8.2  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 (250 mL) should be added slowly by pouring down the side of the
test chamber to avoid excessive turbulence and possible injury to the larvae.

10.9  TERMINATION OF THE TEST

10.9.1  The test is terminated after seven days of exposure.  At test
termination, dead larvae are removed and discarded.  The surviving larvae in
each test chamber (replicate) are counted and immediately prepared as a group
for dry weight determination, or are preserved as a group in 70% ethanol or 4%
formalin before drying and weighing within 7 days.  For safety, formalin
should be used under a hood.

10.9.2  For immediate drying and weighing, place live larvae onto a 500 jum
mesh screen in a large beaker to wash away debris that might contribute to the
dry weight.  Each group of larvae is rinsed with deionized water to remove
food particles, transferred to a tared weighing boat that has been properly
labeled, and dried at 100°C for a minimum of 6 h.   Immediately upon removal
from the drying oven, the weighing boats are placed in a dessicator until
weighed, to prevent the absorption of moisture from the air.  All weights
should be measured to the nearest 0.01 mg and recorded on data sheets
(Figure 3).  Subtract tare weight to determine the dry weight of the larvae in
each replicate.  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 on the data sheet (Figure 3).

10.9.3  Prepare a summary table as illustrated in Figure 4.

11.  SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA

11.1  A summary of test conditions and test acceptability criteria is
presented in Table 1.

12.  ACCEPTABILITY OF TEST RESULTS

12.1  For the test results to be acceptable, survival in the controls must
be at least 80%.  The average dry weight per original control larvae at the
end of the test should equal or exceed 0.30 mg.
                                      75

-------
     TABLE  1.   SUMMARY  OF  TEST  CONDITIONS  AND  TEST  ACCEPTABILITY CRITERIA
               FOR  FATHEAD MINNOW, PIMEPHALES  PROMELAS,  LARVAL  SURVIVAL AND
               GROWTH TOXICITY  TESTS  WITH  EFFLUENTS AND  RECEIVING WATERS
 1.  Test type:

 2.  Temperature  (°C):

 3.  Light quality:

 4.  Light intensity:


 5.  Photoperiod:

 6.  Test chamber  size:

 7.  Test solution volume:

 8.  Renewal  of test
     concentrations:

 9.  Age of test  organisms:
Static renewal

25 ± 1°C

Ambient laboratory illumination

10-20 /iE/m2/s (50-100 ft-c) (ambient
laboratory levels)

16 h light, 8 h darkness

500 mL (minimum)

250 mL (minimum)


Daily

Newly hatched larvae less than 24 h old.
If shipped, not more than 48 h old, and a
window of 24 h age
10.  No.  larvae per test chamber:     15  (minimum  of  10)
11.  No.  replicate chambers
     per concentration:

12.  No.  larvae per concentration:

13.  Feeding regime:
14.  Cleaning:
4 (minimum of 3)

60 (minimum of 30)

Feed 0.1 g newly hatched  (less than 24-h
old) brine shrimp nauplii three times
daily at 4-h intervals or, as a minimum,
0.15 g twice daily, 6 h between feedings
(at the beginning of the  work day prior to
renewal, and at the end of the work day
following renewal).  Sufficient nauplii
are added to provide an excess.  Larvae
fish are not fed during the final 12 h of
the test

Siphon daily, immediately before test
solution renewal
                                      76

-------
    TABLE 1.   SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA
              FOR FATHEAD MINNOW,  PIMEPHALES PROMELAS,  LARVAL SURVIVAL AND
              GROWTH TOXICITY TESTS WITH EFFLUENTS AND RECEIVING WATERS
              (CONTINUED)
15.  Aeration:
16.  Dilution water:
17. Effluent concentrations;

18. Test dilution factor:


19. Test duration:

20. Endpoints:

21. Test acceptability
     criteria:


22. Sampling requirement:
None, unless DO concentration falls below
4.0 mg/L.  Rate should not exceed
100 bubbles/min

Uncontaminated source of receiving or
other natural water, synthetic water
prepared using MILLIPORE MILLI-QR or
equivalent deionized water and reagent
grade chemicals, or DMW (see Section 7)

Minimum of 5 and a control
Effluents:  > 0.5
Receiving waters:

7 days
None or > 0.5
23.  Sample volume required:
Survival and growth (weight)

80% or greater survival in controls;
Average dry weight per original animal  in
control equals or exceeds 0.30 mg

For on-site tests, samples 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 collected on days one, three
and five with a maximum holding time of
36 h before first use

2.5 L/day
                                      77

-------
      Figure 1,
Discharger:
Location:
Data form for the fathead minnow, Pimephales promelas,  larval
survival  and growth test.  Routine chemical and physical
determinations.
                         Analyst:
                         Dates:
Dav
Control :
Terno.
D.O. Initial
Final
pH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










2










3










4










5










6










7










Remarks










                                     Day
Cone:
Terno.
D.O. Initial
Final
pH Initial
Final
Alkal initv
Hardness
Conductivity
Chlorine

1










2










3










4










5










6










7










Remarks










                                     Day
Cone:
Temo.
D.O. Initial
Final
pH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










2










3










4










5










6










7










Remarks










                                      78

-------
      Figure 1.   Data form for the fathead minnow, Pimephales promelas, larva
                 survival  and growth test.  Routine chemical and physical
                 determinations. (Continued).
Discharger:
Location:
Analyst:
Dates:
                                     Day
Cone:
Temp.
D.O. Initial
Final
pH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










2










3










4










5










6










7










Remarks










                                     Day
Cone:
Temp.
D.O. Initial
Final
pH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










2










3










4










5










6










7










Remarks










                                     Day
Cone:
Temp.
D.O. Initial
Final
DH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










2










3










4










5










6










7










Remarks










                                      79

-------
      Figure 2.   Mortality data for the fathead minnow, Pimephales promelas,
                 larval  survival  and growth test.
Discharger:
Location:
Dates: _
Analyst:
No. Survivinq Organisms
Cone: Reo. Day
No.
Control :



Cone:



Cone:



Cone:



Cone:



Cone:



Cone:



Cone:



1
































2
































3
































4
































5
































6
































7
































Remarks
































Comments:
                                          80

-------
      Figure 3.  Weight data for the fathead minnow, Pimephales promelas, larval survival and growth test
   Discharge:
   Location:
   Analyst:
Test Date(s):
Weighing Date:
Drying Temperature (°C):
Drying Time (h): 	
Cone: Rep.
No.
Control



Cone:



Cone:



Cone:



Cone:



Cone:



A
Wgt. of
Tare
(mg)
























B
Dry wgt:
tare and
larvae
(mq)
























B-A
Total dry
wgt of
larvae
(mq)
























C
No. of
original
larvae
(mq)
























(B-AJ/C
Mean dry wgt
of larvae
(mg)
























Remarks
























CO
   1From USEPA (1989a).

-------
      Figure 4.  Summary data for the fathead minnow, Pimephales promelas,
                 larval survival and growth test1'
Discharger:

Location: _
Test Dates:
     Analyst:
Treatment
No. live
larvae
Survival
IO/\
(/o)
Mean dry wgt
of larvae (mg)
± SD
Temperature
range (degrees C)
Dissolved
oxygen range
(mg/L)
Hardness
Conductivity
Control























































Comments:
1Adapted  from  Norberg  and  Mount  (1985a).
                                      82

-------
13.   DATA ANALYSIS

13.1  GENERAL

13.1.1  Tabulate and summarize the data.
response data is shown in Table 2.
A sample set of survival and growth
13.1.2  The endpoints of toxicity tests using the fathead minnow, Pimephales
promelas, larvae are based on the adverse effects on survival and growth.  The
LC50, the EC50, the IC25, and the IC50 are calculated using point estimation
techniques, and LOEC and NOEC values for survival and growth are obtained
using a hypothesis testing approach such as Dunnett's Procedure  (Dunnett,
1955) or Steel's Many-one Rank Test (Steel, 1959; Miller, 1981)  (see Section
9, Chronic Toxicity Endpoints and Data Analysis).  Separate analyses are
performed for the estimation of the LOEC and NOEC endpoints and  for the
estimation of the LC50, EC50, IC25 and IC50.  Concentrations at which there is
no survival in any of the test chambers are excluded from the statistical
analysis of the NOEC and LOEC for survival and growth, but included in the
estimation of the LC50, IC25, IC50, and EC50.  See the Appendices for examples
of the manual computations, program listings, and examples of data input and
program output.
  TABLE 2.  SUMMARY OF SURVIVAL AND GROWTH DATA FOR FATHEAD MINNOW, PIMEPHALES
            PROMELAS, LARVAE EXPOSED TO A REFERENCE TOXICANT FOR SEVEN DAYS1
            Proportion of
  NaPCP  Survival in Replicate    Mean
  Cone.        Chambers           Prop.
  (fig/L)    A    B    C    D      Surv
  Avg Dry Wgt (mg) In       Mean
  Replicate Chambers2       Dry Wgt
   A     B     CD      (mg)
0
32
64
128
256
512
1.0
0.8
0.9
0.9
0.7
0.4
1.0
0.8
1.0
0.9
0.9
0.3
0.9
1.0
1.0
0.8
1.0
0.4
0.9
0.8
1.0
1.0
0.5
0.2
0
0
0
0
0
0
.95
.85
.975
.90
.775
.325
0.711
0.646
0.669
0.629
0.650
0.358
0.662
0.626
0.669
0.680
0.558
0.543
0.718
0.723
0.694
0.513
0.606
0.488
0.767
0.700
0.676
0.672
0.508
0.495
0.714
0.674
0.677
0.624
0.580
0.471
Vour replicates  of 10 larvae each.
 Values  are as  mean per surviving larvae;  not per original  number.
                                      83

-------
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 FATHEAD MINNOW, PIMEPHALES PROMELAS, SURVIVAL
      DATA

13.2.1  Formal statistical analysis of the survival data is outlined on the
flowchart in Figure 5.  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,  EC50, and 1C endpoints.  Concentrations at which there
are 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 1C, EC,
and LC 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 (see Appendix
D).   The Wilcoxon Rank Sum Test with the Bonferroni adjustment is the
nonparametric alternative (see Appendix F).

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.  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 (see USEPA, 1991b).

13.2.5  Example of Analysis of Survival Data

13.2.5.1  This example uses the survival data from the Fathead Minnow Larval
Survival and Growth Test (Table 2).  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 toxicant
concentration and control are listed in Table 3.  A plot of the survival
proportions is provided in Figure 6.
                                      84

-------
          STATISTICAL ANALYSIS OF FATHEAD MINNOW LARVAL
                     SURVIVAL AND GROWTH TEST

                            SURVIVAL
                             SURVIVAL DATA
                         PROPORTION SURVIVING
                             ARC SINE
                          TRANSFORMATION
 ENDPOINT ESTIMATE
       LC50
SHAPIRO-WILK'S TEST
             NORMAL DISTRIBUTION
                                             NON-NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
                           BARTLETT'S TEST
                         HETEROGENEOUS
                            VARIANCE
NO
•
EQUAL NUMBER OF
REPLICATES?
YES
i
r
                                     EQUAL NUMBER OF
                                       REPLICATES?
                                                       NO
                                       YES
TWITH
.RRONI
TMENT


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 fathead
                   minnow,  Pimephales promelas, larval survival  data
                                 85

-------
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 4.
13.2.6.2  Calculate the denominator, D,  of the 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
        TABLE 3.   FATHEAD MINNOW,  PIMEPHALES PROMELAS,  SURVIVAL DATA
NaPCP Concentration (uq/L)
Replicate
A
RAW B
C
D
ARC SINE A
TRANS- B
FORMED C
D
Mean(Y,-)
S?
i
Control
1
1
0
0
1
1
1
1
1
0
1
.0
.0
.9
.9
.412
.412
.249
.249
.330
.0088

0
0
1
0
1
1
1
1
1
0
2
32
.8
.8
.0
.8
.107
.107
.412
.107
.183
.0232


0
1
1
1
1
1
1
1
1
0
3
64
.9
.0
.0
.0
.249
.412
.412
.412
.371
.0066

128
0
0
0
1
1
1
1
1
1
0
4
.9
.9
.8
.0
.249
.249
.107
.412
.254
.0155

256
0
0
1
0
0
1
1
0
1
0
5
.7
.9
.0
.5
.991
.249
.412
.785
.109
.0768

512
0.4
0.3
0.4
0.2
0.685
0.580
0.685
0.464
0.604
0.0111
6
                                      86

-------
CD
               §
               §
                   1.0 -
                   0.9
                   0.8
                   0.7
                   0.6 -
                   0.5 H
                   0.4
                   0.3 -
                                                                                   CONNECTS THE MEAN VALUE FOR EACH CONCENTRATION
                                                                                   REPRESENTS THE CRITICAL VALUE FOR DUNNETT'S TEST
                                                                                   (ANY PROPORTION BELOW THIS VALUE WOULD BE
                                                                                   SIGNIFICANTLY DIFFERENT FROM THE CONTROL)
                   0.2
                   0.1
                  o.o -
— 1
 32
                                                              64                 128
                                                    SODIUM PENTACHLOROPHENATE (UG/L)
                                                           256
512
                        Figure  6.   Plot  of  mean  survival  proportion data  in  Table 3.

-------
        TABLE 4.  CENTERED OBSERVATIONS FOR SHAPIRO-MILK'S EXAMPLE
     Replicate     Control
   NaPCP Concentration
32      64     128     256     512
A 0.
B 0.
C -0.
D -0.
082
082
081
081
-0
-0
0
-0
.076
.076
.229
.076
-0
0
0
0
.122
.041
.041
.041
-0
-0
-0
0
.005
.005
.147
.158
-0
0
0
-0
.118
.140
.303
.324
0.
-0.
0.
-0.
081
024
081
140
13.2.6.3  For this set of data:     n = 24
                                   X = _!_ (0.000) = 0.000
                                       24
                                   D = 0.4265
13.2.6.4  Order the centered observations from smallest to largest
               X(1)   X(2)       - X
-------
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 = 24  and k = 12.   The  a,- values  are
listed in Table 6.

13..2.6.6  Compute the test statistic, W, as follows:

                       k
               W = I [ s a, (X(n'i+1)   X(i)) ]2
                   D  1 = 1

The differences x(n"i+1)   X(i) are listed in Table  6.   For  the  data
in this example,

               W =    1      (0.6444)2 = 0.974
                    0.4265


TABLE 6.  COEFFICIENTS AND DIFFERENCES FOR SHAPIRO-WILK'S EXAMPLE
                  a,        X(n-'+1) - X(n
1
2
3
4
5
6
7
8
9
10
11
12
0.4493
0.3098
0.2554
0.2145
0.1807
0.1512
0.1245
0.0997
0.0764
0.0539
0.0321
0.0107
0.627
0.376
0.298
0.262
0.200
0.163
0.162
0.157
0.117
0.117
0.065
0.0
X(24)
X(23)
X(22)
X(21)
X(20)
X(19)
X(18)
X(17)
X(16)
X(15)
X(14)
v(13)
A
- x(1)
v(2)
A
v(3)
A
X(4)
X(5)
v(6)
A
x<7)
X(8)
X(9)
x<10)
X(11)
X(12)
13.2.6.7  The decision rule for this test is to compare W as calculated  in
section 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 = 24 observations is 0.884.  Since  W =  0.974
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 mean  proportion
surviving is the same across all toxicant concentrations including  the

                                      89

-------
control,  is Bartlett's Test (Snedecor and Cochran, 1980).  The test  statistic
is as follows:
                   p             p
               [ ( E V,)  In S2   s V, In S,.2  ]
           B      i=l _ id _
                             C

  Where:   V;  =  degrees of freedom  for each toxicant concen-
                 tration and control, Vs  =  (n^    1)

          nf  =  the number of replicates  for  concentration i.

          In = loge

          i  = 1, 2, ..., p where  p is the  number of concentrations
               including the control


                      ( S V, S-2)
           S2 =        1=1
          C  = 1 + ( 3CP-1))-1  [ SP1/V,   ( 2 V,)'1  ]
                               i=l        i=l

13.2.7.2  For the data in this example, (See Table 3) all toxicant
concentrations including the control  have the same number of replicates
(n,  = 4  for all  i).   Thus,  V,  = 3  for  all  i.

13.2.7.3  Bartlett's statistic is therefore:

                                P     2
       B =  [(18)ln(0.0236)   3 X ln(S,-)]/1.1296
                               i = l

         =  [18(-3.7465)   3(-24. 7516)]/1. 1296

         =  6.8178/1.1296

         =  6.036

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 (from a table of chi-square distribution), at a
significance level of 0.01 with five degrees of freedom, is 15.086.  Since B =
6.036 is less than the critical value of 15.086,  conclude that the variances
are not  different.
                                      90

-------
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 7.
                            TABLE 7.  ANOVA TABLE
Source
Between
Within
Total
Where: p
df
P 1
N - p
N 1
= number toxi
Sum of Squares
(SS)
SSB
SSW
SST
cant concentrations
Mean Square(MS)
(SS/df)
Sg = SSB/(p-l)
Sy = SSW/(N-p)

including the control
         N  = total number of observations n1 + n?  ...  +n
         n,-  = number of observations in concentration i

              P  2       7
        SSB = Z 1,/n,.  -  GYM          Between Sum of Squares
              P   ni    ,    ,
        SST = S   S Y--2   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"
        YJJ  = the  jth  observation  for concentration  "i"  (represents
              the proportion surviving for toxicant concentration
              i in test chamber j)
13.2.8.2   For the data in this example:

    n, = n2
    N  =  24
    n,  = n2 =  n3 = n4 = n5  = n6 = 4
T.
T2
T3

T5
T6
=
=
=
=
=
=
Yn "
Y21 -
Y31 H
Y41 <
Y51 H
Y6i -
h Y
h Y
h Y
h Y
h Y
h Y
12 "*
22 "*
32 H

52 "*
62 H
r Y
r Y
- Y
H Y
h Y
h Y
13 ^
23 ^
33 H
43 H
53 H
63 H
h Y
h Y
h Y
h Y
h Y
h Y
H
24
34

54
64
=
=
=
=
=
=
5
4
5
5
4
2
.322
.733
.485
.017
.437
.414
                                      91

-------
    G  = T,  +  T2 + T3 + T4 + T5  +  T6 = 27.408

          p
    SSB = X T,.2/n,-   G2/N
        = _1_(131.495)   (27. 408)2  = 1.574
           4                 24
                  •>    7
    SST = E   2 Yn2   G2/N
         i=l j=l

        = 33.300   (27.408)2  = 2.000
                      24

    SSW = SST   SSB =2.000   1.574 = 0.4260

    S2  = SSB/(p-l)  = 1.574/(6-l)  = 0.3150

    S2  = SSW/(N-p)  = 0.426/(24-6)  = 0.024

13.2.8.3  Summarize these calculations in the ANOVA table  (Table  8)


            TABLE 8.  ANOVA TABLE FOR DUNNETT'S PROCEDURE  EXAMPLE
Source
Between
Within
Total
df
5
18
23
Sum of Squares
(SS)
1.574
0.426
2.002
Mean Square(MS)
(SS/df)
0.315
0.024

13.2.8.4  To perform the individual comparisons, calculate  the  t statistic for
each concentration, and control combination as follows:

                                    ( Y,   Y,  )
                                 Sw  V  (1/n,)  + (1/n,-)
                                      92

-------
Where:  Y,.   = mean proportion surviving for concentration i
        Y1   = mean proportion surviving for the control
        Su   = square root of within mean square
        n1   = number of replicates for the control
        nt-   = number of replicates for concentration i.

13.2.8.5  Table 9 includes the calculated t values for each concentration and
control combination.  In this example, comparing the 32 M9/L concentration
with the control the calculation is as follows:

                             ( 1.330   1.183 )
                  t2 =   	 = 1.341
                       [ 0.155 V (1/4) + (1/4)  ]


                    TABLE 9.  CALCULATED T VALUES
           NaPCP Concentration(Mg/L)
32
64
128
256
512
2
3
4
5
6
1.341
-0.374
0.693
2.016
6.624
13.2.8.6  Since the purpose of this test is to detect a significant reduction
in proportion surviving, 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, 18 degrees of freedom for error and five concentrations
(excluding the control) the critical value is 2.41.  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.  Since t6 is greater than 2.41,  the  512 M9/L concentration has
significantly lower survival than the control.  Hence the NOEC and the LOEC
for survival are 256 /^g/L and 512 M9/L, respectively.

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 V (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)

                                      93

-------
        n,  =  the  number  of  replic^oa  in  the  control.

13.2.8.8  In  this example:
                   MSD = 2.41  (0.155)  V (1/4)  + (1/4)
                       = 2.41  (0.155)(0.707)
                       = 0.264

13.2.8.9  The MSD (0.264)  is in transformed units.   To determine the MSD in
terms of percent survival,  carry out the following  conversion.

    1.  Subtract the MSD from the transformed  control  mean.

                            1.330 -  0.264 = 1.066

    2.  Obtain the untransformed values for the control  mean and the difference
       calculated in 1.

                         [Sine ( 1.330) ]2 = 0.943
                         [Sine ( 1.066) ]2 = 0.766

    3.  The untransformed MSD (MSD )  is  determined by subtracting  the
       untransformed values from 2.

                        MSDU = 0.943 -  0.766 =  0.177

13.2.8.10  Therefore,  for this set of  data, the minimum difference in mean
proportion surviving between the control  and  any toxicant concentration that
can be detected as statistically significant  is 0.177.

13.2.8.11  This represents a decrease  in survival of 19% from the control.

13.2.9  Probit Analysis

13.2.9.1  The data used for the Probit Analysis is  summarized in Table 10. To
perform the Probit Analysis, run the EPA Probit Analysis Program.  An example
of the program input and output is supplied in Appendix I.   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


                     TABLE 10.  DATA FOR PROBIT ANALYSIS

Number
Number

Dead
Exposed
Control
2
40
NaPCP
32
6
40
Concentration
64
1
40
128
4
40
(UQ/L)
256
9
40

512
27
40
                                      94

-------
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 appropriate for this data.

13.2.9.3.  Table 11 shows the output data for the Probit Analysis of the data
from Table 10 and Figure 7 is a plot of the adjusted probits and the predicted
regression line from the EPA Probit Program.


TABLE 11.  OUTPUT FOR EPA PROBIT ANALYSIS PROGRAM FOR EC50 VALUES, VERSION  1.4
Probit Analysis of Fathead Minnow Larval Survival Data
 Cone.

  Control
  32.0000
  64.0000
 128.0000
 256.0000
 512.0000
              Number
             Exposed

               40
               40
               40
               40
               40
               40
            Number
          Responding

              2
              6
              1
              4
              9
             27
 Observed
Proportion
Responding

  0.0500
  0.1500
  0.0250
  0.1000
  0.2250
  0.6750
 Adjusted
Proportion
Responding

  0.0000
  0.0779
  -.0577
  0.0237
  0.1593
  0.6474
 Predicted
Proportion
Responding

  0.0782
  0.0000
  0.0001
  0.0101
  0.1650
  0.6472
Chi-square Heterogeneity
Mu      =  2.626029
Sigma   =  0.223555
                           4.522
 Parameter
                Estimate
                 Std.  Err.
            95% Confidence Limits
Intercept
Slope
Spontaneous
Response Rate
-6.746692
4.473178
0.078182
3.112017
1.196026
0.022541
( -12.846246,
( 2.128967,
( 0.034002,
-0.647139)
6.817389)
0.122363)
         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.

 127.6359
 181.2605
 218.5347
 247.9341
 422.6964
 720.6437
 817.5905
 985.7196
1399.8575
                                        Lower            Upper
                                        (95% Confidence  Limits)
   34.5885
   71.4914
  104.8182
  135.2143
  345.7290
  562.5553
  616.4054
  702.9269
  893.3054
     195.4335
     248.7074
     284.0806
     311.8864
     531.0254
    1420.7512
    1836.3506
    2696.8005
    5581.7588
                                       95

-------
Probit  Analysis of Fathead Minnow Larval  Survival  Data

        PLOT OF ADJUSTED PROBITS AND PREDICTED  REGRESSION LINE
Probit
   10+
    8+
    7+
    6+



                                           .0.


    5+
                         O. . . .
    4+
     -o              ....

     -o        .  .

    3+
    2+
    1+
    O+o
      -+	+	+	+	+	+	+_
      EC01           EC10     EC25      EC50      EC75     EC90           EC99
   Figure  7.   Plot of adjusted probits and  predicted regression line
               from EPA Probit Program.
                                       96

-------
13.3  EXAMPLE OF ANALYSIS OF FATHEAD MINNOW, PIMEPHALES PROMELAS, GROWTH DATA

13.3.1  Formal statistical analysis of the growth data is outlined in
Figure 8.  The response used in the statistical analysis is mean weight per
replicate.  An 1C estimate can be calculated for the growth data via a point
estimation technique (see Section 9).  Hypothesis testing can be used to
obtain a 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 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 (see Appendix D).  The Wilcoxon Rank Sum Test with the Bonferroni
adjustment is the nonparametric alternative (see Appendix F).

13.3.4  The data, mean and standard deviation of the observations at each
concentration including the control are listed in Table 12.  A plot of the
mean weights for each treatment is provided in Figure 9.  Since there is
significant mortality in the 512 M9/L concentration, its effect on growth is
not considered.
          TABLE 12.  FATHEAD MINNOW, PIMEPHALES PROMELAS, GROWTH DATA
Replicate    Control
                                   NaPCP Concentration (uq/L)
                   32
         64
128    256
512
A
B
C
D
0
0
0
0
.711
.662
.646
.690
0.
0.
0.
0.
517
501
723
560
0.
0.
0.
0.
669
669
694
676
0
0
0
0
.602
.612
.440
.672
0
0
0
0
.455
.502
.606
.254
-
-
-
~
Mean(Y,.)
jj
0.714
0.0018
1
0.674   0.677
0.0020  0.0001
2       3
0.624  0.580   -
0.0059 0.0037  -
456
                                      97

-------
           STATISTICAL ANALYSIS OF FATHEAD 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
        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?
i
YES
r

DUNNETT'S
  TEST
STEEL'S MANY-ONE
   RANKTEST
  WILCOXON RANK SUM
      TEST WITH
BONFERRONI ADJUSTMENT
                           ENDPOINT ESTIMATES
                               NOEC.LOEC
 Figure  8.  Flowchart  for statistical  analysis of fathead minnow,
           Pimephales promelas,  larval  growth data.
                                 98

-------
vo
                                                     CONNECTS THE MEAN VALUE FOR EACH CONCENTRATION
                                                     REPRESENTS THE CRITICAL VALUE FOR DUNNETT'S TEST
                                                     (ANY MEAN WEIGHT BELOW THIS VALUE WOULD BE
                                                      SIGNIFICANTLY DIFFERENT FROM THE CONTROL)
                                                                              128
                                                                                                   256
                                            SODIUM PENTACHLOROPHENATE(UG/L)
             Figure 9.   Plot of  mean weight data  from fathead minnow,  Pimephales promelas, larval  survival
                         and growth test.

-------
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 13.
         TABLE 13.  CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S EXAMPLE
NaPCP Concentration (uq/L)
Replicate
A
B
C
D
Control
-0.003
-0.052
0.004
0.053
32
-0.028
-0.048
0.049
0.026
64
-0.008
-0.008
0.017
-0.001
128
0.005
0.056
-0.111
0.048
256
0.070
-0.022
0.026
-0.072
13.3.5.2  Calculate the denominator, D, of the test statistic:
                     D = S (X,    X)!
    Where X-  = the ith  centered  observation
          X
          n
the overall mean of the centered observations
the total number of centered observations.
For this set of data,            n = 20

                                 X = _L_ (0.000) = 0.000
                                     20
                                 D = 0.0412

13.3.5.3  Order the centered observations from smallest to largest

                  X(1)   X(2)    ... - X(n)

Where X(1) is the ith ordered observation.  These ordered observations are
listed in Table 14.

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 = 20, k = 10.  The a- values are
listed in Table 15.                                          '

                                      100

-------
  TABLE  14.   ORDERED CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S  EXAMPLE
i
1
2
3
4
5
6
7
8
9
10
X(i)
-0.111
-0.072
-0.052
-0.048
-0.028
-0.022
-0.008
-0.008
-0.003
-0.001
i
11
12
13
14
15
16
17
18
19
20
X(i>
0.004
0.005
0.017
0.026
0.026
0.048
0.049
0.053
0.056
0.070
13.3.5.5  Compute the test statistic, W, as follows
               W = I [ s a, (X(n'i+1)
                   D  1 = 1
the differences x(n'l>1) - X
For this


TABLE
i
1
2
3
4
5
6
7
8
9
10
set of data:
W =

15. COEFFICIENTS
ai
0.4734
0.3211
0.2565
0.2085
0.1686
0.1334
0.1013
0.0711
0.0422
0.0140
(i) are listed in Tabl

1 (0.1988)2 =
0.0412
AND DIFFERENCES FOR
«(n-i+1) _ x/(i)
0.181
0.128
0.105
0.097
0.076
0.048
0.034
0.025
0.008
0.005
e 15.

0.959





SHAPIRO-WILK'S EXAMPLE

X(20)
Y(19)
A
Y(18)
A
x<17)
X(16)
X(15)
X(U)
X(13)
x<12)
v(H)
A

Xd)
- x(2)
- x(3>
- x(4)
- x(5>
- x(6)
- x(7)
X(8)
X(9)
X(10)
                                      101

-------
13.3.5.6  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 20
observations (n) is 0.868.  Since W = 0.959 is greater than the critical
value, the conclusion of the test is that the data are normally distributed.

13.3.6  Test for Homogeneity of Variance

13.3.7.1  The test used to examine whether the variation in mean dry weight is
the same across all toxicant concentrations including the control, is
Bartlett's Test (Snedecor and Cochran,  1980).   The test statistic is as
follows:
           B
                   P        _    P       ,
               [ ( Z V,)  In  S2  - S V,- In S? ]
    Where V,-  =   degrees  of freedom for each toxicant concen-
                 tration and control, V,. = (n,.  -  1)
          n,.  = the  number of replicates for concentration i.
          In = loge

          i   = 1, 2, ..., p where p is the number of concentrations
               including the control
          ?
          C  = 1 + (  3(p-l))-1  [ S1/V, - (
13.3.6.2  For the data in this example,  (See Table 12) all  toxicant
concentrations including the control  have the same number of replicates
(n,  =  4  for  all  i).   Thus,  V, = 3 for all i.

13.3.7.3  Bartlett's statistic is therefore:
                                P
       B =  [(15)ln(0.0027) - 3 2 ln(S?)]/1.133
                               i=l

         =  [15(-5.9145) -  3(-32.4771]/1.133

         =  8.7138/1.133

         =  7.691

                                     102

-------
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 four
degrees of freedom, is 13.277.  Since B = 7.691 is less than the critical
value of 13.277, conclude that the variances are not different.

13.3.7  Ounnett'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 16.

                          TABLE 16.  ANOVA TABLE
Source
Between
Within
Total
df
p - 1
N - p
N - 1
Sum of Squares
(SS)
SSB
SSW
SST
Mean Square(MS)
(SS/df)
S* = SSB/(p-l)
Sy = SSW/(N-p)

Where:      p  = number toxicant concentrations including the control
            N  = total number of observations n1  + n2  ... +np

            ni  = number of observations in concentration  i
SSB =
                        , - G2/N
Between Sum of Squares
                 P   ni   ,    ,
           SST = S   S Yn2 - G2/N
                              Total  Sum of Squares
           SSW = SST - SSB
                               Within Sum of Squares
             G = the grand total of all sample observations, G = s T,-
                                                                i=l
            T,- = the total  of the replicate measurements for
                 concentration "i"
           (,-j  =  the  jth observation for concentration  "i"  (represents
                 the mean dry weight of the fish for toxicant
                 concentration i in test chamber j)
                                      103

-------
13.3.7.2  For the data in this example:

    n.  = n2  =  n, = n4 = n5 = 4
    N  = 2T)
    TI  = YH + Y12  +  Y13 + Yu = 2.858
    T  =.Y  + Y   +  Y  + Y24 = 2.695
    T  = Y  + Y2.  +  YJJ + Yj, = 2.708
    T  = Y  + Y   +  Y  + Y* = 2.494
    TS  - Y;; + Y^  +  Y5^ + Y54 = 2.322

    G  - T,  + T2 + T3 + T4 + T5 - 13.077

    SSB = S T,.2/nf   G2/N
         i = l

        =  1  (34.376) -  (13.077)2  = 0.044
           4                20


    SST = 2   sY-:2  - G2/N
         1-1 j=l

        =  8.635  - (13.077)2  = 0.085
                      20

    SSW = SST   SSB = 0.085 - 0.044  =  0.0410

    Sg  = SSB/(p-l) = 0.044/(5-l) = 0.0110

    Sy  = SSW/(N-p) = 0.041/(20-5) = 0.0027

13.3.7.3  Summarize these calculations in  the ANOVA table (Table 17)


          TABLE 17.   ANOVA TABLE FOR DUNNETT'S PROCEDURE EXAMPLE
Source
Between
Within
df
4
15
Sum of Squares
(SS)
0.044
0.041
Mean Square(MS)
(SS/df)
0.011
0.0027
    Total         19             0.085
                                      104

-------
13.3.7.4  To perform the individual comparisons, calculate the t statistic for
each concentration, and control combination as follows:
                                         - Y
                                Sw V (1/n,)  +  (1/n,-)
Where Y,-   = mean dry weight for toxicant concentration i
      Y1   = mean dry weight for the control
      SH   = square root of within mean square
      n1   = number of replicates for the control
      n,   = number of replicates for concentration i.

13.3.7.5  Table 18 includes the calculated t values for each concentration
and control combination.   In this example, comparing the 32 Atg/L concentration
with the control the calculation is as follows:
                            ( 0.714 - 0.674)
                                                  = 1.081
                        [ 0.052 V (1/4) + (1/4)  ]
                      TABLE 18.  CALCULATED T VALUES
                     NaPCP
                 Concentration
                     (M9/L)
t.-
32
64
128
256
2
3
4
5
1.081
1.000
2.432
3.622
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, 15 degrees of freedom for error and four concentrations  (excluding the
control) the critical value is 2.36.  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 t4 and t5  are greater than 2.36,
the 128 M9/L and 256 M9/L concentrations have significantly  lower growth than
the control.  Hence the NOEC and the LOEC for growth are 64 ng/L  and 128 M9/L,
respectively.

                                      105

-------
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  V (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 = 2.36 (0.052) V (1/4) + (1/4)
                       = 2.36 (0.052)(0.707)
                       = 0.087

13.3.7.9  Therefore, for this set of data, the minimum difference that can be
detected as statistically significant is 0.087 mg.

13.3.7.10  This represents a 12% reduction in mean weight from the control.

13.3.8  Example of Calculation of the 1C

13.3.8.1  The growth data in Table 2 are utilized in this example.  As seen in
Table 2 and Figure 10, the observed means are not monotonically non-increasing
with respect to concentration (the mean response for each higher concentration
is not less than or equal to the mean response for the previous concentration,
and the responses between concentrations do not follow a linear trend).
Therefore,  the means are smoothed prior to calculating the 1C.  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 meaji, Y. = 0_.714, we see that
Y, > Y2.  Set M, - Y,.  Comparing 72 to  Y3, Y2 < Y3.

13.3.8.3  Calculate the smoothed means:
                  M, = M,  =  (Y, + Y,)/2 = 0.675
13.3.8.4
M4 becomes 74,
For the remaining observed means,  M3 > Y4  >  Y5 > Y6.  Thus,
    M5 becomes Y5,  etc.,  for the remaining concentrations.
Table 19 contains the smoothed means, and Figure 10 provides a plot of the
smoothed concentration response curve.

13.3.8.5  An IC25 and an IC50 can be estimated using the Linear  Interpolation
Method.   A 25% reduction in weight, compared to the controls, would result  in
a mean weight of 0.536 mg, where M?(l -  p/100)  = 0.714(1 - 25/100).  A 50%
reduction in weight, compared to the controls, would result in a mean weight
of 0.357 mg, where M.(l  -  p/100)  = 0.714(1 - 50/100).   Examining the
smoothed means and their associated concentrations (Table 19), the response
                                      106

-------
0.77
0.76
0.73
0.71
0.60
0.67
0.65
0.63

0.50
0.57-
0.88-
0.83-
0.81
0.40-
0.47-
0.48-
0.43-
0.41-
0.30-
0.37-
0.35-
                                                          •***      INDIVIDUAL BXPLICATB UEAM VIGHT
                                                         	     COKNXCTS TOT OBSERVED HEAN VALUE
                                                         	     COMNXCT3 THE SMOOTHED MEAN VALUE
                                                  64                128
                                             SODIUM PENTACHLOROPHENATE (UG/L)
                                                                             256
512
Figure 10.
    Plot of  raw data,  observed means,  and smoothed means for the  fathead  minnow,  Pimephales
    pro/ne/as,  growth data  in Tables 2  and 19.

-------
            TABLE 19.  FATHEAD MINNOW, PIMEPHALES PROMELAS,
                       MEAN GROWTH RESPONSE AFTER SMOOTHING
NaPCP
Cone
(M9/L)
Control
32
64
128
256
512

i

1
2
3
4
5
6

M,-
(mg)
0.714
0.675
0.675
0.624
0.580
0.471
0.536 mg is bracketed by C5 = 256 ^g/l and C, =  512 jug/L.   For  the  50%
reduction (0.357 mg), the response (0.471 /ng) at the highest toxicant
concentration (512 M9/L) is greater than 50% of the control (0.357 mg).  Thus
the IC50 is specified as greater than 512 M9/L.

13.3.8.6  Using Equation 1 from Appendix J, the estimate of the IC25  is
calculated as follows:

               ICp = Cj  + [M,(l    p/100)    Mj](CJ+1   Cj)
              IC25 = 256 + [0.714(1 - 25/100) - 0.5801(512 - 256)
                                                     (0.471 - 0.580)

                   = 360 ug/L


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 349.7
M9/L, with a standard deviation of 62.3 jitg/L (coefficient of variation =
17.8%).  The empirical 95% confidence interval  for the true mean was  (253.3-
451.3).  The BOOTSTRP computer program output for the IC25 for this data set
is shown in Figure 11.

13.3.8.8  When BOOTSTRP was used to analyze this set of data for the  IC50,
requesting 80 resamples, the output indicated that the response of the highest
concentration of toxicant exceeded 50% of the control.  Thus, the IC50 could
not be calculated.  The BOOTSTRP computer program output is shown in
Figure 12.
                                      108

-------
THE NUMBER OF RESAMPLES IS     80


*** LISTING OF GROUP CONCENTRATIONS (% EFF.) AND RESPONSE MEANS ***

CONC.  (%EFF)              RESPONSE MEAN           MEAN AFTER POOLING
        .000                       .715                        .715
      32.000                       .674                        .675
      64.000                       .677                        .675
     128.000                       .623                        .623
     256.000                       .581                        .581
     512.000                       .471                        .471
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
  FROM THE INPUT SAMPLE IS 360.3287.
            BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY
                       OF THE ESTIMATED ICp
THE MEAN OF THE BOOTSTRAP ESTIMATES IS 349.6990.

THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS 62.3066.

AN EMPRICAL 94.9% CONFIDENCE INTERVAL FOR THE
     BOOTSTRAP ESTIMATE IS (253.2589,451.3332).
*** NOTE:  THE ABOVE BOOTSTRAP CALCULATIONS WERE BASED ON   79
    INSTEAD OF  80 RESAMPLINGS.  THOSE RESAMPLES NOT
    USED HAD ESTIMATES ABOVE THE HIGHEST CONCENTRATION % EFF.
       Figure 11.  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
        .000                       .715                        .715
      32.000                       .674                        .675
      64.000                       .677                        .675
     128.000                       .623                        .623
     256.000                       .581                        .581
     512.000                       .471                        .471
*** NO LINEAR INTERPOLATION ESTIMATE CAN BE CALCULATED FROM THE INPUT
    DATA, SINCE NONE OF THE (POSSIBLY POOLED) GROUP RESPONSE MEANS
    WERE LESS THEN 50.0% OF THE CONTROL RESPONSE MEAN.
            BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY
                       OF THE ESTIMATED ICp
                      k********************j

*** BOOTSTRAP ESTIMATES OF ICp FOR ALL RESAMPLES WERE ABOVE THE
    HIGHEST CONCENTRATION  % EFF.
       Figure 12.  BOOTSTRP program output for the IC50.
                                 110

-------
14.   PRECISION AND ACCURACY

14.1  PRECISION

14.1.1  Single-Laboratory Precision

14.1.1.1  Information on the single-laboratory precision of the fathead minnow
larval survival and growth test is presented in Table 20.  The range of NOECs
was only two concentration intervals, indicating good precision.

14.1.2  Multilaboratory Precision

14.1.2.1  An interlaboratory study of Method 1000.0 described in the first
edition of this manual (USEPA, 1985c), was performed using seven blind samples
over an eight month period (DeGraeve, et. al., 1988).  In this study, each of
the 10 participating laboratories was to conduct two tests simultaneous with
each sample, each test having two replicates of 10 larvae for each of five
concentrations and the control.  Of the 140 tests planned, 135 were completed.
Only nine of the 135 tests failed to meet the acceptance criterion of 80%
survival in the controls.  Of the 126 acceptable survival NOECs reported, an
average of 41% were median values, and 89% were within one concentration
interval of the median (Table 21).  For the growth (weight) NOECs, an average
of 32% were at the median, and 84% were within one concentration interval of
the median (Table 22).  Using point estimate techniques, the precision (CV) of
the IC50 was 19.5% for the survival data and 19.8% for the growth data.  If
the mean weight acceptance criterion of 0.25 mg for the surviving control
larvae, which was included in the second edition of the method (USEPA, 1989a),
had been applied to the results of the interlaboratory study, 40 of the 135
completed tests would have been considered unacceptable (Norberg-King, 1989).
Had the acceptance criterion of 0.30 mg which is included in this revised
edition of the method been applied to the results of the interlaboratory
study, an additional 6 tests would have failed to meet the test criteria.

14.2  ACCURACY

14.2.1  The accuracy of toxicity tests can not be determined.
                                      Ill

-------
TABLE 20.  PRECISION OF THE FATHEAD MINNOW, PIMEPHALES PROMELAS, LARVAL
           SURVIVAL AND GROWTH TEST, USING NAPCP AS A REFERENCE TOXICANT3'6


Test
1
2
3
4
5
n:
Mean:
CV(%):

LCI
mg/L
0.58
2.31
1.50
1.71
1.43
5
1.51
41.3

NOEC
(M9/L)
256
128
256
128
128
5
NA
NA

LOEC
(M9/L)
512
256
512
256
256
5
NA
NA
Chronic
Value
(M9/L)
362
181
362
181
181
5
253.4
NA
aFrom Pickering, 1988.
bFor a discussion of the precision of data from chronic toxicity
 tests see Section 4, Quality Assurance.
       TABLE 21.  COMBINED FREQUENCY DISTRIBUTION FOR SURVIVAL NOECs
                  FOR ALL LABORATORIES8
                                           NOEC Frequency (%) Distribution
                                  Tests with Two Reps     Tests with Four Reps
  Sample
Median  ± I1
>2C
Median  ± lb  >2C
1.
2.
3.
4.
5.
6.
7.
Sodium Pentachlorophenate (A)
Sodium Pentachlorophenate (B)
Potassium Dichromate (A)
Potassium Dichromate (B)
Refinery Effluent 301
Refinery Effluent 401
Utility Waste 501
35
42
47
41
26
37
56
53
42
47
41
68
53
33
12
16
6
18
6
10
11
57
56
75
50
78
56
56
29
44
25
50
22
44
33
14
0
0
0
0
0
11
^From DeGraeve  et  a).,  1988.
percent of values within  one concentration intervals of the median.
cPercent of values within  two or more concentrations intervals of the median.
                                      112

-------
      TABLE 22.  COMBINED FREQUENCY DISTRIBUTION FOR WEIGHT NOECs
                 FOR ALL LABORATORIES8
                                          NOEC Frequency (%) Distribution
                                 Tests with Two Reps     Tests with Four Reps

 Sample                            Median  ± lb   >2C       Median   ±  lb   >2C
1.
2.
3.
4.
5.
6.
7.
Sodium Pentachlorophenate (A)
Sodium Pentachlorophenate (B)
Potassium
Potassium
Refinery
Refinery
Dichromate (A)
Dichromate (B)
Effl
Effl
Utility Waste
uent 301
uent 401
501
59
37
35
12
35
37
11
41
63
47
47
53
47
61
0
0
18
41
12
16
28
57
22
88
63
75
33
33
43
45
0
25
25
56
56
0
33
12
12
0
11
11
From DeGraeve et al.,  1988.
Percent of values within oni
Percent of values within two or more concentrations intervals of the median,
bPercent  of  values within  one  concentration  intervals  of the  median.
                                     113

-------
                                  SECTION 12

                                 TEST METHOD

                    FATHEAD MINNOW, PIHEPHALES PRONELAS,
                EMBRYO-LARVAL SURVIVAL AND TERATOGENICITY TEST
                                METHOD 1001.0

1.  SCOPE AND APPLICATION

1.1  This method estimates the chronic toxicity of whole effluents and
receiving water to the fathead minnow, Pimephales promelas, using embryos
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 organisms.   The test is useful  in screening for
teratogens because organisms are exposed during embryonic development.

1.2  Detection limits of the toxicity of an effluent  or pure substance are
organism dependent.

1.3  Brief 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
degradable and highly volatile toxicants, such as chlorine, 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  Fathead minnow, Pimephales promelas, embryos are exposed to different
concentrations of effluent or to receiving water in a static renewal system
for seven days, starting shortly after fertilization  of the eggs.  Test
results are based on the total frequency of both mortality and gross
morphological deformities (terata).

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), high concentrations of
suspended and/or dissolved solids, and extremes of pH may mask the presence of
toxic substances.

3.3  Improper effluent sampling and sample handling may adversely affect test
results (see Section 8, Effluent and Receiving Water Sampling, Sample

                                      114

-------
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  Fathead minnow and brine shrimp culture units -- See Section 11 and
USEPA, 1991b.  To test effluent toxicity on-site or in the laboratory,
sufficient numbers of newly fertilized eggs must be available, preferably from
a laboratory fathead minnow culture unit.  If necessary, embryos can be
shipped in well oxygenated water in insulated containers.  In cases where
shipping is necessary, up to 48-h old embryos may be used for the test.

5.2  Samplers -- automatic sampler, preferably with sample cooling
capability, that can collect a 24-h composite sample of 2 L or more.

5.3  Sample containers -- for sample shipment and storage (see Section 8,
Effluent and Receiving Water Sampling, Sample Handling, and Sample Preparation
for Toxicity Tests).

5.4  Environmental chamber or equivalent facility with temperature control
(25 ± 1°C).

5.5  Water purification system -- MILLIPORE MILLI-QR deionized water or
equivalent (see Section 5, Facilities, Equipment, and Supplies).

5.6  Balance -- analytical, capable of accurately weighing to 0.00001 g.

5.7  Reference weights, Class S -- for checking performance of balance.
Weights should bracket the expected weights of material to be weighed.

5.8  Test chambers -- four (minimum of three) borosilicate glass or
disposable, nontoxic plastic labware, per test solution, such as: 500-mL
beakers; 100 mm x 15 mm or 100 mm x 20 mm glass or disposable polystyrene
Petri dishes; or 12-cm OD, stackable "Carolina" culture dishes.  The chambers
should be covered with safety glass plates or sheet plastic during the test to
avoid potential contamination from the air and excessive evaporation of the
test solutions during the test.

5.9  Dissecting microscope, or long focal length magnifying lens, hand or
stand supported -- for examining embryos and larvae in the test chambers.

5.10  Light box, microscope lamp, or flashlight -- for illuminating chambers
during examination and observation of embryos and larvae.

5.11  Volumetric flasks and graduated cylinders -- Class A, borosilicate glass
or nontoxic plastic labware, 10-1000 ml, for making test solutions.

                                      115

-------
5.12  Volumetric pipets -- Class A,  1-100 ml.
5.13  Serological pipets -- 1-10 ml, graduated.
5.14  Pipet bulbs and fillers -- PROPIPETR,  or equivalent.
5.15  Droppers, and glass tubing with fire polished edges,  2-mm ID -- for '
transferring embryos, and 4-mm ID -- for transferring larvae.
5.16  Wash bottles -- for washing embryos from substrates and containers and
for rinsing small glassware and instrument electrodes and probes.
5.17  Glass or electronic thermometers -- for measuring water temperatures.
5.18  Bulb-thermograph or electronic-chart type thermometers -- for
continuously recording temperature.
5.19  National Bureau of Standards Certified thermometer (see EPA Method
170.1, USEPA 1979b).
5.20  Meters,  pH, DO, and specific conductivity -- for routine physical and
chemical measurements.
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 recording data.
6.3  Tape, colored -- for labeling test chambers.
6.4  Markers,  water-proof -- for marking containers, etc.
6.5  Reagents for hardness and alkalinity tests (see EPA Methods 130.2 and
310.1, USEPA 1979b).
6.6  Membranes and filling solutions for DO probe (see USEPA Method 360.1,
USEPA 1979b),  or reagents for modified Winkler analysis.
6.7  Standard pH buffers 4, 7, and 10 (or as per instructions of instrument
manufacturer)  -- for instrument calibration (see USEPA Method 150.1, USEPA
1979b).
6.8  Specific conductivity standards (see EPA Method 120.1, USEPA 1979b).
6.9  Laboratory quality control samples and standards -- for calibration of
the above methods.
6.10  Reference toxicant solutions (see Section 4, Quality Assurance).
                                      116

-------
6.11  Reagent water -- defined as distilled or deionized water which does not
contain substances which are toxic to the test organisms (see Subsection 5.5
above).

6.12  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  TEST ORGANISMS FATHEAD MINNOWS, PIMEPHALES PROMELAS

6.13.1  Fathead minnow embryos, less than 36-h old, are used for the test.
The test is conducted with four (minimum of three) test chambers at each
toxicant concentration and control.  Fifteen (minimum of ten) embryos are
placed in each replicate test chamber.  Thus 60 (minimum of 30) embryos are
exposed at each test concentration and 360 (minimum of 180) embryos would be
needed for a test consisting of five effluent concentrations and a control.

6.13.2  Sources of organisms

6.13.2.1  It is recommended that the embryos be obtained from inhouse cultures
or other local sources if at all possible, because it is often difficult to
ship the embryos so that they will be less than 36 h old for beginning the
test.  Receipt of embryos via Express Mail, air express, or other carrier,
from a reliable outside source is an acceptable alternative, but they must not
be over 48 h old when used to begin the test.

6.13.2.2  Culturing methods for fathead minnows, Pimephales promelas, are
described in Section 11, Subsection 6.17.5 and in USEPA, 1991b.

6.13.2.3  Fish obtained from outside sources (see Section 5) such as
commercial biological supply houses for use as brood stock should be
guaranteed to be (1) of the correct species, (2) disease free, (3) in the
requested age range, and (4) in good condition.  This can be done by providing
the record of the date on which the eggs were laid and hatched, and
information on the sensitivity of the contemporary fish to reference
toxicants.

6.13.3  Obtaining embryos for toxicity tests from inhouse cultures.

6.13.3.1  Spawning substrates with the newly-spawned, fertilized embryos are
removed from the spawning tanks or ponds, and the embryos are separated from
the spawning substrate by using the index finger and rolling the embryos
gently with a circular movement of the finger (See Gast and Brungs, 1973).
The embryos are then combined and washed from the spawning substrate onto a
400 /urn NITEXR screen,  sprayed with a stream of deionized water to remove
detritus and food particles, and back-washed with dilution water into a
crystallizing dish for microscopic examination.  Damaged and infertile eggs
are discarded.

6.13.3.2  The embryos from three or more spawns are pooled in a single
container to provide a sufficient number to conduct the tests.  These embryos
may be used immediately to start a test inhouse or may be transported for use

                                      117

-------
at a remote location.  When transportation is required, embryos should be
taken from the substrates within 12 h of spawning.  This permits off-site
tests to be started with less than 36-h old embryos.  Embryos should be
transported or shipped in clean, opaque, insulated containers, in well aerated
or oxygenated fresh culture or dilution water, 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 water, should be less than 2°C.   Sudden changes in
pH, dissolved ions, osmotic strength, and DO should be avoided.

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.

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.
Receiving water toxicity is determined with samples used directly as
collected or after samples are passed through a  60 /urn 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 one liter, or more, of sample per test day.

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%).  Improvements in precision
decline rapidly if the dilution factor is increased beyond 0.5 and precision
declines rapidly if a smaller dilution factor is 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%).  If a high rate of mortality  is observed during the first
1 to 2 h of the test, additional dilutions should be added at the lower range

                                      118

-------
of effluent concentrations.

10.1.2.3  The volume of effluent required for daily renewal of four replicates
per concentration, each containing 100 ml of test solution, is 1.5 L.
Sufficient test solution (approximately 1000 mL) is prepared at each effluent
concentration to provide 400 ml additional volume for chemical analyses.   If
the sample is used for more than one daily renewal of test solutions, the
volume must be increased proportionately.

10.1.2.4  Tests 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 the 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.

10.1.2.5  Just prior to test initiation (approximately one h) the temperature
of sufficient quantity of the sample to make the test solutions should be
adjusted to the test temperature and maintained at that temperature during the
addition of dilution water.

10.1.2.6  The DO of the test solutions should be checked prior to test
initiation.  If any of the solutions are supersaturated with oxygen or any
solution has a DO below 4.0 mg/L, all of the solutions and the control must be
gently aerated.

10.1.3  Dilution Water

10.1.3.1  Dilution water may be uncontaminated receiving water, a standard
synthetic (reconstituted) water, or some other uncontaminated natural water
(see Section 7, Dilution Water).

10.1.3.2  If the hardness of the test solutions (including the control) does
not equal or exceed 25 mg/L as CaCO,,  it may be necessary to adjust the
hardness by adding reagents for synthetic softwater as listed in Table 3,
Section 7.  In this case parallel tests should be conducted, one with the
hardness adjusted and one unadjusted.

10.2  START OF THE TEST

10.2.1  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 are used for each effluent test.  Each treatment
(including the control) should have four (minimum of three) replicates.

10.2.2  Tests performed in laboratories that have inhouse fathead minnow
breeding cultures must initiate tests with embryos less than 36 h old.  When
the embryos must be shipped to the test site from a remote location, it may be
necessary to use embryos older than 36 h because of the difficulty of
coordinating test organism shipments with field operations.  However,  in the
latter case, the embryos must not be more than 48 h old at the start of the
test and should all be within 24 h of the same age.


                                      119

-------
10.2.3  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.

10.2.4  The test organisms should come from a pool of embryos consisting of at
^ast three separate spawnings.  Gently agitate and mix the embryos to be used
in the test in a large container so that eggs from different spawns are
thoroughly mixed.

10.2.5  Using a small bore (2mm) glass tube, the embryos are placed one or two
it a time into each randomly arranged test chamber or intermediate container
in sequential order, until each chamber contains 15 (minimum of 10) embryos,
For a total of 60 (minimum of 30) embryos for each concentration (see Appendix
\).  The amount of water added to the chambers when transferring the embryos
to the compartments 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 new undamaged embryos.  Placing the test chambers on a light
table may facilitate examining and counting the embryos.

10.3  LIGHT, PHOTOPERIOD 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 of light and 8 h of darkness.  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 satisfactory DO concentrations.  The DO
concentrations should be measured in the new solutions at the start of the
test (Day 0) and before daily renewal of the new solutions on subsequent days.
The DO concentrations 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 concentrations 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 equivalent.   Care should be taken
to ensure that turbulence resulting from the aeration does not cause undue
physical  stress to the embryos.

10.5  FEEDING

10.5.1  Feeding is not required.
                                      120

-------
10.6  OBSERVATIONS DURING THE TEST

10.6.1  Minimum Routine Chemical and Physical Observations

10.6.1.1  DO is measured at the beginning and end of each 24-h exposure period
in at least one test chamber at each test concentrations and in the control

10.6.1.2  Temperature and pH are measured at the end of each 24-h exposure
period in at least one test chamber at each test concentration and in the
control.  Temperature should also be monitored continuously or observed and
recorded daily for at least two locations in the environmental control system
or the samples.  Temperature should be measured in a sufficient number of test
vessels, at least at the end of the test, to determine temperature variation
in the environmental chamber.

10.6.1.3  The pH is measured in the effluent sample each day before new test
solutions are made.

10.6.1.4  Conductivity, alkalinity and hardness are measured in each new
sample (100% effluent or receiving water) and in the control.

10.6.2  Record all the measurements on the data sheet (Figure 1).

10.6.3  Routine Biological Observations

10.6.3.1  At the end of the first 24 h of exposure, before renewing the test
solutions, examine the embryos.  Remove the dead embryos (milky colored and
opaque) and record the number (Figure 2).  If the rate of mortality (including
those with 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.

10.6.3.2  At 25°C,  hatching may begin on the fourth day.   After hatching
begins, count the number of dead and live embryos and the number of hatched,
dead, live, and deformed larvae, daily.  Deformed larvae are those with gross
morphological abnormalities such as lack of appendages,  lack of fusiform shape
(nondistinct mass), lack of mobility, a colored, beating heart in an opaque
mass, or other characteristics that preclude survival.  Count and remove dead
embryos and larvae as previously discussed and record the numbers for all  of
the test observations (Figure 2).  Upon hatching, deformed larvae are counted
as dead.

10.6.3.3  Protect the embryos and larvae from unnecessary disturbance during
the test by carrying out the daily test observations, solution renewals,  and
removal of dead organisms carefully.  Make sure the test organisms remain
immersed during the performance of the above operations.

10.7  DAILY CLEANING OF TEST CHAMBERS

10.7.1  Since feeding is not required, test chambers are not cleaned daily
unless accumulation of particulate matter at the bottom of the chambers
causes a problem.

                                      121

-------
10.8  TEST SOLUTION RENEWAL

10.8.1  Freshly prepared solutions are used to renew the tests daily.  For
on-site toxicity studies,  fresh effluent or receiving water samples should be
collected daily, and no more than 24 h should elapse between collection of the
samples and their use in the tests (see Section 8, Effluent and Receiving
Water Sampling, Sample Holding and Sample preparation for Toxicity Tests).
For off-site tests, a minimum of three samples are collected, preferably on
days one, three, and five.  Maintain the samples in the refrigerator at 4°C
until used.

10.8.2  The test solutions are renewed immediately after removing dead embryos
and/or larvae.  During the daily renewal process, 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 should be added slowly by pouring down the
side of the test chamber to avoid excessive turbulence and possible injury to
the embryos or larvae.

10.9  TERMINATION OF THE TEST

10.9.1  The test is terminated after seven days of exposure.  Count the
number of surviving, dead, and deformed larvae, and record the numbers of
each (Figure 2).  The deformed larvae are treated as dead in the analysis of
the data.  Keep a separate record of the total number and percent of deformed
larvae for use in reporting the teratogenicity of the test solution.

10.9.2  Prepare a summary of the data as illustrated in Figure 3.

11.  SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA

11.1  A summary of test conditions and test acceptability criteria is
presented in Table 1.

12.  ACCEPTABILITY OF TEST RESULTS

12.1  For the test results to be acceptable,  survival in the controls
must be at least 80%.

13.  DATA ANALYSIS

13.1  GENERAL

13.1.1  Tabulate and summarize the data (Figure 3).

13.1.2  The endpoints of this toxicity test are based on total mortality,
combined number of dead embryos, and dead and deformed larvae.  The EC1 is
calculated using Probit Analysis (Finney, 1971).  Separate analyses are
performed for the estimation of LOEC and NOEC endpoints and for the estimation
of the EC1 endpoint.  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 EC1 endpoint.  See the Appendices
for examples of the manual computations and examples of data input and output

                                      122

-------
  TABLE  1.   SUMMARY  OF  TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA FOR
             FATHEAD  MINNOW, PIMEPHALES PROMELAS,  EMBRYO-LARVAL SURVIVAL AND
             TERATOGENICITY TOXICITY TESTS WITH EFFLUENTS AND RECEIVING
             WATERS
 1.  Test  type:
 2.  Temperature:
 3.  Light quality:
 4.  Light intensity:

 5.  Photoperiod:
 6.  Test  chamber  size:
 7.  Test  solution volume:
 8.  Renewal  of  test concentration:
 9.  Age of test organisms:

10.  No. embryos per test chamber:
11.  No. replicate test
    chambers per  concentration:
12.  No. embryos per concentration:
13.  Feeding regime:
14.  Aeration:
15.  Dilution water:
Static renewal
25 ± 1°C
Ambient laboratory illumination
10-20 /iE/m2/s or 50-100  ft-c  (ambient
laboratory levels)
16 h light, 8 h dark
150 mL (Minimum)
70 mL (Minimum)
Daily
Less than 36-h  old embryos (Maximum of
48-h if shipped)
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 receiving or
other natural water, synthetic water
prepared using  MILLIPORE MILLI-QR or
equivalent deionized water and reagent
grade chemicals or DMW (see Section
7).  The hardness of the test solutions
should equal or exceed 25 mg/L (CaC03) to
ensure hatching success
16.  Effluent test concentrations:    5 and a control
                                      123

-------
    TABLE 1.  SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA FOR
              FATHEAD MINNOW, PIHEPHALES PROMELAS, EMBRYO-LARVAL SURVIVAL AND
              TERATOGENICITY TOXICITY TESTS WITH EFFLUENTS AND RECEIVING
              WATERS (CONTINUED)
17. Test dilution factor:


18. Test duration:

19. Endpoint:


20. Test acceptability criteria:

21. Sampling requirement:
22.  Sample volume required:
Effluents:  > 0.5
Receiving waters:

7 days
None, or > 0.5
Combined mortality (dead and deformed
organisms)

80% or greater survival in controls

For on-site tests, samples 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 collected on days one, three,
and five with a maximum holding time of
36 h before first use.

1.5 to 2.5 L/day depending on volume of
test solutions used
                                     124

-------
      Figure 1
Discharger:
Location:
Data form for the fathead minnow, Pimephales promelas,
embryo-larval survival and teratogenicity test.  Routine
chemical and physical determinations.
                         Analyst:
                         Dates:
                                     Dav
Control :
Temo.
D.O. Initial
Final
pH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










2










3










4










5










6










7










Remarks










                                     Day
Cone:
Terno.
D.O. Initial
Final
pH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










2










3










4










5










6










7










Remarks










                                     Dav
Cone:
Temo.
D.O. Initial
Final
oH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










2










3










4










5










6










7










Remarks










                                      125

-------
      Figure 1.   Data form for the fathead minnow,  Pimephales promelas,
                 embryo-larval  survival  and teratogenicity test.  Routine
                 chemical  and physical  determinations (Continued).
Discharger:
Location:
Analyst:
Dates:
Dav
Cone:
Terno.
D.O. Initial
Final
pH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










2










3










4










b










6










7










Remarks










                                     Dav
Cone:
Temo.
D.O. Initial
Final
pH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










, 2










3










4










5










6










7










Remarks










                                     Day
Cone:
Terno.
D.O. Initial
Final
oH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










2










3










4










5










6










7










Remarks










                                      126

-------
Discharger:
Location: _
            Figure  2.   Data form for the fathead minnow, Pimephales promelas,
                       embryo-larval survival and teratogenicity test. Survival
                       and terata data.
Test Dates:.
Analyst: 	
Condition of
Cone: Rep. Embryos/larvae
No.
Control: 1 Live/dead
Terata
2 Live/dead
Terata
3 Live/dead
Terata
4 Live/dead
Terata
Treatment: 1 Live/dead
Terata
2 Live/dead
Terata
3 Live/dead
Terata
4 Live/dead
Terata
Treatment: 1 Live/dead
Terata
2 Live/dead
Terata
3 Live/dead
Terata
4 Live/dead
Terata
Treatment: 1 Live/dead
Terata
2 Live/dead
Terata
3 Live/dead
Terata
4 Live/dead
Terata
1
































2
































Day
3
































4
































5
































6
































7

































































                                          127

-------
    Figure 2.  Data form for the fathead minnow, Pimephales promelas,
               embryo-larval survival and teratogenicity test. Survival
               and terata data (Continued).

                               	   Test Dates:	
Discharger:

Location:
                                          Analyst:
Condition of
Cone: Rep. Embryos/larvae
No.
Treatment: 1 Live/dead
Terata
2 Live/dead
Terata
3 Live/dead
Ferata
4 Live/dead
Terata
Treatment: 1 Live/dead
Terata
2 Live/dead
Terata
3 Live/dead
Terata
4 Live/dead
Terata
1
















2
















Day
3
















4
















5
















6
















7

































Comments:
                                     128

-------
      Figure 3.  Summary data for the fathead minnow, Pimephales promelas,
                 embryo-larval survival and teratogenicity test.
Discharger:

Location:
Test Dates:
     Analyst:
Treatment
No. dead embryos
and larvae
No. terata
Total mortality
(dead and
deformed)
Total mortality
(%)
Terata (%)
Hatch (%)
Control
















































Comments:
                                      129

-------
for the computer programs.

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 FATHEAD MINNOW EMBRYO-LARVAL SURVIVAL AND
      TERATOGENICITY DATA

13.2.1  Formal statistical analysis of the total mortality data is outlined on
the flowchart in Figure  4.  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 endpoint.  Concentrations at which there is 100% total
mortality in all of the  test chambers are excluded from statistical analysis
of the NOEC and LOEC, but included in the estimation of the EC1 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 (see Appendix
D).  The Wilcoxon Rank Sum Test with the Bonferroni adjustment is the
nonparametric alternative (see Appendix F).

13.2.4  Probit Analysis  (Finney, 1971) is used to estimate the concentration
that causes a specified  percent decrease in total survival from the control.
In this analysis, the total mortality data from all test replicates at a given
concentration are combined.

13.2.5  The data for this example are listed in Table 2.  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 effluent concentration and control are
listed in Table 3.  A plot of the data is provided in Figure 5.  Since there
is 100% total mortality  in both replicates for the 16.0% concentration, it  is
not included in this statistical analysis and is considered a qualitative
mortality effect.
                                      130

-------
       STATISTICAL ANALYSIS OF FATHEAD MINNOW EMBRYO-LARVAL
                 SURVIVAL AND TERATOGENECITY TEST
                           TOTAL MORTALITY:
                     TOTAL NUMBER OF DEAD EMBRYOS,
                    DEAD LARVAE, AND DEFORMED LARVAE
 ENDPOINT ESTIMATE
        EC1
ARC SINE
TRANSFORMATION
i
i
        SHAPIRO-WILK'S TEST
                                             NON-NORMAL DISTRIBUTION
             NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
                            BARTLETTS 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 4.  Flowchart for statistical  analysis of fathead
                 minnow,  Pimephales promelas, embryo-larval data,
                                131

-------
           TABLE 2.   DATA FROM FATHEAD MINNOW,  PIMEPHALES PROMELAS,
                     EMBRYO-LARVAL TOXICITY TEST WITH TRICKLING FILTER WASTE

             A.  REPLICATES A AND B (USED IN DUNNETT'S PROCEDURE)
Repl.
A





B





Effl. No. Dead at
Cone. Eggs at Hatching
(%) Start No. (%)
Cont.
3
5
7
11
16
Cont.
3
5
7
11
16
51
50
52
50
50
49
49
50
50
50
49
50
5
5
5
2
10
39
9
6
10
6
30
29
10
10
10
4
20
80
18
12
20
12
61
58
Dead + Deform.
at Hatching
No. (%)
6
5
6
8
25
39
9
6
10
10
37
34
12
10
12
16
50
80
18
12
20
20
76
68
. Dead at Test
Termination
No. (%)
6
5
5
9
17
49
10
9
10
16
33
45
12
10
10
18
34
100
20
18
20
32
66
90
Dead + Deform.
at Termination
No. (%)
7
5
6
15
32
49
10
9
10
20
40
50
14
10
12
30
64
100
20
18
20
40
82
100
     B.  COMBINED DATA FROM REPLICATES A AND B (USED IN PROBIT ANALYSIS)
Repl.   Effl.    No.     Dead at   Dead + Deform.
       Cone.  Eggs at  Hatching   at Hatching
        (%)    Start   No.   (%)    No.   (%)
Dead at Test  Dead + Deform.
Termination   at Termination
No.  (%)        No.  (%)
A&B Cont.
3
5
7
11
16
100
100
102
100
99
99
14
11
15
8
40
68
14
11
15
8
40
69
15
11
16
18
62
73
15
11
16
18
62
74
16
14
15
25
50
94
16
14
15
25
50
95
17
14
16
35
72
99
17
14
16
35
73
100
                                      132

-------
TABLE 3.  FATHEAD MINNOW, PIMEPHALES PROMELAS, EMBRYO-LARVAL TOTAL
          MORTALITY DATA
Replicate
                                Effluent Concentration (%)
Control
3.0
5.0
7.0
11.0
16.0
RAW
ARC SINE
TRANS-
FORMED


A
B
A
B
M|AN(Yf)
i
1
0
0
0
0
0
0
1
.14
.20
.384
.464
.424
.003
0
0
0
0
0
0
2
.10
.18
.322
.438
.380
.007
0.12
0.20
0.354
0.464
0.409
0.006
3
0.30
0.40
0.580
0.685
0.632
0.006
4
0
0
0
1
1
0
5
.64
.82
.927
.133
.030
.021
1.0
1.0
-

                             133

-------
        0.0 i
                                                   CONNECTS THE MEAN VALUE FOR EACH CONCENTRATION
                                                   REPRESENTS THE CRITICAL VALUE  TOR DUNNETT'S TEST
                                                   (ANY MEAN OF TOTAL MORTALITY ABOVE THIS  VALUE WOULD
                                                    BE SIGNIFICANTLY DIFFERENT FROM THE  CONTROL)
                                              EFFLUENT CONCENTRATION («)
Figure  5.   Plot of fathead minnow, Pimephales promelas,  total mortality data from  the  embryo-larval
            test.

-------
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
       df
Sum of Squares
     (SS)
Mean Square(MS)
    (SS/df)

Between

Within
Total

P -

N -
N

1

P
1

SSB

SSW
SST
2
SB = SSBX(p-l)
2
Su = SSW/(N-p)

Where:      p  = number of effluent concentration levels including the
                 control
            N  = total number of observations n1  + n,  ... +np
            n,-  = number of observations in concentration i
SSB = Z V/n, - G2/N
                                         Between Sum of Squares
      P   ni   ,    ,
SST = Z   Z Y;i2 - G2/N
                                         Total Sum of Squares
           SSW = SST - SSB
                               Within Sum of Squares
                                      135

-------
                                                                   p
            G  =  the  grand  total  of all  sample observations,  6=21,-
                                                                  i=l
            T,. = the total  of the replicate  measurements for
                  concentration  "i"
           YJJ  =  the jth observation for concentration  "i"  (represents
                  the  proportion of  total  mortality for effluent
                  concentration  i  in test  chamber j)

13.2.7.2  For the data  in this  example:

    n1 = n2  =  n3 = n4 =  n5 = 2
    N  = 1 Q
    T, = Ylt + Y12 = 0.848
    T2 = Y21 + Y22 = 0.760
    T, = Y31 + Y32 = 0.818
    T4 = Y41 + Y42 = 1.265
    T5 = Y   + Y   = 2.060
= Y51 +  Y52
= T, + T2 +
    G  = T  + T  + T + T  + T  = 5.751
    SSB = X I,-2/!!,-  - G2/N
        = _J_(7.810) -  (5.751)2  = 0.598
           2              10
               ,
    SST = S   s Y,,2  - G2/N
         1=1 j-1

        = 3.948 - (5.751)2  = 0.640
                     10

    SSW = SST - SSB = 0.640 - 0.598 = 0.042

    S2  = SSB/(p-l)  = 0.598/(5-l) = 0.1495

    Sj  = SSW/(N-p)  - 0.042/(10-5) = 0.008
13.2.7.3  Summarize these calculations  in  an  ANOVA table (Table 5).
                                      136

-------
          TABLE 5.   ANOVA TABLE FOR DUNNETT'S PROCEDURE EXAMPLE
Source
Between
Within
Total
df
4
5
9
Sum of Squares
(SS)
0.598
0.042
0.640
Mean Square(MS)
(SS/df)
0.1495
0.008

13.2.7.4  To perform the individual comparisons, calculate the t statistic for
each concentration,  and control combination as follows:

                              Su V (1/n,) +  (1/n,)
Where Y,-   = mean  proportion of total  mortality for concentration i
      Y1   = 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
each concentration.

13.2.7.5   Table 6 includes the calculated t values for each concentration
and control  combination.  In this example, comparing the 3.0%
concentration with  the control, the calculation is as follows:


                            ( 0.380 - 0.424 )
                  t2 = 	 - - 0.494
                       [ 0.089 V (1/2) + (1/2) ]
                                      137

-------
                      TABLE 6.   CALCULATED T VALUES
           Effluent Concentration (%)         i           t,
3.0
5.0
7.0
11.0
2
3
4
5
-0.494
-0.168
2.337
6.809
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 11.0% concentration has  a significantly higher mean proportion
of total mortality than the control.   Hence the NOEC is 7.0% and the LOEC
is 11.0%.

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 V (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.

13.2.7.8  In this example:
                   MSD = 2.85 (0.089) V (1/2) + (1/2)
                       = 2.85 (0.089)(1.0)
                       * 0.254

13.2.7.9  The MSD (0.254) is in transformed units.  To determine the MSD
in terms of percent total mortality, carry out the following conversion.

    1. Add the MSD to the transformed control mean.


                            0.424 + 0.254 = 0.678


                                      138

-------
    2.  Obtain the untransformed values for the control mean and the sum
       calculated in 1.

                       [Sine ( 0.424) ]2 = 0.169
                       [Sine ( 0.678) ]2 = 0.393

    3.  The untransformed MSD (MSD )  is determined by subtracting the
       untransformed values from 2.

                      MSD,, = 0.393 -  0.169 = 0.224
                         u
13.2.7.10  Therefore,  for this set of data, the minimum difference in
mean proportion of total  mortality between the control and any effluent
concentration that can be detected as statistically significant is 0.224.

13.2.7.11  This represents a 132% 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 effluent concentration with 100% mortality in
both replicates is considered.  To perform the Probit Analysis, run the EPA
Probit Analysis Program provided in Appendix I.  Examples of the program
output are illustrated in Table 8 and Figure 6.

12.2.8.2  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
                                     Effluent Concentration (%)
                  Control    3.0     5.0    7.0    11.0   16.0
Number Dead
Number Exposed
17
100
14
100
16
102
35
100
72
99
99
99
                                      139

-------
    TABLE 8.  OUTPUT FOR  EPA  PROBIT  ANALYSIS PROGRAM, VERSION 1.4
    Cone.
 Number
Exposed
Control
3.0000
5.0000
7.0000
11.0000
16.0000
100
100
102
100
99
99
    Number
    Resp.

        17
        14
        16
        35
        72
        99
 Observed
Proportion
Responding

  0.1700
  0.1400
  0.1569
  0.3500
  0.7273
  1.0000
            Mjusted
           Proportion
           Responding

             0.0000
             -.0190
             0.0010
             0.2298
             0.6769
             1.0000
         Predicted
         Proportion
         Responding

           0.1560
           0.0000
           0.0174
           0.1765
           0.7449
           0.9759
Chi - Square Heterogeneity =     5.286
Mu
Sigma

Parameter
    0.959956
    0.123640

    Estimate
        Std. Err.
             95% Confidence Limits
Intercept
Slope

Spontaneous
Response Rate
   -2.764127
    8.088003

    0.156014
        1.002530
        0.990954

        0.022593
       (   -4.729086,
       (    6.145732,

       (    0.111732,
                      -0.799168)
                      10.030273)

                       0.200296)
      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.
        4.7025
        5.7093
         ,3314
         .7892
         ,1192
       12.2489
       13.1345
       14.5657
       17.6840
6.
6.
9.
                  Lower       Upper
                95% Confidence Limits
    3.
    4.
    5.
    5.
  .6073
  .6408
  .3031
  .7994
 8.3614
11.4157
12.1697
13.3302
15.7134
7.
7.
 5.5567
 6.5196
  .1058
  .5354
 9.7763
13.3942
14.5708
16.5676
21.2145
                                     140

-------
Probit Analysis  of  Fathead Minnow Embryo-Larval Survival  and Teratogenecity Data


        PLOT  OF  ADJUSTED PROBITS AND PREDICTED REGRESSION LINE


Probit
   10+
    9+
    8+
    7+
    6+
    3+
    2+
    1+
    O+o  o
      -+	^	+	+	H	1	+_

      ECOl           EC10     EC25      EC50       EC75     EC90           EC99
   Figure 6.  Plot  of adjusted  probits and predicted regression line.

                                       141

-------
14.  PRECISION AND ACCURACY

14.1  PRECISION

14.1.1  Data shown in Tables 9 and 10 indicate that the precision of the
embryo-larval  survival and teratogenicity test, expressed as the relative
standard deviation (or coefficient of variation, CV) of the LCI values,
was 62% for cadmium (Tables 9) and 41% for Diquat (Table 10).

14.1.2  Precision data are also available from four embryo-larval
survival and teratogenicity tests on trickling filter pilot plant
effluent (Table 11).   Although the data could not be analyzed by Probit
Analysis,  the  NOECs and LOECs obtained using Dunnett's Test were the same
for all  four tests, 7% and 11% effluent,  respectively, indicating maximum
precision  in terms of the test design.

14.2  ACCURACY

14.2.1  The accuracy  of toxicity tests cannot be determined.
                                     142

-------
  TABLE 9.   PRECISION OF THE FATHEAD MINNOW, PIMEPHALES PROMELAS,
            EMBRYO-LARVAL SURVIVAL AND TERATOGENICITY TEST, USING
            CADMIUM AS A REFERENCE TOXICANT8'6
Test
1
2
3
4
5
N
Mean
SD
CV(%)
LC1C
(mg/L)
0.014
0.006
0.005
0.003
0.006
5
0.0068
0.0042
62
95% Confidence NOECd
Limits (mg/L)
0.009 0.018 0.012
0.003 - 0.010 0.012
0.003 - 0.009 0.013
0.002 0.004 0.011
0.003 0.009 0.012
5
NA
NA
aTests  conducted  by Drs.  Wesley Birge and Jeffrey Black,
 University of Kentucky,  Lexington, under a cooperative
 agreement with the Aquatic Biology Branch, Environmental
 Monitoring Systems Laboratory, U. S. Environmental
 Protection Agency, Cincinnati, Ohio.
bCadmium chloride was  used as the reference toxicant.
 The nominal concentrations, expressed as cadmium (mg/L), were:
 0.01,  0.032, 0.100, 0.320, and 1.000.  The dilution water was
 reconstituted water with a hardness of 100 mg/L as calcium
 carbonate, and a pH of 7.8.
Determined by Probit  Analysis.
dHighest no-observed-effect concentration determined
 by independent statistical analysis (2X2 Chi-square Fisher's
 Exact Test).  NOEC range of 0.011 - 0.013 represents a difference
 of one exposure concentration.
                                  143

-------
TABLE 10.  PRECISION OF THE FATHEAD MINNOW, PIMEPHALES
           PROMELAS, EMBRYO-LARVAL, SURVIVAL AND
           TERATOGENICITY TOXICITY TEST, USING DIQUAT
           AS A REFERENCE TOXICANT8'6
Test

1
2
3
4
5
N
Mean
SD
CV(%)
LC1C
(mg/L)
0.58
2.31
1.50
1.71
1.43
5
1.51
0.62
41.3
95% Confidence
Limits
0.32 0.86
d
1.05 1.87
1.24 2.09
0.93 - 1.83




  aTests  conducted  by Drs. Wesley  Birge  and Jeffrey  Black,
   University of Kentucky,  Lexington,  under  a cooperative
   agreement with the Aquatic Biology  Branch, Environmental
   Monitoring Systems Laboratory,  U. S.  Environmental
   Protection Agency, Cincinnati,  Ohio.
   The Diquat  concentrations  were  determined  by  chemical
   analysis.  The dilution  water was reconstituted water
   with  a hardness  of 100 mg/L as  calcium carbonate, and
   a pH  of 7.8.
  cDetermined  by  Probit  analysis.
  dCannot  be calculated.
                             144

-------
TABLE 11.  PRECISION OF THE FATHEAD MINNOW, PIMEPHALES PROMELAS,
           EMBRYO-LARVAL SURVIVAL AND TERATOGENICITY STATIC-RENEWAL
           TEST CONDUCTED WITH TRICKLING FILTER EFFLUENTa'b'c
Test
No.
1
2
3
4
NOEC
(% Effl)
7
7
7
7
LOEC
(% Effl)
11
11
11
11
         a
          Data provided by Timothy Neiheisel, Aquatic
          Biology Branch,  Environmental Monitoring
          Systems Laboratory, U. S.  Environmental
          Protection Agency, Cincinnati, Ohio.
         bEffluent concentrations used:  3, 5, 7, 11 and 16%
         cMaximum precision achieved in terms of
          NOEC-LOEC interval.   For a discussion  of  the
          precision of data from chronic toxicity tests
          see Section 4,  Quality Assurance.
                                145

-------
                                 SECTION 13

                                 TEST METHOD

        DAPHNID,  CERIODAPHHIA DUBIA, SURVIVAL AND REPRODUCTION TEST
                                METHOD 1002.0


1.   SCOPE AND APPLICATION

1.1  This method  measures the chronic toxicity of effluents and receiving
water to the daphnid,  Ceriodaphm'a dubia, using less than 24 h old neonates
during a three-brood (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 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, and 96-h LCSOs).

1.3  Detection limits  of the toxicity of an effluent or pure substance are
organism dependent.

1.4  Brief 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
degradable or highly volatile toxicants in the source may not be detected in
the test.

1.5  This test 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  Ceriodaphm'a dubia are exposed in a static renewal system to different
concentrations of effluent, or to receiving water,  until 60% of surviving
control organisms have three broods of offspring or for a maximum of seven
days.  Test results are based on survival and reproduction.  If the test is
conducted as described, the control organisms should produce 20 young in three
broods during a seven-day period and at least 80% of the controls should have
a brood.

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
(see Section 8,  Effluent and Receiving Water Sampling, Sample Handling, and

                                      146

-------
Sample Preparation for Toxicity Tests).

3.3  Pathogenic and/or predatory organisms in the dilution water and effluent
may affect test organism survival and confound test results.

3.4  The amount and type of natural food in the effluent or dilution water may
confound test results.

3.5  Food added during the test may sequester metals and other toxic
substances and confound test results.  Daily renewal of solutions, however,
will reduce the probability of reduction of toxicity caused by feeding.

4.  SAFETY

4.1  See Section 3, Health and Safety.

5.  APPARATUS AND EQUIPMENT

5.1  Ceriodaphm'a and algal culture units -- See Ceriodaphnia and algal
culturing methods below and algal culturing methods in Section 14 as well as
USEPA, 1991b.

5.2  Samplers -- automatic sampler, preferrably with sample cooling
capability, capable of collecting a 24-h composite sample of 2 L or more.

5.3  Sample containers -- for sample shipment and storage (See Section 8,
Effluent and Receiving Water Sampling, Sample Handling, and Sample Preparation
for Toxicity Tests).

5.4  Environmental chambers, incubators, or equivalent facilities with
temperature control (25 ± 1°C).

5.5  Water purification system -- MILLIPORE MILLI-QR deionized water or
equivalent (see Section 5, Facilities, Equipment, and Supplies).

5.6  Balance -- analytical, capable of accurately weighing 0.00001 g.

5.7  Reference weights, Class S -- for checking performance of balance.
Weights should bracket the expected weights of the weighing pans and the
expected weights of the weighing pans plus the daphnids.

5.8  Test Vessels -- 10 test vessels are required for each concentration and
control.  Test vessels such as 30-mL borosilicate glass beakers or disposable
polystyrene cups are recommended because they will fit in the viewing  field of
most stereoscopic microscopes.  The glass beakers and plastic cups are rinsed
thoroughly with dilution water before use.  To avoid potential contamination
from the air and excessive evaporation of the test solutions during the test,
the test vessels should be covered with safety glass plates or sheet plastic
(6 mm thick).

5.9  Mechanical shaker or magnetic stir plates -- for algal cultures.


                                      147

-------
5.10  Light meter --  with a range of  0-200 \it/m2/s  (0-1000 ft-c).

5.11  Fluorometer (optional)  --  equipped with chlorophyll detection light
source,  filters,  and  photomultiplier tube (Turner Model 110 or equivalent).

5.12  UV-VIS spectrophotometer (optional) -- capable of accommodating 1-5 cm
cuvettes..

5.13  Cuvettes for spectrophotometer --  1-5 cm light path.

5.14  Electronic  particle counter (optional) --  Coulter Counter, ZBI, or
equivalent, with  mean cell  (particle) volume determination.

5.15  Microscope  with 10X,  45X,  and 100X objective lenses, 10X ocular
lenses,  mechanical  stage, substage condenser, and light source (inverted or
conventional microscope)  -- for  determining sex and  verifying identification.

5.16  Dissecting  microscope,  stereoscopic,  with zoom objective, magnification
to 50X --  for examining and counting the neonates in the test vessels.

5.17  Counting chamber -- Sedgwick-Rafter,  Palmer-Maloney, or hemocytometer.

5.18  Centrifuge  (optional) -- plankton, or with swing-out buckets having a
capacity of 15-100 ml.

5.19  Centrifuge  tubes -- 15-100 ml, screw-cap.

5.20  Filtering apparatus --  for membrane and/or glass fiber filters.

5.21  Racks (boards)  -- to hold  test chambers.  It is convenient to use a
piece of styrofoam insulation board, 50  cm x 30 cm x 2.5 cm (20 in x 12 in x 1
in), drilled to hold  60 test  chambers,  in six rows of 10 (see Figure 1 in this
Section).

5.22  Light box -- for illuminating organisms during examination.

5.23  Volumetric  flasks and graduated cylinders -- class A, borosilicate glass
or non-toxic plastic  labware, 10-1000 mL, for culture work and preparation of
test solutions.

5.24  Pipettors,  adjustable volume repeating dispensers -- for feeding.
Pipettors  such as the Gil son  REPETMAN*,  Eppendorf, Oxford,  or equivalent,
provide  a  rapid and accurate  means of dispensing small volumes (0.1 mL) of
food to  large numbers of test chambers.

5.25  Volumetric  pipets -- class A, 1-100 mL.

5.26  Serological pipets -- 1-10 mL, graduated.

5.27  Pipet bulbs and fillers -- PROPIPETR,  or equivalent.

5.28  Disposable  polyethylene pipets, droppers, and glass tubing with

                                     148

-------
fire-polished edges, > 2-mm ID -- for transferring organisms.
5.29  Wash bottles -- for rinsing small glassware and instrument electrodes
and probes.
5.30  Glass or electronic thermometers -- for measuring water temperatures.
5.31  Bulb-thermograph or electronic-chart type thermometers -- for
continuously recording temperature.
5.32  National Bureau of Standards Certified thermometer -- see EPA Method
170.1, USEPA 1979b.
5.33  Meters, DO, pH, and specific conductivity -- for routine physical and
chemical  measurements.
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 recording the data.
6.3  Vials, marked -- for preserving specimens for verification (optional).
6.4  Tape, colored -- for labeling test vessels.
6.5  Markers, water-proof -- for marking containers.
6.6  Reagents for hardness and alkalinity tests (see EPA Methods 130.2 and
310.1, USEPA 1979b).
6.7  Standard pH buffers 4, 7, and 10 (or as per instructions of instrument
manufacturer) for instrument calibration (see USEPA Method 150.1, USEPA
1979b).
6.8  Specific conductivity standards (see EPA Method 120.1,  USEPA 1979b).
6.9  Membranes and filling solutions for DO probe (see Method 360.1, USEPA,
1979b), or reagents for modified Winkler analysis.
6.10  Laboratory quality control samples and standards -- for calibration of
the above methods.
6.11  Reference toxicant solutions (see Section 4, Quality Assurance).
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, surface water, and dilution water -- see Section 7, Dilution
                                      149

-------
Water, and Section 8,  Effluent and Receiving Water Sampling, Sample
Handling, and Sample Preparation for Toxicity Tests.

6.14  Trout chow,  yeast,  and CEROPHYL"  (or  substitute food)  --  for feeding the
cultures and test  organisms.

6.14.1  Digested trout chow, or substitute  flake food (TETRAMINR), is prepared
as follows:

  1. Preparation of trout chow or substitute flake food requires one week.
     Use starter or No.  1 pellets prepared  according to current U.S. Fish and
     Wildlife Service  specifications.  Suppliers of trout chow include
     Zeigler Bros., Inc., P.O. Box 95,  Gardners, Pennsylvania,  17324
     (717-780-9009); Glencoe Mills,  1011 Elliott, Glencoe, Minnesota, 55336
     (612-864-3181); and  Murray Elevators,  118 West 4800 South, Murray, Utah
     84107 (800-521-9092).
  2. Add 5.0 g of  trout  chow pellets or substitute flake food to 1 L of
     MILLI-Q" water.   Mix well  in a  blender and  pour into a  2-1 separatory
     funnel.  Digest prior  to use by aerating continuously from the bottom of
     the vessel  for one  week at ambient laboratory temperature.  Water lost
     due to evaporation  is  replaced  during  digestion.  Because of the
     offensive odor usually produced during digestion, the vessel should be
     placed in a fume  hood  or other  isolated, ventilated area.
  3. At the end  of digestion period, place  in a refrigerator and allow to
     settle for  a  minimum of 1 h.  Filter the supernatant through a fine mesh
     screen (i.e., NITEXR 110 mesh).  Combine with equal  volumes of
     supernatant from CEROPHYLLR and yeast  preparations (below).   The
     supernatant can be  used fresh,  or  frozen until use.  Discard the
     sediment.

6.14.2  Yeast is prepared as follows:

  1. Add 5.0 g of  dry  yeast, such as FLEISCHMANN'SR (or Lake State Yeast)  to
     1 L of MILLI-QR water.
  2. Stir with a magnetic stirrer, shake vigorously by hand, or mix with a
     blender at  low speed,  until the yeast  is well dispersed.
  3. Combine the yeast suspension immediately (do not allow to settle) with
     equal volumes of supernatant from  the  trout chow (above) and CEROPHYLLR
     preparations  (below).   Discard  excess  material.
6.14.3  CEROPHYLL* is prepared as follows:

  1. Place 5.0 g of dried, powdered, cereal leaves (or substitute alfalfa
     rabbit pellets) in a blender.  (Available as "CEREAL LEAVES," from Sigma
     Chemical Company, P.O. Box 14508, St.  Louis, Missouri, 63178,
     (800-325-3010); or as CEROPHYLLR, from Ward's Natural  Science
     Establishment, Inc., P.O. Box 92912, Rochester, New York, 14692-9012,
     (716-359-2502).  Dried, powdered, alfalfa leaves, obtained from  health
     food stores, have been found to be a satisfactory substitute for cereal
     leaves.
  2. Add 1 L of MILLI-QR water.
  3. Mix in a blender at high speed for 5 min, or stir overnight at medium

                                      150

-------
     speed  on  a magnetic stir plate.
  4.  If a blender is used to suspend  the material,  place in a refrigerator
     overnight to settle.  If a magnetic stirrer is used,  allow to settle for
     1  h.  Decant the supernatant and combine with  equal volumes of
     supernatant from trout chow and  yeast preparations (above).  Discard
     excess material.

6.14.4   Combined yeast-cerophyl-trout chow (YCT) is mixed  as follows:

  1.  Thoroughly mix equal (approximately 300 ml) volumes of the three foods as
     described above.
  2.  Place  aliquots of the mixture in small  (50 ml  to 100  mL) screw-cap
     plastic bottles and freeze until needed.
  3.  Freshly prepared food can be used immediately, or it  can be frozen until
     needed.  Thawed food is stored in the refrigerator between feedings, and
     is used for a maximum of one week.   Do  not store frozen over three
     months.
  4.  It is  advisable to measure the dry weight of solids in each batch of YCT
     before use.  The food should contain 1.7 - 1.9 g solids/L.  Cultures or
     test solutions should contain 12-13 mg  solids/L.

6.15  Algal Food

6.15.1   Algal  Culture Medium is prepared as  follows:

  1.  Prepare (five) stock nutrient solutions using  reagent grade chemicals as
     described in Table 1.
  2.  Add 1  ml  of each stock solution, in the order  listed  in Table 1, to
     approximately 900 ml of MILLI-Q   water.   Mix well  after the addition of
     each solution.  Dilute to 1 L, mix well.  The  final concentration of
     macronutrients and micronutrients in the culture medium is given in
     Table  2.
  3.  Immediately filter the medium through a 0.45 ^m pore  diameter membrane at
     a  vacuum  of not more than 380 mm (15 in.) mercury, or at a pressure of
     not more  than one-half atmosphere (8 psi).  Wash the  filter with 500 mL
     deionized water prior to use .
  4.  If the filtration is carried out with sterile  apparatus, filtered medium
     can be used immediately, and no  further sterilization steps are required
     before the inoculation of the medium.  The medium can also be sterilized
     by autoclaving after it is placed in the culture vessels.
  5.  Unused sterile medium should not be stored more than  one week prior to
     use, because there may be substantial loss of  water by evaporation.

6.15.2   Algal  Cultures

6.15.2.1  See  Section 6, Test Organisms, for information on sources of
"starter" cultures of Selenastrum capricornutum, S. minutum, and Chlamydomonas
reinhardti.

6.15.2.2  Two  types of algal cultures are maintained:  (1) stock cultures,
and (2) "food" cultures.


                                      151

-------
   TABLE 1.  NUTRIENT STOCK SOLUTIONS FOR MAINTAINING ALGAL STOCK  CULTURES
             AND TEST CONTROL CULTURES.
STOCK COMPOUND
SOLUTION
1. MACRONUTRIENTS
A. MgCl2-6H20
CaCl2-2H20
NaN03
B. MgS04-7H20
C. K2HP04
D. NaHC03
2. MICRONUTRIENTS
H,B03
MnCl2 4H20
ZnCl2
FeCl,-6H20
CoCl2-6H20
Na2MoO,-2H20
CuCl2-2H20
Na2EDTA-2H20
Na2Se04
AMOUNT DISSOLVED IN
sooML MILLI-Q" WATER

6.08 g
2.20 g
12.75 g
7.35 g
0.522 g
7.50 g

92.8 mg
208.0 mg
1.64 mga
79 . 9 mg
0.714 mgb
3.63 mgc
0.006 mgd
1 50 . 0 mg
1.196 mge
aZnC!2   Weigh out  164 mg  and dilute  to  100 mL.   Add  1  mL  of this
 solution to Stock #1.

bCoCl2-6H,0 - Weigh out 71.4 mg and dilute to 100 mL.   Add  1  mL  of
 this solution to Stock #1.

cNa,Mo04-2H20 - Weigh out 36.6 mg and dilute to 10 mL.  Add 1 mL
 of this solution to Stock #1.

dCuCl2-2H20   Weigh out 60.0 mg and dilute to 1000 mL.  Take  1 mL
 of this solution and dilute to 10 mL.  Take 1 mL of the second  dilution
 and add to Stock #1.

6Na2Se04 - Weigh out  119.6 mg and dilute to 100 mL.  Add 1  mL of this
 solution to Stock #1.
                                      152

-------
TABLE 2.  FINAL CONCENTRATION OF MACRONUTRIENTS AND MICRONUTRIENTS
          IN THE CULTURE MEDIUM
MACRONUTRIENT
NaN03
MgCl2-6H20
CaCl2-2H20
MgS04-7H20
K2HP04
NaHC03


MICRONUTRIENT
H3B03
MnCl2-4H20
ZnCl2
CoCl2-6H20
CuCl2-2H20
Na2Mo04-2H20
FeCl3-6H20
Na2EDTA-2H20
Na,SeO,
CONCENTRATION
(ma/L)
25.5
12.2
4.41
14.7
1.04
15.0


CONCENTRATION
(UQ/U
185
416
3.27
1.43
0.012
7.26
160
300
2.39
ELEMENT
N
Mg
Ca
S
P
Na
K
C
ELEMENT
B
Mn
Zn
Co
Cu
Mo
Fe
--
Se
CONCENTRATION
(ma/L)
4.20
2.90
1.20
1.91
0.186
11.0
0.469
2.14
CONCENTRATION
(ua/L)
32.5
115
1.57
0.354
0.004
2.88
33.1
	
0.91
                                 153

-------
6,15.2.2.1   Establishing and Maintaining Stock Cultures of Algae:

  1.  Upon receipt of the "starter"  culture (usually about 10 ml), a stock
     culture is initiated by aseptically transferring one milliliter to each
     of several 250-mL culture flasks containing 100 ml algal culture medium
     (prepared as described above).   The remainder of the starter culture can
     be held in reserve for up to six months in a refrigerator (in the dark)
     at 4°C.
  2.  The stock cultures are used as  a source of algae to initiate "food"
     cultures for Ceriodaphm'a dubia toxicity tests.  The volume of stock
     culture maintained at any one  time will depend on the amount of algal
     food required for the Ceriodaphm'a dubia cultures and tests.  Stock
     culture volume may be rapidly  "scaled up" to several liters, if
     necessary, using 4-L serum bottles or similar vessels, each containing
     3 L of growth medium.
  3.  Culture temperature is not critical.   Stock cultures may be maintained at
     25°C in  environmental  chambers  with cultures  of other organisms  if the
     illumination is adequate (continuous  "cool-white" fluorescent lighting of
     approximately 86 ± 8.6 |iE/m2/s,  or 400  ft-c).
  4.  Cultures are mixed twice daily  by hand.
  5.  Stock  cultures can be held in  the refrigerator until used to start "food"
     cultures, or can be transferred to new medium weekly.  One-to-three
     milliliters of 7-day old algal  stock  culture, containing approximately
     1.5 X  106 cells/ml,  are transferred to  each  100 ml of fresh  culture
     medium.   The inoculum should provide  an initial cell density of
     approximately 10,000-30,000 cells/ml  in the new stock cultures.   Aseptic
     techniques should be used in maintaining the stock algal cultures, and
     care should be exercised to avoid contamination by other microorganisms.
  6.  Stock  cultures should be examined microscopically weekly, at transfer,
     for microbial contamination.  Reserve quantities of culture organisms  can
     be maintained for 6-12 months  if stored in the dark at 4°C.   It  is
     advisable to prepare new stock  cultures from "starter" cultures obtained
     from established outside sources of organisms (see Section 6) every 4-6
     months.

6.15.2.2.2   Establishing and Maintaining "Food" Cultures of Algae:

  1.  "Food" cultures are started seven days prior to use for Ceriodaphm'a
     dubia  cultures and tests.  Approximately 20 ml of 7-day-old algal stock
     culture (described in the previous paragraph), containing 1.5 X 106
     cells/ml, are added to each liter of  fresh algal culture medium
     (i.e., 3 L of medium in a 4-L  bottle, or 18 L in a 20-L bottle).  The
     inoculum should provide an initial cell density of approximately 30,000
     cells/ml.  Aseptic techniques  should  be used in preparing and maintaining
     the cultures, and care should  be exercised to avoid contamination by
     other  microorganisms.  However, sterility of food cultures is not as
     critical as in stock cultures  because the food cultures are terminated in
     7-10 days.  A one-month supply of algal food can be grown at one time,
     and stored in the refrigerator.
  2.  Food cultures may be maintained at 25°C in environmental chambers with
     the algal stock cultures or cultures  of other organisms if the
     illumination is adequate (continuous  "cool-white" fluorescent lighting of

                                      154

-------
     approximately 86 ± 8.6 ^E/m2/s,  or  400  ft-c).
  3.  Cultures  are mixed continuously on  a magnetic stir plate (with a medium
     size  stir bar)  or in a moderately aerated separatory funnel, or are mixed
     twice daily by hand.  If the cultures are placed on a magnetic stir
     plate,  heat generated by the stirrer might elevate the culture
     temperature several  degrees.  Caution should be exercised to prevent the
     culture temperature from rising more than 2-3°C.

6.15.2.3  Preparing Algal Concentrate for Use as Ceriodaphnia dubia Food:

  1.  An algal  concentrate containing 3.0 to 3.5 X 107 cells/ml is prepared
     from  food cultures by centrifuging  the algae with a plankton or
     bucket-type centrifuge, or by allowing the cultures to settle in a
     refrigerator for at least three weeks and siphoning off the supernatant.
  2.  The cell  density (cells/ml) in the  concentrate is measured with an
     electronic particle counter, microscope and hemocytometer, fluorometer,
     or spectrophotometer (see Section 14),  and used to determine the
     dilution  (or further concentration) required to achieve a final cell
     count of 3.0 to 3.5 X 107/mL.
  3.  Assuming  a cell density of approximately 1.5 X 106 cells/ml in the
     algal food cultures at 7 days, and  100% recovery in the concentration
     process,  a 3-L, 7-10 day culture will provide 4.5 X 109 algal  cells.
     This  number of cells would provide  approximately 150 ml of algal cell
     concentrate (1500 feedings at 0.1 ml/feeding) for use as food.  This
     would be  enough algal food for four Ceriodaphnia dubia tests.
  4.  Algal concentrate may be stored in  the refrigerator for one month.

6.15.3  FOOD QUALITY

6.15.3.1  EPA recommends Fleishman's" yeast,  Cerophyll",  trout  chow,  and
Selenastrum capricornutum as the preferred Ceriodaphnia dubia food
combination.  This recommendation is based on extensive data developed by many
laboratories which indicated high Ceriodaphnia dubia survival and reproduction
in culturing and testing.  The use of substitute food(s) is acceptable only
after side-by-side tests are conducted to determine that the quality of the
substitute food(s) is equal to the EPA recommended food combination based on
survival and reproduction of Ceriodaphnia dubia.

6.15.3.2  The quality of food prepared with newly acquired supplies of yeast,
trout chow,  dried cereal leaves, algae,  and/or any substitute food(s) should
be determined in side-by-side comparisons of Ceriodaphnia dubia survival and
reproduction,  using the new food and food of known, acceptable quality, over a
seven-day  period in control medium.

6.16   TEST ORGANISMS, DAPHNIDS, CERIODAPHNIA DUBIA

6.16.1  Cultures of test organisms should be started at least three weeks
before the brood animals are needed, to ensure an adequate supply of neonates
for the test.   Only a few individuals are needed to start a culture because  of
their prolific reproduction.

6.16.2  Neonates used for toxicity tests must be obtained from  individually

                                      155

-------
cultured organisms.   Mass cultures may be maintained, however, to serve as a
reserve source of organisms for use in initiating individual cultures and in
case of loss of individual  cultures.

6.16.3  Starter animals may be obtained from commercial sources and may be
shipped in polyethylene bottles.   Approximately 40 animals and 3 ml of food
(see below) are placed in a 1-L bottle filled full with culture water for
shipment.  Animals received from an outside source should be transferred to
new culture media gradually over a period of 1-2 days to avoid mass mortality.

6.16.4  It is best to start the cultures with one animal, which is sacrificed
after producing young, mounted on a microscope slide, and retained as a
permanent slide mount to facilitate identification and permit future
reference.  The species identification of the stock culture should be verified
by preparing slide mounts,  regardless of the number of animals used to start
the culture.  The following procedure is recommended for making slide mounts
of Cen'odaphm'a dubia (modified from Beckett and Lewis, 1982):

     1. Pipet the animal onto a watch glass.
     2. Reduce the water volume by withdrawing excess water with the
        pi pet.
     3. Add a few drops of carbonated water (club soda or seltzer
        water) or 70% ethanol to relax the specimen so that the
        post-abdomen is extended.  (Optional:  with practice,
        extension of the postabdomen  may be accomplished by putting
        pressure on  the cover slip).
     4. Place a small amount (one to  three drops) of mounting medium
        on a glass microscope slide.   The recommended mounting
        medium is CMCP-9/10 Medium1,  prepared by mixing two parts
        of CMCP-9 with one part of CMCP-10 stained with enough acid
        fuchsin dye  to color the mixture a light pink.  For more viscosity
        and faster drying,  CMC-10 stained with acid fuchsin may be used.
     5. Using forceps or a pipet, transfer the animal to the drop
        of mounting  medium on the microscope slide.
     6. Cover with a 12 mm round cover slip and exert minimum pressure to
        remove any air bubbles trapped under the cover slip.  Slightly more
        pressure will extend the postabdomen.
     7- Allow mounting medium to dry.
     8. Make slide permanent by placing varnish around the edges of
        the coverslip.
     9. Identify to  species (see Pennak, 1978 or 1989, and/or Berner, 1986).
    10. Label with waterproof ink or  diamond pencil.
    11. Store for permanent record.

6.16.5  MASS CULTURE

6.16.5.1  Mass cultures are used only as a "backup" reservoir of organisms.
Neonates from mass cultures are not to be used directly in toxicity tests (see
'CMCP-9,  CMCP-10  and  Acid  Fuchsin  are  available  from Polysciences, Inc., Paul
 Valley Industrial  Park,  Warrington,  Pennsylvania, 18976 (215-343-6484).
                                      156

-------
Paragraph 10.2.3 below).

6.16.5.2  One-liter or 2-L glass beakers, crystallization dishes, "battery
jars," or aquaria may be used as culture vessels.  Vessels are commonly filled
to three-fourths capacity.  Cultures are fed daily.  Four or more cultures are
maintained in separate vessels and with overlapping ages to serve as back-up
in case one culture is lost due to accident or other unanticipated problems,
such as low DO concentrations or poor quality of food or laboratory water.

6.16.5.3  Mass cultures which will serve as a source of brood organisms for
individual culture should be maintained in good condition by frequent renewal
with new culture medium at least twice a week for two weeks.  At each renewal,
the adult survival is recorded, and the offspring and the old medium are
discarded.  After two weeks, the adults are also discarded, and the culture is
re-started with neonates in fresh medium.  Using this schedule, 1-L cultures
will produce 500 to 1000 neonate Ceriodaphm'a dubia each week.

6.16.6  INDIVIDUAL CULTURE

6.16.6.1  Individual  cultures are used as the immediate source of neonates for
toxicity tests.

6.16.6.2  Individual  organisms are cultured in 15 mL of culture medium in
30-mL (1 oz) plastic cups or 30-mL glass beakers.  One neonate is placed in
each cup.  It is convenient to place the cups in the same type of board used
for toxicity tests (see Figure 1 in this Section).

6.16.6.3  Organisms are fed daily (see paragraph 6.16.9) and are transferred
to fresh medium a minimum of three times a week, typically on Monday,
Wednesday, and Friday.  On the transfer days, food is added to the new medium
immediately before or after the organisms are transferred.

6.16.6.4  To provide cultures of overlapping ages, new boards are started
weekly, using neonates from adults which produce at least eight young in their
third or fourth brood.  These adults can be used as sources of neonates until
14 days of age.  A minimum of two boards are maintained concurrently to
provide backup supplies of organisms in case of problems.

6.16.6.5  Cultures which are properly maintained should produce at least 20
young per adult in three broods (seven days or less).  Typically, 60 adult
females (one board) will produce more than the minimum number of neonates
(120) required for two tests.

6.16.6.6  Records should be maintained on the survival of brood organisms and
number of offspring at each renewal.  Greater than 20% mortality of adults, or
less than an average of 20 young per female would indicate problems, such as
poor quality of culture media or food.  Cultures that do not meet these
criteria should not be used as a source of test organisms.

6.16.7  CULTURE MEDIUM
6.16.7.1  Moderately hard synthetic water prepared using MILLIPORE MILLI-QR or
equivalent deionized water and reagent grade chemicals or 20% DMW is

                                      157

-------
recommended as a standard culture medium (see Section 7, Dilution Water).

6.16.8  CULTURE CONDITIONS

6 16.8.1  The daphnid,  Cen'odaphm'a dubia,  should be cultured at a temperature
of 25 ± 1°C.

6.16.8.2  Day/night cycles prevailing in most laboratories will provide
adequate illumination for normal  growth and reproduction.  A 16-h/8-h
day/night cycle is recommended.   Light intensity should be 10-20 tiE/m/s, or
50 to 100 foot candles.

6.16.8.3  Clear, double-strength  safety glass or 6 mm plastic panels are
placed on the culture vessels to  exclude dust and dirt, and reduce
evaporation.

6.16.8.4  The organisms  are delicate and should be handled as carefully and as
little as possible so that they  are not unnecessarily stressed.  They are
transferred with a pipet of approximately 2-mm bore, taking care to release
the animals under the surface of  the water.  Any organism that is injured
during handling should be discarded.

6.16.9  FOOD AND FEEDING

6.16.9.1  Feeding the proper amount of the right food is extremely important
in Cen'odaphm'a dubia culturing.   The key is to provide sufficient nutrition
to support normal reproduction without adding excess food which may reduce the
toxicity of the test solutions,  clog the animal's filtering apparatus, or
greatly decrease the DO  concentration and increase mortality.  A combination
of Yeast, CEROPHYLLR,  and Trout chow (YCT),  along with  the unicellular green
alga, Selenastrum capricornutum,  will provide suitable nutrition if fed daily.

6.16.9.2  Other algal species (such as 5. minutum or Chlamydomonas
reinhardti),  other substitute food combinations (such as Flake Fish Food), or
different feeding rates  may be acceptable as long as performance criteria are
met and side-by-side comparison  tests confirm acceptable quality (see
Paragraph 6.15.3 above).

6.16.9.3  Cultures should be fed  daily to maintain the organisms in optimum
condition so as to provide maximum reproduction.  Stock cultures which are
stressed because they are not adequately fed may produce low numbers of young,
large numbers of males,  and/or ephippial females.  Also, their offspring may
produce few young when used in toxicity tests.

6.16.9.4  Feed as follows:

  1. If YCT is frozen,  remove a  bottle of food from the freezer 1 h before
     feeding time, and allow to  thaw.
  2. YCT food mixture and algal  concentrates should both be thoroughly mixed
     by shaking before dispensing.
  3. Mass cultures are fed daily  at the rate of 7 mL YCT and 7 mL algae
     concentrate/L culture.

                                      158

-------
  4.  Individual  cultures are fed at the rate of 0.1 ml YCT and 0.1 ml algae
     concentrate per 15 ml culture.
  5.  Return unused YCT food mixture and algae concentrate to the refrigerator.
     Do not refreeze YCT.  Discard unused portion after one week.

6.16.10  It is recommended that chronic toxicity tests be performed monthly
with  a reference toxicant.  Daphnid, Ceriodaphnia dubia, neonates less than 24
h old, and all within 8 h of the same age are used to monitor the chronic
toxicity of the reference toxicant to the Ceriodaphnia dubia produced by the
culture unit.   See Section 4, Quality Assurance, Subsection 4.17, Reference
Toxicants.

6.16.11  Record Keeping

6.16.11.1  Records, kept in a bound notebook, include (1) source of organisms
used  to start  the cultures, (2) type of food and feeding times, (3) dates
culture were thinned and restarted, (4) rate of reproduction in individual
cultures, (5)  daily observations of the condition and behavior of the
organisms in the cultures, and (6) dates and results of reference toxicant
tests performed (see Section 4, Quality Assurance).

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.

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.
Receiving water toxicity is determined with samples used directly as collected
or after samples are passed through a 60 \im NITEXR filter and compared without
dilution, against a control.  For a test consisting of single receiving water
and control, approximately 600 mL of sample would be required for each test,
assuming 10 replicates of 15 ml, and sufficient additional sample for chemical
analysis.

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

                                      159

-------
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%).  Improvements in precision
decline rapidly if the dilution factor is increased beyond 0.5, and precision
declines rapidly if a smaller dilution factor is 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%).  If a high rate  of mortality is observed during the first
1  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 10 replicates
per concentration, each containing  15 ml of test solution, with a dilution
series of 0.5, is approximately 1 L/day.  A volume of 15 ml of test solution
is adequate for the organisms, and  will  provide a depth in which it is
possible to count the animals under a stereomicroscope with a minimum of
refocusing.  Ten test chambers are  used  for each effluent dilution and for the
control.  Sufficient test solution  (approximately 550 ml) is prepared at each
effluent concentration to provide 400 ml additional volume for chemical
analyses at the high, medium,  and low test concentrations.

10.1.2.4  Tests 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.

10.1.2.5  Just prior to test initiation  (approximately one h) the temperature
of sufficient quantity of the sample  to  make the test solutions should be
adjusted to the test temperature and  maintained at that temperature during the
preparation of the test solutions.

10.1.2.6  The DO of the test solutions should be checked prior to test
initiation.  If any of the solutions  are supersaturated with oxygen or any
solution has a DO concentration below 4.0 mg/L, all the solutions and the
control must be gently aerated.

10.1.3  Dilution Water

10.1.3.1  Dilution water may be uncontaminated receiving water, a standard
synthetic (reconstituted) water, or some other uncontaminated natural water
(see Section 7, Dilution Water).

10.2  START OF THE TEST

10.2.1  Label the test chambers with  a marking pen.  Use of color-coded tape
to identify each treatment and replicate is helpful.  A minimum of five
effluent concentrations and a control are used for each effluent test.  Each
treatment (including the control) must have ten replicates.


                                      160

-------
            0
DDD

D P D -O
Pppb
PPPP
nj  w  CM  «
D.O
                           ODD
                           b p P p p p .o
                            U3  C*  <0  U3  T-  C*   CO
                      v^cni-^cnioi-
                            v  «  «
                            O«  
-------
10.2.2  The test solutions can be randomly assigned to a board using a
template (Figure 1) or by using a table of random numbers (see Appendix A).
When using the randomized block design, test chambers are randomized only
once, at the beginning of the test.  A number of different templates should be
prepared, so that the same template is not used for every test.

10.2.3  Neonates less than 24 h old, and all within 8 h of the same age, are
required to begin the test.   The neonates are obtained from individual
cultures using brood boards,  as described above (paragraph 6.16.6).  Neonates
are taken only from adults that have eight or more young in their third or
subsequent broods.   These adults can be used as brood stock until they are 14
days old.  If the neonates are held more than one or two hours before using in
the test, they should be fed (0.1 mL YCT and 0.1 ml algal concentrate/15 ml of
media).

10.2.4  Ten brood cups, each with 8 or more young, are randomly selected from
a brood  board for use in setting up a test.  To start the test, one neonate
from the first brood cup is  transferred to each of the six test chambers in
the first row on the test board (Figure 1-3).  One neonate from the second
brood cup is transferred to  each of the six test chambers in the second row on
the test board.  This process is continued until each of the 60 test chambers
contains one neonate.

10.2.4.1  The cups and test  chambers may be placed on a light table to
facilitate counting the neonates.  However, care must be taken to avoid
temperature increase due to  heat from the light table.

10.2.5  This blocking procedure allows the performance of each female to be
tracked.  If a female produces one weak offspring or male, the likelihood of
producing all weak offspring or all males is greater.  By using this known
parentage technique, poor performance of young from a given female can be
omitted from all concentrations.

10.3  LIGHT, PHOTOPERIOD, AND TEMPERATURE

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.

10.3.2  It is critical that  the test water temperature be maintained at 25 ±
1°C to obtain three broods in seven days.

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 concentrations.  The DO
concentrations should be measured in the new solutions at the start of the
test (Day 0) and before daily renewal of the test solutions on subsequent
days.  The DO concentration  should not fall below 4.0 mg/L (see Section 8,
Effluent and Receiving Water Sampling, Sample Handling, and Sample Preparation
for Toxicity Tests).  Aeration is generally not practical during the daphnid,
Ceriodaphnia dubia, test.  If the DO in the effluent and/or dilution water  is

                                      162

-------
low,  aerate gently before preparing the test solutions.  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 physical stress to the organisms.

10.5  FEEDING

10.5.1  The organisms are fed when the test is initiated, and daily
thereafter.  Food is added to the fresh medium immediately before or
immediately after the adults are transferred.  Each feeding consists of 0.1 ml
YCT and 0.1 ml Selenastrum capn'cornutum concentrate/15 ml test solution (0.1
ml of algal concentrate containing 3.0-3.5 X 10  cells/ml will  provide 2-2.3 X
105 cells/ml in  the  test chamber).

10.5.2  The YCT and algal suspension can be added accurately to the test
chambers by using automatic pipettors, such as Gil son, Eppendorf, Oxford, or
equivalent.

10.6  OBSERVATIONS DURING THE TEST

10.6.1  Routine Chemical and Physical Observations

10.6.1.1  DO is measured at the beginning and end of each 24-h exposure period
in at least one test chamber at each test concentration and in the control.

10.6.1.2  Temperature and pH are measured at the end of each 24-h exposure
period in at least one test chamber at each test concentration and in the
control.  Temperature should be monitored continuously or observed and
recorded daily for at least two locations in the environmental  control system
or the samples.   Temperature should be measured in sufficient number of test
vessels at least at the end of the test to determine the temperature variation
in the environmental chamber.

10.6.1.3  The pH is measured in the effluent sample each day before new test
solutions are made.

10.6.1.4  Conductivity, alkalinity and hardness are measured in each new
sample (100% effluent or receiving water) and in the control.

10.6.1.5  Record the data on data sheet (Figure 2).

10.6.2  Routine Biological Observations

10.6.2.1  Three or four broods are usually obtained in the controls in a 7-day
test conducted at 25 ± 1°C.   A brood is a group of offspring released from the
female over a short period of time when the carapace is discarded during
molting.  In the controls, the first brood of two-to-five young is usually
released on the third or fourth day of the test.  Successive broods are
released every 30 to 36 h thereafter.  The second and third broods usually
consist of eight to 20 young each.  The total number of young produced by  a
healthy control  organism in three broods often exceeds 30 per female.

                                      163

-------
10.6.2.2  The release of a brood may be inadvertently interrupted during the
daily transfer of organisms to fresh test solutions, resulting in a split in
the brood count between two successive days.  For example, four neonates of a
brood of five might be released on Day 3, just prior to test solution renewal,
and the fifth released just after renewal, and counted on Day 4.  Partial
broods, released over a two-day period, should be counted as one brood.

10.6.2.3  Each day, the live adults are transferred to fresh test solutions,
and the numbers of live young are recorded (see data form. Figure 3). The
young can be counted with the aid of a stereomicroscope with substage
lighting.  Place the test chambers on a light box over a strip of black tape
to aid in counting the neonates.  The young are discarded after counting.

10.6.2.4  Some of the effects caused by toxic substances include, (1) a
reduction in the number of young produced, (2) young may develop in the brood
pouch of the adults, but may not be released during the exposure period, and
(3) partially or fully developed young may be released, but are all dead at
the end of the 24-h period.  Such effects should be noted on the data sheets
(Figure 3).

10.6.2.5  Protect the daphnids, Ceriodaphnia dubia, from unnecessary
disturbance during the test by carrying out the daily test observations,
solution renewals, and transfer of females carefully.  Make sure the females
remain immersed during the performance of these operations.

10.7  DAILY CLEANING OF TEST CHAMBERS

10.7.1  The test is started (Day 0) with new disposable polystyrene cups or
precleaned 30-mL borosilicate glass beakers.  On day one a second set of new
cups or glass beakers is prepared and labeled and color-coded with tape
similar to the original set.  The new solutions are placed in the new set of
test vessels and the females are transferred from the original test vessel to
the new ones with the corresponding label and color-code.  After counting any
young, completing the chemical analyses, and discarding the contents, the old
vessels may be rinsed twice in deionized water and saved for reuse on day two.
Alternate sets of test vessels may be used each day in the same manner for the
remainder of the test with no additional cleaning of the test vessels needed.

10.8  TEST SOLUTION RENEWAL

10.8.1  Freshly prepared solutions are used to renew the test daily.  For
on-site toxicity studies, fresh effluent or receiving water samples should be
collected daily, and no more than 24 h should elapse between collection of the
samples and their use in the tests (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 is collected, preferrably on
days one, three, and five.  No more than 36 h should elapse between collection
of the sample and the first use in the test.  Maintain the samples in the
refrigerator at 4°C until  used.

10.8.2  New test solutions are prepared daily, and the test organisms are


                                      164

-------
transferred to the freshly prepared solutions using a small-bore (2 mm) glass
or polyethylene dropper or pipet.  The animals are released under the surface
of the water so that air is not trapped under the carapace.  Organisms that
are dropped or injured are discarded.

10.9  TERMINATION OF THE TEST

10.9.1  Tests should be terminated when 80% of the controls have a brood and
at least 60% have produced their third brood or by the end of 7 days at the
latest.  Because of the rapid rate of development of Ceriodaphm'a dubia, at
test termination all observations on organism survival and numbers of
offspring should be completed within two hours.  An extension of more than a
few hours in the test period would be a significant part of the brood
production cycle of the animals, and could result in additional broods.

10.9.2  Count the young, conduct required chemical measurements, and complete
the data sheets (Figure 3).

10.9.3  Any animal not producing young should be examined to determine if it
is a male (see Berner, 1986).  In most cases, the animal will need to be
placed on a microscope slide before examining (see subsection 6.16.4).

11.  SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA

11.1  A summary of test conditions and test acceptability criteria is
presented in Table 3.

12.  ACCEPTABILITY OF TEST RESULTS

12.1  For the test results to be acceptable, at least 80% of the control
animals must survive and have young.  At least 60% of the original  ten
neonates in the controls must have three broods producing an average of 20 or
more young per surviving female in seven days.

13.  DATA ANALYSIS

13.1  GENERAL

13.1.1  Tabulate and summarize the data.  A sample set of survival  and
reproduction data is listed in Table 4.

13.1.2  The endpoints of toxicity tests using the daphnid, Ceriodaphnia dubia,
are based on the adverse effects on survival and reproduction.  The LC50, the
IC25, the IC50 and the EC50 are calculated using point estimation techniques,
and LOEC and NOEC values for survival and reproduction are obtained using a
hypothesis test approach such as Fisher's Exact Test (Finney, 1948; Pearson
and Hartley, 1962), Dunnett's Procedure (Dunnett, 1955) or Steel's Many-one
Rank Test (Steel, 1959; Miller, 1981) (see Section 9, Chronic Toxicity
Endpoints and Data Analysis).  Separate analyses are performed for the
estimation of the LOEC and NOEC endpoints and for the estimation of the LC50,
IC25, IC50 and EC50.  Concentrations at which there is no survival  in any of
the test chambers are excluded from the statistical analysis of the NOEC and

                                      165

-------
TABLE 3.  SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA FOR
          DAPHNID, CERIODAPHNIA DUBIA,  SURVIVAL AND REPRODUCTION TOXICITY
          TESTS WITH EFFLUENTS AND RECEIVING WATERS
   1. Test type:
   2. Temperature (°C):
   3. Light quality:
   4. Light intensity:
   5. Photoperiod:
   6. Test chamber  size:
   7. Test solution volume:
   8. Renewal  of test solutions:
   9. Age of test organisms:
  10. No. neonates per
       test chamber:
  11. No. replicate test
       chambers per concentration;
  12. No. neonates per
       test concentration:
  13. Feeding regime:
  14. Cleaning

  15. Aeration:
  16. Dilution water:
Static renewal
25 ± 1°C
Ambient laboratory illumination
10-20 jiE/m2/s, or 50-100 ft-c
(ambient laboratory levels)
16 h light, 8 h dark
30 mL (minimum)
15 mL (minimum)
Daily
Less than 24 h; and all released
within a 8-h period
10
10
Feed 0.1 mL each of YCT and algal
suspension per test chamber
daily
Rinse chambers twice in deionized water
before re-use
None
Uncontaminated source of receiving or
other natural water, synthetic water
prepared using MILLIPORE MILLI-QR
or equivalent deionized water and
reagent grade chemicals or DMW
(see Section 7)
                                      166

-------
   TABLE 3.   SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA FOR
             DAPHNID, CERIODAPHNIA DUBIA, SURVIVAL AND REPRODUCTION TOXICITY
             TESTS WITH EFFLUENTS AND RECEIVING WATERS (CONTINUED)
17.  Effluent concentrations:


18.  Test dilution factor:


19.  Test duration:


20.  Endpoints:

21.  Test acceptability criteria:
22. Sampling requirements:
Minimum of 5 effluent concentrations
and a control
Effluents:  > 0.5
Receiving Waters:
None, or > 0.5
22. Sample volume required:
Until 60% of control animals have three
broods (not more than 7 days)

Survival  and reproduction

80% or greater survival and an average of
20 or more young/surviving animal in the
control solutions.  At least 80% of
control animals must have a brood
and at least 60% of the control animals
should have produced their third brood in
7 days

For on-site tests, samples 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 collected on days one,
three, and five with a maximum holding
time of 36 h before first use

1 L/day/test
                                    167

-------
   Figure 2.  Data form for the daphnid, Cen'odaphm'a dubia, survival and
              reproduction test.  Routine chemical  and physical
              determinations.
Di scharger:
Location:
Template No. :
Analyst:
Dates:
Food:
Dav
Control :
Temp.
D.O. Initial
Final
oH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










2










3










4










5










6










7










Remarks










                                     Day
Cone:
Temp.
D.O. Initial
Final
oH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










2










3










4










5










6










7










Remarks










                                     Dav
Cone:
Temo.
D.O. Initial
Final
pH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










2










3










4










5










6










7










Remarks










V = Test organism alive,  no young        x
0 = Number of live young               (-0)
L =  Lost or missing test organism       M
Test organism dead
Number of dead young
Male
                                      168

-------
Figure 2.   Data form for the daphnid, Ceriodaphm'a dubia, survival and
           reproduction test.  Routine chemical and physical
           determinations (Continued).
Discharger:
Location:
Template No. :
Analyst:
Dates:
Food:
Day
Cone:
Terno.
D.O. Initial
Final
pH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










2










3










4










5










6










7










Remarks










                                  Day
Cone:
Temp.
D.O. Initial
Final
oH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










2










3










4










5










6










7










Remarks










                                  Day
Cone:
Terno.
D.O. Initial
Final
DH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine

1










2










3










4










5










6










7










Remarks










                                   169

-------
   Figure 3.   Data form for the daphnid,  Ceriodaphm'a dubia,  survival  and
              reproduction test.   Daily summary of data.
Discharger:
Location:
Date Sample Collected:
Analyst:	
Test Start-Date/Time:
Test Stop -Date/Time:
Cone.

Dav
1
2
3
4
5
6
7

Total
Replicate
1









2









3









4









5









6









7









8









9









10









Number
of
Young









Number
of
Adults









Young
per
Adult



















Cone.

Dav
1
2
3
4
5
6
7

Total
Replicate
1









2









3









4









5









6









7









8









9









10









Number
of
Young









Number
of
Adults









Young
per
Adult



















Cone.

Dav
1
2
3
4
5
6
7

Total
Replicate
1









2









3









4









5









6









7









8









9









10









Number
of
Young









Number
of
Adults









Young
per
Adult



















                                       170

-------
  Figure 3.   Data form for the daphnid,  Cen'odaphm'a dubia,  survival  and
             reproduction test.   Daily summary of data (Continued).
Discharger:
Location:
Date Sample Collected:
Analyst:	
Test Start-Date/Time:
Test Stop -Date/Time:
Cone.

Dav
1
2
3
4
5
6
7

Total
Repl icate
1









2









3









4









5









6









7









8









9









10









Number
of
Young









Number
of
Adults









Young
per
Adult



















Cone.

Dav
1
2
3
4
5
6
7

Total
Replicate
1









2









3









4









5









6









7









8









9









10









Number
of
Young









Number
of
Adults









Young
per
Adult



















Cone.

Dav
1
2
3
4
5
6
7

Total
Repl icate
1









2









3









4









5









6









7









8









9









10









Number
of
Young









Number
of
Adults









Young
per
Adult



















                                       171

-------
LOEC for reproduction, but included in the estimation of the LC50, IC25, IC50,
and EC50.  See the Appendicies 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 THE DAPHNID, CERIODAPHNIA DUBIA, SURVIVAL DATA

13.2.1  Formal statistical analysis of the survival data is outlined on the
flowchart in Figure 4.  The response used in the analysis is the number of
animals surviving at each test concentration.  Separate analyses are performed
for the estimation of the NOEC and LOEC endpoints and for the estimation of
the EC50, LC50, IC25, or IC50 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 LC, EC,
and 1C endpoints.

13.2.2  Fisher's Exact Test is used to determine the NOEC and LOEC endpoints.
It provides a conservative test of the equality of any two survival
proportions assuming only the independence of responses from a Bernoulli
(binomial) population.  Additional information on Fisher's Exact Test is
provided in Appendix G.

13.2.3  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 number dead at a given concentration is the response.

13.2.4  Example of Analysis of Survival Data

13.2.4.1  The data in Table 4 will be used to illustrate the analysis of
survival data from the daphnid, Cen'odaphnia dubia, Survival and Reproduction
Test.  As can be seen from the data in Table 4, there were no deaths in the
1.56%, 3.12%, 6.25%, and 12.5% concentrations.  These concentrations are
obviously not different from the control in terms of survival.  This leaves
only the 25% effluent concentration to be tested statistically for a
difference in survival from the control.

13.2.5  Fisher's Exact Test

13.2.5.1  The basis for Fisher's Exact Test is a 2X2 contingency table.  From
the 2X2 table prepared by comparing the control and the effluent
concentration, determine statistical significance by looking up a value in the
table provided in Appendix G (Table G.5).  However, to use this table the
contingency table must be arranged in the format illustrated in Table 5.
                                      172

-------
          STATISTICAL ANALYSIS OF CERiODAPHNIA
            SURVIVAL AND REPRODUCTION TEST
                       SURVIVAL DATA
                    PROPORTION SURVIVING
                                         i
                                     FISHER'S EXACT
                                         TEST
     ENDPOiNT ESTIMATE
           LC50
ENDPOINT ESTIMATES
    NOEC.LOEC
Figure 4.   Flowchart  for  statistical  analysis of the daphnid,
           Ceriodaphm'a dubia,  survival data.
                           173

-------
    TABLE 4.  SUMMARY OF SURVIVAL AND REPRODUCTION DATA FOR THE DAPHNID,
              CERIODAPHNIA DUBIA, EXPOSED TO AN EFFLUENT FOR SEVEN DAYS

Effluent
Concentration
Control
1 . 56%
3.12%
6.25%
12.5%
25.0%


No.
of Young per Adult
Reolicate
1
27
32
39
27
10
0
2
30
35
30
34
13
0
3
29
32
33
36
7
0
4
31
26
33
34
7
0
5
16
18
36
31
7
0
6
15
29
33
27
10
0
7
18
27
33
33
10
0
8
17
16
27
31
16
0
9
14
35
38
33
12
0
10
27
13
44
31
2
0
No.
Live
Adults
10
10
10
10
10
3
             TABLE 5.   FORMAT OF THE 2X2 CONTINGENCY TABLE

Condition 1
Condition 2
Number
Successes
a
b
of
Fail
A
B

ures
- a
- b
Number of
Observations
A
B
              Total
a + b    [(A+B) - a - b]
A + B
13.2.5.2  Arrange the table so that the total  number of observations for
row one is greater than or equal  to the total  for row two (A ;> B).
Categorize a success such that the proportion  of successes for row one is
greater than or equal to the proportion of successes for row two (a/A z
b/B).  For these data, a success  may be 'alive'  or 'dead' whichever causes
a/A ;> b/B.  The test is then conducted by looking up a value in the table
of significance levels of b and comparing it to  the b value given  in the
contingency table.  The table of  significance  levels of b is included in
Appendix G, Table G.5.  Enter Table G.5 in the section for A, subsection
for B, and the line for a.  If the b value of  the contingency table is
equal to or less than the integer in the column  headed 0.05 in Table G.5,
then the survival proportion for  the effluent  concentration is
significantly different from that of the control.  A dash or absence of
entry in Table G.5 indicates that no contingency table in that class is
significant.

                                      174

-------
13.2.5.3  To compare the control and the effluent concentration of 25%, the
appropriate contingency table for the test is given in Table 6.


          TABLE 6.  2X2 CONTINGENCY TABLE FOR CONTROL AND 25% EFFLUENT



Control
25% Effluent
Number of

Alive
10
3


Dead
0
7
Number of
Observations

10
10
           Total             13             7              20
13.2.5.4  Since 10/10 * 3/10, the category 'alive' is regarded as a success.
For A = 10, B = 10 and, a = 10, under the column headed 0.05, the value from
Table G.5 is b = 6.  Since the value of b (b = 3) from the contingency table
(Table 6), is less than the value of b (b = 6) from Table G.5 in Appendix G,
the test concludes that the proportion surviving in the 25% effluent
concentration is significantly different from the control.  Thus the NOEC for
survival is 12.5% and the LOEC is 25%.

13.2.6  Probit Analysis

13.2.6.1  The data used for the Probit Analysis are summarized in Table 7.
For the Probit Analysis, the data from all concentrations are considered.  To
perform the Probit Analysis, run the EPA Probit Analysis Program.  An example
of the program input and output is supplied in Appendix I.

13.2.6.2  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.  For this example there is only one partial
mortality, and Probit Analysis is not appropriate.  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 (see USEPA, 1991b).
                                      175

-------
                      TABLE 7.   DATA FOR PROBIT ANALYSIS
                                     Effluent Concentration (%)

                       Control    1.56   3.12   6.25   12.5   25.0
Number Dead
Number Exposed
0
10
0
10
0
10
0
10
0
10
7
10
13.3  EXAMPLE OF ANALYSIS OF THE DAPHNID,  CERIODAPHNIA DUBIA, REPRODUCTION
      DATA

13.3.1  Formal statistical analysis of the reproduction data is outlined on
the flowchart in Figure 5.  The response used in the statistical analysis is
the number of young produced per adult female,  which is determined by taking
the total  number of young produced until either the time of death of the adult
or the end of the experiment, whichever comes first.  An animal that dies
before producing young, if it has not been identified as a male, would be
included in the analysis with zero entered as the number of young produced.
The subsequent calculation of the mean number of live young produced per adult
female for each toxicant concentration provides a combined measure of the
toxicant's effect on both mortality and reproduction.  An 1C estimate can be
calculated for the reproduction data using a point estimation technique (see
Section 9).  Hypothesis testing can be used to obtain an NOEC for
reproduction.  Concentrations above the NOEC for survival are excluded from
the hypothesis test for reproduction 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 using the
Shapiro-Wilk's Test for normality, and Bartlett's Test for homogeneity of
variance.   If either of these tests fail,  a nonparametric test, Steel's
Many-one Rank Test, is used to determine the NOEC and LOEC.  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 (see Appendix D).  The Wilcoxon Rank Sum Test with the Bonferroni
adjustment is the nonparametric alternative (see Appendix F).

13.3.4  The data, mean, and standard deviation of the observations at each
concentration including the control are listed in Table 8.  A plot of the
number of young per adult female for each concentration is provided in

                                      176

-------
STATISTICAL ANALYSIS OF CERIODAPHNIA
SURVIVAL AND REPRODUCTION TEST
REPRODUCTION DATA
NO. OF YOUNG PRODUCED

(i

POINT ESTIMATION (^™™*
ABOVE NOEC F
\ f
ENDPOINT ESTIMATE 1
IC25, IC50 f SHAPIRO-V
NORMAL DISTRIBUTION
i
rBARTLE
r
S TESTING
)NCENTRATIONS
OR SURVIVAL)

INUN-IN
L/ll l/'C TCCT
vlLr\ o I Lo 1
1
ITT'O TCTCT ^h.
. 1 1 O 1 LO 1 ^^

Nin
EQUAL NUMBER OF
REPLICATES?
YES
EQUAL NUMBER 0
REPLICATES?
YES
poIFclLSE DUNNETT'S STEEL'S MANY-ONE WIL
BONFERRONI TF«?T RANK TEST
An.HJRTMFNIT TE5T HAINMtC.1 RnNpF





\
ENDPOINT ESTIMATES
NOEC.LOEC
DRMAL DISTRIBUTION
HETEROGENEOUS
VARIANCE
NO

1
30XON RANK SUM
TEST WITH
ERRONI ADJUSTMENT




Figure 5.   Flowchart for statistical  analysis of the daphnid,
           Cen'odaphnia dubia,  reproduction data.
                             177

-------
00
            50
            40 -
            30 -
            20-
            10 -
             0 -
               0.00
                                                                      CONNECTS UEAN VALUE FOR EACH CONCENTRATION
                                                                      REPRESENTS THE CRITICAL VALUE FOR DUNNETT8 TEST
                                                                      (ANY MEAN NO OF OFFSPRING BELOW THIS VALUE WOULD BE
                                                                      SIGNIFICANTLY DIFFERENT FROU THE CONTROL)
1.56                    3.12
         EFFLUENT CONCENTRATION (%)
6.25
  r
12.50
Figure 6.   Plot  of number of young per adult female from  a daphnid, Cen'odaphnia dubia, survival  and
             reproduction  test.

-------
       TABLE  8.   THE DAPHNID,  CERIODAPHNIA DUBIA, REPRODUCTION DATA
Replicate    Control
                                   Effluent Concentration (%)
1.56
3.12
6.25
12.5
1
2
3
4
5
6
7
8
9
10
Mean(Yi)
s?
i
27
30
29
31
16
15
18
17
14
27
22.4
48.0
1
32
35
32
26
18
29
27
16
35
13
26.3
64.0
2
39
30
33
33
36
33
33
27
38
44
34.6
23.4
3
27
34
36
34
31
27
33
31
33
31
31.7
8.7
4
10
13
7
7
7
10
10
16
12
2
9.4
15.1
5
Figure 6.   Since there is significant mortality in the 25% effluent
concentration,  its effect on reproduction is not considered.

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 9.

13.3.5.2  Calculate the denominator, D, of the test statistic:
                         n
                         L (X
  X)2
    Where X,.  =  the  ith  centered observation
          X  = the overall mean of the centered observations
          n  = the total number of centered observations.
                                      179

-------
        TABLE 9.  CENTERED OBSERVATIONS  FOR SHAPIRO-MILK'S EXAMPLE
                                    Effluent Concentration (%)
Replicate    Control              1.56      3.12      6.25      12.5
1
2
3
4
5
6
7
8
9
10
4.6
7.6
6.6
8.6
-6.4
-7.4
-4.4
-5.4
-8.4
4.6
5.7
8.7
5.7
-0.3
-8.3
2.7
0.7
-10.3
8.7
-13.3
4.4
-4.6
-1.6
-1.6
1.4
-1.6
-1.6
-7.6
3.4
9.4
-4.7
2.3
4.3
2.3
-0.7
-4.7
1.3
-0.7
1.3
-0.7
0.6
3.6
-2.4
-2.4
-2.4
0.6
0.6
6.6
2.6
-7.4
For this set of data,            n = 50

                                 X = _L(0.0) =  0.0
                                     50
                                 D = 1433.4

13.3.5.3  Order the centered observations from smallest  to  largest

                  X(1)  . x<2>  -      -  X(n)

    Where XO) is the ith ordered observation.  These  ordered
    observations are listed in Table 10.

13.3.5.4  From Table 4, Appendix B, for the number of observations,  n,  obtain
the coefficients av  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 = 50, k = 25.   The a-  values are
listed in Table 11.                                           '

13.3.5.5  Compute the test statistic, W, as follows:

                       k
               W = 1 [ 2 a,. (x(n-j+1> - X(i)) ]2
                   D   i=l

the differences X(n'1+1) - X(i) are listed in Table 11.

For this set of data:

                     W =    1    (37.3)2 = 0.97
                          1433.4

                                      180

-------
13.3.5.6  The decision rule for this test is to compare W with the critical
value found in Table 6,  Appendix B.   If the computed V 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 50
observations (n) is 0.930.   Since W = 0.97 is greater than the critical value,
the conclusion of the test  is that the data are normally distributed.
  TABLE 10.   ORDERED CENTERED OBSERVATIONS FOR SHAPIRO-MILK'S EXAMPLE
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
X(i>
-13.3
-10.3
-8.4
-8.3
-7.6
-7.4
-7.4
-6.4
-5.4
-4.7
-4.7
-4.6
-4.4
-2.4
-2.4
-2.4
1.6
1.6
-1.6
-1.6
-0.7
-0.7
-0.7
-0.3
0.6
i
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
X(i)
0.6
0.6
0.7
1.3
1.3
1.4
2.3
2.3
2.6
2.7
3.4
3.6
4.3
4.4
4.6
4.6
5.7
5.7
6.6
6.6
7.6
8.6
8.7
8.7
9.4
                                      181

-------
TABLE 11.   COEFFICIENTS AND DIFFERENCES FOR SHAPIRO-MILK'S EXAMPLE
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
ai
0.3751
0.2574
0.2260
0.2032
0.1847
0.1691
0.1554
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
A "" A
22.7
19.0
17.1
16.9
15.2
14.0
14.0
12.1
11.1
9.3
9.3
9.0
8.7
6.0
5.8
5.1
4.2
3.9
3.9
3.0
2.0
2.0
1.4
0.9
0.0

x
C)
X
XC47>
X(46,
C!
X
x(43)
X(42)
X(41)
Y<40)
A
XC39)
X(38)
X(37)
X(36)
X(35)
x<34)
x(33)
x<32)
X(31)
X(30)
X(29>
X(28)
x<27)
x<26)

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(14)
- x(15>
x<16)
X(17)
- x(18)
- x<19>
- x<20)
- x(21)
- XC22)
v<23)
"" A
X(24>
- x(25>
                                  182

-------
13.3.6  Test for Homogeneity of Variance

13.3.6.1  The test used to examine whether the variation in number of
young produced 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
               [ ( E VJ In S2    E  V.. In S,2 ]
           B =    i=l	        i = l
    Where Vf  =    degrees of freedom for each effluent concen-
                  tration and control, V, = (n,-  -  1)
          p  =    number of levels of effluent concentration and control
          n;  =    the number of replicates for concentration i
          In =    loge

          i  =    1, 2,  ..., p where p is the number of concentrations
                  including the control

                   P     2
          _2      ( 2 V,- S,)
          S2  =    1=1
          C  = 1 + ( 3(p-l))-  [  EPl/Vi
 13.3.6.2  For the data in this example (See Table 8), all effluent
 concentrations including the control have the same number of replicates (nf
 = 10 for all i).  Thus, V,-  = 9 for all  i.

 13.3.6.3  Bartlett's statistic is therefore:

                              P      ->
       B =  [(45)ln(31.8) - 9 £ ln(S,.2)]/l .04
                             1-1

         =  [45(3.5) - 9(16.1)]/1.04

         =  12.6/1.04

         =  12.1

 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  four

                                      183

-------
degrees of freedom,  is 13.3.   Since B = 12.1 is less than the critical value
of 13.3, 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 12.
                            TABLE 12.   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
S* = SSB/(p-l)
Sy = SSW/(N-p)
Where:
P
N
ni
               = number effluent concentrations including the control
               = total  number of observations n1  +  n,  . . . +np
               = number of  observations  in  concentration i
           SSB
P
S
                       ,  - G2/N
                                    Between Sum of Squares
                      -
           SST = 21   I Yn2   G2/N
                                                Total  Sum of Squares
           SSW = SST - SSB
                                                Within Sum of Squares
            T,-  =
                  the grand total  of all  sample observations, G = E Tj
                                                                 i = l
                  the total  of  the  replicate  measurements  for
                  concentration "i"
                  the jth observation for concentration  "i"  (represents
                  the number of young produced by female j in
                  effluent  concentration  i)
                                      184

-------
13.3.7.2  For the data  in  this  example:

    n.  = n2 =  n, = n, =  n= = 10
    N = 50
    TI  = YH + YIZ  + .  .  . + Y110  =  224
    T2  - Y21 + Y22  + .  .  . * Y210  -  263
    T3  - Y31 + Y32  + .  •  • + Y310  =  346
    T4  = Y41 + Y42  + .  .  . + Y410  =  317
    T5  = Y51 + Y52  + .  .  . + Y510  -   94

    G  = T,  + T2 + T3 +  T4  + T5 = 1244

          P   .       ,
    SSB = E T,2/n, - G2/N
        =  1  (348,386)  -  (1244)2   =  3887.88
          10                 50

          P   n,-
    SST = I   I Yn2  -  G2/N
         1-1 j-1

        = 36,272  -  (1244)2  =  5321.28
                       50

    SSW = SST - SSB  =  5321.28  - 3887.88 = 1433.40

    S2  = SSB/(p-l)  = 3887.88/(5-l) = 971.97

    S2  = SSW/(N-p)  = 1433.40/(50-5)  = 31.85

13.3.7.3  Summarize  these  calculations in an ANOVA table  (Table  13)


           TABLE 13.  ANOVA  TABLE  FOR DUNNETT'S PROCEDURE EXAMPLE
Source
Between
Within
df
4
45
Sum of Squares
(SS)
3887.88
1433.40
Mean Square(MS)
(SS/df)
971.97
31.85
    Total         49              5321.28
                                       185

-------
13.3.7.4  To perform the individual  comparisons, calculate the t statistic for
each concentration and control  combination as follows:
                         t.
                            SH y  (1/n,) +  (1/n,)

Where Y,-  =  mean  number  of young produced  for  effluent concentration i
      YI  =  mean  number  of young produced  for  the  control
      Su  =  square  root  of within mean  square
      n1  =  number  of  replicates for  the control
      n,-  =  number  of  replicates for  concentration i.

Since we  are looking  for a decrease  in reproduction from the control,  the mean
for concentration  i  is  subtracted from the control mean in the t statistic
above.  However, if we  were looking  for an increased response over the
control,  the control  mean would be subtracted from the mean at a
concentration.

13.3.7.5   Table 14 includes the calculated t  values for each concentration
and control combination.  In this example, comparing the 1.56% concentration
with the  control the  calculation is  as follows:

                              ( 22.4 - 26.3  )
                  t2  =  	 =  -1.55
                       [ 5.64 y  (1/10) + (1/10)  ]


                      TABLE 14.   CALCULATED T-VALUES
           Effluent Concentration (%)
1.56
3.12
6.25
12.5
2
3
4
5
-1.55
-4.84
-3.69
5.16
13.3.7.6  Since the purpose of this test is to detect a significant reduction
in mean reproduction, a one-sided test is appropriate.  The critical value
for this one-sided test is found in Table 5, Appendix C.  Since an entry for
45 degrees of freedom for error is not provided in the table, the entry for 40
degrees of freedom for error, an alpha level of 0.05 and four concentrations
(excluding the control) will be used, 2.23.  The mean reproduction for
concentration "i" is considered significantly less than the mean reproduction
for the control if t, is  greater than the critical  value.   Since t5  is  greater
than 2.23, the 12.5% concentration has significantly lower reproduction than

                                      186

-------
the control.   Hence the NOEC and the LOEC for reproduction are 6.25%
12.5%,  respectively.

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
       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 = 2.23 (5.64) /  (1/10) +  (1/10)
                       = 2.23 (5.64)(0.45)
                       = 5.66

13.3.7.9  Therefore, for this set of data, the minimum difference  that  can  be
detected as statistically significant is 5.66.

13.3.7.10  This represents a 25% decrease in mean reproduction  from  the
control.

13.3.8  Calculation of the 1C

13.3.8.1  The reproduction data in Table 4 are utilized  in this example.  As
can be seen from  Figure 7, the observed means are not monotonically
non-increasing with respect to concentration.  Therefore, the means  must  be
smoothed prior to calculating the 1C.

13.3.8.2  Starting with the observed control_mean, Y,= 22.4, and the observed
mean for the lowest effluent concentration, Y2 = 26.3, we see that 7,, is less
than Y2.

13.3.8.3  Calculate the smoothed means:

                M, = M2  =  (?, + Y2)/2 =  24.35

13.3.8.4  Since Y3 = 34.6 is larger than M2,  average  Y3 with  the previous
concentrations:

                Mj = M2  =  M3 = (M, + M2 + 73)/3  =  27.7.

13.3.8.5  Additionally, Y4 = 31.7 is larger than M3,  and  is pooled with the
first three means.  Thus:

       (M,  + M2 +  MS + 7J/4 = 28.7 = M1  = M2  =  M3 = M4


                                      187

-------
13.3.8.6  Since M4 > ?5  =  9.4,  set  M5 = 9.4.  Likewise, M5 > 7, = 0,
and M6 becomes 0.   Table 15 contains the  smoothed means and Figure 7
gives a plot of the smoothed means and the interpolated response curve.
                  TABLE 15.  DAPHNID, CERIODAPHNIA DUBIA, REPRODUCTION
                             MEAN RESPONSE AFTER SMOOTHING
Cone
Control
1.56
3.12
6.25
12.5
25.0
i
1
2
3
4
5
6
MI
28.75
28.75
28.75
28.75
9.40
0.00
13.3.8.7  Estimates of the IC25 and IC50 can be calculated using the Linear
Interpolation Method.   A 25% reduction in reproduction, compared to the
controls, would result in a mean reproduction of 21.56 young per adult, where
M^l  -  p/100)  = 28.75(1    25/100).   A  50% reduction  in reproduction,
compared to the controls, would result in a mean reproduction of 14.38 young
per adult, where M,(l  -  p/100)  = 28.75(1  -  50/100).   Examining  the smoothed
means and their associated concentrations (Table 15), the two effluent
concentrations bracketing 21.56 young  per adult are  C4 =  6.25%  effluent and
C5 =  12.5% effluent.   The two effluent  concentrations bracketing a response
of 14.38 young per adult are also C4 =  6.25  and C5 =  12.5.

13.3.8.8  Using the the Equation from  paragraph 4.2  in Appendix J, the IC25
estimate is 8.6% effluent:
               ICp
                               - p/100) - MJ(CJ+1
                                              (M,
                                                                6.25)
              IC25 = 6.25 + [28.75(1  -  25/100)  -  28.751(12.5
                                                       (9.40 - 28.75)
                   = 8.57% effluent

13.3.8.9  The IC50 estimate is  10.9%  effluent:

               ICp = Cj +  [Mjd - p/100)   MJ(CJ+1  - CJ
              IC50 = 6.25 + [28.75(1  -  50/100)  -  28.751(12.5 -  6.25)
                                                       (9.40 - 28.75)
                   = 10.89% effluent
                                     188

-------
       50-
       40
       30
     w
oo
to
       20-
       10-
       0 -\
         0.00
                           1.56
                                                 ****   INDIVIDUAL NUMBER OF YOUNG

                                                 	   CONNECTS THE OBSERVED UEAN VALUE

                                                 	   CONNECTS THE SMOOTHED MEAN VALUE
 3.12              8.25


EFFLUENT CONCENTRATION (X)
12.50
25.00
    Figure  7.   Plot  of raw  data, observed means, and  smoothed means for  the daphnid,  Ceriodaphnia dubia,

                reproductive data.

-------
13.3.8.10   When the Bootstrap program (BOOTSTRP) was usea to analyze this
data set for the IC25,  requesting 80 resamples, the mean estimate of the IC25
was 8.6% effluent,  with a standard deviation of 0.11% effluent (coefficient of
variation = 1.3%).   The empirical 95% confidence interval for the true mean
was (8.4 - 8.9% effluent).   The computer output for this data set is provided
in Figure 8.

13.3.8.11   When BOOTSTRP was used to analyze this data set for the IC50,
requesting 80 resamples,  the mean estimate of the IC50 was 11.1% effluent,
with a standard deviation of 0.22% effluent (coefficient of variation = 2%).
The empirical 95% confidence interval for the true mean was (10.6 - 11.5%
effluent).  The computer output for this data set is provided in Figure 9.
                                     190

-------
THE NUMBER OF RESAMPLES IS   80

*** LISTING OF GROUP CONCENTRATIONS (%EFF.) AND RESPONSE MEANS***

CONC. (%EFF)             RESPONSE MEAN         MEAN AFTER POOLING
    .000                   22.400                   28.750

   1.560                   26.300                   28.750

   3.120                   34.600                   28.750

   6.250                   31.700                   28.750

  12.500                    9.400                    9.400

  25.000                     .000                     .000
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
   FROM THE INPUT SAMPLE IS   8.5715.
             BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY
                        OF THE ESTIMATED ICp
THE MEAN OF THE BOOTSTRAP ESTIMATE IS  8.6486.

THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS   .1102.

AN EMPIRICAL 95.0% CONFIDENCE INTERVAL FOR THE
     BOOTSTRAP ESTIMATE IS (  8.4150,  8.8677).
          Figure 8.  Example of BOOTSTRP program output
                     for the IC25.
                                191

-------
THE NUMBER OF RESAMPLES IS   80

*** LISTING OF GROUP CONCENTRATIONS (%EFF.)  AND RESPONSE MEANS***

CONC.  (%EFF)             RESPONSE MEAN         MEAN AFTER POOLING
    .000                   22.400                   28.750

   1.560                   26.300                   28.750

   3.120                   34.600                   28.750

   6.250                   31.700                   28.750

  12.500                    9.400                    9.400

  25.000                     .000                     .000
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
   FROM THE INPUT SAMPLE IS  10.8931.
    *        BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY        *
    *                   OF THE ESTIMATED ICp                    *
    *************************************************************

THE MEAN OF THE BOOTSTRAP ESTIMATE IS   11.0473.

THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS   .2205.

AN EMPIRICAL 95.0% CONFIDENCE INTERVAL FOR THE
     BOOTSTRAP ESTIMATE IS (   10.5800,  11.4854).
          Figure 9.   Example of BOOTSTRP program output
                     for the IC50.
                                192

-------
14.   PRECISION AND ACCURACY

14.1  PRECISION

14.1.1  Single Laboratory Precision

14.1.1.1  Information on the single-laboratory precision of the daphnid,
Ceriodaphm'a dubia, reproduction test based on the NOEC and LOEC values from
nine tests with the reference toxicant sodium pentachlorophenate (NaPCP) is
provided in Table 16.  The NOECs and LOECs of all tests fell in the same
concentration range, indicating maximum possible precision.  Table 17 gives
precision data for the IC25 and IC50 values for seven tests with the reference
toxicant NaPCP.  Coefficient of variation was 41% for the  IC25 and 28% for the
IC50.

14.1.2  Multilaboratory Precision

14.1.2.1  An interlaboratory study was performed by the Aquatic Biology
Branch, EMSL-Cincinnati in 1985, involving a total of 11 analysts in 10
different laboratories (Neiheisel et. al., 1988a; USEPA, 1988e).  Each analyst
performed one-to-three seven-day tests using aliquots of a copper-spiked
effluent sample, for a total of 25 tests.  The tests were  performed on the
same day in all participating laboratories, using a prepubl ication draft of
Method 1002.  Some deviations from the standard protocol were reported by the
participating laboratories.

14.1.2.1.1  Ten sets of data from six laboratories met the acceptability
criteria, and were statistically analyzed using nonparametric procedures to
determine NOECs and LOECs.  The NOECs and LOECs for these  tests were within
one concentration interval which, with a dilution factor of 0.5, is equivalent
to a twofold range in concentration  (Table 18).

14.1.2.2  A second interlaboratory study of Method 1002.0  (using the first
edition of this manual; USEPA, 1985c), was coordinated by  Battelle,
Columbus Division, and involved 11 participating laboratories (Table 19)
(DeGraeve et al., 1989).  All participants used 10% DMW (10% PERRIER* Water)
as the culture and dilution water, and used their own formulation of food for
culturing and testing the Ceriodaphm'a dubia.  Each laboratory was to conduct
at least one test with each of eight blind samples.  Each  test consisted of 10
replicates of one organism each for  five toxicant concentrations and a
control.  Of the 116 tests planned,  91 were successfully initiated, and 70
(77%) met the survival and reproduction criteria for acceptability of the
results (80% survival and nine young per initial female).  If the reproduction
criteria of 20 young/female, used in this edition of the method, had been
applied to the results of the interlaboratory study, 22 additional tests would
have been unacceptable.  The overall precision (CV) of the test was 27% for
the survival data (7-day LC50s) and  37% for the reproduction data (ICSOs).

14.2  ACCURACY

14.2.1  The accuracy of toxicity tests cannot be determined.


                                      193

-------
TABLE 16.  SINGLE LABORATORY PRECISION OF' THE DAPHNID, CERIODAPHNIA DUBIA,
           SURVIVAL AND REPRODUCTION TEST, USING NAPCP AS A REFERENCE
           TOXICANTa'b
NOEC
Test (mq/L)
lc 0.25
2d 0.20
3 0.20
4e 0.30
5 0.30
6 0.30
7 0.30
8 0.30
9 0.30
LOEC
(mq/L)
0.50
0.60
0.60
0.60
0.60
0.60
0.60
0.60
0.60
Chronic
Value
(mq/L)
0.35
0.35
0.35
0.42
0.42
0.42
0.42
0.42
0.42
aFor a discussion  of the  precision  of data from chronic toxicity
 tests see Section 4, Quality Assurance.
bData from tests  performed  by Philip  Lewis,  Aquatic Biology Branch,
 EMSL-Cincinnati.   Tests  were conducted in reconstituted hard water
 (hardness = 180  mg CaCOyL;  pH  = 8.1).
Concentrations used in Test  1 were:  0.03,  0.06,  0.12,  0.25,  0.50,
 1.0 mg NaPCP/L.
Concentrations used in Tests 2  and 3 were:  0.007,  0.022,  0.067,
 0.20, 0.60 mg NaPCP/L.
Concentrations used in Tests 4  through  9  were:  0.0375,  0.075,
 0.150, 0.30,  0.60 mg NaPCP/L.
 TABLE 17.  THE DAPHNID,  CERIODAPHNIA DUBIA,  SEVEN-DAY SURVIVAL AND
            REPRODUCTION TEST PRECISION FOR A SINGLE LABORATORY
            USING NAPCP AS THE REFERENCE TOXICANT1
Test Number         NOEC (mg/L)         IC25  (mg/L)          IC50 (mg/L)
19
46A
46B
49
55
56
n
Mean
CV(%)
0.30
0.20
0.20
0.20
0.20
0.10
7
NA
NA
0.3754
0.0938
0.2213
0.2303
0.2306
0.2241
7
0.2157
41.1
0.4508
0.2608
0.2879
0.2912
0.3177
0.2827
7
0.2953
27.9
Vrom USEPA  (1991c).
                                      194

-------
     TABLE  18.   INTERLABORATORY  PRECISION  FOR THE  DAPHNID,  CERIODAPHNIA
                 DUBIA,  SURVIVAL  AND  REPRODUCTION TEST  WITH  COPPER SPIKED
                 EFFLUENT1
Analyst
3
4
4
5
5
6
6
10
10
11
Test
1
1
2
1
2
1
2
1
2
1
Endpoi
Reproducti
NOEC
12
6
6
6
12
12
6
6
6
12
nts (% Effluent)
on
LOEC
25
12
12
12
25
25
12
12
12
25
Survi
NOEC
25
12
25
12
12
25
25
12
12
25
val
LOEC
50
25
50
25
25
50
50
25
25
50
1From USEPA  (1988e).
                                     195

-------
TABLE 19.  INTERLABORATORY PRECISION DATA FOR THE DAPHNID, CERIODAPHNIA
           DUBIA, SUMMARIZED FOR EIGHT REFERENCE TOXICANTS AND EFFLUENTS3
Test Material
Sodium chloride
Industrial
Sodium chloride
Pulp and Paper
Potassium dichromate
Pulp and Paper
Potassium dichromate
Industrial
n
Mean
Standard Deviation
Mean IC50
1.34
3.6
0.96
60.0
35.8
70.2
53.2
69.8



CV%
29.9
83.3
57.4
28.3
30.8
7.5
25.9
37.0
8
37.5
23.0
Mean IC25
1.00
3.2
0.09
47.3
23.4
55.7
29.3
67.3



CV%
34.3
78.1
44.4
27.0
32.7
12.2
46.8
36.7
8
39.0
19.1
3From  USEPA  (1991c).
                                     196

-------
                                 SECTION 14

                                 TEST HETHOD

                GREEN ALGA, SELENASTRUM CAPRICORNUTUM, GROWTH TEST
                                METHOD 1003.0

1.  SCOPE AND APPLICATION

1.1  This method measures the chronic toxicity of effluents and
receiving water to the freshwater green alga, Selenastrum capn'cornutum, in a
four-day static 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.

1.2  Daily observations on growth make it possible to also calculate acute
toxicity for desired exposure periods (i.e., 24-h, 48-h, 96-h ECSOs).

1.3  Detection limits of the toxicity of an effluent or pure substance are
organism dependent.

1.4  Brief 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
degradable or highly volatile toxicants present in the source may not
be detected in the test.

1.5  This test 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.

1.6  This test is very versatile because it can also be used to identify
wastewaters which are biostimulatory and may cause nuisance growths of algae,
aquatic  weeds, and other organisms at higher trophic levels.

2.  SUMMARY OF METHOD

2.1  A green alga, Selenastrum capricornutum, population is exposed in a
static system to a series of concentrations of effluent, or to receiving
water, for 96 h.  The response of the population is measured in terms of
changes  in cell density (cell counts per mL), biomass, chlorophyll content, or
absorbance.

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).


                                      197

-------
3.2  Adverse effects of high concentrations of suspended and/or dissolved
solids, color, and extremes of pH may mask the presence 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 organisms and/or planktivores in the dilution water and
effluent may affect test organism survival and growth, and confound test
results.

3.5  Nutrients in the effluent or dilution water may confound test results.

4.  SAFETY

4.1  See Section 3, Safety and Health.

5.  APPARATUS AND EQUIPMENT

5.1  Laboratory Selenastrum capn'cornutum culture unit -- see culturing
methods below.  To test effluent toxicity, sufficient numbers of
log-phase-growth organisms must be available.

5.2  Samplers -- automatic sampler, preferably with sample cooling capability,
that can collect a 24-h composite sample of 1 L or more.

5.3  Sample containers -- for sample shipment and storage (see Section 8,
Effluent and Receiving Water Sampling, Sample Handling, and Sample Preparation
for Toxicity Tests).

5.4  Environmental chamber, incubator, or equivalent facility -- with
"cool-white" fluorescent illumination (86 ± 8.6 /uE/m2/s,  400  ± 40 ft-c,  or
4306 lux) and temperature control (25 ± 1°C).

5.5  Mechanical shaker -- capable of providing orbital motion at the rate of
100 cycles per minute (cpm).

5.6  Light meter -- with a range of 0-200 p.E/m*/s (0-1000 ft-c)..

5.7  Water purification system -- MILLIPORE MILLI-QR deionized water or
equivalent (see Section 5, Facilities, Equipment, and Supplies).

5.8  Balance -- analytical, capable of accurately weighing 0.00001 g.

5.9  Reference weights, class S -- for .checking performance of balance.

5.10  Volumetric flasks and graduated cylinders -- class A, 10-1000 mL,
borosilicate glass, for culture work and preparation of test solutions.

5.11  Volumetric pipets -- class A, 1-100 mL.

5.12  Serological  pipets -- 1-10 mL, graduated.

                                      198

-------
5.13  Pipet bulbs and fillers -- PROPIPETR,  or equivalent.

5.14  Wash bottles -- for rinsing small glassware, instrument electrodes, and
probes.

5.15  Test chambers -- four (minimum of three) 125 or 250 mL borosilicate,
Erlenmeyer flasks, with foam plugs or stainless steel or Shumadzu closures.
For special glassware cleaning requirements see Section 5, Facilities,
Equipment, and Supplies.

5.16  Culture chambers -- 1-4 L borosilicate, Erlenmeyer flasks.

5.17  Glass or electronic thermometers -- for measuring water temperatures.

5.18  Bulb-thermograph or electronic-chart type thermometers -- for
continuously recording temperature.

5.19  National Bureau of Standards Certified thermometer (see EPA Method
170.1, USEPA 1979b).

5.20  Meters, pH and specific conductivity -- for routine physical and
chemical  measurements.

5.21  Tissue grinder -- for chlorophyll extraction.

5.22  Fluorometer (Optional) -- equipped with chlorophyll detection light
source,  filters, and photomultiplier tube (Turner Model 110 or equivalent).

5.23  UV-VIS spectrophotometer -- capable of accommodating 1-5 cm cuvettes.

5.24  Cuvettes for spectrophotometer -- 1-5 cm light path.

5.25  Electronic particle counter (Optional) -- Coulter Counter, Model ZBI,
or equivalent, with mean cell (particle) volume determination.

5.26  Microscope -- with 10X, 45X, and 100X objective lenses, 10X ocular
lenses,  mechanical stage, substage condenser, and light source (inverted or
conventional microscope).

5.27  Counting chamber -- Sedgwick-Rafter, Palmer-Maloney, or hemocytometer.

5.28  Centrifuge -- with swing-out buckets having a capacity of 15-100 ml.

5.29  Centrifuge tubes -- 15-100 ml, screw-cap.

5.30  Filtering apparatus -- for membrane and/or glass fiber filters.

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

                                      199

-------
for Toxicity Tests).

6.2  Data sheets (one set per test) -- for recording data.

6.3  Tape, colored -- for labeling test chambers.

6.4  Markers,  water-proof -- for marking containers, etc.

6.5  Reagents  for hardness and alkalinity tests (see EPA Methods 130.2 and
310.1,  USEPA 1979b).

6.6  Standard  pH buffers 4,  7, and 10 (or as per instructions of instrument
manufacturer)  for instrument calibration (see USEPA Method 150.1, USEPA
1979b).

6.7  Specific  conductivity standards (see EPA Method 120.1, USEPA 1979b).

6.8  Standard  particles -- such as chicken or turkey fibroblasts or polymer
microspheres,  5.0 ± 0.03 im diameter, 65.4 jum3 volume,  for calibration of
electronic particle counters (available from Duke Scientific Co., 1135D, San
Antonio Road,  Palo Alto, California 94303).

6.9  Membranes and filling solutions for DO probe (see Method 360.1, USEPA
1979b), or reagents for modified Winkler analysis.

6.10  Laboratory quality control samples and standards -- for calibration of
the above methods.

6.11  Reference toxicant solutions (see Section 4,  Quality Assurance).

6.12  Reagent  water -- defined as distilled or deionized water that does not
contain substances which are toxic to the test organisms (see Subsection 5.7
above).

6.13  Effluent or 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).

6.14  Acetone  -- pesticide-grade or equivalent.

6.15  Dilute (10%) hydrochloric acid  -- carefully add 10 mL of concentrated
HC1 to 90 ml of MILLI-QR water.

6.16  TEST ORGANISMS, GREEN ALGA, SELENASTRUM CAPRICORNUTUM

6.16.1   Selenastrum capricornutum, a unicellular coccoid green alga is the
test organism.

6.16.2   Algal  Culture Medium is prepared as follows:

  1. Prepare (five) stock nutrient solutions using reagent grade chemicals as
     described in Table 1.  Cautionary note:  EDTA may affect metal toxicity.

                                      200

-------
     It  is  recommended that tests be conducted with and without EDTA in the
     culture media if metals are suspected in the effluent or receiving
     water.
  2.  Add 1  mL of each stock solution,  in the order listed in Table 1,  to
     approximately 900 mL of MILLI-Q  water.   Mix well  after the  addition  of
     each solution.   Dilute to 1 L,  mix well, and adjust the pH to 7.5 ±
     0.1, using 0.1N NaOH or HC1, as appropriate.  The final concentration of
     macronutrients  and micronutrients in the culture medium is given  in
     Table  2.
  3.  Immediately filter the pH-adjusted medium through a 0.45 urn pore  diameter
     membrane at a vacuum of not more  than 380 mm (15 in.) mercury,  or at  a
     pressure of not more than one-half atmosphere (8 psi).   Wash the  filter
     with 500 ml deionized water prior to use.
  4.  If  the filtration is carried out  with sterile apparatus, filtered medium
     can be used immediately,  and no further sterilization steps are required
     before the inoculation of the medium.  The medium can also be sterilized
     by  autoclaving  after it is placed in the culture vessels.  If a 0.22  p,g
     filter is used  (ASTM, 1988) no sterilization is needed.
  5.  Unused sterile  medium should not  be stored more than one week prior to
     use, because there may be substantial loss of water by evaporation.

6.16.3  Stock Algal  Cultures

6.16.3.1  See Section 6, Test Organisms, for information on  sources  of
"starter" cultures of the green alga,  Selenastrum capricornutum.

6.16.3.2  Upon receipt of the "starter" culture (usually about 10 ml),  a stock
culture  is  initiated by aseptically transferring 1 mL to a culture flask
containing  control algal culture medium (prepared as described above).   The
volume of stock culture medium initially prepared will  depend upon the number
of test  flasks to be inoculated later  from the stock, or other planned uses,
and may  range from 25 mL in a 125 mL flask to 2 L in a 4-L flask.  The
remainder of the starter culture can be held in reserve for up to six  months
in a refrigerator (in the dark) at 4°C.

6.16.3.3  Maintain the stock cultures  at 25 ± 1°C,  under continuous
"Cool-White" fluorescent lighting of 86 ± 8.6 ^E/m2/s (400 ± 40  ft-c).
Shake continuously at 100 cpm or twice daily by hand.

6.16.3.4  Transfer 1 to 2 mL of stock  culture weekly to 50   100 mL  of new
culture  medium to maintain a continuous supply of "healthy"  cells for  tests.
Aseptic  techniques should be used in maintaining the algal cultures, and
extreme  care should  be exercised to avoid contamination.  Examine the  stock
cultures with a microscope for contaminating microorganisms  at each  transfer.

6.16.3.5  Viable unialgal culture material may be maintained for long  periods
of time  if  placed in a refrigerator at 4°C.

6.16.4  It  is recommended that chronic toxicity tests be performed monthly
with  a reference toxicant.  Algal cells four to seven days old are used to
monitor  the chronic  toxicity (growth)  of the reference toxicant to the algal
stock produced by the culture unit.   See Section 4, Quality Assurance,

                                      201

-------
   TABLE 1. NUTRIENT STOCK SOLUTIONS FOR MAINTAINING ALGAL  STOCK  CULTURES
            AND TEST CONTROL CULTURES.
STOCK COMPOUND
SOLUTION
1. MACRONUTRIENTS
A. MgCl2-6H20
CaCl,-2H20
NaN03
B. MgS04-7H20
C. K2HP04
D. NaHC03
2. MICRONUTRIENTS
H3BO,
MnCl2 4H20
ZnCl2
FeCl,-6H20
CoCl2-6H20
Na2Mo04-2H20
CuCl2-2H20
Na2EDTA-2H20
Na2Se04
AMOUNT DISSOLVED IN
500ML MILLI-Q" WATER

6.08 g
2.20 g
12.75 g
7.35 g
0.522 g
7.50 g

92.8 mg
208.0 mg
1.64 mga
79.9 mg
0.714 mgb
3.63 mgc
0.006 mgd
150.0 mg
1.196 mge
3ZnCl2  - Weigh out  164 mg  and dilute to  100 mL.  Add  1  mL  of this
 solution to Stock #1.

bCoCl2-6H,0 - Weigh out 71.4 mg and dilute to 100 mL.  Add  1  mL of
 this solution to Stock #1.

cNa,Mo04-2H20 - Weigh out 36.6 mg and dilute to 10 mL.  Add 1 mL
 of this solution to Stock #1.
dCuCl2-2H20   Weigh out 60.0 mg and dilute to 1000 mL.  Take  1 mL
 of this solution and dilute to 10 mL.  Take 1 mL of the second di
 and add to Stock #1.
dilution
sNa2Se04 - Weigh out 119.6 mg and dilute to 100 mL.  Add  1 mL of this
 solution to Stock #1.
                                      202

-------
TABLE 2.  FINAL CONCENTRATION OF MACRONUTRIENTS AND MICRONUTRIENTS
          IN THE CULTURE MEDIUM
MACRONUTRIENT
NaN03
MgCl2-6H20
CaCl2-2H20
MgS04-7H20
K2HP04
NaHC03


MICRONUTRIENT
H3B03
MnCl2-4H20
ZnCl2
CoCl2-6H20
CuCl2-2H20
Na2Mo04-2H20
FeCl3-6H20
Na2EDTA-2H20
Na,SeO,
CONCENTRATION
(mq/L)
25.5
12.2
4.41
14.7
1.04
15.0


CONCENTRATION
(UQ/l)
185
416
3.27
1.43
0.012
7.26
160
300
2.39
ELEMENT
N
Mg
Ca
S
P
Na
K
C
ELEMENT
B
Mn
Zn
Co
Cu
Mo
Fe
--
Se
CONCENTRATION
(mq/L)
4.20
2.90
1.20
1.91
0.186
11.0
0.469
2.14
CONCENTRATION
fuq/L)
32.5
115
1.57
0.354
0.004
2.88
33.1
	
0.91
                                 203

-------
Subsection 4.17, Reference Toxicants.

6.16.5  Record Keeping

6.16.5.1  Records, kept in a bound notebook, include (1) dates culture media
was prepared, (2) source of "starter" cultures, (3) date stock cultures were
started,. (4) cell density in stock cultures, and (5) dates and results of
reference toxicant tests performed (see Section 4, Quality Assurance).

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.

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.
Receiving water toxicity is determined with samples used directly as collected
or after samples are passed through a 60 urn NITEXR filter and compared without
dilution against a control.  Using four replicate chambers per test, each
containing 100 mL and 400 ml for chemical analyses, would require
approximately 1 L or more of sample for the 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 five effluent
concentrations (6.25%, 12.5%, 25%, 50%, and 100%).  Improvements in precision
decline rapidly if the dilution factor is increased beyond 0.5 and precision
declines rapidly if a smaller dilution factor is 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%).  If a high rate of mortality is observed during the first
1 to 2 h of the test, additional dilutions should be added at the lower  range
of the effluent concentrations.
                                      204

-------
10.1.2.3   The volume of effluent required for the test is 1 to 2 L.
Sufficient test solution (approximately 900 or 1500 ml) is prepared at each
effluent  concentration to provide 400 ml additional volume for chemical
analyses  at the high,  medium,  and low test concentrations.  There is no daily
renewal  of test solution.

10.1.2.4   Tests should begin as soon as possible, preferably within 24h 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.

10.1.2.5   Just prior to test initiation (approximately one h) the temperature
of sufficient quantity of the sample to make test solutions should be adjusted
to the test temperature and maintained at that temperature during the addition
of dilution water.

10.1.2.6   The DO of the test solutions should be checked prior to test
initiation.  If any of the solutions are supersaturated with oxygen or any
solution  has a DO concentration below 4.0 mg/L, all of the solutions and the
control  must be gently aerated.

10.1.2.7   Effluents may be toxic and/or nutrient poor.  "Poor" growth in an
algal toxicity test, therefore, may be due to toxicity or nutrient limitation,
or both.   To eliminate false negative results due to low nutrient
concentrations, 1 ml of each stock nutrient solution is added per liter of
effluent  prior to use  in preparing the test dilutions.  Thus, all test
treatments and controls will contain at a minimum the concentration of
nutrients in the stock culture medium.

10.1.2.8   If samples contain volatile substances, the test sample should be
added below the surface of the dilution water towards the bottom of the test
container through an appropriate delivery tube.

10.1.3  Dilution Water

10.1.3.1   Dilution water may be stock culture medium, any uncontaminated
receiving water, a standard synthetic (reconstituted) water,  or some other
natural  water (see Section 7,  Dilution Water).  However,  if water other than
the stock culture medium is used for dilution water, 1 ml of each stock
nutrient  solution should be added per liter of dilution water.  Natural waters
used as  dilution water must be filtered through a prewashed filter, such as a
GF/A, GF/C, or equivalent filter, that provides 0.45 /xm particle size
retention.

10.1.3.2   If the growth of the algae in the test solutions is to be measured
with an  electronic particle counter, the effluent and dilution water must be
filtered  through a GF/A or GF/C filter, or other filter providing 0.45 jum
particle  size retention, and checked for "background" particle count before it
is used  in the test.  Glass-fiber filters generally provide more rapid
filtering rates and greater filtrate volume before plugging.


                                      205

-------
10.1.4  Preparation of Inoculum

10.1.4.1  The inoculum is prepared no more than 2 to 3 h prior to the
beginning of the test, using Selenastrum capricornutum harvested from a four-
to-seven-day stock culture.   Each millilHer of inoculum must contain enough
cells to provide an initial  cell  density of approximately 10,000 cells/ml
(± 10%) in the test flasks.   Assuming the use of 250 ml flasks,  each
containing 100 ml of test solution,  the inoculum must contain 1,000,000
cells/ml.

10.1.4.2  Estimate the volume of  stock culture required to prepare the
inoculum.  As an example, if the  four-to-seven-day-old stock culture used as
the source of the inoculum has a  cell  density of 2,000,000 cells/ml, a test
employing 24 flasks, each containing 100 ml of test medium and inoculated with
a total of 1,000,000 cells,  would require 24,000,000 cells or 15 ml of stock
solution (24,000,000/2,000,000) to provide sufficient inoculum.   It is
advisable to prepare a volume 20% to 50% in excess of the minimum volume
required, to cover accidental  loss in transfer and handling.

10.1.4.3  Prepare the inoculum as follows:

    1. Centrifuge 15 ml of stock  culture at 1000 x g for 5 min.
       This volume will provide a 50% excess in the number of cells.
    2. Decant the supernatant and resuspend the cells in 10 ml of
       control medium.
    3. Repeat the centrifugation  and decantation step, and resuspend
       the cells in 10 ml control  medium.
    4. Mix well  and determine the cell  density in the algal
       concentrate.  Some cells will be lost in the concentration
       process.
    5. Determine the density of cells (cells/ml) in the stock
       culture (for this example,  assume 2,000,000 per ml).
    6. Calculate the required volume of stock culture as follows:


Volume (ml) of      Number of test      Volume of Test
Stock Culture   =   flasks to be  used X Solution/Flask X 10,000 cells/ml
Required             Cell density (cells/mL) in the stock culture

                =       24 flasks X 100 mL/flask X 10.000 cells/ml
                                 2,000,000 cells/mL

                    12.0 ml Stock Culture


    7. Dilute the cell concentrate as needed to obtain a cell
       density of 1,000,000 cells/mL,  and check the cell density in
       the final inoculum.
    8. The volume of the algal inoculum should be considered in
       calculating the dilution of toxicant in the test flasks.
                                     206

-------
10.2  START OF THE TEST

10.2.1  Label  the test chambers with a marking pen and use the color-coded
tape to identify each treatment and replicate.  A minimum of five effluent
concentrations and a control are used for each effluent test.  Each treatment
(including the control) should have four (minimum of three) replicates.

10.2.2  Randomize the position of the test flasks at the beginning of the test
(see Appendix A).  Preparation of a position chart may be helpful.

10.2.3  The test begins when the algae are added to the test flasks.  Mix the
inoculum well, and add 1 ml to the test solution in each randomly arranged
flask.  Make a final check of the cell density in three of the test solutions
at time "zero" (within 2 h of the inoculation).

10.2.3.1  Alkalinity, hardness, and conductivity are measured at the
beginning of the test in the high, medium, and low effluent concentrations and
control before they are dispensed to the test chambers and the data recorded
on the data sheet (Figure 1).

10.3  LIGHT, PHOTOPERIOD, AND TEMPERATURE

10.3.1  Test flasks are incubated under continuous illumination at 86 ± 8.6
ME/nr/s (400 ± 40 ft-c),  at 25 ± 1°C,  and  should  be  shaken  continuously  at  100
cpm on a mechanical shaker or twice daily by hand.  Flask positions in the
incubator should be randomly rotated each day to minimize possible spatial
differences in illumination and temperature on growth rate.  If it can be
verified that test specifications are met at all  positions, this need not be
done.

10.4  DISSOLVED OXYGEN (DO) CONCENTRATION

10.4.1  Because of the continuous illumination of the test flasks, DO
concentration should never be a problem during the test and no aeration will
be required.

10.5  OBSERVATIONS DURING THE TEST

10.5.1  Minimum Routine Chemical and Physical Observations

10.5.1.1  Temperature should be monitored continuously or observed and
recorded daily for at least two locations in the environmental control system
or the samples.  Temperature should be checked in a sufficient number of test
vessels at least at the end of the test to determine variability in the
environmental  chamber.

10.5.1.2  Temperature and pH are measured at the end of each 24-h exposure
period in at least one test flask at each concentration and in the control.

10.5.1.3  Record all the measurements on the data sheet (Figure 1).
                                      207

-------
10.5.2  Biological  Observations

10.5.2.1  Toxic substances in the test solutions may degrade or volatilize
rapidly, and the inhibition in algal  growth may be detectable only during the
first one or two days in the test.   It may be desirable, therefore, to
determine the algal growth response daily.  Otherwise, biological observations
are not required until  the test is  terminated and the test solutions are not
renewed during the  test period.

10.6  TERMINATION OF THE TEST

10.6.1  The test is terminated 96 h after initiation.  The algal growth in
each flask is measured  by one of the following methods:  (a) cell counts, (b)
chlorophyll content, or (c) turbidity (light absorbance).

10.6.2  Cell counts

10.6.2.1  Automatic Particle Counters

10.6.2.1.1  Several types of automatic electronic and optical particle
counters are available  for use in the rapid determination of cell density
(cells/mL) and mean cell volume (MCV) in /Lon3/cell.   The Coulter Counter is
widely used and is  discussed in detail by Miller et al. (USEPA, 1978b).

10.6.2.1.2  If biomass  data are desired for algal growth potential
measurements, a Model ZM Coulter Counter is used.  However, the instrument
must be calibrated  with a reference sample of particles of known volume.

10.6.2.1.3  When the Coulter Counter is used, an aliquot (usually 1 mL) of the
test culture is diluted 10X to 20X  with a 1% sodium chloride electrolyte
solution, such as ISOTONR,  to facilitate counting.   The resulting dilution is
counted using an aperture tube with a 100-/im diameter aperture.  Each cell
(particle) passing  through the aperture causes a voltage drop proportional to
its volume.  Depending  on the model,  the instrument stores the information on
the number of particles and the volume of each, and calculates the mean cell
volume.  The following  procedure is used:

    1. Mix the algal culture in the flask thoroughly by swirling the contents
       of the flask approximately six times in a clockwise direction, and then
       six times in the reverse direction; repeat the two-step process at
       least once.
    2. At the end of the mixing process, stop the motion of the liquid in the
       flask with a strong brief reverse mixing action, and quickly remove
       1 ml of cell culture from the flask with a sterile pipet.
    3. Place the aliquot in a counting beaker, and add 9 ml (or 19 ml) of
       electrolyte  solution (such as Coulter ISOTONR).
    4. Determine the cell density (and MCV, if desired).

10.6.2.2  Manual microscope counting method

10.6.2.2.1  Cell counts may be determined using a Sedgwick-Rafter,
Palmer-Maloney, hemocytometer, inverted microscope, or similar methods.   For

                                      208

-------
details on microscope counting methods, see APHA, 1985, and USEPA, 1973.
Whenever feasible, 400 cells per replicate are counted to obtain ± 10%
precision at the 95% confidence level.  This method has the advantage of
allowing for the direct examination of the condition of the cells.

10.6.3  Chlorophyll Content

10.6.3.1  Chlorophyll may be estimated in-vivo fluorometrically, or in-vitro
either fluorometrically or spectrophotometrically.  In-vivo fluorometric
measurements are recommended because of the simplicity and sensitivity of the
technique and rapidity with which the measurements can be made  (Rehnberg et
al., 1982).

10.6.3.2  The in-vivo chlorophyll measurements are made as follows:

    1. Adjust the "blank" reading of the fluorometer using the  filtrate from
       an equivalent dilution of effluent filtered through a 0.45 jum particle
       retention filter.
    2. Mix the contents of the test culture flask by swirling successively in
       opposite directions (at least three times), and remove 1 ml of culture
       from the flask with a sterile pipet.
    3. Place the aliquot in a small disposable vial and record  the
       fluorescence as soon as the reading stabilizes.  (Do not allow the
       sample to stand in the instrument more than 1 min).
    4. Discard the sample.

10.6.3.3  For additional information on chlorophyll measurement methods, see
APHA, 1985.

10.6.4  Turbidity (Absorbance)

10.6.4.1  A second rapid technique for growth measurement involves the use of
a spectrophotometer to determine the turbidity, or absorbance,  of the cultures
at a wavelength of 750 nm.  Because absorbance is a complex function of the
volume, size, and pigmentation of the algae, it would be useful to construct a
calibration curve to establish the relationship between absorbance and cell
density.

10.6.4.2  The algal growth measurements are made as follows:

    1. A blank is prepared as described for the fluorometric analysis.
    2. The culture is thoroughly mixed as described above.
    3. Sufficient sample is withdrawn from the test flask with  a sterile pipet
       and transferred to a 1- to 5-cm cuvette.
    4. The absorbance is read at 750 nm and divided by the light path length
       of the cuvette, to obtain an "absorbance-per-centimeter" value.
    5. The 1-cm absorbance values are used in the same manner as the cell
       counts.

10.6.5  Record the data as indicated in Figure 2.
                                      209

-------
11.  SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA

11.1  A summary of test conditions and test acceptability criteria is
presented in Table 3.

12.  ACCEPTABILITY OF TEST RESULTS

12.1  For the test results to be acceptable, the algal cell density in the
control flasks must exceed 2 X 10  cells/mL at the end of the test,  and not
vary more than 20% among replicates.

13.  DATA ANALYSIS

13.1  GENERAL

13.1.1  Tabulate and summarize the data.   A sample set of algal growth
response data is shown in Table 4.

13.1.2  The endpoints of toxicity tests using the green alga, Selenastrum
capn'cornutum, are based on the adverse effects on cell growth (see Section
9).  The EC50, the IC25, and the IC50 are calculated using the point
estimation techniques, and LOEC and NOEC  values for growth are obtained using
a hypothesis testing approach such as Dunnett's Procedure (Dunnett,  1955) or
Steel's Many-one Rank Test (Steel, 1959;  Miller, 1981).  Separate analyses are
performed for the estimation of the LOEC  and NOEC endpoints and for the
estimation of the EC50, IC25, and IC50.  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.
                                      210

-------
 TABLE 3.   SUMMARY  OF  TEST  CONDITIONS  AND  TEST  ACCEPTABILITY  CRITERIA FOR
            GREEN ALGA,  SELENASTRUM CAPRICORNUJUM,  GROWTH  TOXICITY TESTS
            WITH EFFLUENTS AND  RECEIVING WATERS
     1. Test  type:
     2. Temperature:
     3. Light quality:
     4. Light intensity:

     5. Photoperiod:
     6. Test  chamber  size:
     7. Test  solution volume:
     8. Renewal  of  test  solutions;
     9. Age of test organisms:
     9. Initial  cell  density in
         test chambers:
    10. No. replicate chambers
         per  concentration:
    11. Shaking rate:

    12. Aeration:
    13. Dilution water:
    14.  Effluent concentrations:
Static nonrenewal
25 ± 1°C
"Cool white" fluorescent lighting
86 ± 8.6 nE/mz/s (400 ± 40  ft-c
4306 lux)
Continuous illumination
125 mL or 250 mL
50 mL or 100 mL1
None
4 to 7 days

10,000 cells/mL
4 (minimum of 3)
100 cpm continuous, or twice daily
by hand
None
Algal stock culture medium, enriched
uncontaminated source of receiving or
other natural water, synthetic water
prepared using MILLIPORE MILLI-QR or
equivalent deionized water and reagent
grade chemicals, or DMW (see Section
7, Dilution Water)
Minimum of 5 and a control
1For tests not continuously  shaken use  25 mL  in  125 mL  flasks  and  50 mL  in  250
 mL flasks.
                                     211

-------
TABLE 3.  SUMMARY OF TEST CONDITIONS AND TEST ACCEPTABILITY CRITERIA FOR
          GREEN ALGA,  SELENASTRUM CAPRICORNUTUM, GROWTH TOXICITY TESTS
          WITH EFFLUENTS AND RECEIVING WATERS (CONTINUED)
  15.  Test dilution factor:


  16.  Test duration:

  17.  Endpoint:


  18.  Test acceptability:



  19.  Sampling requirement:
Effluents:  > 0.5
Receiving Waters:

96 h
None or > 0.5
 20.  Sample volume required:
Growth (cell counts, chlorophyll
fluorescence, absorbance, biomass)

2 X 105 cells/mL in the controls;
Variability of controls should not
exceed 20%

For on-site tests, one sample
collected at test initiation, and used
within 24 h of the time it is removed
from the sampling device.  For off-
site tests, holding time must not
exceed 36 h

1 or 2 L depending on test volume
                                   212

-------
      Figure 1.   Data form for the green alga, Selenastrum cap.ricornutum,
                 growth test.  Routine chemical and physical determinations,
Discharger:
Location:  _
Test Dates:
Analyst: 	
                          Effluent Concentration
Parameter
Temperature
oH
Alkalinity
Hardness
Conductivity
Chlorine

Control























































Remarks







       Figure 2.   Data form for the green alga, Selenastrum capricornutum,
                  growth test.  Cell density determinations.
Discharger:
Location:
Test Dates:
Analyst: 	
Concentration
Control
Cone:
Cone:
Cone:
Cone:
Cone:
Cell
1






Density Measurement
Replicate
2






3






4






Treatment
Mean






Comments






Comments:
                                      213

-------
            STATISTICAL ANALYSIS OF ALGAL GROWTH TEST
                       GROWTH RESPONSE DATA
                             CELLS / ML
  POINT ESTIMATION
 ENDPOINT ESTIMATE

     IC25, IC50
                        SHAPIRO-WILK'S TEST
               NORMAL DISTRIBUTION
                 NON-NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
        NO
                           BARTLETTS TEST
                      HETEROGENEOUS
                         VARIANCE
               EQUAL NUMBER OF
                 REPLICATES?
                 YES
 T-TESTWITH

ADJUSTMENT
         EQUAL NUMBER OF
           REPLICATES?
                            NO
           YES
                 niiwwirrrQ
                 DUNNtl I S
                   TP
-------
  TABLE 4.   GREEN  ALGA,  SELENASTRUM CAPRICORNUTUM,  GROWTH RESPONSE DATA
                                 Toxicant Concentration (ug Cd/L)
     Replicate   Control
10
20
40
80

A
B
C
Log10 A
Trans- B
formed C



Mean(Y,.)
s?
i
1209
1180
1340
3.082
3.072
3.127
3.094
0.0009
1
1212
1186
1204
3.084
3.074
3.081
3.080
0.00003
2
826
628
816
2.917
2.798
2.912
2.876
0.0045
3
493
416
413
2.693
2.619
2.616
2.643
0.0019
4
127
147
147
2.104
2.167
2.167
2.146
0.0013
5
49.3
40.0
44.0
1.693
1.602
1.643
1.646
0.0021
6
13.2  EXAMPLE OF ANALYSIS OF ALGAL GROWTH DATA

13.2.1  Formal  statistical  analysis of the growth data is outlined on the
flowchart in Figure 3.   The response used in the statistical  analysis is the
number of cells per milliliter per replicate.  Separate analyses are performed
for the estimation of the NOEC and LOEC endpoints and for the estimation of
the IC25 and IC50 endpoints.

13.2.2  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.

13.2.3  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 (see Appendix D).  The Wilcoxon Rank Sum Test with the Bonferroni
adjustment is the nonparametric alternative (see Appendix F).
                                     215

-------
13.2.4  Data from an algal growth test with cadmium chloride will be used
to illustrate the statistical analysis.  The cell counts were Iog10
transformed in an effort to stabilize the variance for the ANOVA
analysis.  The raw data, Iog10 transformed data,  mean  and  standard
deviation of the observations at each concentration including the control
are listed in Table 4.  A plot of the Iog10  transformed cell  counts  for
each treatment is provided in Figure 4.

13.2.5  Test for Normality

13.2.5.1  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 5.
        TABLE 5.  CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S EXAMPLE
                                 Toxicant Concentration (ua Cd/L)
     Replicate    Control
     10
20
40
80
A
B
C
-0.012
-0.022
0.033
0.004
-0.006
0.001
0.041
-0.078
0.036
0.050
-0.024
-0.027
-0.042
0.021
0.021
0.047
-0.044
-0.003
13.2.5.2  Calculate the denominator,  D,  of the test statistic:

                     D - Z (X,  -  X)2
    Where:   Xf  =  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 - _1 (0.000) = 0.000
    18
D - 0.0214
                                     216

-------
ro
                                                  CONNECTS THE MEAN  VALUE  FOR EACH CONCENTRATION
                                                  REPRESENTS THE  CRITICAL  VALUE FOR OUNNETT'S TEST
                                                  (ANY MEAN GROWTH BELOW THIS VALUE WOULD BE
                                                   SIGNIFICANTLY  DIFFERENT  FROM THE CONTROL)
                                        TOXICANT CONCENTRATION (UG CD/L)
  Figure  4.   Plot of  the Iog10  transformed cell  count data from the  green alga, Selenastrum capricornutum,
              growth response test  (Table 4).

-------
13.2.5.3  Order the centered  observations  from smallest to largest:

                  v(1)  _  v(2) _       x
-0.078
-0.044
-0.042
-0.027
-0.024
-0.022
-0.012
-0.006
-0.003
i
10
11
12
13
14
15
16
17
18
X(t>
0.001
0.004
0.021
0.021
0.033
0.036
0.041
0.047
0.050
13.2.5.4  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 =  18,  k = 9.  The a,.
values are listed in Table 7.

13.2.5.5  Compute the test statistic, W, as  follows:
               W - _L_ [ Z a, (X(n-'"*1> - X(i>)  ]2
                    D   i=l

The differences x(n'i+1) - X(i)  are  listed  in  Table 7.

For this set of data:
                        W =    1       (0.1436)2  = 0.964
                              0.0214
                                      218

-------
  TABLE  7.   COEFFICIENTS AND DIFFERENCES FOR SHAPIRO-WILK'S EXAMPLE
                                 -  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
0.128
0.091
0.083
0.063
0.057
0.043
0.033
0.010
0.004
X(18)
x<17)
X(16>
X(15>
X(14)
x<13)
X(12)
xcii>
x<10)
X(1)
- x(2)
- x(3)
X(4)
x<5)
X(6)
X(7)
X(8)
X(9)
13.2.5.6   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 = 0.964 is
greater than  the critical  value,  the conclusion of the test is that the
data are  normally  distributed.

13.2.6 Test  for Homogeneity of Variance

13.2.6.1   The  test used to examine whether the variation in mean cell
count is  the  same  across all toxicant concentrations including the
control,  is Bartlett's Test (Snedecor and Cochran, 1980).  The test
statistic is  as follows:
B =
[

P
< z

v,

)

In

S"2 -
C
P
f-lVf

In Sj2

]

    Where:   V^  = degrees  of freedom for each  toxicant  concen-
                 tration  and control, V,- =  (n(  -  1)

            p  =  number of levels of toxicant concentration
                 including the control

            n{  = the  number of replicates for concentration i

            In  =  loge

            i  =  1,  2,  ..., p, where p is the number of concentrations
                 including the control
                                      219

-------
                  ( s v,  s?)
                   i=l
          C  = 1 + ( 3(p-l))'1  [ X  1/V, -  ( S V
                                1=1        1=1
13.2.6.2  For the data in this example, (see Table 4) all toxicant
concentrations including the control have the same number of replicates
(n,.  = 3 for all  i).   Thus,  V,  - 2  for  all  i.

13.2.6.3  Bartlett's statistic is therefore:

                                P     ,
       B =  [(12)ln(0.0018) - 2 2 ln(S?)]/1.194
                               i=l

         =  [12(-6.3200)   2(-41.9082)]/1.194

         =  7.9764/1.194

         =  6.6804

13.2.6.4  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 five
degrees of freedom,  is 15.09.  Since B = 6.6804 is less than the critical
value of 15.09,  conclude that the variances are not different.

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 8.
                                      220

-------
                          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)
S2 = SSB/(p-l)
S2 = SSW/(N-p)

Where:
      number of toxicant concentrations  including the control
            N  = total number of observations  n, + n2 ...  H

            n,-  = number of observations in concentration  i

                 P  ,
SSB
                         -  G/N
                                  Between Sum of Squares
           SST = I   E'Y*J  -  G2/N
                              Total  Sum of Squares
           SSW = SST - SSB
                             Within  Sum  of Squares
                                                                  H
            G  = the grand total of all sample  observations,  G  = S Ts

            I,-  = the total  of the replicate measurements for
                 concentration  "i"

           Y,-j  =  the jth  observation  for concentration "i" (represents
                 the cell count for toxicant  concentration  i  in test
                 chamber j)

 13.2.7.2  For the data in this example:
                       n5 = n6  =  3
= Y
T3 = Y
T4 = Y
T5 = Y
T  - Y
n
21
31
41
51
61
         12
         22
         32
         42
         52
         62
         Y13 =
         Y23 =
         Y33 =
         Y43 =
         Y53 =
         Y63 =
                         9.281
                         9.239
                         8.627
                         7.928
                         6.438
                         4.938
                                      221

-------
    G  = T,  +  T2 + T3 + T4 + T5 +  T6 = 46.451


    SSB = S T2/n,- - G2/N


        - JL(374.606) - (46.451)2  =  4.997
           3                 18


    SST = S   s'Y2;- G2/N


        = 124.890  -  (46.451)2  = 5.018
                         18

    SSW = SST  - SSB = 5.018 - 4.997 = 0.0210

    S2   = SSB/(p-l) =  4.996/(6-l) = 0.9990

    Sj   = SSW/(N-p) =  0.021/(18-6) = 0.0018

13.2.7.3  Summarize these calculations in the ANOVA table (Table 9)



            TABLE 9.   ANOVA TABLE FOR DUNNETT'S PROCEDURE EXAMPLE
Source
Between
Within
Total
df
5
12
17
Sum of Squares
(SS)
4.997
0.021
5.017
Mean Square (MS)
(SS/df)
0.999
0.0018

13.2.7.4  To perform the individual comparisons, calculate the t
statistic for each concentration, and control combination as  follows:
                                         - Yf
                                Sw V
                                      222

-------
Where:
£
        n
mean cell count for toxicant concentration  i
mean cell count for the control
square root of within mean square
number of replicates for the control
number of replicates for concentration  i.
13.2.7.5  Table 10 includes the calculated t values for each
concentration and control combination.  In this example, comparing the
5 jig/L concentration with the control the calculation is as follows:
                            ( 3.094 - 3.080 )
                                                    = 0.405
                        [ 0.0424  V (1/3) + (1/3)  ]
                      TABLE 10.  CALCULATED T VALUES
           Toxicant Concentration
                  (lig Cd/L)
5
10
20
40
80
2
3
4
5
6
0.405
6.300
13.035
27.399
41.850
13.2.7.6  Since the purpose of this test is to detect a significant reduction
in mean cell  count, 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.  The mean count for
concentration "i" is considered significantly less than the mean count for the
control if t,  is greater than the critical  value.   Since t3,
                                               V  *5
                                                            and  t6 are
greater than 2.50, the 10, 20, 40 and 80 [ig/L concentrations have'
significantly lower mean cell counts than the control.  Hence the NOEC and the
LOEC for the test are 5 jig/L and 10 ng/L, respectively.

13.2.7.7  To quantify the sensitivity of the test, the minimum significant
difference (MSD) that can be statistically detected may be calculated.
Where:   d

        n
           MSD = d Sw  / (1/n,) + (1/n)

     the critical value  for  the  Dunnett's  procedure
     the square root of the within mean square
     the common number of  replicates  at each  concentration
                                      223

-------
             (this assumes equal replication at each concentration)
             the number of replicates in the control.
13.2.7.8  In this example:
                   MSD = 2.50 (0.0424)  V (1/3) + (1/3)
                       = 2.50 (0.0424)(0.8165)
                       = 0.086

13.2.7.9  The MSD (0.086) is in transformed units.  An approximate MSD in
terms of cell count per 100 ml may be calculated via the following
conversion.

    1. Subtract the MSD from the transformed control mean.

          3.094 - 0.086 = 3.008

    2. Obtain the untransformed values for the control mean and the
       difference calculated in 1.
          10<3.oo8)
    3. The untransformed MSD (MSDU)  is  determined by subtracting the

       untransformed values from 2.

        MSUU = 1241.6 -  1018.6  = 223

13.2.7.10  Therefore, for this  set of data, the minimum difference in
mean cell count between the control  and any toxicant concentration that
can be detected as statistically significant is 223.

13.2.7.11  This represents a decrease in growth of 18% from the control.

13.2.8  Calculation of the 1C

13.2.8.1  The growth data in Table 4 are utilized in this example.  Table 11
contains the means 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.  See figure 5 for a plot of the response curve.
                                      224

-------
PO
ro
en
        1400 -
        1300 -
*  *
                                                                             INDIVIDUAL REPUCATH rm I COUNT

                                                                             CONNECTS THE OBSERVED MEAN VALUE
                                        TOXICANT CONCENTRATION (ug Cd/L)
    Figure  5.   Plot of raw data  and  observed  means for the  green alga,  Selenastrum  capricornutum,  growth

                data.

-------
                  TABLE  11 ALGAL  GROWTH  MEANS
Toxicant
Cone.
(M9 Cd/L)
Control
5
10
20
40
80
i
1
2
3
4
5
6
M-
(cel Is/mi)
1243
1201
757
441
140
44
13.2.8.2  An IC25 and IC50 can be estimated  using  the  Linear Interpolation
Method.  A 25% reduction in cell count, compared to  the  controls,  would result
in a mean count of 932 cells, where FUl-p/lOQ) =  1243(1-25/100).   A  50%
reduction in cell count, compared to the controls, would result in a  mean
count of 622 cells.  Examining the means and their associated concentrations
(Table 11), the response, 932 cells, is bracketed  by C2  = 5 jug  Cd/L and  C3 =
10 /jg Cd/L.  The response, 622 cells,  is bracketed by  u = 10 ug Cd/L and C, =
20 M9 Cd/L.                                                                4

13.2.8.3  Using Equation (1) from paragraph 4.2 of Appendix  J,  the estimate of
the IC25 is calculated as follows:
ICp = Cj  +  [M,(l    p/100)    Mj]  (C
                                             J+1
                                           (MJ+1    Mj)

          IC25 = 5 + [1243(1 - 25/100)   1201]   (10  -  5)
                                               (757  -  1201)

               = 8 ug Cd/L.

13.2.8.4  The IC50 estimate is 14 jug Cd/L:

           ICp = Cj + [M,(l  -  p/100)  -  MJ  (CJ+1 -  Cj)

                                           (MJ+1  -  Mj)

          IC50 = 10 + [1243(1 - SO/100) -  757]  (20  -  10)

                                               (441  -  757)

               = 14 M9 Cd/L.
                                      226

-------
13.2.8.5  When the Bootstrap program ^BuOTSTRP) was used to analyze this set
of data, requesting 80 resamples, the mean estimate of the IC25 was 8.0698 ^9
Cd/L,  with a standard deviation of 0.5076 M9 Cd/L (coefficient of variation =
6.3%).   The empirical 95% confidence interval for the true mean was (7.1079 ^.
Cd/L,  8.9753 M9 Cd/L).  The BOOTSTRP computer program output for the IC25 for
this data set is shown in Figure 6.

13.2.8.6  When the Bootstrap program (BOOTSTRP) was used to analyze this set
of data, requesting 80 resamples, the mean estimate of the IC50 was 14.1877 /x
Cd/L,  with a standard deviation of 1.2037 ug Cd/L (coefficient of variation =
8.5%).   The empirical 95% confidence interval for the true mean was (11.5473
/ig Cd/L, 15.8451 M9 Cd/L).  The BOOTSTRP computer program output for the IC50
for this data set is shown in Figure 7.

13.3  BIOSTIMULATION

13.3.1  Where the growth response in effluent (or surface water) exceeds
growth in the control flasks, the percent stimulation, S(%), is
calculated as shown below.  Values which are significantly greater than
the control indicate a possible degrading enrichment effect on the
receiving water (Walsh et al., 1980):


                             S(%) = T - C
                                            X 100
                                      227

-------
THE NUMBER OF RESAMPLES IS   80


*** LISTING OF GROUP CONCENTRATIONS (% EFF.) AND RESPONSE MEANS ***

CONC. (%EFF)             RESPONSE MEAN            MEAN AFTER POOLING
      .000                 1243.000                   1243.000

     5.000                 1200.667                   1200.667

    10.000                  756.667                    756.667

    20.000                  440.667                    440.667

    40.000                  140.333                    140.333

    80.000                   44.433                     44.433
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
   FROM THE INPUT SAMPLE IS   8.0227.
    ************************************************************
    *        BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY       *
    *                   OF THE ESTIMATED ICp                   *
    ************************************************************

THE MEAN OF THE BOOTSTRAP ESTIMATES IS   8.0698.

THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS    .5076

AN EMPIRICAL 95.0% CONFIDENCE INTERVAL FOR THE
     BOOTSTRAP ESTIMATE IS (   7.1079,   8.9753).
     Figure 6.   BOOTSTRP program output for the IC25.


                                     228

-------
THE NUMBER OF RESAMPLES IS   80


*** LISTING OF GROUP CONCENTRATIONS (% EFF.) AND RESPONSE MEANS ***

CONC.  (%EFF)             RESPONSE MEAN            MEAN AFTER POOLING
      .000                 1243.000                   1243.000

     5.000                 1200.667                   1200.667

    10.000                  756.667                    756.667

    20.000                  440.667                    440.667

    40.000                  140.333                    140.333

    80.000                   44.433                     44.433
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
   FROM THE INPUT SAMPLE IS  14.2774.
    *        BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY       *
    *                   OF THE ESTIMATED ICp                   *
    ************************************************************

THE MEAN OF THE BOOTSTRAP ESTIMATES IS  14.1877.

THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS   1.2037

AN EMPIRICAL 95.0% CONFIDENCE INTERVAL FOR THE
     BOOTSTRAP ESTIMATE IS ( 11.5473,  15.8451).
     Figure 7.   BOOTSTRP program output for the IC50.
                                     229

-------
14.  PRECISION AND ACCURACY

14.1  PRECISION

14.1.1  Data from repetitive 96-h toxicity tests conducted with cadmium
chloride as the reference toxicant,  using medium containing EDTA, are shown in
Table 12.  The relative standard deviation (coefficient of variation) of the
ECBOs was 44.1%.


TABLE 12.  SINGLE-LABORATORY PRECISION OF THE GREEN ALGA, SELENASTRUM
           CAPRICORNUTUM, 96-H TOXICITY TESTS, USING THE REFERENCE TOXICANT
           CADMIUM CHLORIDE8
Test Number
1
2
3
4
5
6
7
8
9
10
n
Mean
CV (%)
EC50 (mg/L)
2.3
2.4
2.3
2.8
2.6
2.1
2.1
2.1
2.6
2.4
10
2.37
10.2
CV (%)
4.8
9.6
5.5
13.3
4.4
8.2
14.4
7.1
11.9
5.0
10
8.42
44.1
aFrom USEPA (1991c).


14.2  ACCURACY

14.2.1  The accuracy of toxicity tests cannot be determined.

                                      230

-------
                             SELECTED REFERENCES


Adams,  N.; K. H. Golding; and A. J. Dodds.  1985.  Toxicity of eight
      water-soluble organic chemicals to Selenastrum capn'cornutum: A study
      of methods for calculating toxic values using different growth
      parameters.  Arch. Environ. Contam. Toxicol. 14:333.

Adams,  N.; K. H. Golding; and A. J. Dodds.  1986.  Effect of acetone on the
      toxicity of four chemicals to Selenastrum capn'cornutum.  Bull. Environ
      Contam. Toxicol. 36:254-259.

Akasenova, Y. I.; G. I. Bogucharskova; and M. I. Zozulina.  1969.  The
      role of phytoplankton and bacterioplankton in the food of the dominant
      Cladocera of the Lower Don.  Hydrobiologia 5(5):19-26.

APHA.  1985.  Standard Methods for the Examination of Water and Wastewater.
      16th Ed. American Public Health Association, Washington, DC.

APHA.  1989.  Standard Methods for the Examination of Water and Wastewater.
      17th Ed.  American Public Health Association, Washington, DC.

ASTM.  1988a.  Standard guide for conducting early life-stage toxicity tests
      with fishes.  ASTM E-1241-88.  American Society for Testing and
      Materials, Philadelphia, PA.

ASTM.  1988b.  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, PA.

Atkins, W. R. G.  1923.  The phosphate content of fresh and salt waters in
      its relationship to the growth of algal plankton.  J. Mar. Biol. Assoc.
      13:119-150.

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 of
      concentrations in short-term fathead minnow toxicity tests.  EPA
      Contract No. 68-03-3534. Criteria and Standards Division, U.S.
      Environmental Protection Agency, Washington, DC.

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.

Benoit, D. A.; F. A. Puglisi; and D. L. Olson.  1982.  A fathead minnow,
      Pimephales promelas, early life stage toxicity test method evaluation
      and exposure to four organic chemicals.  Environ. Pollut. (Series A)
      28:189-197.
                                      231

-------
Berner, D. D.  1986.  The taxonomy of Ceriodaphnia (CrustacearCladocera)  in
      U. S. Environmental Protection Agency Cultures.  Environmental
      Monitoring and Support Laboratory, U. S. Environmental Protection
      Agency, Cincinnati, OH 45268. EPA/600/4-86/032.

Birge, W. J.; J. A. Black; and B. A. Ramey.  1981.  The reproductive
      toxicology of aquatic contaminants.  In: Hazard Assessments of
      Chemicals, Current Developments, Vol. 1.  Academic Press, Inc.  pp.
      59-114.

Birge, W. J.; J. A. Black; T. M. Short; and A. G. Westerman.  1989.
      A comparative ecological and toxicological  investigation of an STP
      effluent  and its receiving stream.  Environ. Toxicol. Chem. 8(5):437-
      450.

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.  pp. 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. Toxicol. 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.

Blaise, C. R.  1986.  Micromethod for acute aquatic toxicity assessment using
      the green alga Selenastrum capricornutum.  Toxic. Assess. 1:377-385.

Brungs, W. A.  1971.  Chronic effects of low dissolved oxygen concentrations
      on fathead minnows (Pimephales promelas),  J. Fish. Res. Board Can.
      28:1119-1123.

Burton, G. A., Jr.; D. Nimmo; D. Murphey; and F.  Payne.  1987.  Stream profile
      determinations using microbial activity assays and Ceriodaphnia.
      Environ. Trwirnl  T.hpm  fi'RnR-513
ue tci in i M
-------
Collins,  M.  T.; J. B. Gratzer; D. L. Dawe; and T. G. Nemetz.  1976.  Effects
      of antibacterial agents on nitrification in aquatic recirculating
      systems.  J. Fish. Res. Board Can. 33.215-218.

Conit, R. J.  1972.  Phosphorus and algal growth in the Spokane River.
      Northwest Sci.  46(3):177-189.

Conover,  W.  J.  1980.  Practical Nonparametric Statistics.  Second edition.
      John Wiley and Sons, New York, NY. pp. 466-467.

Cooney,  J. D.; G.  M.  DeGraeve; E. L. Moore; T. L. Pollock; and W. D. Palmer.
      1986.   Effects of food and water quality on the culturing and toxicity
      testing of Ceriodaphnia.  Draft Report, No. 526-P-5372, Battelle
      Laboratories, Columbus, OH.

Cowgill,  U.  M.  1987.  Critical analysis of factors affecting the sensitivity
      of zooplankton and the reproducibility of toxicity test results.  Water
      Res. 21(12):1453-1462.

Cowgill,  U.  M.; K. I. Keating; and I. T. Tokahashi.  1985.  Fecundity and
      longevity of Ceriodaphnia dubia/affim's in relation to diet at two
      different temperatures.  J. Crust. Biol. 5(3):420-429.

Cowgill,  U.  M.; D. P. Milazzo; and C. K. Meagher.  1988.  New diet for
      Ceriodaphnia dubia.  Bull. Environ. Contain. Toxicol. 41:304-309.

Cowgill,  U.  M.; I. T. Takahashi; and S. L. Applegath.  1985.  A comparison of
      the effect of four benchmark chemicals on Daphm'a magna and Ceriodaphnia
      dubia-affinis tested at two different temperatures.  Environ.
      Toxicol. Chem.   4:415-422.

DeGraeve, G. M. and J. D. Cooney.  1987.  Ceriodaphnia: An update on effluent
      toxicity testing and research needs.  Environ. Toxicol. Chem.
      6:331-333.

DeGraeve, G. M.; J. D. Cooney; T. L. Pollock; N. G. Reichenbach; J. H. Dean;
      M.  D.  Marcus; and D. 0. Mclntyre.  1988.  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.  Columbus, Battelle, Columbus, OH.  Available from:
      American Petroleum Institute, 1220 L Street, NW, Washington, DC  20005,
      Rept.  No. 4486.

DeGraeve, G. M.; J. D. Cooney; B. H. Marsh; T. L. Pollock; and N. G.
      Reichenbach.  1989.  Precision of the EPA seven-day Ceriodaphnia dubia
      test:  Intra- and Interlaboratory study to determine the reproducibility
      of the seven-day Ceriodaphnia dubia survival and reproduction tests.
      EPRI EN-6469. Electric Power Institute, Palo Alto, CA 94369.
                                      233

-------
DeGraeve, G. M.; J. D.  Cooney; D. 0. Mclntyre; T. L. Pollock; N. G.
      Reichenbach; J. H. Dean; and M. D. Marcus.  1991.  Variability in the
      performance of the 7-day fathead minnow (Pimephales promelas) larval
      survival  and growth test:  An intra- and interlaboratory study.
      Environ.  Toxicol.  Chem. 10:1189-1201.

DeWoskin, R. S.  1984.   Good laboratory practice regulations: A comparison.
      Research  Triangle  Institute, Research Triangle Park, NC. 63 pp.

Dixon,  W. J. and F- J.  Massey, Jr.  1983.   Introduction to Statistical
      Analysis.  Fourth  edition.   McGraw Hill, New York, NY.

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.

Eagleson, K. W.; D. L.  Lenat; L.  Ausley; and F.  Winborne.  1990.  Comparison
      of measured in-stream biological responses with responses predicted  by
      Ceriodaphnia chronic toxicity test.   Environ.  Toxicol. Chem. 9:1019-
      1028.

Efron,  B.  1982.  The Jackknife,  the Bootstrap,  and  other resampling
      plans.  CBMS 38,  Soc. Industr. Appl. Math., Philadelphia, PA.

Eloranta, V. and 0. Laitinen.  1982.  The usefulness of the Selenastrum
      capricornutum algal assay to evaluate the toxic effects of pulp and
      paper mill effluents in lake water.   Vatten 38:317-331.

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):2380-2383.

FDA.  1978.  Good laboratory practices for nonclinical  laboratory
      studies.   Part 58, Fed. Reg. 43(247):60013-60020, December 22, 1978.

Ferris,  J. J.;  S. Kobyashi; and N. L. Clesceri.   1974.   Growth of
      Selenastrum capricornutum in natural waters augmented with detergent
      products  in wastewaters.  Water Res. 8(12):1013-1020.

Filip,  D. S. and E. J.  Middelbrooks.  1975.  Evaluation of sample
      preparation techniques for algal bioassays.  Water Res. 9:581-585.

Finney,  D. J.  1948.  The Fisher-Yates test of significance in 2X2
      contingency tables.  Biometrika 35:145-156.
                                      234

-------
Finney,  D.  J.   1971.   Probit Analysis. Third Edition.  Cambridge Press,
      New York,  NY.   668 pp.

Flickinger,  S.  A.   1966.  Determination of sexes in fathead minnow.  Trans.
      Am.Fish.  Soc.  98(3):526-527.

Forsberg,  C.  G.   1972.   Algal assay procedure.   J.  Wat. Pollut. Control Fed.
      44(8):1623-1628.

Forsberg,  C.  G.  and  A.  Forsberg.  1972.  Algal  growth potential test
      improves  sewage settlement control.  Ambio l(l):26-29.

Gast,  M. H.  and W.  A. Brungs.  1973.  A procedure for separating eggs of
      the fathead  minnow.   Prog. Fish-Cult. 35:54.

Goldman, C.  R.;  M.  G. Tunzi; and R. Armstrong.   1969.  Carbon-14 uptake as
      a sensitive  measure of the growth of algal cultures.   In: Middlebrooks,
      E. J.;  T.  E.  Maloney; C. F. Powers; and L. M. Kaack (eds.).   Proe. of
      the  Eutroph.  Bioassess. Workshop, 19-21  June, 1969.   U.S. Pacific
      Northwest Water Laboratory, Corvallis, OR.  pp. 158-170.

Gophen,  M.   1976.   Temperature dependence on food intake, ammonia
      excretion and  respiration in Cen'odaphm'a reticulata  (Jurine) (Lake
      Kinneret,  Israel).  Freshwat. Biol. 6(5):451-455.

Gray,  R. H.;  R.  W.  Hanf; D. D. Dauble; and J. R. Skalaski.   1982.   Chronic
      effects  of a coal  liquid on a freshwater  alga, Selenastrum
      capn'cornutum.   Environ. Sci. Technol. 16(4):225-229.

Greene,  J.  C.;  W.  E.  Miller; T. Shiroyama; and  E. Merwin.  1975.  Toxicity
      of zinc  to the green alga Selenastrum capn'cornutum as a  function of
      phosphorus or  ionic strength.  In: Middlebrooks,  E. J.; T. E. Maloney;
      C. F.  Powers;  and L. M. Kaack (eds.).  Proc.  of the Eutroph.
      Bioassessment  Workshop, 19-21 June, 1969.   U. S.  Pacific  Northwest Water
      Laboratory,  Corvallis, OR.  pp.  28-43.  EPA-660/3-75-034.

Hamilton,  M.  A.   1986.   Statistical analysis of cladoceran  reproductivity
      tests.   Environ.  Toxicol. Chem.  5:205-212.

Hedtke,  S.  F.;  C.  W.  West; K. N. Allen; T. J. Norberg-King;  and D.  I. Mount.
      1986.   Toxicity of pentachlorophenol to aquatic organisms under
      naturally varying and controlled environmental conditions.  Environ.
      Toxicol.  Chem.  5:531-542.

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.

Hoffman, G.  L.  and A. J. Mitchell.  1980.  Some chemicals that  have been used
      for fish  diseases and pests. Fish Farming  Exp. Station,  Stuttgart,
      AR 72160.   8 pp.


                                      235

-------
Jensen, A. L.  1972.  Standard error of LC50 and sample size in fish bioassay.
      Water Res.  6:85-89.

Keller, A. E.; R. J. Dutton; G. Bitton; and T. L. Crisman.  1988.  Chronic
      toxicity of hydrothol-191 to Cen'odaphm'a dubia at 25 and 15°C.   Bull.
      Environ. Contam. Toxicol. 41:233-240.

Knight, J. T. and W. T. Waller.  1987.  Incorporating Daphm'a magna into the
      seven-day Cen'odaphm'a effluent toxicity test method.  Environ.  Toxicol.
      Chem. 6:635-645.

Leger, P.; D. A.  Bengston; K. L.  Simpson;  and P. Sorgeloos.  1986.  The use
      and nutritional value of Artemia as  a food source.  Oceanorg. Mar. Biol.
      Ann. Rev. 24:521-623.

Leger, P.; P. Sorgeloos; 0. M. Millamena;  and K. L. Simpson.  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. 93:71-82.

Lewis, M. A.   1986  Comparison of the effects of surfactants on freshwater
      phytoplankton  communities in experimental enclosures and on algal
      population  growth in the laboratory.  Environ. Toxicol.  Chem.
      5:319-332.

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, PA.   pp.
      137-146.

Marking, L. L. and V. K. Dawson.   1973.  Toxicity of quinaldine sulfate to
      fish.  Invest. Fish Contr.  No. 48,  U. S. Fish and Wildlife Service,
      Washington, DC.  8 pp.

Marcus, A. H. and A. P. Holtzman.  1988.   A robust 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.

Mayer, F. L., Jr.; K. S. Mayer; and M..R.  Ellersieck.  1986.  Relation of
      survival to other endpoints in chronic toxicity tests with fish.
      Environ. Toxicol. Chem. 5:737-748.
Mayes,  M.  A.; H. C.  Alexander; D. C. Hopkins; and P. B. Latavitis.
      Acute and chronic toxicity of ammonia to freshwater fish: A
      site-specific  study.  Environ. Toxicol. Chem. 5:437-442.
1986.
                                      236

-------
McKim,  J.  M.  1977.  Evaluation of tests with the early life stages of
      fish for predicting long-term toxicity.  J. Fish. Res. Board Can.
      34:1148-1154.

McNaught,  D. C.  and D. I. Mount.  1985.  Appropriate durations and measures
      for Ceriodaphnia toxicity tests.  In: R. C. Banner and D. J. Hansen
      (eds.).  Aquatic Toxicology and Hazard Assessment: Eighth Symposium.
      ASTM STP 891.  American Society for Testing and Materials, Philadelphia,
      PA.   pp. 375-381.

McPhee, C.  1961.  Bioassay of algal production in chemically altered
      waters.  Limnol. Oceanogr. 6:416-422.

Miller, R. G.  1981.  Simultaneous statistical inference.  Springer-Verlag,
      New York,  NY.  299 pp.

Mount,  D.  I.; T.  J. Norberg-King; R. Keen; and J. T. Taraldsen.  1987.  A
      reference test water for cladocerans.  Abstract, llth Annual Symposium,
      Aquatic Toxicology and Hazard Assessment.  American Society for Testing
      and Materials, May 10-12, 1987. Cincinnati, OH.

Mount,  D.  I.  1968.  Chronic toxicity of copper to fathead minnows (Pimephales
      promelas,  Rafinesque).  Water Res. 2:215-223.

Mount,  D.  I. and T. J. Norberg.  1984.  A seven-day life-cycle cladoceran
      test.   Environ. Toxicol. Chem.  3:425-434.

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.

Neiheisel, T. M.; W. B. Horning, II; B. M. Austern; D. F. Bishop; T.  L. Reed;
      and J. F.  Estenik.  1988.  Toxicity reduction at municipal wastewater
      treatment plants.  J. Wat. Pollut. Control Fed. 60(l):57-67.

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, VA.

Norberg, T.  J. and D. I. Mount.  1985a.  A new fathead minnow (Pimephales
      promelas)  subchronic toxicity test.  Environ. Toxicol. Chem.
      4(5):711-718.

Norberg, T.  J. and D. I. Mount.  1985b.  Diets for Ceriodaphnia reticulata
      life-cycle tests.  In: Cardwell, R. D.; R. Purdy; and R. C. Bahner
      (eds.).  Aquatic Toxicology and Hazard Assessment, Seventh Symposium,
      ASTM STP 854, American Society for Testing and Materials, PA.  pp.
      42-52.

Norberg-King, T.  J.  1987.  Relative sensitivity of the fathead minnow
      (Pimephales promelas) seven day subchronic test to the early life stage
      and life cycle tests. M. S. Thesis, Montana State Univ, Bozeman, MT.

                                      237

-------
Norberg-King, T. J.  1989.  An evaluation of the fathead minnow seven-day
      subchronic test for estimating chronic toxicity.  Environ. Toxicol.
      Chem. 8:1075-1089.

Norberg-King, T. J.  1991.  Calculation of ICp values of IC15, IC20, IC25,
      IC30, and IC50 for Appendix A of the revised Technical Support Document.
      Memorandum to M.  Heber,  USEPA.

Owsley,  J. A. and D. E. McCauley.  1986.   Effect of extended sublethal
      exposure to sodium selenite on Cen'odaphm'a affim's.   Bull.  Environ.
      Contam. Toxicol.  36:876-880.

Parker,  M.  1977.  The  use of algal bioassays to predict the short and
      long term changes in algal  standing crop which result from altered
      phosphorus and nitrogen loadings.  Water Res. 11:719-725.

Pearson, E. S. and T. 0. Hartley.  1962.   Biometrika tables for statisticians.
      Vol. 1.  Cambridge Univ. Press, England,  pp. 65-70.

Pennak,  R. W.  1978.  Freshwater Invertebrates of the United States.  Second
      Edition.  Ronald  Press,  New York, NY.   769 pp.

Pennak,  R. W. 1989.  Fresh-water Invertebrates of the United States.  Protozoa
      to Mollusca.  Third Edition.  John  Wiley & Sons, Co., New York, NY.

Pickering, Q. H.  1988.  Evaluation and comparison of two short-term fathead
      minnow tests for  estimating chronic toxicity.  Water Res. 22(7):883-893.

Rehnberg, B. G.; D. A.  Schultz; and R. L. Raschke.  1982.  Limitations  of
      electronic counting in reference to algal assays.   J. Wat. Pollut.
      Control Fed. 54(2):181-186.

Sachdev, D. R. and N. L. Clesceri.  1978.  Effects of organic fractions
      from secondary effluent on Selenastrum capn'cornutum (Kutz.).   J. Wat.
      Pollut. Control Fed. 52:1810-1820.

Scheffe, H.  1959.  The Analysis of Variance.  John Wiley and Sons,
      New York, NY.  477 pp.

Shapiro, S. S. and M. B. Wilk.  1965.  An analysis of variance test for
      normality (complete samples).  Biometrika 52:591-611.

Shuba, T. and R. R. Costa.  1982.  Development and growth of Cen'odaphm'a
      reticulata embryos.  Trans. Amer. Microsc. Soc. 9(3):429-435.

Skulberg, 0.  1964.  Algal problems related to the eutrophication of
      European water supplies, and a bioassay method to assess fertilizing
      influences of pollution on inland waters.  In: D.  F.  Jackson (ed.).
      Algae and man.  Plenum Press, New York, NY.  pp. 262-299.

Snedecor, G. W. and W.  G. Cochran.  1980.  Statistical Methods.  Seventh
      Edition.  Iowa State University Press, Ames, Iowa.  593 pp.

                                      238

-------
Soderberg, R. W.  1982.  Aeration of water supplies for fish culture in
      flowing water.  Prog. Fish-Cult. 44(2):89-93.

Steel,  R.  G. D.  1959.  A multiple comparison rank sum test: treatments
      versus control.  Biometrics 15:560-572.

Steel,  R.  G. D. and J. H. Torrie.  1960.  Principles and procedures of
      statistics with special reference to biological sciences.  McGraw-Hill
      Publ., New York, NY.

Takahashi, I. T.; U. M. Cowgill; and P. G. Murphy.  1987.  Comparison of
      ethanol toxicity to Daphm'a magna and Cen'odaphm'a dubia tested at two
      different temperatures: Static acute toxicity test results.  Bull.
      Environ. Contam. Toxicol. 39:229-236.

Tallqvist, T.  1973.  Algal assay procedure (Bottle Test) at the  Norwegian
      Institute for Water Research.  In:  Algal Assays in Water Pollution
      Research:  Proceedings from a Nordic Symposium, Oslow. October, 25-29,
      1972.  Nordforsk. Secretariat of Environm. Sci., Helsinki.  Publ.
      1973-2.  pp. 5-17.

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.

Taylor J.  K.  1987.  Quality Assurance of chemical measurements.  Lewis Publ.,
      Inc. Chelsea, MI.

Thomas, R. E. and R. L. Smith.  1975.  Assessing treatment process
      efficiency with the algal assay test.  In:  Middlebrooks, E. J.;  T. E.
      Maloney; C. F. Powers; and L. M. Kaack (eds.).  Proc.  of the Eutroph.
      Bioassessment Workshop, 19-21 June, 1969.  U. S. Pacific Northwest Water
      Laboratory, Corvallis, OR.  pp. 244-248.  EPA-660/3^75-034.

Thurston,  R. V.; R. C. Russo; and K. Emerson.  Aqueous ammonia equilibrium
      calculations.  Tech. Rep. No. 741.  Fish Bioassay Laboratory,  Montana
      State Univ., Bozeman, MT.  18 pp.

Thurston,  R. V.; R. C. Russo; E. L. Meyn; and R. K. Zaidel.   1986.  Chronic
      toxicity of ammonia to fathead minnows.  Trans. Amer.  Fish. Soc.
      115:196-207.

USEPA.   1971.  Algal assay procedures: Bottle test.  Miller, W. E.;  J.  C.
      Greene; and T. Shiroyama (eds.).  National Eutrophication Program.
      Environmental Research Laboratory, U. S. Environmental Protection
      Agency, Corvallis, OR.  82 pp.

USEPA.   1973.  Biological field and laboratory methods for measuring the
      quality of surface waters and effluents.  Weber, C. I. (ed.).
      Environmental Monitoring Series.  Office of Research and Development,
      U.S. Environmental Protection Agency, Cincinnati, OH.   EPA/670/4-73/001.

                                      239

-------
USEPA.  1975.  Methods for acute toxicity tests with fish, macroinvertebrates,
      and amphibians.  Environmental Research Laboratory, U. S. Environmental
      Protection Agency, Duluth, MN.  EPA-660/3-75-009.

USEPA.  1978a  Methods for measuring the acute toxicity of effluents to
      aquatic organisms. Second edition.  Peltier, W. (ed.).  Environmental
      Monitoring and Support Laboratory - Cincinnati.  U. S. Environmental
      Protection Agency, Cincinnati, OH. EPA-600/4-78-012.

USEPA.  1978b.  The Selenastrum capn'cornutum Printz algal assay bottle test.
      Miller, W. E.; J. C. Greene;, and T. Shiroyama (eds.).  Environmental
      Research Laboratory, U. S. Environmental Protection Agency, Corvallis,
      OR.  EPA-600/9-78-018.

USEPA.  1979a.  Handbook for analytical quality control in water and
      wastewater laboratories.   U. S. Environmental Protection Agency,
      Environmental Monitoring and Support Laboratory,  Cincinnati, OH.
      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, OH.  EPA-600/4-79-020.

USEPA.  1979c.  Interim NPDES compliance biomonitoring inspection manual.
      MCD-62.  Office of Water Enforcement,  U. S. Environmental Protection
      Agency, Washington, DC.

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.  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.
      OWEP-82-001.  Office of Water Enforcement and Permits, U. S.
      Environmental Protection Agency,  Washington, DC.

USEPA.  1982a.  User's guide for conducting life-cycle chronic toxicity tests
      with fathead minnows (Pimephales promelas).  Benoit, D. A. (ed.).
      Environmental Research Laboratory, U.  S. Environmental Protection
      Agency, Duluth, MN.  EPA-600/8-81-011.

                                     240

-------
USEPA.   1982b.   Methods for organic chemical analysis of municipal and
      industrial  wastewater.  Environmental Monitoring and Support
      Laboratory, U.S. Environmental Protection Agency, Cincinnati, OH.
      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, OH 45268.  EPA-600/8-83-020.

USEPA.   1984.   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. (eds.).
      Environmental  Research Laboratory, U. S. Environmental Protection
      Agency,  Duluth, MN.  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.

USEPA.   1985b.   Ambient water quality criteria for ammonia - 1984.  Office of
      Water Regulations and Standards Criteria and Standards Division,
      Washington, DC. 20480.  EPA/440/5-85-001.

USEPA.   1985c.   Short-term methods for estimating the chronic toxicity of
      effluents and receiving waters to freshwater organisms.  Horning, W. B.,
      II  and C. I.  Weber (eds.).  Environmental Monitoring and Support
      Laboratory, U.S. Environmental Protection Agency, Cincinnati, OH 45268.
      EPA/600/4-85/014.

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, OH.  EPA/600/4-85-013.

USEPA.   1985e.   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, MN.  EPA/600/8-85/015.

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, MN.  EPA/600/3-85/044.

USEPA.   1985g.   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, MN.
      EPA/600/3-85/071.
                                      241

-------
USEPA.   1986a.  Ambient water quality criteria for pentachlorophenol.
      Criteria and Standards Division, Office of Water Regulations and
      Standards, U. S. Environmental Protection Agency, Washington,  DC.
      EPA/440/5-86/009.

USEPA.   1986b.  Validity of effluent and ambient toxicity tests for  predicting
      biological impact, Kanawha River, Charleston, West Virginia.   Mount, D.
      I. and T. J. Norberg-King. (eds.j.  Environmental Research Laboratory,
      U. S. Environmental Protection Agency, Duluth, MN.  EPA/600/3-86/006.

USEPA.   1986c.  Validity of effluent and ambient toxicity tests for
      predicting  biological impact, Naugatuck River, Connecticut.   Mount, D.
      I.; T. J. Norberg-King; and A. E. Steen. (eds.j.  Environmental Research
      Laboratory, U. S. Environmental Protection Agency, Duluth, MN.
      EPA/600/8-86/005.

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. J. Norberg-King. (eds.).  Environmental Research
      Laboratory, U. S. Environmental Protection  Agency, Duluth, MN.   EPA/600/8-86/OC

USEPA.   1986e.  Validity of effluent and ambient toxicity tests for  predicting
      biological impact, Skeleton Creek, Enid, Oklahoma.  Norberg-King,  T. J.
      and D. I. Mount, (eds.).   Environmental Research Laboratory, U. S.
      Environmental Protection  Agency, Duluth, MN.  EPA/600/8-86/002.

USEPA.   1986f.  Occupational health and safety manual.  Office of
      Administration, U. S. Environmental Protection Agency, Washington, DC.

USEPA.   1987a.  Ambient water quality criteria for zinc.  Criteria and
      Standards Division, Office of Water Regulations and Standards, U.  S.
      Environmental Protection  Agency, Washington, DC.  EPA/440/5-87/003.

USEPA.   1987b.  Permit writer's guide to water quality-based permitting for
      toxic pollutants.  Office of Water, U. S. Environmental Protection
      Agency,  Washington, DC.  EPA/440/4-87/005.

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,  pp.
      153-164.

USEPA.   1988c.  Short-term methods for estimating the chronic toxicity of
      effluents and receiving waters to marine and  estuarine organisms.
      Weber, C. I.; W. B. Horning; D. J. Klemm; T. W. Neiheisel; P.  A. Lewis;
      E. L. Robinson; J. Menkedick; and F. Kessler (eds.).  Environmental
      Monitoring Systems Laboratory, U. S. Environmental Protection  Agency,
      Cincinnati, OH 45268.  EPA/600/4-87/028.
                                      242

-------
USEPA.   1988d.  Protocols for short-term toxicity screening of hazardous waste
      sites.  Greene, J. C.; C. L. Bartels; W. J. Warren-Hicks; B. R.
      Parkhurst; G. L. Under; S. A. Peterson; and W. E. Miller (eds.).
      Environmental Research Laboratory, U. S. Environmental Protection
      Agency, Con/all is, OR.  EPA/600/3-88/029.

USEPA.   1988e.  Multi-laboratory study of Cen'odaphm'a chronic toxicity
      test.  Neiheisel, T. W.; J. R. Menkedick; F. A. Kessler; W. B. Horning;
      and C. I. Weber.  In House Report. Environmental Monitoring Systems
      Laboratory, U.S. Environmental Protection Agency, Cincinnati, OH.

USEPA.   1988f.  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, MN.

USEPA.   1989a.  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. Norberg-king; W. B. Horning, II; F. A.
      Kessler; J. R. Menkedick; T. W. Neiheisel; 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, OH 45268.
      EPA/600/4-89/001.

USEPA.   1989b.  Toxicity reduction evaluation protocol for municipal
      wastewater treatment plants.  Botts, J. A.; J. W. Braswell; J. Zyman; W.
      L. Goodfellow; and S. B. Moore (eds.).  Reduction Engineering
      Laboratory, U.S. Environmental Protection Agency, Cincinnati, OH 45268.
      EPA/600/2-88/062.

USEPA.   1989c.  Generalized methodology for conducting industrial toxicity
      reduction evaluations (TREs).  Fava, J. A.; 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, OH 45268.  EPA/600/2-88/070

USEPA.   1990.   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 (eds.)  Environmental Monitoring Systems
      Laboratory, U. S. Environmental Protection Agency, Cincinnati, OH 45268.
      EPA/600/4-90/030.

USEPA.   1991a.  Manual for the evaluation of laboratories performing aquatic
      toxicity tests.  Klemm, D. J.; L. B. Lobring; and W. H. Horning, II.
      Environmental Monitoring and Support Laboratory, U.S. Environmental
      Protection Agency, Cincinnati, OH 45268.  EPA/600/4-90/031.
                                      243

-------
US£PA.  1991b.  Methods for measuring the acute toxicity of effluents to
      freshwater and marine organisms.  Weber, C.I. (ed.).  Environmental
      Monitoring Systems Laboratory, U. S. Environmental Protection Agency,
      Cincinnati, OH 45268.  EPA/600/4-90/027.

USEPA.  1991c  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.  1991d.  Toxicity identification evaluation: Characterization of
      chronically toxic effluents, Phase I.  Norberg-King, T. J.; D. I. Mount;
      J. R. Amato; D. A. Jensen; and J. A. Thompson (eds.).  Office of
      Research and Development, U. S. Environmental Protection Agency,
      Duluth, MN 55804.  EPA-600/6-91/005.

USEPA.  1992.  Short-term methods for estimating the chronic toxicity of
      effluents and receiving waters to marine and estuarine organisms
      (Second Edition).  Klemm, D. J. and 6. E. Morrison (eds.).
      Environmental Monitoring Systems Laboratory, U.S. Environmental
      Protection Agency, Cincinnati, OH 45268.  EPA/600/4-91/021.

Vanhaecke, P- and P. Sorgeloos.  1980.  International  study on Artemia. IV.
      The biometrics of Artemia strains from different geographical origin.
      In: Persoone, G.; P.  Sorgeloos; 0. Roels; and E. Jaspers (eds.).  The
      brine shrimp, Artemia.  Vol. 3, Ecology, culturing, use in aquaculture.
      Universa Press, Wettersen, Belgium,  pp. 393-405.

Vanhaecke, P.; H. Steyaert; and P. Sorgeloos.  1980.  International study on
      Artemia. III.  The use of Coulter Counter equipment for the biometrical
      analysis of Artemia cysts.  Methodology and mathematics.  In: Persoone,
       P.; P. Sorgeloos; 0. Roels; and E. Jaspers (eds.).  The brine shrimp,
       Artemia. Vol. I, Morphology, radiobiology, and  toxicology.  Universa
       Press, Wettersen, Belgium,  pp. 107-115.

Walsh, G. E.  and S. V.  Alexander.  1980.  A marine algal bioassay method:
      Results with pesticides and industrial wastes.  Wat. Air Soil Pollut.
      13:45-55.

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, MA.

Ward, G. S. and P. R. Parrish.  1980.  Evaluation of early life stages
      toxicity tests with embryos and juveniles of sheepshead minnows.  In:
      Eaton,  J. G.; P.  R. Parrish; and A. C. Hendricks (eds.).  Aquatic
      Toxicology, ASTM STP 707, American Society for Testing and Materials,
      Philadelphia, PA.  pp. 143-247.
                                      244

-------
Weiss,  C.  M.  1975.  Field investigation of the algal assay procedure on
      surface waters of North Carolina.  In: Middlebrooks, E. J.; E. H.
      Falkenborg; and T. E. Maloney (eds.).  Biostimulation and Nutrient
      Assessment.  Ann Arbor Science Publ., Ann Arbor, MI.  pp. 93-126.

Weiss,  C.  M.  1976.  Evaluation of the algal assay procedure.  U. S.
      Environmental Protection Agency, Corvallis, OR.  EPA-660/3-76-064.

Weiss,  C.  M. and R. W. Helms.  1971.   Interlaboratory precision test:  An
      eight-laboratory evaluation of the Provisional Algal Assay Procedure
      Bottle Test.  National Eutrophication Research Program.  U. S.
      Environmental Protection Agency, Corvallis, OR.

Winner, R. W.  1988.  Evaluation of the relative sensitivities of 7-D
      Daphm'a magna and Ceriodaphm'a dubia toxicity tests for cadmium and
      sodium pentachlorophenate.  Environ. Toxicol. Chem. 7:153-159.

Woltering, D. M.  1984.  The growth response in fish chronic and early life
      stage toxicity tests: A critical review.  Aquat. Toxicol. 5:1-21.

Wong, P. T. S.; Y. K. Chau; and P. L.  Luxon.  1978.  Toxicity of a mixture
      of metals on freshwater algae.   J. Fish. Res. Board Can. 35:479-481.

Yoshioka,  Y.; Y. Ose; and T. Sato.  1986.  Correlation of the five test
      methods to assess chemical toxicity  and relation to physical properties
      Ecotoxicol. Environ. Safety 12:15.
                                      245

-------
                                  APPENDICES

A. Independence, Randomization, and Outliers   	   248
   1. Statistical Independence  	   248
   2. Randomization	248
   3. Outliers	253
B. Validating Normality and Homogeneity of Variance
     Assumptions	255
   1. Introduction	255
   2. Test for Normal  Distribution of Data	255
   3. Test for Homogeneity of Variance	272
   4. Transformations  of Data	273
C. Dunnett's Procedure  	   276
   1. Manual Calculations 	   276
   2. Computer Calculations 	   282
D. T-test with Bonferroni's Adjustment	288
E. Steel's Many-one Rank Test	295
F. Wilcoxon Rank Sum Test	300
G. Fisher's Exact Test	306
H. Single Concentration Toxicity Test - Comparison of Control with
     100% Effluent or Receiving Water	315
I. Probit Analysis	319
J. Linear Interpolation Method  	   324
   1.  General Procedure  	   324
   2.  Data Summary and Plots	324
   3.  Monotonicity	324
   4.  Linear Interpolation Method  	   325

                                      246

-------
5.   Confidence Intervals 	  326
6.   Manual  Calculations  	  326
7.   Computer Calculations  	  330
                                   247

-------
                                  APPENDIX A

                 INDEPENDENCE,  RANDOMIZATION,  AND OUTLIERS


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,
nonnormality, 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 ensure
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 Fathead  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

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.
                                      248

-------
     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.
6.
6.
6.
12
12
12
12
25
25
25
25
50
50
50
50
100
100
100
100
25%
25%
25%
25%
.5%
.5%
.5%
.5%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
effl
effl
effl
effl
effl
effl
effl
effl
effl
effl
effl
effl
effl
effl
effl
effl
effl
effl
effl
effl
uent,
uent,
uent,
uent,
uent,
uent,
uent,
uent,
uent,
uent,
uent,
uent,
uent,
uent,
uent,
uent,
uent,
uent,
uent,
uent,
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
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.
                                      249

-------
TABLE A.2.  TABLE OF RANDOM NUMBERS
10 09 73
37 54 20
08 42 26
99 01 90
12 80 79
66 06 57
31 06 01
85 26 97
63 57 33
73 79 64
98 52 01
11 80 50
83 45 29
88 68 54
99 59 46
65 48 11
80 12 43
74 35 09
69 91 62
09 89 32
91 49 91
80 33 69
44 10 48
12 55 07
63 60 64
61 19 69
15 47 44
94 55 72
42 48 11
23 52 37
04 49 35
00 54 99
35 96 31
59 80 80
46 05 88
32 17 90
69 23 46
19 56 54
45 15 51
94 86 43
98 08 62
33 18 51
80 95 10
79 75 24
18 63 33
74 02 94
54 17 84
11 66 44
48 32 47
69 07 49

25 33
48 05
89 53
25 29
99 70
47 17
08 05
76 02
21 35
57 53
77 67
54 31
96 34
02 00
73 48
76 74
56 35
98 17
68 03
05 05
45 23
45 98
19 49
37 42
93 29
04 46
52 66
85 73
62 13
83 17
24 94
76 54
53 07
83 91
52 36
05 97
14 06
14 30
49 38
19 94
48 26
62 32
04 06
91 40
25 37
39 02
56 11
98 83
79 28
41 38

76
64
19
09
80
34
45
02
05
03
14
39
06
86
87
17
17
77
66
14
68
26
85
11
16
26
95
67
97
73
75
64
26
45
01
87
20
01
19
36
45
41
96
71
98
77
80
52
31
87

52 01
89 47
64 50
37 67
15 73
07 27
57 18
05 16
32 54
52 96
90 56
80 82
28 89
50 75
51 76
46 85
72 70
40 27
25 22
22 56
47 92
94 03
15 74
10 00
50 53
45 74
27 07
89 75
34 40
20 88
24 63
05 18
89 80
42 72
39 09
37 92
11 74
75 87
47 60
16 81
24 02
94 15
38 27
96 12
14 50
55 73
99 33
07 98
24 96
63 79

35 86
42 96
93 03
07 15
61 47
68 50
24 06
56 92
70 48
47 78
86 07
77 32
80 83
84 01
49 69
09 50
80 15
72 14
91 48
85 14
76 86
68 58
79 54
20 40
44 84
77 74
99 53
43 87
87 21
98 37
38 24
81 59
93 45
68 42
22 86
52 41
52 04
53 79
72 46
08 51
84 04
09 49
07 74
82 96
65 71
22 70
71 43
48 27
47 10
19 76

34 67
24 80
23 20
38 31
64 03
36 69
35 30
68 66
90 55
35 80
22 10
50 72
13 74
36 76
91 82
58 04
45 31
43 23
36 93
46 42
46 16
70 29
32 97
12 86
40 21
51 92
59 36
54 62
16 86
68 93
45 86
96 11
33 35
83 60
77 28
05 56
15 95
40 41
43 66
34 88
44 99
89 43
20 15
69 86
31 01
97 79
05 33
59 38
02 29
35 58

35 43
52 40
90 25
13 11
23 66
73 61
34 26
57 48
35 75
83 42
94 05
56 82
67 00
66 79
60 89
77 69
82 23
60 02
68 72
75 67
28 35
73 41
92 65
07 46
95 25
43 37
78 38
24 44
84 87
59 14
25 10
96 38
13 54
94 97
14 40
70 70
66 00
92 15
79 45
88 15
90 88
54 85
12 33
10 25
02 46
01 71
51 29
17 15
53 68
40 44
250
76
37
60
65
53
70
14
18
48
82
58
48
78
51
28
74
74
10
03
88
54
35
75
97
63
29
48
31
67
16
25
96
62
00
77
07
00
85
43
53
96
81
87
91
74
19
69
39
70
01

80 95 90
20 63 61
15 95 33
88 67 67
98 95 11
65 81 33
86 79 90
73 05 38
28 46 82
60 93 52
60 97 09
29 40 52
18 47 54
90 36 47
93 78 56
73 03 95
21 11 57
45 52 16
76 62 11
96 29 77
94 75 08
53 14 03
57 60 04
96 64 48
43 65 17
65 39 45
82 39 61
91 19 04
03 07 11
26 25 22
61 96 27
54 69 28
77 97 45
13 02 12
93 91 08
86 74 31
18 74 39
66 67 43
59 04 79
01 54 03
39 09 47
88 69 54
25 01 62
74 85 22
05 45 56
52 52 75
56 12 71
09 97 33
32 30 75
10 51 82

91 17
04 02
47 64
43 97
68 77
98 85
74 39
52 47
87 09
03 44
34 33
42 01
06 10
64 93
13 68
71 86
82 53
42 37
39 90
88 22
99 23
33 40
08 81
94 39
70 82
95 93
01 18
25 92
20 59
96 63
93 35
23 91
00 24
48 92
36 47
71 57
24 23
68 06
00 33
54 56
34 07
19 94
52 98
05 39
14 27
80 21
92 55
34 40
75 46
16 15

39 29
00 82
35 08
04 43
12 27
11 19
23 40
18 62
83 49
35 27
50 50
52 77
68 71
29 60
23 47
40 21
14 38
96 28
94 40
54 38
37 08
42 05
22 22
28 70
07 20
42 58
33 21
92 92
25 70
05 52
65 33
23 28
90 10
78 56
70 61
85 39
97 11
84 96
20 82
05 01
35 44
37 54
94 62
00 38
77 93
80 81
36 04
88 46
15 02
01 84

27 49 45
29 16 65
03 36 06
62 76 59
17 68 33
92 91 70
30 97 32
38 85 79
12 56 24
38 84 35
07 39 98
56 78 51
17 78 17
91 10 62
83 41 13
81 65 44
55 37 63
60 26 55
05 64 18
21 45 98
92 00 48
08 23 41
20 64 13
72 58 15
73 17 90
26 05 27
15 94 66
74 59 73
14 66 70
28 25 62
71 24 72
72 95 29
33 93 33
52 01 06
74 29 41
41 18 38
89 63 38
28 52 07
66 95 41
45 11 76
13 18 80
87 30 43
46 11 71
75 95 79
89 19 36
45 17 48
09 03 24
12 33 56
00 99 94
87 69 38


-------
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.  EXAMPLE OF RANDOM ASSIGNMENT OF FIRST TEN FISH TO REPLICATE
                 CHAMBERS
Fish
Assignment
First
Second
Third
Fourth
Fifth
Sixth
Seventh
Eighth
Ninth
Tenth
fi
fi
fi
fi
fi
fi
fi
fi
fi
fi
sh
sh
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%
6.25%
50.0%
100.0%
6.25%
25.0%
50.0%
100.0%
50.0%
100.0%
effluent,
effluent,
effluent
effluent,
effluent,
effluent,
effluent,
effluent,
effluent,
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.

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.
                                      251

-------
   TABLE A.4.  RANDOM ASSIGNMENT OF REPLICATE CHAMBERS TO POSITIONS EXAMPLE
               LABELLING THE POSITIONS WITHIN THE WATER BATH
1
7
13
19
2
8
14
20
3
9
15
21
4
10
16
22
5
11
17
23
6
12
18
24
   TABLE A.5.  RANDOM ASSIGNMENT OF REPLICATE CHAMBERS TO POSITIONS EXAMPLE
               ASSIGNED NUMBERS FOR EACH POSITION
Assigned
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,
Numbers
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
Position
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
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.

                                     252

-------
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.  RANDOM ASSIGNMENT OF REPLICATE CHAMBERS TO POSITIONS EXAMPLE
               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
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.

                                      253

-------
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).
                                     254

-------
                                  APPENDIX B

       VALIDATING NORMALITY AND HOMOGENEITY OF VARIANCE ASSUMPTIONS

1.   INTRODUCTION

1.1  Dunpett'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 judgment call, and a statistician
should be consulted in selecting the analysis.

2.   TEST  FOR NORMAL DISTRIBUTION OF DATA

2.1  Shapiro-Wilk's Test

2.1.1  One formal test for normality is the Shapiro-Wilk's Test.  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 is recommended.
An example of the Shapiro-Wilk's test is provided below.

2.2  The  example uses growth data from the Fathead 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 the 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:


                     D =  5 (X,-   X)2
    Where:  Xi  =  the centered observations and X is the overall  mean of
                  the centered observations.  For this set of data, X = 0,
                  and D = 0.0412.
                                      255

-------
2.5  Order the centered observations from smallest to largest.
                           (1)     Y<2>           _    (n)
                          f\    ~  A      •  •  •     A


    where X(i) denotes the ith ordered observation.   The ordered  observations
are 1isted in Table B.3.

2.6  From Table B.4, 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 = 20, k = 10.  The a,- values are listed
in Table B.5.

2.7  Compute the test statistic, W, as follows:

                          r   k
                 W = jl    S  a,-  (X(n-i+1)      X(i))
                      D   L  i = l


    The differences, X(n"i+1) - X(O, 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 at a significance level  of 0.01 and 20 observations (n)  is
0.868.   The calculated value, 0.959, 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.
                                      256

-------
        TABLE B.I.   FATHEAD LARVAL GROWTH DATA (WEIGHT IN MG)
                    FOR THE SHAPIRO-WILK'S TEST
                                   NaPCP Concentration
Replicate      Control             32        64       128       256
A
B
C
D
Mean(Y,.)
s?
i
0.711
0.662
0.718
0.767
0.714
0.0018
1
0.646
0.626
0.723
0.700
0.674
0.0020
2
0.669
0.669
0.694
0.676
0.677
0.0001
3
0.629
0.680
0.513
0.672
0.624
0.0059
4
0.650
0.558
0.606
0.508
0.580
0.0037
5
    TABLE B.2.   EXAMPLE OF SHAPIRO-WILK'S TEST:   CENTERED OBSERVATIONS
                                       NaPCP Concentration (uq/L)
 Replicate           Control        32        64        128        256
A
B
C
D
-0.003
-0.052
0.004
0.053
-0.028
-0.048
0.049
0.026
-0.008
-0.008
0.017
-0.001
0.005
0.056
-0.111
0.048
0.070
-0.022
0.026
-0.072
                                      257

-------
TABLE B.3.  EXAMPLE OF THE SHAPIRO-WILK'S TEST:  ORDERED OBSERVATIONS
                                                      X(f)
1
2
3
4
5
6
7
8
9
10
-0.111
-0.072
-0.052
-0.048
-0.028
-0.022
-0.008
-0.008
-0.003
-0.001
11
12
13
14
15
16
17
18
19
20
0.004
0.005
0.017
0.026
0.026
0.048
0.049
0.053
0.056
0.070
                                   258

-------
        TABLE B.4.  COEFFICIENTS'FOR THE SHAPIRO-WILK'S TEST1
i
1
2
3
4
5
i
1
2
3
4
5
6
7
8
9
10
2 3
0.7071 0.7071
n nnnn
	 u . uuuu



11 12
0.5601 0.5475
0.3315 0.3325
0.2260 0.2347
0.1429 0.1586
0.0695 0.0922
0.0000 0.0303




Number
4
0.6872 0
0.1667 0
n
u


Number
13
0.5359 0.
0.3325 0.
0.2412 0.
0.1707 0.
0.1099 0.
0.0539 0.
0.0000 0.



of observations
5 6
.6646
.2413
.0000


0.6431
0.2806
0.0875




of observations
14 15
5251
3318
2460
1802
1240
0727
0240



0.5150
0.3306
0.2495
0.1878
0.1353
0.0880
0.0433
0.0000


0
0
0
0
0
0
0
0


7
0.6233
0.3031
0.1401
0.0000

16
.5056
.3209
.2521
.1939
.1447
.1005
.0593
.0196


8
0.6052
0.3164
0.1743
0.0561

17
0.4968
0.3273
0.2540
0.1988
0.1524
0.1109
0.0725
0.0359
0.0000

9
0.5888
0.3244
0.1976
0.0947
0.0000
18
0.4886
0.3253
0.2553
0.2027
0.1587
0.1197
0.0837
0.0496
0.0163

10
0.5739
0.3291
0.2141
0.1224
0.0399
19
0.4808
0.3232
0.2561
0.2059
0.1641
0.1271
0.0932
0.0612
0.0303
0.0000



0
0
0
0
0
0
0
0
0
0


20
.4734
.3211
.2565
.2085
.1686
.1334
.1013
.0711
.0422
.0140
                   Number of observations
   21      22      23      24      25      26      27      28      29      30

074643  0.4590  0.4542  0.4493  0.4450  0.4407  0.4366  0..4328  0.4291  0.4254
0.3185  0.3156  0.3126  0.3098  0.3069  0.3043  0.3018  0.2992  0.2968  0.2944
0.2578  0.2571  0.2563  0.2554  0.2543  0.2533  0.2522  0.2510  0.2499  0.2487
0.2119  0.2131  0.2139  0.2145  0.2148  0.2151  0.2152  0.2151  0.2150  0.2148
0.1736  0.1764  0.1787  0.1807  0.1822  0.1836  0.1848  0.1857  0.1864  0.1870
0.1399
0.1092
0.0804
0.0530
0.0263
0.0000



0
0
0
0
0
0



.1443
.1150
.0878
.0618
.0368
.0122



0
0
0
0
0
0
n
u


.1480
.1201
.0941
.0696
.0459
.0228
nnnn



0.1512
0.1245
0.0997
0.0764
0.0539
0.0321
n m n?



0.1539
0.1283
0.1046
0.0923
0.0610
0.0403
0 0200
0 0000


0.1563
0.1316
0.1089
0.0876
0.0672
0.0476
0 0284
0 0094


0.1584
0.1346
0.1128
0.0923
0.0728
0.0540
0 0358
0 0178
0 0000

0.1601
0.1372
0.1162
0.0965
0.0778
0.0598
0 0424
0 0253
0 0084

0.1616
0.1395
0.1192
0.1002
0.0822
0.0650
0 0483
0.0320
0 0159
0.0000
0.1630
0.1415
0.1219
0.1036
0.0862
0.0697
0 0537
0 0381
0 0227
0.0076
                 1Taken from Conover (1980).

                                    259

-------
TABLE B.4.  COEFFICIENTS FOR THE SHAPIRO-WILK'S TEST (Continued)
i
i
i
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
7D
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







0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0






32
.4188
.2898
.2462
.2141
.1878
.1651
.1449
.1265
.1093
.0931
.0777
.0629
.0485
.0344
.0206
.0068






Number of observations
33 34 35 36
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




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




0.4096
0.2834
0.2427
0.2127
0.1883
0.1673
0.1487
0.1317
0.1160
0.1013
0.0873
0.0739
0.0610
0.0484
0.0361
0.0239
0.0119
0.0000


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
0.1663
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
n nnnn
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
n nrua
                                   260

-------
TABLE B.4.  COEFFICIENTS FOR THE SHAPIRO-MILK'S TEST (Continued)
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
9R
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






Number of observations
43 44 45 46
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0




.3894
.2684
.2334
.2078
.1871
.1695
.1539
.1398
.1269
.1149
.1035
.0927
.0824
.0724
.0628
.0534
.0442
.0352
.0263
.0175
.0087
.0000




0.3872
0.2667
0.2323
0.2072
0.1868
0.1695
0.1542
0.1405
9,1278
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




0.3850
0.2651
0.2313
0.2065
0.1865
0.1695
0.1545
0.1410
0.1286
9.1170
0.1062
0.0959
0.0860
0.0765
0.0673
0.0584
0.0497
0.0412
0.0328
0.0245
0.0163
0.0081
0.0000


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
n nnnn
50
0.3751
0.2574
0.2260
0.2032
0.1847
0.1691
0.1554
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
n nmc;
          TABLE B.5.  EXAMPLE OF THE SHAPIRO-WILK'S TEST:
                      TABLE OF COEFFICIENTS AND DIFFERENCES
i a,-
1 0.4734
2 0.3211
3 0.2565
4 0.2085
5 0.1686
6 0.1334
7 0.1013
8 0.0711
9 0.0422
10 0.0140
X( n* i *1 ) v ( i )
A
0.181
0.128
0.105
0.097
0.076
0.048
0.034
0.025
0.008
0.005

x(20>
Y ( 19)
X(18)
X( 1 7)
,,..
X(16)
v C 1 5 )
x(14)
x(13)
X(12)
x<11)

- x^1>
- x(2)
- x(3)
- x(4)
y(5)
- x(6)
- x(7)
X<8)
- x(9)
X(10)
                                   261

-------
TABLE B.6.  QUANTILES OF THE SHAPIRO-MILK'S TEST STATISTIC (Conover,  1990)
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
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
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
0.929
0.929
0.930
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
0.937
0.937
0.938
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
0.947
0.947
0.947
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
0.954
0.955
0.955
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
0.974
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
0.985
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
0.988
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
0.990
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
0.991
0.991
0.991
                                     262

-------
2.10  Kolmogorov "D" Test

2.10.1  A formal two-sided test for normality is the Kolmogorov "D" Test.  The
test statistic is calculated by obtaining the difference between the
cumulative distribution function estimated from the data and the standard
normal cumulative distribution function for each standardized observation.
This test is recommended for a sample size greater than 50.  If the sample
size is less than or equal to 50, then the Shapiro-Wilk's Test is recommended.
An example of the Kolmogorov "D" test is provided below.

2.10.2  The example uses reproduction data from the daphnid, Ceriodaphm'a
dubia, Survival and Reproduction Test.  The observed data and the mean of the
observations at each concentration, including the control, are listed in Table
B.7.

2.10.3  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 for the
example are listed in Table B.8.

2.10.4  Order the centered observations from smallest to largest:

                           X(1) < X(2)  <  ... < X(n)

where X(O denotes the ith ordered observation, and n denotes the total number
of centered observations.  The ordered observations for the example are listed
in Table B.9.

2.10.5  The next step is to standardize the ordered observations.  Let z,-
denote the standardized value of the  ith ordered observation.  Then,

                         xci)                  E [x
-------
2.10.7  Next, calculate the following differences for each ordered
observation:

                         D,-+  =  (i/n)  -  p,.

                         Dr  =  Pj  - [(i-D/n]

The differences for the example are listed in Table B.9.

2.10.8  Obtain the maximum of the D,-+,  and denote it  as  D+.   Obtain the
maximum of the D,--,  and denote it as D-.   For the example,  D+ = 0.0525, and D-
= 0.0597.

2.10.9  Next, obtain the maximum of D+  and D-, and denote it as D.  For the
example, D = 0.0597.

2.10.10  The test statistic, D*, is calculated as follows:

                         D*  =  D(Vn -  0.01 + 0.85/Vh")

For the example, D* = 0.4684.

2.10.11  The decision rule for the two  tailed test is to compare the critical
value from Table B.ll to the computed D*.  If the computed value is greater
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 is 1.035.
The calculated value, 0.4684, is not greater than the critical value.   Thus,
the conclusion of the test is that the  data are normally distributed.

2.10.12  In general, if the data fail the test for normality, a transformation
such as the log transformation may normalize the data.  After transforming the
data, repeat the Kolmogorov D test for  normality.
                                     264

-------
       TABLE B.7.  CERIODAPHNIA DUBIA REPRODUCTION DATA
                   FOR THE KOLMOGOROV D TEST
Effluent Concentration (%)
Replicate
1
2
3
4
5
6
7
8
9
10
Mean
Control
27
30
29
31
16
15
18
17
14
27
22.4
1 . 56%
32
35
32
26
18
29
27
16
35
13
26.3
3.12%
39
30
33
33
36
33
33
27
38
44
34.6
6.25%
27
34
36
34
31
27
33
31
33
31
31.7
12.5%
19
25
26
17
16
21
23
15
18
10
19.0
25.0%
10
13
7
7
7
10
10
16
12
2
9.4
TABLE B.8.  CENTERED OBSERVATIONS FOR KOLMOGOROV D EXAMPLE
Effluent Concentration (%)
Replicate
1
2
3
4
5
6
7
8
9
10
Control
4.6
7.6
6.6
8.6
-6.4
-7.4
-4.4
-5.4
-8.4
4.6
1 . 56%
5.7
8.7
5.7
-0.3
-8.3
2.7
0.7
-10.3
8.7
-13.3
3.12%
4.4
-4.6
-1.6
-1.6
1.4
-1.6
-1.6
-7.6
3.4
9.4
6.25%
-4.7
2.3
4.3
2.3
-0.7
-4.7
1.3
-0.7
1.3
-0.7
12.5%
0.0
6.0
7.0
-2.0
-3.0
2.0
4.0
-4.0
-1.0
-9.0
25.0%
0.6
3.6
-2.4
-2.4
-2.4
0.6
0.6
6.6
2.6
-7.4
                           265

-------
TABLE B.9.  EXAMPLE CALCULATION OF THE KOLMOGOROV D STATISTIC
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
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
X(i)
-13.3
-10.3
-9.0
-8.4
-8.3
-7.6
-7.4
-7.4
-6.4
-5.4
-4.7
-4.7
-4.6
-4.4
-4.0
-3.0
-2.4
-2.4
-2.4
-2.0
-1.6
-1.6
-1.6
-1.6
-1.0
-0.7
-0.7
-0.7
-0.3
0.0
0.6
0.6
0.6
0.7
1.3
1.3
1.4
2.0
2.3
2.3
2.6
2.7
3.4
3.6
4.0
4.3
4.4
zi
-2.51
-1.94
-1.70
-1.58
-1.57
-1.43
1.40
1.40
1.21
1.02
-0.89
-0.89
-0.87
-0.83
-0.75
-0.57
-0.45
-0.45
-0.45
-0.38
-0.30
-0.30
-0.30
-0.30
-0.19
-0.13
-0.13
-0.13
-0.06
0.00
0.11
0.11
0.11
0.13
0.25
0.25
0.26
0.38
0.43
0.43
0.49
0.51
0.64
0.68
0.75
0.81
0.83
Pi
0.0060
0.0262
0.0446
0.0571
0.0582
0.0764
0.0808
0.0808
0.1131
0.1539
0.1867
0.1867
0.1922
0.2033
0.2266
0.2843
0.3264
0.3264
0.3264
0.3520
0.3821
0.3821
0.3821
0.3821
0.4247
0.4483
0.4483
0.4483
0.4761
0.5000
0.5438
0.5438
0.5438
0.5517
0.5987
0.5987
0.6026
0.6480
0.6664
0.6664
0.6879
0:6950
0.7389
0.7517
0.7734
0.7910
0.7967
Di+
0.0107
0.0071
0.0054
0.0096
0.0251
0.0236
0.0359
0.0525
0.0369
0.0128
-0.0034
0.0133
0.0245
0.0300
0.0234
-0.0176
-0.0431
-0.0264
-0.0097
-0.0187
-0.0321
-0.0154
0.0012
0.0179
-0.0080
-0.0150
0.0017
0.0184
0.0072
0.0000
-0.0271
-0.0105
0.0062
0.0150
-0.0154
0.0013
0.0141
-0.0147
-0.0164
0.0003
-0.0046
0.0050
-0.0222
-0.0184
-0.0234
-0.0243
-0.0134
"V
0.0060
0.0095
0.0113
0.0071
-0.0085
-0.0069
-0.0192
-0.0359
-0.0202
0.0039
0.0200
0.0034
-0.0078
-0.0134
-0.0067
0.0343
0.0597
0.0431
0.0264
0.0353
0.0488
0.0321
0.0154
-0.0012
0.0247
0.0316
0.0150
-0.0017
0.0094
0.0167
0.0438
0.0271
0.0105
0.0017
0.0320
0.0154
0.0026
0.0313
0.0331
0.0164
0.0212
0.0117
0.0389
0.0350
0.0401
0.0410
0.0300
                            266

-------
TABLE B.9.  EXAMPLE CALCULATION OF THE KOLMOGOROV D STATISTIC (CONTINUED)
i
48
49
50
51
52
53
54
55
56
57
58
59
60
xci>
4.6
4.6
5.7
5.7
6.0
6.6
6.6
7.0
7.6
8.6
8.7
8.7
9.4
zi
0.87
0.87
1.08
1.08
1.13
1.25
1.25
1.32
1.43
1.62
1.64
1.64
1.77
Pi
0.8078
0.8078
0.8599
0.8599
0.8708
0.8944
0.8944
0.9066
0.9236
0.9474
0.9495
0.9495
0.9616
Di+
-0.0078
0.0089
-0.0266
-0.0099
-0.0041
-0.0111
0.0056
0.0101
0.0097
0.0026
0.0172
0.0338
0.0384
Dr
0.0245
0.0078
0.0432
0.0266
0.0208
0.0277
0.0111
0.0066
0.0069
0.0141
-0.0005
-0.0172
-0.0217
                                     267

-------
TABLE B.
z
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
0.11
0.12
0.13
0.14
0.15
0.16
0.17
0.18
0.19
0.20
0.21
0.22
0.23
0.24
0.25
0.26
0.27
0.28
0.29
0.30
0.31
0.32
0.33
0.34
0.35
0.36
0.37
0.38
0.39
0.40
0.41
0.42
0.43
0.44
0.45
0.46
0.47
10

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
. P IS THE VALUE
P
.5000
.5040
.5080
.5120
.5160
.5199
.5239
.5279
.5319
.5359
.5398
.5438
.5478
.5517
.5557
.5596
.5636
.5675
.5714
.5753
.5793
.5832
.5871
.5910
.5948
.5987
.6026
.6064
.6103
.6141
.6179
.6217
.6255
.6293
.6331
.6368
.6406
.6443
.6480
.6517
.6554
.6591
.6628
.6664
.6700
.6736
.6772
.6808

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
z
.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
OF THE STANDARD NORMAL CUMULATIVE DISTRIBUTION AT
P
0.6844
0.6879
0.6915
0.6950
0.6985
0.7019
0.7054
0.7088
0.7123
0.7157
0.7190
0.7224
0.7257
0.7291
0.7324
0.7357
0.7389
0.7422
0.7454
0.7486
0.7517
0.7549
0.7580
0.7611
0.7642
0.7673
0.7704
0.7734
0.7764
0.7794
0.7823
0.7852
0.7881
0.7910
0.7939
0.7967
0.7995
0.8023
0.8051
0.8078
0.8106
0.8133
0.8159
0.8186
0.8212
0.8238
0.8264
0.8289
z
0.
0.
0.
0.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
96
97
98
99
00
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
P
0.8315
0.8340
0.8365
0.8389
0.8413
0.8438
0.8461
0.8485
0.8508
0.8531
0.8554
0.8577
0.8599
0.8621
0.8643
0.8665
0.8686
0.8708
0.8729
0.8749
0.8770
0.8790
0.8810
0.8830
0.8849
0.8869
0.8888
0.8907
0.8925
0.8944
0.8962
0.8980
0.8997
0.9015
0.9032
0.9049
0.9066
0.9082
0.9099
0.9115
0.9131
0.9147
0.9162
0.9177
0.9192
0.9207
0.9222
0.9236

1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
z
.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
P
0.9251
0.9265
0.9279
0.9292
0.9306
0.9319
0.9332
0.9345
0.9357
0.9370
0.9382
0.9394
0.9406
0.9418
0.9429
0.9441
0.9452
0.9463
0.9474
0.9484
0.9495
0.9505
0.9515
0.9525
0.9535
0.9545
0.9554
0.9564
0.9573
0.9582
0.9591
0.9599
0.9608
0.9616
0.9625
0.9633
0.9641
0.9649
0.9656
0.9664
0.9671
0.9678
0.9686
0.9693
0.9699
0.9706
0.9713
0.9719
268

-------
TABLE B.10.
z
1.92
1.93
1.94
1.95
1.96
1.97
1.98
1.99
2.00
2.01
2.02
2.03
2.04
2.05
2.06
2.07
2.08
2.09
2.10
2.11
2.12
2.13
2.14
2.15
2.16
2.17
2.18
2.19
2.20
2.21
2.22
2.23
2.24
2.25
2.26
2.27
2.28
2.29
2.30
2.31
2.32
2.33
2.34
2.35
2.36
2.37
2.38
2.39
2.40

0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0,
p
9726
9732
9738
9744
9750
9756
9761
9767
9772
9778
9783
9788
9793
9798
9803
9808
9812
9817
9821
9826
9830
9834
9838
9842
9846
9850
9854
9857
9861
9864
9868
9871
9875
9878
9881
9884
9887
9890
9893
9896
9898
9901
9904
9906
9909
9911
9913
9916
9918

2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
z
.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

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
p
.9920
.9922
.9925
.9927
.9929
.9931
.9932
.9934
.9936
.9938
.9940
.9941
.9943
.9945
.9946
.9948
.9949
.9951
.9952
.9953
.9955
.9956
.9957
.9959
.9960
.9961
.9962
.9963
.9964
.9965
.9966
.9967
.9968
.9969
.9970
.9971
.9972
.9973
.9974
.9974
.9975
.9976
.9977
.9977
.9978
.9979
.9979
.9980
.9981
(CONTINUED)

2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
z
.90
.91
.92
.93
.94
.95
.96
.97
.98
.99
.00
.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
P
0.9981
0.9982
0.9982
0.9983
0.9984
0.9984
0.9985
0.9985
0.9986
0.9986
0.9987
0.9987
0.9987
0.9988
0.9988
0.9989
0.9989
0.9989
0.9990
0.9990
0.9990
0.9991
0.9991
0.9991
0.9992
0.9992
0.9992
0.9992
0.9993
0.9993
0.9993
0.9993
0.9994
0.9994
0.9994
0.9994
0.9994
0.9995
0.9995
0.9995
0.9995
0.9995
0.9995
0.9996
0.9996
0.9996
0.9996
0.9996
0.9996
z
3.39
3.40
3.41
3.42
3.43
3.44
3.45
3.46
3.47
3.48
3.49
3.50
3.51
3.52
3.53
3.54
3.55
3.56
3.57
3.58
3.59
3.60
3.61
3.62
3.63
3.64
3.65
3.66
3.67
3.68
3.69
3.70
3.71
3.72
3.73
3.74
3.75
3.76
3.77
3.78
3.79
3.80
3.81
3.82
3.83
3.84
3.85
3.86
3.87
P
0.9997
0.9997
0.9997
0.9997
0.9997
0.9997
0.9997
0.9997
0.9997
0.9997
0.9998
0.9998
0.9998
0.9998
0.9998
0.9998
0.9998
0.9998
0.9998
0.9998
0.9998
0.9998
0.9998
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
269

-------
                        TABLE B.10.   (CONTINUED)
3.88  0.9999        3.91  1.0000        3.94  1.0000        3.97  1.0000
3.89  0.9999        3.92  1.0000        3.95  1.0000        3.98  1.0000
3.90  1.0000        3.93  1.0000        3.96  1.0000        3.99  1.0000
                                   270

-------
TABLE B.ll.  CRITICAL VALUES FOR THE KOLMOGOROV D TEST

              Alpha          Critical
              Level             Value

              0.010             1.035
              0.025             0.955
              0.050             0.895
              0.100             0.819
              0.150             0.775
                            271

-------
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 Fathead Minnow
Larval  Survival  and Growth Test, and are the same data used in Appendices
C and D.  These data are listed in Table B.12, together with the calculated
variance for the control and each toxicant concentration.

3.3  The test statistic for Bartlett's Test (Snedecor and Cochran, 1980) is as
follows:

               [(2 V,)  In S2 -  2 V,  In S,2]
                i=l           i=l
            B =         	
Where:  V,-   =  Degrees of freedom for each toxicant concentration and control
        p   =  Number of levels of toxicant concentration including the
               control
         ,      (S V, S,-2)
        S   =   1=1
                  [3(p-l)r1  [S1/V,   (2PVi)'1]
                             1=1      1=1
       In   = Log
 3.4  Since B is approximately distributed as chi-square with p - 1 degrees of
freedom when the variances are equal, the appropriate critical value is
obtained from a table of the chi-square distribution for p - 1 degrees of
freedom and a significance level of 0.01.  If B is less than the critical
value then the variances are assumed to be equal.

3.5  For the data in this example, V,- = 3,  p  = 5,  S  = 0.0027, and
C = 1.133.  The calculated B value is:
                                           P     2
                      (15)[ln(0.0027)] - 3 2 ln(Sf)
                                          i = l
                  B = _
                                 1.133
                                       272

-------
                       15(  5.9145) - 3(- 32.4771)
                                 1.133
                   = 7.691
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 13.277 for a significance level of 0.01.  Since B = 7.691 is less than
the critical value of 13.277, conclude that the variances are not different.
    TABLE B.12.   FATHEAD LARVAL GROWTH DATA (WEIGHT IN MG) USED FOR
                 BARTLETT'S TEST FOR HOMOGENEITY OF VARIANCE
                                   NaPCP Concentration (ug/L)
Replicate
Control
32
64
128
256
A
B
C
D
Mean(Y,)
s

i
0.711
0.662
0.718
0.767
0.714
0.0018

1
0.646
0.626
0.723
0.700
0.674
0.0020

2
0.669
0.669
0.694
0.676
0.677
0.0001

3
0.629
0.680
0.513
0.672
0.624
0.0059

4
0.650
0.558
0.606
0.508
0.580
0.0037

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 by 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
reciprocals.  After the data have been transformed, 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.

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 ith treatment is
proportional to P,- (1    P,),  where  P, is  the expected proportion for the
1From USEPA (1985d).
                                       273

-------
treatment.  This clearly violates the homogeneity of variance  assumption  required
by parametric procedures such as Dunnett's Procedure or the t-test with
Bonferroni's adjustment, 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
normality assumption may be invalid.  The arc sine square root  (arc sine  vP)
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
calculated, 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

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

                                       274

-------
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
                                        275

-------
                                  APPENDIX C

                             DUNNETT'S PROCEDURE
1.  MANUAL CALCULATIONS

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, a t-test with
Bonferroni's adjustment is used (see Appendix D).

1.2  The data used in this example are growth data from a Fathead 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.
  TABLE C.I.   FATHEAD MINNOW,  PIMEPHALES PROMELAS,  LARVAL GROWTH DATA
              (WEIGHT IN MG)  USED FOR DUNNETT'S PROCEDURE
Replicate
                                   NaPCP Concentration (uq/L)
Control
32
64
128
256
A
B
C
D
Mean(Yj)
Total (Ts)
0.711
0.662
0.718
0.767
0.714
2.858
0.646
0.626
0.723
0.700
0.674
2.695
0.669
0.669
0.694
0.676
0.677
2.708
0.629
0.680
0.513
0.672
0.624
2.494
0.650
0.558
0.606
0.508
0.580
2.322
                                       276

-------
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 = E  Y,.j      G /N
                              ij


Between Sum of Squares: SSB = 2 T./n.     G /N
                              i

Within Sum of Squares:  SSW = SST   SSB

    Where: G = The grand total of all sample observations; G = E T,-
                                                               i
           N = The total sample size; N = Z n,.
                                        i

          n^ = The number of replicates for concentration "i11

          I,- = The total  of the replicate measurements for concentration "i"

         YJJ  =  The jth observation  for concentration  "i"


1.4  Calculations:


Total Sum of Squares:    SST = s Y^-      G /N
                               ij


                             = 8.635  - (13.077)2/20

                             = 0.085
Between Sum of Squares:  SSB = 2  T/n,-     G /N
                             = 8.594  -  (13.077)2/20

                             = 0.044

Within Sum of Squares:   SSW = SST  -  SSB

                             = 0.085    0.044

                             = 0.041
                                        277

-------
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
                S  = 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      0.044

Within    20   5 = 15      0.041
                                  0.011

                                  0.0027
Total
  19
0.085
                                       278

-------
1.7  To perform the individual comparisons, calculate the t statistic for
each concentration and control combination, as follows:


                                   (Yi - V,  )
   Where:  Y,- =  Mean for each concentration

           Y1 =  Mean for the control

           SH =  Square root of the within mean square

           n1 =  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.
              NaPCP             i                   t,-
          Concentration
             (M9/L)
32
64
128
256
2
3
4
5
1.081
1.000
2.432
3.622
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 (2.36), with an overall alpha level of 0.05, 15 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).  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 t4
and t5 are greater than 2.36, the 128 ug/L and 256 ug/L concentrations have
significantly lower growth than the control.  Hence the NOEC and LOEC for

                                       279

-------
growth are 64 ug/L and 128 ug/L, respectively.
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 V(l/n1) +  (1/n)

      Where: d   = Critical value for the Dunnett's Procedure
             Su  = 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
             n.,  = Number of replicates  in the  control
    For example:
          MSD = 2.36 (0.052) [3 V(l/4) + (1/4)]  = 2.36 (0.052)(V 2/4 )
              = 2.36 (0.052)(0.707)
              = 0.087
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
0.087 mg.  This represents a decrease in growth of 12% 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.
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

                                       280

-------
1.11.2.2  Calculate the percent reduction from the control that MSD
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:

                            0.714 - 0.087 = 0.627

    Step 2. Obtain untransformed values for the control mean (0.714) and the
            difference (0.627) obtained in Step 1, above.

                [Sine(0.714)]2  =  0.429
                [Sine(0.627)]2  =  0.344

    Step 3. The untransformed MSD (MSDJ is determined by subtracting the
            untransformed values obtained in Step 2.

            MSDU  =  0.429 - 0.344  =  0.085

            In this case, the MSD would represent a 19.8% decrease in survival
            from the control  [(0.085/0.429)(100)].
                                        281

-------
                    TABLE C.5.  DUNNETT'S "T" VALUES1
                                    (One-tailed) d

\*
J
g
7
g
9
10
11
12
13
14
IS
16
17
18
19
30
34
30
40
60
110


1
2.02
.94
.89
.38
.63
.81
.80
.78
.77
.78
.75
.75
.74
.73
.73
.72
.71
.70
.68
.87
.66
.64

























2
1.44
1.34
.27
.22
.18
.15
. 13
. 11
.09
.08
.07
.06
.05
.04
.03
.03
.01
.99
.97
.99
.93
.92

3
3.68
2.56
3.48^
3.43
3.37
3.34
3.31
3.39
3.27
2.39
2.24
2.23
2.22
2.21
1.10
1.19
1.17
1.13
1.13
1.10
1.08
1.08

4
2.85
2.71
3.63
3.55
1.50
3.47
2.44
3.41
2.39
1.37
3.36
2.34
2.33
1.31
1.31
1.30
1.38
1.19
3.23
3.11
3.18
1.16
a -























• .05
5
1.98
1.83
.73
.66
.60
.56
.53
.50
.48
.48
.44
.43
.42
.41.
.40
.39
.36
.33
.31
.38
1.36
1.33

























6
1.08
1.91
1.81
.74
.68
.64
.60
.58
.55
.53
.51
.50
.49
.48
.47
.46
.43
.40
.37
.33
.32
.M

























7
.16
.00
.89
.81
.75
.70
.67
.64
.61
.39
.57
.56
.54
.53
.52
.51
.48
.45
.42
.39
.37
.34

























a
.34
.07
.95
.87
.81
.76
.73
.69
.68
.64
.62
.61
.59
.58
.57
.56
.53
.50
.47
1.44
1.41
1.38

, 9
3'. 30
3.13
3.01
3.92
3.86
3.81
3.77
3.74
3.71
1.69
1.67
1.65
1.64
1.62
1.61
1.60
1.37
1.54
2.51
2.48
1.4S
1.41

























1
.37
.14
.00
.90
.82
.76
.71
.68
.69
.61
.60
.58
.97
.35
.94
.S3
.4f
.48
L42
1.39
.36
.33


3.90
1.61
3.41
3.2*
3. U
3.11
3.06
3.81
2.91
1.94
1.91
l.M
1.16
1.84
2.U
1.11
2.T7
1.72
1.6*
2.M
1.80
1.S6


























1.11
.88
.66
.51
.40
.31
.19
.19
.19
.11
.01
.09
.03
.01
L69
1.97
.92
.87
.82
I.7S
1.73
1.68


























1.43
.07
.83
.67
.55
.45
.38
.32
.17
.13
.20
.17
.14
.11
.10
.08
.03
.97
.92
.87
1.82
1.71
a •























.01

1.80
1.31
.96
.79
.66
.58
.48
.42
.37
.32
.39
.36
.13
.31
.18
.17
.11
.OS
.99
.94
.89
.84


























.73
.33
.07
.88
.75
.64
.58
.50
.44
.40
.36
.33
.30
.27
.35
.33
.17
.11
.09
.00
.94
.89


























.99
.43
.19
.96
.82
.71
.63
.56
.51
.46
.43
.39
.36
.33
.31
.39
.23
.16
.10
,04
.99
.93


























.94
.91
.13
.03
.89
.78
.69
.61
.96
.31
.47
.44
.41
.38
.16
.34
.37
.31
.14
.08
.03
.97


























.03
.59
.30
.09
.H
.83
.74
.87
.61
.56
.31
.48
.45
.41
.40
.38
.31
.34
.18
.11
.06
.00
 1From Miller (1981).
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.
                                       282

-------
2.4  If the number of replicates at each toxicant concentration and control
are not equal,  a t-test with Bonferroni's 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 452.68.  A complete listing of the program is contained in
EPA/600/4-91/021 (USEPA, 1992).  A compiled version of the program can be
obtained from EMSL-Cincinnati by sending a diskette with a written request.

2.6  Data Input and Output

2.6.1  Reproduction data from a daphnid, Cen'odaphm'a dubia,  survival and
reproduction test (Table C.6) 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
                                        283

-------
                TABLE C.6.  SAMPLE DATA FOR DUNNETT'S PROGRAM
                            CERIODAPHNIA DUBIA REPRODUCTION DATA

Replicate
1
2
3
4
5
6
7
8
9
10

Control
27
30
29
31
16
15
18
17
14
27
Effl
1.56
32
35
32
26
18
29
27
16
35
13
uent
Concentration (
3.12 6.25
39
30
33
33
36
33
33
27
38
44
27
34
36
34
31
27
33
31
33
31
%)
12.5
10
13
7
7
7
10
10
16
12
2
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 Figure C.I below.
                                       284

-------
                         MAIN MENU AND DATA INPUT
  1)  Great a data file
  2)  Edit a data file
  3)  Perform ANOVA on existing data

Your  choice ? 1
Number of groups, including control ? 5

Number of observations for group  1 ? 10

Enter the data for group  1 one observation at a time.

NO.  1?  27

NO.  2?  30

NO.  3?  29

NO.  4?  31

NO.  5?  16

NO.  6?  15

NO.  7?  18

NO.  8?  17

NO.  9?  14

NO. 10?  27

Number of observations for group  2  ?  10


Do you wish to save the data on disk  ?y

Disk file for output  ?  cerio
Figure C.I.  Sample Data Input for Dunnett's Program for Reproduction
             Data from Table C.6.

                                       285

-------
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 Figure C.2. below).
  1) Create a data file
  2) Edit a data file
  3) Perform ANOVA on existing data file
  4) Stop
You' choice  ?  3


File name  ?  cerio
Available Transformations
    1)  no transform
    2)  square root
    3)  loglO
    4)  arcsine square root

Your choice ? 1
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 Dunnett's test :  L=less than, G=greater than ? 1
Figure C.2.   Example of Choosing Option 3 from the Main Menu of
             the Dunnett Program.
                                       286

-------
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.
  TABLE C.7.   PROGRAM OUTPUT FOR THE DUNNETT'S PROGRAM USING THE
              REPRODUCTION DATA FROM TABLE C.6.
                          Summary Statistics  and ANOVA

                 Transformation =      None
   Group
Mean
s.d.
CV%
1 = control
2
3
4
5*
10
10
10
10
10
22.4000
26.3000
34.6000
31.7000
9.4000
6.9314
8.0007
4.8351
2.9458
3.8930
30.9
30.4
14.0
9.3
41.4
*) the mean for this group is significantly less than
   the control mean at alpha = 0.05 (1-sided) by Dunnett's test
Minimum detectable difference for Dunnett's test =        -5.628560
This difference corresponds to   -25.13 percent of control


Between groups sum of squares =     3887.880000 with  4 degrees of freedom

Error mean square =        31.853333 with 45 degrees of freedom.

Bartlett's test p-value for equality of variances =   .029
                                       287

-------
                                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 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 256 ng/l concentration 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.  FATHEAD MINNOW,  PIMEPHALES PROMELAS,  LARVAL GROWTH DATA
                    (WEIGHT IN MG)  USED FOR THE T-TEST WITH BONFERRONI'S
                    ADJUSTMENT
Replicate
Control
                                   NaPCP Concentration (ug/L)
32
64
128
256
A
B
C
D
Mean(Y,.)
Total (T,-)
0.711
0.662
0.718
0.767
0.714
2.858
0.646
0.626
0.723
0.700
0.674
2.695
0.669
0.669
0.694
0.676
0.677
2.708
0.629
0.680
0.513
0.672
0.624
2.494
0.650
0.558
(LOST)
0.508
0.572
1.716
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:
                                       288

-------
Total Sum of Squares:   SST - s  Y^-    G2/N
Between Sum of Squares: SSB  = s T^/n,-    G2/N
                              i
Within Sum of Squares:  SSW  = SST  -  SSB

   Where:  G = The grand  total of  all  sample  observations; G = s T,-


           N = The total  sample size;  N  = E  nf
                                           i

          n,- = The number of replicates  for concentration "i"

          T,- = The total of the replicate measurements for concentration "i"

         YJJ  =  The jth observation for concentration "i"


3.2  Calculations:
Total Sum of Squares:     SST  = s Y^-    G /N
                              =  8.268 -  (12.471)2/19
                              = 0.082
 Between  Sum  of  Squares:   SSB = S  1,/n, -  G /N
                                i
                              = 8.228 -  (12.471)2/19
                              = 0.042


 Within  Sum  of Squares:    SSW = SST   SSB

                              = 0.082   0.042

                              = 0.040


                                        289

-------
3.3  Prepare the ANOVA table as follows:
                    TABLE D.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
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
Between 5 1=4 0.042
Within 19 5 = 14 0.040
Mean Square
0.0105
0.0028
Total
  18
0.082
                                       290

-------
3.5  To perform the individual comparisons, calculate the t statistic for
each concentration and control combination, as follows:
                                 [Sw V(l/ni)  + (l/n,)  ]


   Where:  Y,-   =  Mean for each concentration

           Y1   =  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"
3.6  Table D.4 includes the calculated t values for each concentration and
control combination.
                      TABLE D.4.  CALCULATED T VALUES.
NaPCP
Concentration
(M9/L)
32
64
128
256

i

2
3
4
5

t-

1.067
0.987
2.402
3.507
                                        291

-------
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 (2.510), with an overall alpha level  of 0.05, fourteen
degrees of freedom and four concentrations excluding the control, was obtained
from Table D.5.   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 t5  is greater than  2.510,  the  256 ug/L
concentration has  significantly lower growth than the control.  Hence the NOEC
and LOEC for  growth are 128 ug/L and 256 ug/L, respectively.
                                      292

-------
TABLE D.5.  CRITICAL VALUES FOR "T" FOR THE T-TEST WITH BONFERRONI'S ADJUSTMENT
            P = 0.05 CRITICAL LEVEL, ONE TAILED
d.f.
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
26
27
28
29
30
31
K = 1
6.314
2.920
2.354
2.132
2.016
1.944
1.895
1.860
1.834
1.813
1.796
1.783
1.771
1.762
1.754
1.746
1.740
1.735
1.730
1.725
1.721
1.718
1.714
1.711
1.709
1.706
1.704
1.702
1.700
1.698
1.696
K = 2
12.707
4.303
3.183
2.777
2.571
2.447
2.365
2.307
2.263
2.229
2.301
2.179
2.161
2.145
2.132
2.120
2.110
2.101
2.094
2.086
2.080
2.074
2.069
2.064
2.060
2.056
2.052
2.049
2.046
2.043
2.040
K = 3
19.002
5.340
3.741
3.187
2.912
2.750
2.642
2.567
2.510
2.406
2.432
2.404
2.380
2.360
2.343
2.329
2.316
2.305
2.295
2.206
2.278
2.271
2.264
2.258
2.253
2.248
2.243
2.239
2.235
2.231
2.228
K = 4
25.452
6.206
4.177
3.496
3.164
2.969
2.842
2.752
2.686
2.634
2.594
2.561
2.533
2.510
2.490
2.473
2.459
2.446
2.434
2.424
2.414
2.406
2.398
2.391
2.385
2.379
2.374
2.369
2.364
2.360
2.356
K = 5
31.821
6.965
4.541
3.747
3.365
3.143
2.998
2.897
2.822
2.764
2.719
2.681
2.651
2.625
2.603
2.584
2.567
2.553
2.540
2.528
2.518
2.509
2.500
2.493
2.486
2.479
2.473
2.468
2.463
2.458
2.453
K = 6
38.189
7.649
4.857
3.961
3.535
3.288
3.128
3.016
2.934
2.871
2.821
2.730
2.746
2.718
2.694
2.674
2.655
2.640
2.626
2.613
2.602
2.592
2.583
2.574
2.566
2.559
2.553
2.547
2.541
2.536
2.531
K = 7
44.556
8.277
5.138
4.148
3.681
3.412
3.239
3.118
3.029
2.961
2.907
2.863
2.827
2.797
2.771
2.749
2.729
2.712
2.697
2.684
2.672
2.661
2.651
2.642
2.634
2.627
2.620
2.613
2.607
2.602
2.597
K = 8
50.924
8.861
5.392
4.315
3.811
3.522
3.336
3.206
3.111
3.039
2.981
2.935
2.897
2.864
2.837
2.814
2.793
2.775
2.759
2.745
2.732
2.721
2.710
2.701
2.692
2.684
2.677
2.670
2.664
2.658
2.652
K = 9
57.290
9.408
5.626
4.466
3.927
3.619
3.422
3.285
3.185
3.108
3.047
2.998
2.950
2.924
2.895
2.871
2.849
2.830
2.813
2.798
2.785
2.773
2.762
2.752
2.743
2.734
2.727
2.720
2.713
2.707
2.701
K = 10
63.657
9.925
5.841
4.605
4.033
3.708
3.500
3.356
3.250
3.170
3.106
3.055
3.013
2.977
2.947
2.921
2.899
2.879
2.861
2.846
2.832
2.819
2.808
2.797
2.788
2.779
2.771
2.764
2.757
2.750
2.745

-------
 TABLE D.5.  CRITICAL VALUES FOR  "T" FOR THE T-TEST WITH BONFERRONI'S ADJUSTMENT
             P = 0.05 CRITICAL LEVEL, ONE TAILED (CONTINUED)
d.f.
32
33
34
35
36
37
38
39
40
50
60
70
80
90
100
110
120
Infinite
K = 1
1.694
1.693
1.691
1.690
1.689
1.688
1.686
1.685
1.684
1.676
1.671
1.667
1.665
1.662
1.661
1.659
1.658
1.645
K = 2
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
1.984
1.982
1.980
1.960
K = 3
2.224
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
K = 4
2.352
2.349
2.346
2.342
2.340
2.337
2.334
2.332
2.329
2.311
2.300
2.291
2.285
2.280
2.276
2.273
2.270
2.242
K = 5
2.449
2.445
2.442
2.438
2.435
2.432
2.429
2.426
2.424
2.404
2.391
2.381
2.374
2.369
2.365
2.361
2.358
2.327
K = 6
2.527
2.523
2.519
2.515
2.512
2.508
2.505
2.502
2.499
2.478
2.463
2.453
2.446
2.440
2.435
2.432
2.429
2.394
K = 7
2.592
2.587
2.583
2.579
2.575
2.572
2.568
2.565
2.562
2.539
2.324
2.513
2.505
2.499
2.494
2.490
2.487
2.450
K = 8
2.647
2.643
2.638
2.634
2.630
2.626
2.623
2.619
2.616
2.592
2.576
2.564
2.556
2.549
2.544
2.540
2.536
2.498
K = 9
2.696
2.691
2.686
2.682
2.678
2.674
2.670
2.667
2.663
2.638
2.621
2.609
2.600
2.593
2.588
2.583
2.580
2.540
K = 10
2.739
2.734
2.729
2.724
2.720
2.716
2.712
2.708
2.705
2.678
2.661
2.648
2.639
2.632
2.626
2.622
2.618
2.576
d.f. =  Degrees of freedom for MSE  (Mean Square Error) from ANOVA.
  K  =  Number of concentrations to be compared to the control.

-------
                                  APPENDIX E

                        STEEL'S MANY-ONE RANK TEST

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
reproduction data from a Ceriodaphm'a dubia 7-day, chronic test.  The data are
listed in Table E.I.  Significant mortality was detected via Fisher's Exact
Test in the 50% effluent concentration.  The data for this concentration is
not included in the reproduction analysis.

4.   For each control and concentration combination,  combine the data and
arrange the observations 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.

5.   An example of assigning ranks to the combined data for the control and 3%
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.

6.   For this set of data, we wish to determine if the reproduction in any of
the effluent concentrations is significantly lower than the reproduction by
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 reproduction of each of
the various effluent concentrations with some "minimum" or critical rank
sum, at or below which the reproduction would be considered to  be
significantly lower than the control.  At a probability level of 0.05, the

                                      295

-------
critical  rank in a test with four concentrations and ten replicates is 76.
(See Table E.5,  for R=4).

7.  Comparing the rank sums in Table 2.3 to the appropriate critical rank,
the 6%, 12% and  25% effluent concentrations are found to be significantly
different from the control.  Thus the NOEC and LOEC for reproduction are 3%
and 6%, respectively.
    TABLE E.I.   EXAMPLE OF STEEL'S MANY-ONE RANK TEST:  DATA FOR THE
                DAPHNID, CERIODAPHNIA DUBIA,  7-DAY CHRONIC TEST
Effluent
Concentration 1
No.
Live
23456789 10 Adults
Control          20   26   26   23   24   27   26   23   27   24     10
  3%            13   15   14   13   23   26    0   25   26   27      9
  6%            18   22   13   13   23   22   20   22   23   22     10
 12%            14   22   20   23   20   23   25   24   25   21     10
 25%             90976   10   12   14    9   13      8
 50%             0000000000      0
                                     296

-------
TABLE E.2.  EXAMPLE OF STEEL'S MANY-ONE RANK TEST:  ASSIGNING
            RANKS TO THE CONTROL AND 3% EFFLUENT CONCENTRATION
  Rank      Number of Young         Control or % Effluent
             Produced
1
2.5
2.5
4
5
6
8
8
8
10.5
10.5
12
15
15
15
15
15
19
19
19
0
13
13
14
15
20
23
23
23
24
24
25
26
26
26
26
26
27
27
27
3%
3%
3%
3%
3%
Control
Control
Control
3%
Control
Control
3%
Control
Control
Control
3%
3%
Control
Control
3%
                                  297

-------
                         TABLE E.3.   TABLE OF RANKS
Replicate Control8
(Organism)
1 20
2 26
3 26
4 23
5 24
6 27
7 26
8 23
9 27
10 24
Effl
3
(6,4.5,3,11)
(15,17,17,17)
(15,17,17,17)
(8,11.5,8.5,12.5)
(10.5,14.5,12,14.5)
(19,19.5,19.5,19.5)
(15,17,17,17)
(8,11.5,8.5,12.5)
(19,19.5,19.5,19.5)
(10.5,14.5,12,14.5)
13
15
14
13
23
26
0
25
26
27
(2.5)
(5)
(4)
(2.5)
(8)
(15)
(1)
(12)
(15)
(19)
uent Concentration 1%)
6
18
22
13
13
23
22
20
22
23
22
12
(3)
(7.
(1.
(1.
(11
(7.
(4.
(7.
(11
(7.

5)
5)
5)
.5)
5)
5)
5)
.5)
5)
14
22
20
23
20
23
25
24
25
21
(1)
(6)
(3)
(8.
(3)
(8.
(14



5)

5)
.5)
(12)
(14
(5)
.5)

25
9
0
9
7
6
10
12
14
9
13
(5)
(1)
(5)
(3)
(2)
(7)
(8)
(10)
(5)
(9)
Control  ranks  are given  in the order of the concentration with which they
 were ranked.
                         TABLE  E.4.   RANK  SUMS
                  Effluent                       Rank  Sum
               Concentration
                    (*)
                  3                                84
                  6                                63.5
                 12                                76
                 25                                55
                                     298

-------
TABLE E.5.  SIGNIFICANT VALUES OF RANK SUMS: JOINT CONFIDENCE
            COEFFICIENTS OF 0.95 (UPPER) and 0.99 (LOWER) FOR
            ONE-SIDED ALTERNATIVES
n
4
5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

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
k = 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
4
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
treatments
5
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
272
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
--
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
9
--
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
   From Steel  (1959).
                                     299

-------
                               APPENDIX F

                            WILCOXON RANK SUM TEST


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.  An example of the use of the Wilcoxon Rank Sum Test is provided below.
The data used in the example are the same as in Appendix E, except that two
males are presumed to have occurred, one in the control and one in the 25%
effluent concentration.  Thus, there is unequal replication for the
reproduction analysis.

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
3% effluent concentration 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 reproduction in any of
the effluent concentrations is significantly lower than the reproduction by
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 reproduction of each of
the various effluent concentrations with some "minimum" or critical rank sum,
at or below which the reproduction 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 nine replicates in the control is 72 for
those concentrations with ten replicates, and 60 for those concentrations with
nine replicates (See Table F.5, for K = 4).

6.  Comparing the rank sums in Table F.4 to the appropriate critical rank, the
6%, 12% and 25% effluent concentrations are found to be significantly
different from the control.  Thus, the NOEC and LOEC for reproduction are
3% and 6%, respectively.
                                     300

-------
 TABLE F.I.  EXAMPLE OF WILCOXON'S RANK SUM TEST:  DATA FOR THE
             DAPHNID, CERIODAPHNIA DUBIA, 7-DAY CHRONIC TEST

Effluent
Concentration
Cont
3%
6%
12%
25%
50%










Reolicate
1
M
13
18
14
9
0
2
26
15
22
22
0
0
3
26
14
13
20
9
0
4
23
13
13
23
7
0
5
24
23
23
M
6
0
6
27
26
22
23
10
0
7
26
0
20
25
12
0
8
23
25
22
24
14
0
9
27
26
23
25
9
0
10
24
27
22
21
13
0
No.
Live
Adults
10
9
10
10
8
0
TABLE F.2.  EXAMPLE OF WILCOXON'S RANK SUM TEST:  ASSIGNING
            RANKS TO THE CONTROL AND EFFLUENT CONCENTRATIONS
 Rank      Number of Young
            Produced
Control  or % Effluent
1
2.5
2.5
4
5
7
7
7
9.5
9.5
11
14
14
14
14
14
18
18
18
0
13
13
14
15
23
23
23
24
24
25
26
26
26
26
26
27
27
27
3%
3%
3%
3%
3%
Control
Control
3%
Control
Control
3%
Control
Control
Control
3%
3%
Control
Control
3%
                                301

-------
                        TABLE F.3.   TABLE OF RANKS
Replicate Control8
(Organism)
1 M
2 26
3 26
4 23
5 24
6 27
7 26
8 23
9 27
10 24
Effluent Concentration (%)
3

(14,16,15,16)
(14,16,15,16)
(7,10.5,6.5,11.5)
(9.5,13.5,10,13.5)
(18,18.5,17.5,18.5)
(14,16,15,16)
(7,10.5,6.5,11.5)
(18,18.5,17.5,18.5)
(9.5,13.5,10,13.5)
13
15
14
13
23
26
0
25
26
27
(2.5)
(5)
(4)
(2.5)
(7)
(14)
(1)
(11)
(14)
(18)
6
18
22
13
13
23
22
20
22
23
22
(3)
(6.
(1.
(1.
(10
(6.
(4)
(6.
(10
(6.

5)
5)
5)
.5)
5)

5)
.5)
5)
12
14
22
20
23
M
23
25
24
25
21
(1)
(4)
(2)
(6.

(6.
(12



5)

5)
.5)
(10)
(12
(3)
.5)

25
9
0
9
7
6
10
12
14
9
13
(5)
(1)
(5)
(3)
(2)
(7)
(8)
(10)
(5)
(9)
aControl  ranks  are  given  in  the  order  of  the  concentration  with  which  they
 were ranked.
                           TABLE F.4.   RANK SUMS
              Effluent        Rank Sum      No.  of       Critical
           Concentration                    Replicates   Rank Sum
               3                 79           10            72
               6                 57           10            72
              12                 58            9            60
              25                 55           10            72
                                     302

-------
TABLE F.5.  CRITICAL VALUES FOR WILCOXON'S RANK SUM TEST WITH
            BONFERRONI'S ADJUSTMENT OF ERROR RATE FOR COMPARISON
            OF "K" TREATMENTS VS. A CONTROL FIVE PERCENT CRITICAL
            LEVEL (ONE-SIDED ALTERNATIVE: TREATMENT CONTROL)
K No. Reolicates No. of Reolicates Per Effluent
in Control

1 3
4
5
6
7
8
9
10
2 3
4
5
6
7
8
9
10
3 3
4
5
6
7
8
9
10

3
6
6
7
8
8
9
10
10
..
--
6
7
7
8
8
9


--
6
7
7
7
8

4
10
11
12
13
14
15
16
17

10
11
12
13
14
14
15

10
11
11
12
13
13
14

5
16
17
19
20
21
23
24
26
15
16
17
18
20
21
22
23

16
17
18
19
20
21
22

6
23
24
26
28
29
31
33
35
22
23
24
26
27
29
31
32
21
22
24
25
26
28
29
31

7
30
32
34
36
39
41
43
45
29
31
33
34
36
38
40
42
29
30
32
33
35
37
39
41

8
39
41
44
46
49
51
54
56
38
40
42
44
46
49
51
53
37
39
41
43
45
47
49
51
Concentration

9
49
51
54
57
60
63
66
69
47
49
52
55
57
60
62
65
46
48
51
53
56
58
61
63

10
59
62
66
69
72
72
79
82
58
60
63
66
69
72
75
78
57
59
62
65
68
70
73
76
                                  303

-------
TABLE F.5.  CRITICAL VALUES FOR WILCOXON'S RANK SUM TEST WITH BONFERRONI'S
            ADJUSTMENT OF ERROR RATE FOR COMPARISON OF "K" TREATMENTS VS.
            A CONTROL FIVE PERCENT CRITICAL LEVEL (ONE-SIDED ALTERNATIVE:
            TREATMENT CONTROL) (CONTINUED)
K Nn. Rppli rates No. of Reolicates Per Effl
in Control

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
7 3
4
5
6
7
8
9
10

3
..
--
--
6
6
7
7
7
..
--
--
--
6
6
7
7
..
--
--
--
6
6
6
7
..


--
--
6
6
7

4
_ _
--
10
11
12
12
13
14
..
--
10
11
11
12
13
13
..
--
10
11
11
12
12
13
..

--
10
11
11
12
13

5
_ _
15
16
17
18
19
20
21
..
15
16
17
18
19
20
21
..
15
16
16
17
18
19
20
..
--
15
16
17
18
19
20

6
21
22
23
24
26
27
28
30
..
22
23
24
25
27
28
29
..
21
22
24
25
26
27
29
..
21
22
23
25
26
27
28

7
28
30
31
33
34
36
38
40
28
29
31
32
34
35
37
39
28
29
30
32
33
35
37
38
..
29
30
32
33
35
36
38
uent

8
37
38
40
42
44
46
48
50
36
38
40
42
43
45
47
49
36
38
39
41
43
45
47
49
36
37
39
41
43
44
46
48
Concentration

9
46
48
50
52
55
57
60
62
46
48
50
52
54
56
59
61
45
47
49
51
54
56
58
60
45
47
49
51
53
55
58
60

10
56
59
61
64
67
69
72
75
56
58
61
63
66
68
71
74
56
58
60
63
65
68
70
73
56
58
60
62
65
67
70
72
                                     304

-------
TABLE F.5.  CRITICAL VALUES FOR WILCOXON'S RANK SUM TEST WITH BONFERRONI'S
            ADJUSTMENT OF ERROR RATE FOR COMPARISON OF "K" TREATMENTS VS.
            A CONTROL FIVE PERCENT CRITICAL LEVEL (ONE-SIDED ALTERNATIVE:
            TREATMENT CONTROL) (CONTINUED)
K No. Replicates No. of Reolicate Per
in Control

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

3
..
--
--
--
--
6
6
6

--
--
--
--
--
6
6

--

--
--

6
6

4

--

10
11
11
12
12

--

10
10
11
11
12

--
--
10
10
11
11
12

5

--
15
16
17
18
19
19

--
15
16
17
18
18
19
..
--
15
16
16
17
18
19

6

21
22
23
24
25
27
28

21
22
23
24
25
26
28
..
21
22
23
24
25
26
27
Effluent Concentration

7

29
30
31
33
34
36
37

28
30
31
33
34
35
37
..
28
29
31
32
34
35
37

8
36
37
39
40
42
44
46
48

37
39
40
42
44
46
47
..
37
38
40
42
43
45
47

9
45
47
49
51
53
55
57
59
45
46
48
50
52
55
57
59
45
46
48
50
52
54
56
58

10
55
57
59
62
64
67
69
72
55
57
59
62
64
66
69
71
55
57
59
61
64
66
68
71
                                      305

-------
                                  APPENDIX 6

                            FISHER'S EXACT TEST

1.  Fisher's Exact Test (Finney,  1948; Pearson and Hartley, 1962) is a
statistical  method based on the hypergeometric probability distribution that
can be used to test if the proportion of successes is the same in two
Bernoulli (binomial) populations.   When used with the Cen'odaphnia dubia data,
it provides a conservative test of the equality of any two survival
proportions assuming only the independence of responses from a Bernoulli
population.   Additionally, since it is a conservative test, a pairwise
comparison error rate of 0.05 is suggested rather that an experimentwise error
rate.

2.  The basis for Fisher's Exact Test is a 2x2 contingency table.  From the
2x2 table, set up for the control  and the concentration you wish to compare,
you can determine statistical significance by looking up a value in the table
provided (Table G.5).  However, in order to use this table the contingency
table must be arranged in the following format:

                     TABLE 6.1.  FORMAT FOR CONTINGENCY TABLE

Row 1
Row 2
Number
Successes
a
b
of
Failures
A - a
B - b
Number of
Observations
A
B
             Total             a + b    [(A + B)  - a - b]      A + B


3.  Arrange the table so that the total  number of observations for row one is
greater than or equal to the total  for row two (A > B).   Categorize a success
such that the proportion of successes for row one is greater than or equal to
the proportion of successes for row two (a/A > b/B).  For the Cen'odaphnia
dubia survival data, a success may be 'alive' or 'dead',  whichever causes a/A
> b/B.  The test is then conducted by looking up a value  in the table of
significance levels of b and comparing it to the b value  given in the
contingency table.   The table of significance levels of b is Table G.5.  Enter
Table G.5 in the section for A, subsection for B, and the line for a.  If the
b value of the contingency table is equal to or less than the integer in the
column headed 0.05 in Table G.5, then the survival proportion for the effluent
concentration is significantly different from the survival proportion for the
control.  A dash or absence of entry in Table G.5 indicates that no
contingency table in that class is significant.
                                     306

-------
4.  To illustrate Fisher's Exact Test, a set of survival data (Table G.2) from
the daphnid, Ceriodaphnia dubia, survival and reproduction test will be used.

5.  For each control and effluent concentration construct a 2x2 contingency
table.

6.  For the control and effluent concentration of 1% the appropriate
contingency table for the test  is given in Table G.3.

                 TABLE G.2.  EXAMPLE OF FISHER'S EXACT TEST:
                             CERIODAPHNIA DUBIA MORTALITY DATA
Effluent
Concentration (%)
Control
1
3
6
12
25
No. Dead
1
0
0
0
0
10
Total1
9
10
10
10
10
10
1Total  number of live adults at the beginning of the test.
         TABLE G.3.  2X2 CONTINGENCY TABLE FOR CONTROL AND 1% EFFLUENT



1% Effluent
Control
Number of

Alive Dead
10 0
8 1

Number of
Observations
10
9
           Total              18             1              19
                                      307

-------
7.  Since 10/10 > 8/9, the category 'alive' is regarded as a success.
For A = 10,  B = 9 and, a = 10, under the column headed 0.05, the value from
Table G.5 is b = 5.  Since the value of b (b = 8) from the contingency table
(Table G.3), is greater than the value of b (b = 5) from Table G.5, the test
concludes that the proportion of survival is not significantly different for
the control  and 1% effluent.

8.  The contingency tables for the combinations of control and effluent
concentrations of 3%, 6%, 12% are identical to Table G.3.  The conclusion of
no significant difference in the proportion of survival for the control and
the level of effluent would also remain the same.

9.  For the  combination of control and 25% effluent, the contingency table
would be constructed as Table G.4.  The category 'dead' is regarded as a
success, since 10/10 > 1/9.  The b value (b = 1) from the contingency table
(Table G.4)  is less than the b value (b = 5) from the table of significance
levels of b  (Table G.5).  Thus, the percent mortality for 25% effluent is
significantly greater than the percent mortality for the control.  Thus, the
NOEC and LOEC for survival are 12% and 25%, respectively.
          Table G.4.   2X2 CONTINGENCY TABLE FOR CONTROL AND 25% EFFLUENT
                                 Number of	
                                                         Number of
                            Alive	Dead	Observations
            25% Effluent        10            0              10

            Control               1            8               9


             Total               11            8              19
                                     308

-------
TABLE G.5.  SIGNIFICANT LEVELS OF B: VALUES OF B (LARGE TYPE)
            AND CORRESPONDING PROBABILITIES (SMALL TYPE)1


A«3 B=3


A-4 B=4
3


A«5 B = 5

4

3

2



A«6 B=6


5


4

3

2


A-7 B«7




6




5


4


3

2


3


4
4


J
4
S
4
J

J



6
5
4
6
5
4
6
J
£
J
6


7
6
J
4

7
6
5
4

7
6
5
7
«
5
7
<
7
Probability
(KM
0 4»


0 -014
0 -as


I 414
0 -014
1 -041
0 -040
0 -ell

0 44i



2 430
1 -040
0 -OK
1 413+
0 -Oil
0 443*
1 -on
0 -024
0 -012
0 041
0 4M


3 431-
I 4iJ-
0 410+
0 43J-

2 -at
1 -OU+
0 4U
0 44»

2 -Hi*
1 44J»
0 4Z»
1 «M
0 •«»+
0 44J+
0 -an
0 4Ji
0 -OS»
CKJ2S
_


0 414
	


1 414
0 414
0 401
	
0 411

_



1 •««
0 •cot
	
0 413+
0 -oil
_
0 403-
0 -024
0 411

—


2 4ie+
1 413-
0 4io+
	

2 411
0 -OW
0 4I«
_

J 410+
0 4M
—
1 
-------
TABLE G.5.
SIGNIFICANT LEVELS OF B: VALUES OF B (LARGE TYPE)
AND CORRESPONDING PROBABILITIES (SMALL TYPE)1
(CONTINUED)

A = 9 B = S



t



;


2

A-10 B-10






9





8





7





6




5




a
9
g
7
6
9
8
7
6
9
8
7
9

10
9
g
7
6
5
4
10
9
g
7
6
5
10
9
g
7
6
5
10
9
8
7
6
5
10
9
S
7
6
10
9
S
7
<
Probability
(MM
2 42?
1 -023
0 -oio''
0 -021
1 -014
0 -on
0 421
0 449
1 443*
0 -on
0 443*
0 -on

6 -041
4 4»
3 433-
2 4J3-
1 -029
0 -416
0 -04}
5 -93)
4 430-
2 -019
1 415-
1 -040
0 -012
4 -021
3 -012
2 -on
1 423
0 -on
0 -029
3 413-
2 -on
1 -on
1 -OM
0 -017
0 -041
3 -OM
2 -OM
1 -04
0 4io*
0 -02*
2 -022
1 -017
1 -047
0 419
0 -042
0-OZ3
1 403-
1 -021
0 -010+
—
1 -014
0 -an
0 421
—
0 403-
0 411
—
0 -on

S -OK
3 410-
2 -012
1 410-
0 -003*
0 -014
—
4 -on
3 -017
2 419
1 415-
0 -ooi
0 -022
4 -021
2 409
1 -00.
1 423
0 -on
—
3 413-
2 -on
1 -Oil
0 -OM
0 -017
	
2 -OM
1 -on
1 014
0 -oio*
_
2 -022
1 -017
0 -007
0 4i»
^
0-01
1 -003*
0 -003
	
—
0 -ooi
0 -007
—
—
0 403-
—
—
	

4 403*
3 4io-
1 -001
1 410-
0 -003*
—
—
3 -003
2 -ooj-
1 -004
a -002
0 -cot
—
3 -007
2 -009
1 -OOI
0 -004
	
—
2 -001
1 404
0 -002
0 -OM
	
_
2 -ooi
1 -001
0 -001
_
_
1 -004
0 -m
0 407
—
-~
0-005
1 403-
0 -003
—
—
0 -001
—
—
—
0 403-
— -
—
	 .

3 -002
2 -003
1 -003
0 -002
	
—
—
3 -ooi
2 403-
1 404
0 402
	
—
2 402
1 402
0 401
0 404
	
—
2 403
1 404
0 402
	
__
__
1 401
0 401
0 40}
	
_
1 404
0 402
—
—
—

A=IO B = 4



3


2



A-ll B=ll






10






9






S






7





<


a
10
9
g
7
10
9
g
10
9


11
10
9
g
7
«
S
4
11
10
9
g
7
6
5
11
10
9
8
7
6
5
11
10
9
g
7
6
5
11
10
9
S
7
6
11
10
9
Probability
(MS
1 411
1 441
0 413-
0 4)3-
1 431
0 414
0 433-
0 413*
0 443-r


7 443*
5 432
4 440
3 443
2 440
1 432
0 411
0 443*
6 433-
4 421
3 424
2 423
1 417
1 443
0 423
S 42<
4 411
3 440
2 433-
1 423-
0 412
0 430
4 411
3 424
2 422
1 413-
1 437
0 417
0 440
4 443
3 447
2 43»
1 413-
0 410*
0 423-
3 43
24M
1 411
0-025
1 411
0 403-
0 413-
	
0 403
0 414
—
0 413*
	


6 411
4 412
3 413-
2 413-
1 412
0 4M
0 4I(
—
S 411
4 421
3 424
2 423
1 417
0 409
0 423
4 4M
3 412
2 412
1 409
1 413-
0 412
	
4 411
3 424
2 422'
1 413-
0 407
0 417
	
3 411
2 413
1 409
1 413-
0 410*
0 423-
2 4M
1 403*
1 411
(H>1
0 401
0 403-
	
	
0 403
• —
	
	
	


5 4M
3 404
2 404
1 404
0 402
0 40*
—
—
4 404
3 407
2 407
1 40*
0 403
0 409
	
4 401
2 403
1 403
1 409
0 404
	
—
3 403-
2 4M
1 403-
0 402
0 407
	
	
2 402
1 002
1 409
0 404
—
	
2 4M
1 403*
0 401
(K»5
0 401
0 403-
. 	
—
0 403
—
	
	
	


4 402
3 404
2 404
I 404
0 401
—
—
—
4 404
2 402
1 402
0 401
0 403
—
—
3 402
2 40)
1 403
0 401
0 404
	
—
3 403-
1 401
1 403-
0 40]
	
	
—
2 401
1 401
0 401
0 404
—
	
1 401
0 401
0 401
                              310

-------
TABLE G.5.  SIGNIFICANT LEVELS OF B: VALUES OF B (LARGE TYPE)
            AND CORRESPONDING PROBABILITIES (SMALL TYPE)1
            (CONTINUED)

A=ll B = 6


S




4



3



2


A = 12 B = 12







11







10







9




a
8
7
6
11
10
9
8
7
11
10
9
8
11
10
9

11
10

12
11
10
9

8
7
6
5
4
12
11
10
9
g
7
6
5
12
11
10
9
g
7
j
5
12
11
10
9
1
Probability
5
1 443
0 417
0 437
2 411
1 413 -
1 4M
0 411
0 429
1 409
1 411
0 411
0 42*
1 433
0 411
0 427

0 413
0 4U

8 447
6 4«
$ 443-
4 430-

3 430-
2 443-
1 414
0 419
0 447
7 417
5 434
4 429
3 4»
2 43*
1 411
1 443-
0 414
6 419
5 443
4 441
3 444
2 431
1 42*
0 411
0 4M
5 411
4 49
3 42»
2 414
1 414
0-025
0 407
0 417
—
2 411
1 413
0 403-
0 413
	
1 409
0 404
0 411
	
0 403
0 411
__

0 413
_ _

7 41»
J 414
4 411
3 420

2 419
I 414
0 407
0 419

6 414
5 434
3 410*
2 409
1 407
1 4I»
0 409
0 434
5 410-
4 •015'*'
3 41?
2 413-
1 410*
0 403-
0 411
—
S 421
J 409
2 40C
2 414
1 414
0-01
0 407
—
—
1 401
0 401
0 405-
	
	
1 409
0 404
	
—
0 403
__
__

—
^_

6 407
4 403-
3 -OM
2 404

1 403-
0 402
0 407
	
	
5 403-
4 40t
2 40J
2 409
1 407
0 401
0 409
	
5 410-
3 403-
2 403-
1 404
0 401
0 403-
	
—
4 404
3 409
2 40.
1 4S4
0 402
0-005
_
—
—
1 403
0 401
0 403-
—
—
0 401
0 404
—
	
0 401
	
	

	
__

5 402
4 403-
2 403
1 401

1 403-
0 402
	
	
__
5 403-
3 402
2 403
1 401
0 401
0 401
	
	
4 401
3 403-
2 403-
1 404
0 401
0 403-
—
—
3 401
2 402
1 403
0 401
0 401

A=12 B = 9


8






7







6




5





4




3



2



A=13 B-13





a
7
6
5
12
11
10
9
8
7
6
12
11
10
9
8
7
6

12
1 1
10
9
8
7
6

12
11
10
9
8
7

12
11
10
9
8
12
11
10
9
12
11


13
12
11
10
9
1
Probability
0-05
1 437
0 417
0 439
5 449
3 411
2 415*
2 440
1 425-
0 410-
0 424
4 -OM
3 431
2 429
1 417
1 440
0 416
0 4J4

3 433-
2-.^
422
1 411
1 432
0 411
0 425-
0 430-

2 413-
1 410-
1 42t
0 409
0 420
0 441

2 430
1 427
0 401
0 419
0 43<
1 429
0 409
0 422
0 444
0 411
0 411


9 441
7 4J7
6 448
4 434
3 414
2 411
0-025
0 407
0 417
—
4 414
3 411
2 413-
1 410-
1 ou-
0 410-
0 424
3 409
2 410-
1 40t
1 417
0 407
0 -OH
	

3 425-
2_,_
422
1 413
0 403-
0 411
0 413-


2 415-
1 410-
0 403
0 409
0 420

1 407
0 403
0 40(
0 419
	
0 402
0 409
0 412
—
0 411
—


8 4»
6 413*
S 411
4 414
3 414
2 411
0-01
0 -007
	
—
3 -004
2 -004
1 OOJ
1 410-
0 404
—
—
3 409
2 -oio-
1 40«
0 oo:
0 -007
	
	

2 403-
If^LJ
404
0 402
0 403-
—


1 -002
1 410-
0 003
0 409
—

1 407
0 
-------
TABLE G.5.  SIGNIFICANT LEVELS OF B: VALUES OF B (LARGE TYPE)
            AND CORRESPONDING PROBABILITIES (SMALL TYPE)1
            (CONTINUED)


A = 13 B=I3



12








11








10








9






t







7


a
7
<
5
4
13
12
11
10
9
8
7
6
5
13
12
1!
10
9
8
7
6
5
13
12
11
10
9
8
7
6
5
13
12
11
10
9
8
7
6
S
13
12
11
10
9
t
7
6
13
12
Probability
OO5
2 -041
1 -on
0 420
0 -041
8 -0)9
6 -on
5 433
4 -CM
3 -DM
2 -029
1 -030
1 4«
0 -014
7 -031
6 441
4 -021
3 -021
3 430-
2 -040
1 -027
0 -on
0 -030
6 -cat
5 433-
4 -017
3 433
2 -026
1 -017
1 -031
0 -017
0 431
5 -017
4 -023
3 -023
2 -017
2 040
1 423-
0*410*
0 -023
0 -041
5 -042
4 -047
3 -041
2 42»
1 -017
1 -037
0 4i3-
0 -032
4 -031
3 -on
0-025
1 -013+
0 -007
0 -020
	
7 413-
5 410-
4 -013
3 -013
2 -on
1 401
1 -020
0 4io-
0 -024
6 -on
3 -on
4 -021
3 -oil
2 -017
1 -on
0 403-
0 -013
—
6 -024
4 -012
3 -012
2 -010+
1 406
1 -017
0 -007
0 -017

5 -017
4 -013
3 413
2 -017
1 -010*
1 423-
0 -410+
0 -023
—
4 -012
3 -014
2 -on
1 -007
1 417
0 -oo*
0 413-
—
3 407
2 407
(H31
0 -003
0 -007
	
	
6 405*
5 410-
3 404
2 404
1 -003
1 -OM
0 -004
0 410-
_
5 403
4 40*
3 407
2 40*
1 404
0 -002
0 403-
—
—
5 407
3 -003
2 403'
1 402
1 -006
0 403
0 407

—
4 403-
3 407
2 40*
I 404
0 401
0 404
—
—
3 403
2 403
1 402
1 407
0 401
0 40.
—
	
3 407
2 407
0-005
0 403
—
—
—
5 402
4 403
3 404
2 404
1 403
0 401
0 404
—
—
5 403
3 402
2 402
1 401
1 404
0 402
0 403-
—
—
4 402
3 403
2 403
1 402
0 40!
0 403
^_

—
4 403-
2 401
1 401
1 404
0 401
0 404
—
—
3 403
2 403
1 402
0 401
0 401
—
—
—
2 401
1 401


A=13 B-7





6






i





4




3




2

A = 14 B=I4









13






J

11
10
9
8
7
6
13
12
11
10
9
»
7
13
12
11
10
9
8
13
12
11
10
9
13
12
11
10

13
12
14
13
12
11
10
9
8
7
6
5
4
14
13
J2
11
10
9
8
Probability
0-05
2 423
1 412
1 42»
0 410*
0 412
0 444
3 411
2 417
2 44*
1 414
1 430-
0 417
0 434
2 411
2 444
1 412
1 447
0 413-
0 49
2 444
1 412
0 40*
0 413-
0 4»
1 423
0 407
0 411
0 434

0 410-
0 49
10 449
8 43<
6 41)
5 427
4 421
3 427
2 413
1 -01*
1 4M
0 410
0 44»
9 441
7 49
6 417
5 441
4 441
3 4M
2 411
0025
2 4ZZ
1 411
0 404
0 410*
0 411
—
3 411
2 417
1 419-
1 424
0 401
0 417
—
2 412
1 4H
1 411
0 407
0 413-
	
I 40*
1 421
0 40*
0 413-
	
1 -era
0 407
0 411


0 410-
	
9 420
7 41*
6 41]
4 411
3 411
2 401
2 411
1 41*
0 401
0 430

S 41*
6 411
5 413+
4 417
3 4I<
2 411
1 40)
041
1 404
0 401
0 404
—
—
—
2 404
1 403
1 410-
0 403
0 401
—
—
1 401
1 401
0 403
0 407
	
	
1 40*
0 403
0 40*
	
	
0 401
0 407
__


0 410-
	
8 401
« 40*
i 409
3 404
2 40)
2 40»
1 -«>«
0 403
0 40t
	
	
^ 40*
5 404
4 403+
3 40*
2 4M-
1 4U
1 40.
0-005
1 404
0 401
0 404
	
—
	
2 404
1 401
0 401
0 403
	
	
—
1 4O2
0 401
0 401
	
	
—
0 400
0 402
	
	
	
0 403

—
	

	
~~
7 40)
5 401
4 401
3 404
2 401
1 4m
0 401
0 40)
	
	
	
6 402
5 404
3 402
2 401
2 40J-
1 40)
0 491
                              312

-------
TABLE G.5.  SIGNIFICANT LEVELS OF B: VALUES OF B (LARGE TYPE)
            AND CORRESPONDING PROBABILITIES (SMALL TYPE)1
            (CONTINUED)


A = 14 B=13


12









11









10









9








»








a

7
6
5
14
13
12
11
10
9
8
7
6
5
14
13
12
11
10
9
8
7
6
3
14
13
12
11
10
9
g
7
6
5
14
13
12
11
10
9
8
7
6
14
13
12
11
10
9
1
7
6
Probability
0-03
1 411
1 -Ott
0 423-
8 -031
6 421
5 w
4 4«
3 4=4
2 419
2 442
1 -021
0 -on
0 -030
7 42*
6 -on
5 -oo
4 -042
3 -OM
2 -017
1 -017
1 411
0 OI7
0 -oil
6 420
5 -on
4 -021
3 -024
2 -on
2 -040
1 -014
0 4io-
0 -023
0 -047
6 -047
4 oil
3 •on
3 -W2
2 -on
1,417
I'-OM
0 414
0 -OM
5 -DM
4 4J»
3 4w
2 -022
2 *u
1 -OK
0 «o»
0 -029
0 -04*
0-025
1 -021
0 4io*
0 423-
7 -012
6 -021
4 -on
J 4w
3 -024
2 -01?
1 -01:
0 403*
0 -on
—
6 -on
5 -014
4 -01*
3 413-
2 on
1 -007
1 -017
0 407
0 -017
—
6 -020
4 -on
3 -tat
3 -024
2 -on
1 -Oil
1 -024
0 4io-
0 -022
	
5 414
4 -oil
3 -017
2 -012
1 407
1 417
0 40(
0 414
—
4 410-
3 411
2 40*
2 4U
1 412
0 -004
0 40?
0 42*
—
0-01
0 404
—
—
6 404
5 -007
4 -on
3 409
2 407
1 405-
0 402
0 -005*
—
—
6 409
4 404
3 403-
2 404
1 -00}
I -007
0 40)
0 407
—
—
5 4W
4 409
3 409
2 407
1 404
0 402
0 404
0 410-

	
4 404
3 403-
2 404
1 402
1 407
0 402
0 4M
	
	
4 410-
2 402
2 401
1 403-
0 402
0 404
0 40»
_
—
0-005
0 404
—
—
6 -004
4 oo:
3 40!
2 -002
1 402
1 403-
0 402
	
—
—
5 401
4 404
3 003-
2 404
1 40}
0 401
0 40}
	
	
—
4 402
3 -002
2 -oo:
1 -001
1 404
0 40:
0 404
	
__
	
4 404
3 403-
2 404
1 002
0 ooi
0 002
—
—
—
3 402
2 402
1 401
1 403-
0 402
0 404
	
—
—


A=14 B = 7







6







3



1


4






3



2




A=13 B=15











Q

14
13
12
11
10
9
8
7
14
13
12
11
10
9
8
7
14
13
12
11
10
9
8
14
13
12
11
10
9

14
13
12
11
14
13
12


13
14
13
12
11
10
9
8
7
6
5
4
Probability
0-05
4 426
3 423
1 -017
2 441
1 421
1 443
0 413-
0 4}0
*3 411
2 414
2 437
1 411
1 431
0 412
0 -024
0 444
2 4IO-1-
2 437
1 417
1 431
0 411
0 422
0 440
2 43»
1 41*
1 444
0 411
0 423
0 441

1 42
0 406
0 413-
0 42»
0 401
0 423
0 430


11 430-
9 440
7 423*
6 430
5 439
4 413
3 4M
2 423*
1 411
1 440
0 421
0 430-
0-025
3 -CM
2 4«
2 417
1 409
1 421
0 407
0 41!-
—
3 411
2 414
1 407
1 411
0 405-
0 412
0 42<
—
2 410-
1 -OM
I 417
0 40!-
0 411
0 422
	
1 403-
1 419
0 403-
0 411
0 423


I 4=2
0 406
0 413-
	
0 o°»
0 423
	


10 421
8 411
6 410*
5 413
4 411
3 411
2 410*
1 407
1 411
0 401
0 421
^™
0-01
3 -oo«
2 4«
1 403
1 009
0 401
0 407
—
	
2 40]
1 402
1 -007
0 402
0 403*
	
	
—
1 401
1 406
0 402
0 403-
	
_
	
1 403-
0 402
0 403-

	


0 401
0 OM
	
	
0 40*
	
—


9 401
7 407
5 404
4 403-
3 403-
2 404
I 401
1 407
0 003
0 40<
	
^~
0-005
2 401
1 401
1 403
0 401
0 403
—
—
—
1 403
1 402
0 401
0 403
	
	
	
	
1 401
0 401
0 402
0 403-
	
_
	
1 405-
0 402
0 403-

	
	

0 401
—
	
—
—
—
—


8 401
6 401
3 404
4 403-
3 403-
2 404
1 401
0 401
0 401
—
	
~—
                               313

-------
TABLE G.5.  SIGNIFICANT LEVELS OF B: VALUES OF B (LARGE TYPE)
            AND CORRESPONDING PROBABILITIES (SMALL TYPE)1
            (CONTINUED)

A»15 B«I4










13










12










11










10









9

1
a
15
14
13
12
11
10
9
g
7
6
5
\i
14
13
12
11
10
9
t
7
6
S
15
14
13
12
11
10
9
8
7
6
5
15
14
13
12
11
10
9
8
7
6
3
15
14
13
12
11
10
9
t
7
6
15
14
Fro lability
0-05
10 44i
8 -on
7 -041
6 44.
5 44J
4 +tt
3 441
2 4))
1 -(m
1 -049
0 413*
9 433-
7 -013
6 429
5 -on
4 -030
3 -OK
2 433
2 -04J
1 429
0 -013
0 -031
8 -on
7 -04)
6 449
5 -04»
4 443*
3 4)1
2 -021
1 -Oil
1 -031
0 -017
0 -037
7 -012
6 -031
1 4)4
4 -032
3 -021
2 ait
a 4«
1 -024
1 449
0 -022
0 44«
6 -017
5 40)
4 -oxi
3 -OK
3 -042
2 -
-------
                                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-Milk'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.
                                          where S2 > S2
5.  Compare F with the 0.005 level of a tabled F value with n1  -  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 Ceriodaphm'a dubia reproduction data from an effluent  screening
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.  CERIODAPHNIA DUBIA REPRODUCTION DATA
                              FROM AN EFFLUENT SCREENING TEST

Control
100% Effluent

1
36
23

2
38
14

3
35
21

4
35
7
Repl
5
28
12
icate
6
41
17

7
37
23

8
33
8

9 10 X
35.4
18 . 15.9
S2
14.
36.

5
6
                                      315

-------
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.

                           F = 36.61
                               14.55

                             = 2.52
8.  There are 9 replicates for the effluent concentration and 8
replicates for the control.  Thus, the numerator degrees of freedom is 8
and the denominator degrees of freedom is 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 8.68.   Since 2.52 is not greater than 8.68,  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  =

                                 S
                                         "1   n2
   Where:        Y,    =   Mean  for  the  control

                Y2    =   Mean  for  the  effluent  concentration
                          (n,  -  1)  S?   +  (n2  - 1) S2

                              ^    +   n^  -  2
                S   =   Estimate  of  the  variance  for  the  control


                S   =   Estimate  of the  variance  for  the  effluent
                       concentration

                n   =   Number of replicates for  the  control
                n   =  Number of replicates for the effluent
                       concentration

                                     316

-------
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 n, + 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:
                            t  =
                                    35.4 - 15.9
         7.82
                                    '•13 /   1  +  1
                                       V    8     9
     Where:
                            (8    1)  14.5  +   (9   1) 36.6
                                 (8
2)
5.13
9.4  For an 0.05  level of significance test with 15 degrees of freedom
the critical t  is  1.754  (Note:  Table D.5 for K = 1 includes the critical
t values for comparing two groups).  Since 7.82 is greater than 1.754,
the conclusion  is  that the reproduction  in the 100% effluent
concentration is  significantly  lower than the control reproduction.

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
fol1ows:
                                      317

-------
     Where:
                Y1     =   Mean  for  the  control

                Y2     =   Mean  for  the  effluent  concentration

                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
                n   =  Number of replicates for the effluent
                 2     concentration
10.2  Additionally, the degrees of freedom for the test are adjusted
using the following formula:
                              (n,   -  I)(n2  - 1)
               df
                       (nz -  1)C<  +   (1 - C)< (n,
         Where:

                                s2
                             s^
                             n1
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.

                                      318

-------
                                         I

                             PROBIT ANALYSIS

1.   This program calculates the EC1 and EC50 (or LCI and IC50), 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 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 by a set of total mortality data from a
fathead minnow embryo-larval survival and teratogenicity test.  The
program requests the following input (see Table I.I):

    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 original 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 1.2).
    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 1.2).
    3.  Estimates of the mean (mu) and standard deviation (sigma)
        of the underlying Iog10  tolerance  distribution  (see  Table  1.2).
    4.  A table of the estimated EC values and associated 95% confidence
         intervals (see Table 1.2).
    5.  A plot of the fitted regression line with observed data overlaid on
        the plot (see Figure I.I).
                                      319

-------
      TABLE  I.I.  SAMPLE DATA INPUT FOR PROBIT ANALYSIS PROGRAM, VERSION 1.4
     I

     I

     1

     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 = ?   17
Number of  animals exposed in the  concurrent control  group =  ?  100
Number of  administered concentrations  ?  5
                                    320

-------
TABLE  I.I.  SAMPLE DATA  INPUT FOR PROBIT ANALYSIS PROGRAM, VERSION 1.4
           (Continued)
     Input  data starting with the lowest concentration

     Concentration = ? 6.25
     Number responding = ?  14
     Number exposed = ? 100


     Concentration = ? 12.5
     Number responding = ?  16
     Number exposed = ? 102


     Concentration = ? 25
     Number responding = ?  35
     Number exposed = ? 100
     Concentration = ? 50
     Number  responding = ?  72
     Number  exposed = ? 99


     Concentration = ? 100
     Number  responding = ?  99
     Number  exposed = ? 99

Number
1
2
3
4
5

Cone.
6.2500
12.5000
25.0000
50.0000
100.0000
Number
Resp.
14
16
35
72
99
Number
Exposed
100
102
100
99
99
     Do you wish to modify your data  ?  n

     The number of control animals which responded =  17
     The number of control animals exposed = 100
     Do you wish to modify these values ?  n
                                  321

-------
  TABLE 1.2.  SAMPLE DATA OUTPUT FOR EPA PROBIT ANALYSIS PROGRAM, VERSION 1-.4
                   EPA PROBIT ANALYSIS  PROGRAM
                 USED FOR CALCULATING EC VALUES
                          Version  1.4
Example of Probit Analysis
    Cone.

   Control
    6.2500
   12.5000
   25.0000
   50.0000
  100.0000
             Number
            Exposed

               100
               100
               102
               100
                99
                99
      Number
      Resp.

         17
         14
         16
         35
         72
         99
 Observed
Proportion
Responding

  0.1700
  0.1400
  0.1569
  0.3500
  0.7273
  1.0000
 Adjusted
Proportion
Responding

  0.0000
  -.0201
  -.0001
  0.2290
  0.6765
  1.0000
      Predicted
      Proportion
      Responding

         0.1570
         0.0000
         0.0083
         0.1865
         0.7309
         0.9831
Chi - Square Heterogeneity =     3.472
Mu
Sigma

Parameter
                1.575962
                0.199805

                Estimate
          Std. Err.
             95% Confidence Limits
Intercept
Slope

Spontaneous
Response Rate
               -2.887497
                5.004879

                0.156984
          1.040416
          0.630747

          0.021800
            •4.926712,
            3.768614,

            0.114257,
           -0.848282)
            6.241143)

            0.199712)
      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.

 12.9168
 17.6727
 20.8877
 23.3821
 37.6671
 60.6792
 67.9253
 80.2825
109.8419
                                       Lower       Upper
                                     95% Confidence Limits
    8.3881
   12.6476
   15.7183
   18.1828
   32.8982
   53.9806
   59.7534
   69.0899
   89.8905
     16.
     21.
     25,
     27,
     42,
     70,
     81,
    100,
    150,
8877
8381
0859
5756
0806
8083
3411
4315
5078
                                322

-------
Probit
   10+
    9+
    8+
    5+
   4+
2+






1-t-






O+o  o
  -+	

  EC01
BC10     EC25      EC50      EC75      EC90
                                                                        +_

                                                                       CC99
   Figure I.I.   Plot of Adjusted  Probits and Predicted  Regression Line.


                                      323

-------
                                  APPENDIX J

                         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 nonincreasing, 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 nonincreasing,  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
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

                                      324

-------
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 Yj  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.

4.2  To obtain the estimate, determine the concentrations Cd and CJ+1
which bracket the response M^l - p/100),  where M1  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 = Cj + [M^l  -  p/100)  -  MjKCj.,   -  Cj)        (1)
      where:   Cj   =   The tested concentration whose observed
                       mean response is greater than M,(l - p/100).

               CJ+1  =   The  tested  concentration whose observed
                       mean response is less than M.,(l - p/100).

               M1   =  Smoothed mean response for the control.

               Mj   =  Smoothed mean response for concentration J.

               MJ+1  =   Smoothed mean response for concentration  J+l.

               p    =  Percent reduction in response relative to
                       the control response.

               ICp  =  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 Cj".
                                      325

-------
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, YjM, are randomly  resampled with
replacement, to produce a new set of data,  YjM., that is statistically
equivalent to the original 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 "dataj 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 Ceriodaphm'a dubia reproduction
data used in the example in Section 12.  Table J.I includes the raw data and
the mean reproduction for each concentration.  Data are included for all
animals tested regardless of death or survival  of the organism.  If an animal
died during the test without producing young, a zero is entered.  If death
occurred after producing young, the number of young produced prior to death
is entered.  A plot of the data is provided in Figure J.I.
                                      326

-------
            TABLE J.I.  CERIODAPHNIA DUBIA REPRODUCTION  DATA
Effluent Concentration (%)
Replicate
1
2
3
4
5
6
7
8
9
10
Mean (Y,-)
i
Control
27
30
29
31
16
15
18
17
14
27
22.4
1
1.56
32
35
32
26
18
29
27
16
35
13
26.3
2
3.12
39
30
33
33
36
33
33
27
38
44
34.6
3
6.25
27
34
36
34
31
27
33
31
33
31
31.7
4
12.5
10
13
7
7
7
10
10
16
12
2
9.4
5
25.0
0
0
0
0
0
0
0
0
0
0
0
6
6.2  Monotonicity
6.2.1  As can be seen from the plot, Figure J.I,  the  observed  means  are not
monotonically nonincreasing with respect to concentration.   Therefore,  the
means must be smoothed prior to calculating the  1C.
6.2.2_ Starting with the control mean, Y,= 22.4, and Y"2 = 26.3, we see
that Y1  is less than Y2.
6.2.3  Calculate the smoothed means;
                M1  = M2  =  (7, + 72)/2 = 24.35
6.2.4  Since Y3 = 34.6 is larger than M2,  average Y3 with the previous
concentrations:
                M1  = M2  =  M3 =  (M, + M2 + 73)/3 = 27.7.
6.2.5  Additionally, Y4 = 31.7 is larger than M3, and  is pooled with the
first three means.   Thus:
       (M, + M2 + M3 + YJ/4 = 28.7 = M, = M2  = M3 = M4
                                      327

-------
                 50
CO
PO
00
                 0 1
                   0.00
                                                           ****   INDIVIDUAL NUMBER OF YOUNG

                                                           	   CONNECTS THE OBSERVED MEAN VALUE

                                                           	   CONNECTS THE SMOOTHED MEAN VALUE
1.56
  3.12              6.25

EFFLUENT CONCENTRATION (%)
12.50
25.00
               Figure  J.I.  Plot of observed and  smoothed  means  for the daphnid, Cen'odaphm'a  dubia
                             reproduction data.

-------
6.2.6  Since M4 > 7, = 9.4,  set M5  =  9.4.   Likewise, M5 > Y6 = 0,
and M6 becomes 0.  Table J.2 contains the smoothed means and Figure J.I
gives a plot of the smoothed response curve.
                  TABLE J.2.  CERIODAPHNIA DUBIA REPRODUCTION MEAN
                              RESPONSE AFTER SMOOTHING
                         Cone
M..
Control
1.56
3.12
6.25
12.5
25.0
1
2
3
4
5
6
28.75
28.75
28.75
28.75
9.40
0.00
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 reproduction, compared to the
controls, would result in a mean reproduction of 21.56 young per adult, where
M,(l -  p/100)  = 28.75(1 -  25/100).   A 50% reduction in reproduction,
compared to the controls,  would result in a mean reproduction of 14.38 young
per adult, where M^l - p/100)  = 28.75(1  - 50/100).  Examining the smoothed
means and their associated concentrations (Table 15, Section 12), the two
effluent concentrations bracketing 21.56 young per adult are C^ = 6.25%
effluent and C5 = 12.5% effluent.   The two effluent concentrations bracketing
a response of 14.38 young per adult are also C4 = 6.25 and C5 =  12.5.

6.3.2  Using Equation  1 from 4.2, the IC25 estimate is 8.6%  effluent:

               ICp = Cj +  [M,(l  -  p/100)  - MJ(CJ+1     - Cj)
                                              (MJ+1     - Mj)

              IC25 = 6.25 +  [28.75(1  - 25/100)  - 28.751(12.5  -  6.25)
                                                       (9.40  -  28.75)
                   = 8.57% effluent

6.3.3  The IC50 estimate is  10.9% effluent:

               ICp = Cj + [M^l - p/100)  - Mj](CJ+1     Cj)
                                               (M,
                                      329

-------
              IC50 = 6.25 + [28.75(1   50/100)   28.751(12.5 - 6.25)
                                                       (9.40 - 28.75)
                   = 10.89% effluent

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 perform these calculations
manually.  However, a computer program, BOOTSTRP, 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 (ERL-Duluth),  U. S. Environmental Protection
Agency, Duluth, Minnesota (Norberg-King, 1988).   For information concerning
the program and program documentation, contact ERL-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,  Mp  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
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 J.2.

                                      330

-------
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 J.3.
                                     331

-------
     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": S

     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: 27

     YOU HAVE ENTERED THE  FOLLOWING VALUES:
               CONCENTRATION ID = 1
               CONCENTRATION =     .000
               RESPONSE =         27.000

     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 J.2.   Example of BOOTSTRP program data input on the screen.
     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": F

     ENTER THE INPUT FILE NAME (SPECIFYING THE DRIVE AND
        SUBDIRECTORY IF NECESSARY):  cericp.2

     ENTER THE VALUE OF 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 J.3.Example of BOOTSTRP program data input from an existing
              data file.

                                     332

-------
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
J.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 (YJ,
          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 the ICp
          obtained in  Item 2, above.)

7.6  Output From Cen'odaphm'a Data Analysis

7.6.1  BOOTSTRP program output for the analysis of the Cen'odaphm'a
dub!* reproduction data in Table J.I is provided in Figures J.4 and J.5.

7.6.2  Using 80 resamples, the mean of the IC25 estimates was 8.6, with  a
standard deviation of  0.22 (coefficient of variation = 2%), and the empirical
95% confidence interval  was (8.4 - 8.9).
7.6.3  Using 80 resamples,  the mean of the IC50 estimates was 11.0,  with a
standard deviation of 0.22  (coefficient of variation = 2%).   The empirical
confidence interval  was (10.6 - 11.5).
95%
                                     333

-------
THE NUMBER OF RESAMPLES IS   80

*** LISTING OF GROUP CONCENTRATIONS (%EFF.) AND RESPONSE MEANS***

CONC. (%EFF)             RESPONSE MEAN         MEAN AFTER POOLING
    .000                   22.400                   28.750

   1.560                   26.300                   28.750

   3.120                   34.600                   28.750

   6.250                   31.700                   28.750

  12.500                    9.400                    9.400

  25.000                     .000                     .000
THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
   FROM THE INPUT SAMPLE IS   8.5715.

    *************************************************************
    *        BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY        *
                        OF THE ESTIMATED ICp
THE MEAN OF THE BOOTSTRAP ESTIMATE IS  8.6486.

THE STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS   .1102.

AN EMPIRICAL 95.0% CONFIDENCE INTERVAL FOR THE
     BOOTSTRAP' ESTIMATE IS (  8.4150,  8.8677).
          Figure J.4.  Example of BOOTSTRP program output
                       for the IC25.
                                334

-------
        THE NUMBER OF RESAMPLES IS   80

        *** LISTING OF GROUP CONCENTRATIONS (%EFF=) AND RESPONSE MEANS***

        CONC.  (%EFF)             RESPONSE MEAN         MEAN AFTER POOLING
            .000                   22.400                   28.750

           1.560                   26.300                   28.750

           3.120                   34.600                   28.750

           6.250                   31.700                   28.750

          12.500                    9.400                    9.400

          25.000                     .000                     .000
        THE LINEAR INTERPOLATION ESTIMATE OF THE TOTAL IMPACT CONCENTRATION
           FROM THE INPUT SAMPLE IS  10.8931.
                     BOOTSTRAP PROCEDURE TO ESTIMATE VARIABILITY
                                OF THE ESTIMATED ICp
        THE  MEAN OF THE BOOTSTRAP ESTIMATE IS   11.0473.

        THE  STANDARD DEVIATION OF THE BOOTSTRAP ESTIMATES IS   .2205.

        AN EMPIRICAL 95.0% CONFIDENCE INTERVAL FOR THE
             BOOTSTRAP ESTIMATE IS (  10.5800,  11.4854).
                  Figure J.5.   Example of BOOTSTRP program output
                               for the IC50.
                                       335
*U.S. GOVERNMENT PRINTING OFFICE: 3. 9 9 2 . 6 % 8 - 0 o ?tl e 0 e

-------