&EPA
           United States
           Environmental Protection
           Agency
          Environmental Monitoring Systen
          Laboratory
          Cincinnati OH 45268
600489001
March 1989
           Research and Development
Short-Term Methods for
Estimating the Chronic
Toxicity of Effluents and
Receiving Waters to
Freshwater Organisms

Second Edition
XX-3

-------

-------
                                                      EPA/600/4-89/001
                                                      March  1989
              SHORT-TERM METHODS FOR ESTIMATING
    THE CHRONIC TOXICITY OF EFFLUENTS AND RECEIVING WATERS
                   TO FRESHWATER ORGANISMS
                        SECOND EDITION
                          Prepared
                             by


Cornelius I. Weber1, William H.  Peltier2,  Teresa J.
Norberg-King39 William B.  Horning,  II1,  Florence A.  Kessler 4
John R. Menkedick*, Timothy W.  Neiheisel1, Philip A.  Lewis1
Donald J. Klemm1, Quentin H. Pickering1, Ernest L.  Robinson1.
James M. Lazorchak', Larry J. Wymer^, and Ronald W.  Freyberg*
   Mquatic Biology Branch, Qualilty Assurance Research Division
      Environmental Monitoring Systems Laboratory - Cincinnati
   ^Environmental Services Division, Region 4
   ^Environmental Research Laboratory, Duluth, Minnesota
   ^Computer Sciences Corporation, Cincinnati, Ohio
   ENVIRONMENTAL MONITORING SYSTEMS LABORATORY - CINCINNATI
 OFFICE OF MODELING, MONITORING SYSTEMS, AND QUALITY ASSURANCE
              OFFICE OF RESEARCH AND DEVELOPMENT
             U.  S.  ENVIRONMENTAL  PROTECTION AGENCY
                   CINCINNATI, OHIO  45268

-------
                                  NOTICE

    This document has been reviewed  in accordance with U.S.  Environmental
Protection Agency policy and approved for publication.  Mention of trade
names or commercial  products does not constitute endorsement or
recommendation for use.

-------
                                 FOREWORD

    Environmental measurements are required to determine the chemical  and
biological quality of drinking water, surface waters, groundwaters,
wastewaters, sediments, sludges, and solid waste.   The Environmental
Monitoring Systems Laboratory - Cincinnati (EMSL-Cincinnati) conducts
research to:

    0   Develop and evaluate analytical  methods to identify and measure
        the concentration of chemical pollutants.
    0   Identify and quantitate the occurrence of viruses,  bacteria, and
        other human pathogens and indicator organisms.

    0   Measure the toxicity of pollutants to representative species of
        aquatic organisms and determine the effects of pollution on
        communities of indigenous freshwater, estuarine, and marine
        organisms, including the phytoplankton, zooplankton, periphyton,
        macrophyton, macroinvertebrates, and fish.

    0   Develop and operate a quality assurance program to  support
        achievement of data quality objectives for environmental
        measurements.

    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.  Determination of the toxicity of effluents, therefore, plays
an important role in identifying and controlling toxic discharges to
surface waters.  This report is a revision of EPA/600/4-85/014, and
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
                                    ill

-------
                                    PREFACE
    This manual is a revision of EPA/600/4-85/014.   It was reviewed by the
following members of the Bioassay Subcommittee and its parent committee, the
EMSL-Cincinnati Biological  Advisory Committee, representing Agency regional
and headquarters programs,  and research laboratories.

      BIOASSAY SUBCOMMITTEE, EMSL-CIKCINMATI BIOLOGICAL ADVISORY COMMITTEE

    William Peltier, Chairman, Bioassay Subcommittee
      Environmental  Services Division,  Region 4
    Peter Nolan, Environmental Services Division,  Region 1
    Stephen Ward,  Environmental  Services Division,  Region 2
    Roland Hemmett,  Environmental Services Division, Region 2
                    Environmental Services Division,
                    Environmental Services Division,
                    , Environmental  Services Division
                     Environmental  Services Division
                     Environmental  Services Division
Ronald Preston,
Robert Donaghy,
Charles Steiner,
Michael Bastian,
Terry Hollister,
Region 3
Region 3
 Region 5
 Region 6
 Region 6
Region 7
    Michael  Tucker, Environmental  Services Division,
    Loys Parrish, Environmental  Services Division,  Region 8
    Milton Tunzi, Office of Policy and Management,  Region 9
    Peter Husby,  Office of Policy  and Management, Region 9
    Joseph Cummins, Environmental  Services Division,  Region 10
    Bruce Binkley, National  Enforcement Investigations Center, Denver
    Wesley Kinney, Environmental Monitoring Systems Laboratory - Las Vegas
    Steven Schimmel, Environmental Research Laboratory - Narragansett
    Douglas Middaugh, Environmental  Research Laboratory - Gulf Breeze
    Teresa Norberg-King, Environmental  Research Laboratory - Duluth
    Larry Kapustka, Environmental  Research Laboratory - Corvallis
    Richard Swartz, Environmental  Research Laboratory - Mewport
    Margarete Heber, Permits Division,  Office of Water Enforcement and Permits
    Edward Bender, Enforcement Division, Office of  Water Enforcement and
      Permits
    James Plafkin, Monitoring and  Data Support Division, Office of Water
      Regulations and Standards
    Chris Zarba,  Criteria and Standard Division, Office of Water Regulations
      and Standards
    Dan Rieder,  Hazard Evaluation  Division, Office  of Pesticide Programs
    Jerry Smrchek, Health and Environmental Review  Division, Office of Toxic
      Substances
    Gail Hansen,  Office of Solid Waste
    Royal Nadeau, Emergency Response Team, Edison,  NJ
                              Cornelius I.  Weber,  Ph.D.
                              Chairman, Biological  Advisory Committee
                              Acting Chief, Aquatic Biology Branch
                              Environmental Monitoring Systems
                                Laboratory  - Cincinnati
                                       IV

-------
                                 ABSTRACT

    This manual  is a revision of EPA/600/4-85/014,  and  describes
short-term (four- to seven-day) methods for estimating  the chronic
toxicity of effluents and receiving waters  to  the  fathead  minnow
(Plmephales promelas), a cladoceran (Ceriodaphnia  dubia),  and  a green
alga (Selenastrum capricornutum).   Also included are  guidelines on
laboratory safety, quality assurance,  facilities and  equipment, dilution
water, effluent sampling and holding,  data  analysis,  report  preparation,
and organism culturing and handling.  Supplementary information on
statistical techniques for test design and  analysis of  toxicity test data
is provided in the Appendices.

-------

-------
                                 CONTENTS

Foreword ,	    Ill
Preface	     iv
Abstract 	      v
Figures	viil
Tables	viii
Acknowledgments  	     ix

      1. Introduction  	      1
      2. Chronic Toxicity Test End Points and Data Analysis  ...      4
      3. Health and Safety   	     13
      4. Quality Assurance   	     15
      5. Facilities and Equipment	     20
      6. Test Organisms	     22
      7. Dilution Water	     24
      8. Effluent and Receiving Water Sampling and Sample Handling     27
      9. Report Preparation  	     31
     10. Fathead Minnow (Pi'mephales promelas) Larval  Survival
           and Growth Test	     33
     11. Fathead Minnow (Pimephales promelas) Embryo-larval
           Survival and Teratogenicity Test  	     75
     12. Cladoceran (Ceriodaphm'a dubia) Survival  and Reproduction
           Test	~TT~	    105
     13. Algal (Selenastrum capricornutum) Growth  Test 	    147

Selected References  	    175
Appendices	    188
      A. Independence, Randomization, and Outliers 	  .  .    189
      B. Validating Normality and Homogeneity of Variance
           Assumptions	    192
      C. Dunnett's Procedure	    204
      D. Bonferroni's T-test 	    216
      E. Steel's Many-one Rank Test	    221
      F. Wilcoxon Rank Sum Test	    225
      G. Fisher's Exact Test	    231
      H. Toxicity Screening Test - Comparison of Control  with
           100& Effluent or Instream Waste Concentration  	    240
      I. Probit Analysis	    244
                                    vi i

-------
                                  FIGURES
                               (SECTIONS  1-9)

Number                                                             Page
^^_^_^_^_                                                              I - ,ftm\ ii

   1. Flow chart for statistical analysis of test data  	    9
   2. Control charts	   19
                                   TABLES
                               (SECTIONS 1-9)

Number                                                             Page

   1. Preparation of synthetic fresh water using reagent
      grade chemicals	    26
   2. Preparation of synthetic fresh water using diluted
      mineral water  	    26
                                    VI11

-------
                              ACKNOWLEDGMENTS

    Materials for the first (EPA/600/4-85/014) and second editions of
this manual were taken in part from the following sources: USEPA, 1975,
Methods for Acute Toxicity Tests with Fish, Macroinvertebrates, and
Amphibians,  Environmental Research Laboratory, U. S. Environmental
Protection Agency, Duluth, Minnesota, EPA-660/3-75-009; USEPA, 1979,
Handbook for Analytical Quality Control in Water and Wastewater
Laboratories, Environmental Monitoring and Support Laboratory -
Cincinnati, U. S. Environmental Protection Agency, Cincinnati, Ohio,
EPA-600/4-79/019; USEPA, 1979, Interim NPDES Compliance Biomonitoring
Inspection Manual, Enforcement Division, Office of Water Enforcement, U.
S. Environmental Protection Agency, Washington, D.C.; Peltier, W. H., and
C. I. Weber, 1985, Methods for Measuring the Acute Toxicity of Effluents
to Freshwater and Marine Organisms, Environmental Monitoring and Support
Laboratory - Cincinnati, U. S. Environmental Protection Agency,
Cincinnati, Ohio, EPA-600/4-85/013; Mount, D. I., and T. J. Norberg,
1984, A Seven-day Life-cycle Cladoceran Test,  Environ. Toxicol. Chem.
3:425-434; Norberg, T., and D. I. Mount, 1985, A New Subchronic Fathead
Minnow (Pimephales prgmelas) Toxicity Test, Environ. Toxicol. Chem.
4:711-718; Miller, W.E, J. C. Greene, and T. Shiroyama, 1978, The
Selenastrum capricornutum Printz Algal Assay Bottle Test, Environmental
Research Laboratory, U. S. Environmental Protection Agency, Corvallis,
Oregon, EPA-600/9-78-018; Weber, C. I., W. B. Horning, II, D. J. Klemm,
T. W. Neiheisel, P. A. Lewis, E. L. Robinson, J. Menkedick, and F.
Kessler, 1988, 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, Ohio, EPA-600/4-87/028.

    In addition to the contributions of the Biological Advisory Committee
members, review comments received from the following persons are also
gratefully acknowledged:  Max Anderson, Central Regional Laboratory,
USEPA, Chicago, Illinois; Wesley Birge and Jeffrey Black, University of
Kentucky, Lexington, Kentucky; Gary Collins, Environmental Monitoring
Systems Laboratory, USEPA, Cincinnati, Ohio; Robert Elliott, Water
Quality Monitoring Branch, USEPA, San Francisco, California; Charles
Plost, Office of Modeling, Monitoring Systems and Quality Assurance,
USEPA, Washington, DC; Glenn Rodrigeuz, Environmental Services Division,
USEPA, Region 8, Denver, Colorado; Donald Schultz, Environmental Services
Division, Region 4, Athens, Georgia; Thomas Simon, Central Regional
Laboratory, USEPA, Chicago, Illinois; Albert Westerman, Natural Resources
and Environmental Protection Cabinet, Frankfort, Kentucky.

    The graphical displays for the statistical analyses were prepared by
Minghua Grisell, Computer Sciences Corporation, Environmental Monitoring
Systems Laboratory, Cincinnati, Ohio.
                                     IX

-------

-------
                                 SECTION 1

                                INTRODUCTION

1.1  The Federal Water Pollution Control Act (FWPCA)  Amendments of 1972
(PL 92-500), 1977 (Clean Water Act, PL 95-217), and 1987 (Water Quality
Act, PL 100-4), were enacted to restore and maintain the chemical,
physical, and biological integrity of the Nation's waters {Section
101[a]), and contained specific or implied requirements for the
collection of biomonitoring data in at least 15 sections.

1.2  The Declaration of Goals and Policy, Section 101(a)(3), in these
laws, states that "it is the national goal  that the discharge of toxic
pollutants in toxic amounts be prohibited."  To achieve the goals of this
legislation, extensive effluent toxicity screening programs were
conducted during the 1970s by the regions and states.   Acute toxicity
tests (USEPA, 1975; Peltier, 1978) were used to measure effluent toxicity
and to estimate the safe concentration of toxic effluents in receiving
waters.  However, for those effluents that were not sufficiently toxic to
cause mortality in acute {one- to four-day) tests, short-term,
inexpensive methods were not available to detect the more subtle,
low-level, long-term, adverse effects of effluents on aquatic organisms,
such as reduction in growth and reproduction, and occurrence of terata.
Fortunately, rapid developments in toxicity test methodology in this
decade have resulted in the availability of several methods that permit
detection of the low-level, adverse effects (chronic toxicity) of
effluents in seven days or less.

1.3  As a result of the increased awareness of the value of effluent
toxicity test data for toxics control in the water quality program and
the National Pollutant Discharge Elimination System (NPDES) permit
program, which emerged from the extensive effluent toxicity monitoring
activities of the regions and states, and the availability of short-term
chronic toxicity test methods, the U. S. Environmental  Protection Agency
(USEPA) issued a national policy statement entitled,  "Policy for the
Development of Water Quality-Based Permit Limitations for Toxic
Pollutants," in the Federal Register, Vol.  49, No. 48,  p. 9016-9019,
Friday, March 9, 1984.

1.4  This policy proposed the use of toxicity data to assess and control
the discharge of toxic substances to the Nation's waters through the
NPDES permits program.  The policy states that "biological  testing of
effluents is an important aspect of the water quality-based approach for
controlling toxic pollutants.  Effluent toxicity data,  in conjunction
with other data, can be used to establish control priorities, assess
compliance with State water quality standards, and set permit limitations
to achieve those standards."  All states have water quality standards
which include narrative statements prohibiting the discharge of toxic
materials in toxic amounts.

-------
1.5  A technical support document (USEPA,  1985)  and a permit writer's  guide
(USEPA, 1987b) were prepared by the Office of Water to provide  detailed
guidance on the implementation of the biomonitoring policy in the discharge
permit program, and the first edition of this manual  (EPA/600/4-85/014) was
published to provide standardized toxicity test  methodology.  The current
(second) edition of the manual contains many improvements in culturing and
test conditions, and detailed examples of the statistical analysis of  test
data.

1.6  The four short-term tests described in this manual are for use in the
NPDES Program to estimate one or more of the following: (1) the chronic
toxicity of effluents collected at the end of the discharge pipe and tested
with a standard dilution water; (2) the chronic  toxicity of effluents
collected at the end of the discharge pipe and tested with dilution water
consisting of non-toxic receiving water collected upstream from or outside
the influence of the outfall, or with other uncontaminated surface water  or
standard dilution water having approximately the same hardness  as the
receiving water; (3) the toxicity of receiving water downstream from or
within the influence of the outfall; and (4) the effects of multiple
discharges on the quality of the receiving water.  The tests may also  be
useful in developing site-specific water quality criteria.

1.7  These methods were developed to provide the most favorable cost-benefit
relationship possible, and are intended for use  in effluent toxicity tests
performed on-site or off-site.

    The tests include:

    1. A seven-day, sub-chronic, fathead minnow  (Pimephales promelas),
       static renewal, larval survival  and growth test.

    2. A three-brood, seven-day, chronic,  cladoceran (Ceriodaphnia
       dubia), static renewal, survival  and reproduction test.

    3. A seven-day, sub-chronic, fathead minnow  (Pimephales promelas),
       static renewal, embryo-larval survival  and teratogenicity test.

    4. A four-day, chronic, algal, (Selenastrum  capricornutum), static,
       growth test.

1.8  The first two tests were adapted from methods developed by
Dr. Donald Mount and Teresa Norberg-King,  Environmental Research
Laboratory, USEPA, Duluth, Minnesota (Mount and  Norberg, 1984;  Norberg
and Mount, 1985).  The third test was adapted from a method developed  by
Drs. Wesley Birge and Jeffrey Black, Graduate Center for Toxicology,
University of Kentucky, Lexington, Kentucky (Birge and Black, 1981).  The
fourth test, a 96-h, multi-generation test utilizing the freshwater alga,
Selenastrum capricornutum, was adapted from the  publications of the
Environmental Research Laboratory - Corvallis (USEPA, 1971; Miller et
al., 1978).

-------
1.9  The validity of the first two tests  methods  in  predicting adverse
ecological  impacts of toxic discharges  was  demonstrated  in  field  studies
on the Ottawa River, Ohio (Mount et al.,  1984), Scippo Creek, Ohio  (Mount
and Norberg-King, 1985), Five Mile Creek, Alabama (Mount et al.,  1985a),
the Ohio River,  West Virginia (Mount et al.,  1985b),  the Kanawha  River,
West Virginia (Mount and Norberg-King,  1986),  Skeleton Creek, Oklahoma
(Norberg-King and Mount, 1986), the Naugatuck  River,  Connecticut  (Mount
and Norberg-King, 1986a), and the Back  River,  Maryland (Mount et  al.,
1986b).  Other field studies demonstrating  the validity  of  the tests  in
this manual were carried out by Birge et  al.,  (1989), for the Fathead
Minnow Embryo-Larval Survival and Teratogenicity  Test, and  Eagleson,  et
al., (1989), for the Ceriodaphm'a dubia Survival  and Reproduction Test.

1.10  The tests were revised by staff from  EMSL-Cincinnati, Environmental
Research Laboratory-Duluth, and the regional  programs to reflect  the
collective experience of Agency and state programs in the use of  the
methods during the three years since the  first edition of the manual  was
published.   The authority for promulgation  of chemical,  physical, and
biological  test procedures for the analysis of pollutants is contained in
Section 304(h) of the FWPCA.

1.11  The manual was prepared in the established  EMSL-Cincinnati  format
(Kopp, 1983) so that each method can be used independently  of the other
methods.

-------
                                   SECTION  2

                CHRONIC TOXICITY  TEST  ENDPOINTS AND DATA ANALYSIS
2.1  ENDPOINTS

2.1.1  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 parameters 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 parameters,  when compared
to the controls.  The terms currently used to define the endpoints employed in
the rapid, chronic and sub-chronic toxicity tests have been derived from the
terms previously used for full  life-cycle tests.  As shorter chronic tests
were developed, it became common practice to apply the same terminology to the
endpoints.  The primary terms in current use are as follows:

2.1.1.1  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.

2.1.1.2  No-Observed-Effect-Concentration (NOEC) - The highest concentration
of toxicant to which organisms  are exposed in a full  life-cycle or partial
life-cycle 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 parameters are not statistically significantly different from
the controls).  This value is used, along with other factors, to determine
toxicity limits in permits.

2.1.1.3  Lowest-Observed-Effect-Concentration (LOEC)  -The lowest
concentration of toxicant to which organisms are exposed in a life-cycle or
partial life-cycle test, which  causes adverse effects on the test organisms
(i.e., where the values for the  observed parameters are statistically
significantly different from the controls).

2.1.1.4  Maximum Acceptable Toxicant Concentration (MATC) - An  undetermined
concentration within the interval bounded by the NOEC and LOEC  that is
presumed safe by virtue of the  fact that no statistically significant adverse
effect was observed.

2.1.1.5.  Chronic Value (ChV) - A point estimate of the presumably safe
(no-effect) concentration, lying between the NOEC and LOEC, and derived by
calculating the geometric mean  of the NUEC and LOEC.   The ChV has been
referred to as the "Maximum Acceptable Toxicant Concentration."

-------
2.1.1.6  Effective Concentration (EC)  - A point estimate  of  the  toxicant
concentration that would cause an observable adverse affect  (such  as  death,
immobilization, serious incapacitation, reduced fecundity, or reduced growth)
in a given percent of the test organisms, calculated by point estimation
techniques.   For example, the EC50 from a Probit Analysis is the estimated
concentration of toxicant that would cause death, or some other  observable
quanta!, "all or nothing," response, in 50% of the test population.   If the
observable effect is death (mortality), the term LC - Lethal  Concentration,  is
used (see below).  If the observable effect is a non-quantal  biological
measurement, the term, Inhibition Concentration (1C), may be used  (see
below).  A certain EC, LC, or 1C 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 parameters.

2.1.1.7  Lethal Concentration (LC) - Identical to EC when the observable
adverse affect is death or mortality.

2.1.1.8  Inhibition Concentration (1C) - A point estimate of the toxicant
concentration that would cause a given percent reduction  in  a non-quantal
biological measurement such as fecundity or growth.  For  example,  an  IC25
would be the estimated concentration of toxicant that would  cause  a  25%
reduction in mean young per female or  some other non-quantal  biological
measurement.

2.1.2  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  endpoint
of these tests is related to the "safe" or "no-effect" concentration.  NOECs
and LOECs are determined by hypothesis testing, and LCs,  ECs, and  ICs are
determined by point estimation techniques.  There are inherent differences
between the use of an NOEC, LOEC, ChV, or other estimate  derived from
hypothesis testing to estimate a "safe" concentration, and the use of a LC,
EC, 1C, or other point estimate derived from curve fitting,  interpolation, etc

2.1.3  Most point estimates, such as the LC, EC, or 1C are derived from a
mathematical model that assumes a continuous dose-response relationship.  By
definition,  any LC, EC, or 1C value is an estimate of some amount  of  adverse
effect.  Thus the assessment of a safe concentration must be made  from a
biological standpoint.   In this instance, the biologist must determine some
amount of adverse effect that is deemed to be "safe," in  the sense that it
will not from a practical biological viewpoint, affect the normal  propagation
of fish and other aquatic life in receiving waters.  Thus, to use  a  point
estimate such as an LC, EC, 1C to determine a "safe" concentration requires  a
biological judgment of what constitutes an acceptable level  of adverse effect.

2.1.4  The use of NOECs and LOECs, on  the other hand, assumes either  (1)  a
continuous dose-response relationship, or (2) a noncontiguous threshold model
of the dose-response relationship.

2.1.4.1  In  the first case, it is also assumed that adverse  effects  that  are
not "statistically observable" are also not significant from a biological

-------
standpoint, since they are not pronounced enough to test statistically
significant against some measure of the natural  variability of responses.

2.1.4.2  In the second case, 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 effect.  The purpose of the statistical
analysis in this case is to estimate as closely as possible where that
threshold lies.

2.1.4.3  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.   These
aspects include the choice of statistical analysis, the choice of an alpha
level, and the amount of variability between responses at a given
concentration.  The sensitivity of the test, which is related to the magnitude
of the adverse effect that is statistically observable, can be controlled by
the experimental design and by controlling the amount of variability between
responses at the given concentration.

2.1.4.4  In the first case, where the assumption of a continuous dose-response
relationship is made, clearly the NOEC estimate is an estimate of some amount
of adverse effect that is dependent on the experimental design.   In the second
case, the NOEC may be an estimate of a "safe" or "no-effect" concentration but
only 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.  The  NOEC
in that case would indeed be an estimate of a "safe" or "no-effect"
concentration.  If, however, the amount of adverse effect were not great
enough to test as statistically different, then the NOEC might well be an
estimate that again represents some amount of adverse effect which is assumed
safe because it did not test as statistically significant.  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.

2.1.5  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
"no-effect" concentration, but rather a concentration at which the effects are
judged to be of no biological significance.

-------
2.2  DATA ANALYSIS

2.2.1  Role of the Statistician

2.2.1.1  The choice of a statistical  method to analyze toxicity test data  and
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.
Analysts who are 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.

2.2.1.2  The recommended statistical  methods presented in this manual are  not
the only possible methods of statistical analysis.  Many other methods  have
been proposed and considered.  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.  Certainly there are other reasonable and  defensible
methods of statistical analysis of this  kind of toxicity data.  The methods
contained in this manual have been chosen, among other reasons, because they
are (1) well-tested and well-documented, (2) applicable to most different
toxicity test data sets for which they are recommended, but still powerful,
(3) hopefully "easily" understood by non-statisticians, and (4) amenable to
use without a computer, if necessary.

2.2.2  Plotting of the Data

2.2.2.1  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 Appendix.

2.2.3  Data Transformations

2.2.3.1  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.

2.3  INDEPENDENCE, RANDOMIZATION, AND OUTLIERS

2.3.1  Statistical independence among observations is a critical assumption in
the statistical analysis of toxicity data.  One of the best ways to insure
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.  A discussion of statistical independence, outliers and
randomization, and a sample randomization scheme, are included in Appendix A.

2.4  REPLICATION AND SENSITIVITY

2.4.1  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
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.

2.4.2  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.  If there are only two replicates, Dunnett's Procedure
may be used, but it is not possible to check the assumptions of the test.

2.5  CHOICE OF ANALYSIS AND MULTIPLE NOECs

2.5.1  The recommended statistical analysis of most data from chronic toxicity
tests with aquatic organisms follows a decision process illustrated in the
flow chart in Figure 1.  An initial decision is made to use point estimation
techniques and/or to use hypothesis testing.  If hypothesis testing is chosen,
subsequent decisions are made on the appropriate hypothesis testing procedure
for a given set of data, as illustrated in the flow chart.  If point
estimation is chosen, the equivalent of an NOEC can be calculated.  A specific
flow chart is included in the analysis section for each test.

2.5.2  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
parameters 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
parameters would not be subsequently tested for an effect on some other
parameter.  This is one reason for excluding concentrations that have shown a
statistically significant reduction in survival from a subsequent statistical
analysis 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.

-------
                        REPRODUCTION  DATA
                       NO.  OF YOUNG PRODUCED

 POINT ESTIMATION
         HYPOTHESIS TESTING
      (EXCLUDING CONCENTRATIONS
       ABOVE NOEC FOR SURVIVAL)
 ENDPOINT ESTIMATE
    IC25,  IC50
                 1
         SHAPIRO-HILK'S TEST
                              NON-NORMAL DISTRIBUTION
             NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
       NO
                           BARTLETT'S TEST
EQUAL NUMBER OF
  REPLICATES?
                 YES
   T-TEST WITH
   BONFERRONI
   ADJUSTMENT

                                     EQUAL NUMBER OF
                                       REPLICATES?
                                    HETEROGENEOUS
                                       VARIANCE
                                          NO
  DUNNETT'S
    TEST
                      I
            YES
STEEL'S MANY-ONE
    RANK TEST
  NILCOXON RANK SUM
      TEST WITH
BONFERRONI ADJUSTMENT

                          ENDPOINT ESTIMATES
                               NOEC. LOEC
   Figure 1. Flow chart for statistical analysis  of  test data

-------
2.6  ANALYSIS OF GROWTH AND REPRODUCTION DATA

2.6.1  Growth data from the fathead minnow larval  survival  and growth test are
analyzed using hypothesis testing or point estimation techniques according to
the flow chart in Figure 1.  (Note that the nonparametric hypothesis tests can
be used only if at least four replicates were used at each toxicant
concentration).

2.6.2  Reproduction data from the Ceriodaphnia survival  and reproduction test,
after eliminating data from concentrations with a  significant mortality  effect
as determined by Fisher's Exact Test, are analyzed using hypothesis testing or
point estimation techniques according to the flow  chart in Figure 1.  (Note
that the nonparametric hypothesis tests can be used only if at least four
replicates were used at each toxicant concentration).

2.7  ANALYSIS OF ALGAL GROWTH RESPONSE DATA

2.7.1  The growth response data from the algal 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 flow chart in Figure 1.   Point estimates, such as the EC1, EC5, EC10,
or EC50, would also be appropriate in analyzing algal growth data.

2.8  ANALYSIS OF MORTALITY DATA

2.8.1  Mortality data from the  fathead minnow larval  survival and growth test
and the fathead minnow 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  flow chart in Figure 1.

2.8.2  Mortality data from the  Ceriodaphnia 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 (see discussion below).

2.9  DUNNETT'S PROCEDURE

2.9.1  Dunnett's 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 two replicates  per treatment and an equal number of data points
(replicates) for each  concentration.  However, as stated above,  it is not
possible to check the assumptions of the test.  In cases where the number of
data points 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.

2.9.2  The assumptions upon which the use of Dunnett's Procedure is contingent
are that the observations within treatments are independent and normally
distributed, with homogeneity of variance.  Before analyzing the data, the
assumptions must be verified using the procedures  provided in Appendix B.

                                       10

-------
2.9.3  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.

2.9.4  The estimate of the safe concentration derived from this test is
reported in terms of the NOEC.  A step-by-step example of  Dunnett's Procedure
is provided in the Appendix.

2.9.5  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 per toxicant concentration.  If the numbers
of data points (replicates) for each toxicant concentration are not equal, the
Wilcoxon Rank Sum Test with Bonferroni's adjustment  should be used (see
Appendix F).

2.10  BONFERRONI'S T-TEST

2.10.1  Bonferroni's T-test (see Appendix  D) 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.

2.11  STEEL'S MANY-ONE RANK TEST

2.11.1  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
efficient (Hodges and Lehmann, 1956).   It  is necessary to  have at least four
replicates per toxicant concentration  to use Steel's test.  The sensitivity of
this test cannot be stated in terms of the minimum difference between
treatment means and the control mean.

2.11.2  The estimate of the safe concentration is reported as the NOEC.   A
step-by-step example of Steel's Many-One Rank Test is provided in Appendix E.

2.12  WILCOXON RANK SUM TEST

2.12.1  The Wilcoxon Rank Sum Test is  a  nonparametric test for comparing  a
treatment with a control.  The data are  ranked and the analysis proceeds
exactly as in Steel's Test except that Bonferroni's  adjustment for multiple
comparisons is used instead of Steel's tables.  When Steel's test can be  used
(i. e., when there are equal numbers of  data points  per toxicant
concentration), it will be more powerful  (able to detect smaller differences
as statistically significant) than the Wilcoxon Rank Sum Test with
Bonferroni's adjustment.

                                      11

-------
2.12.2  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.

2.13  INTERPOLATION APPROACH

2.13.1  Chronic toxicity test data can be analyzed by an interpolation
approach as described by DeGraeve et al.  (1988, Appendix B; 1989).   Precision
estimates can be calculated using this approach.   The round robin data
(DeGraeve et al., 1988; 1989) show that the endpoints estimated by this
approach are much less variable than those estimated by hypothesis testing.

2.14  PROBIT ANALYSIS

2.14.1  Probit Analysis is used to analyze percentage data from concentration-
response tests.  The analysis can provide an estimate of the concentration of
toxicant affecting a given percent of the test organisms and provide a
confidence interval for the estimate.  Probit Analysis assumes a  normal
distribution of log tolerances and independence of the individual  responses.
To use Probit Analysis, at least two partial  mortalities must  be  obtained.   If
a test results in 100% survival and 100% mortality in adjacent treatments (all
or nothing effect), a LC50 may be estimated using the graphical method,  and
the LC50 and confidence interval may be estimated by the moving average  angle,
Spearman-Karber, or other methods (see Peltier and Weber, 1985).

2.14.2  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.

2.14.3  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.

2.14.4  A discussion of Probit Analysis and the natural occurrence rate, along
with a computer program for performing the Probit Analysis, are included in
Appendix I.
                                       12

-------
                                  SECTION 3

                              HEALTH AND SAFETY^
3.1  GENERAL PRECAUTIONS
3.1.1  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.

3.1.2  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

    Personnel should use safety equipment, as required, such as rubber
aprons, laboratory coats, respirators, gloves, safety glasses,  hard hats,
and safety shoes.

3.2.2  Laboratory Safety Equipment

    Each laboratory (including mobile laboratories) should be provided with
safety equipment such as first aid kits, fire extinguishers, fire blankets,
emergency showers, and eye fountains.

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
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 should be used
whenever necessary.

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.
lAdapted from: Peltier and Weber (1985).
                                     13

-------
3.3.5.  Good housekeeping contributes to safety and reliable  results.

3.3.6.  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,

3.3.7.  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, and  polio.

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 (1977)  and  Walters and
Jameson (1984).
                                     14

-------
                                   SECTION 4

                               QUALITY ASSURANCE^

4.1  INTRODUCTION

4.1.1  Quality Assurance (QA) practices for effluent toxicity tests consist of
all aspects of the test that affect data quality,  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.  For general guidance on good laboratory practices  related to
toxicity testing, see:  FDA, 1978; USEPA, 1979d,  1980b, and 1980c; and
DeWoskin, 1984.

4.2  EFFLUENT AND RECEIVING WATER SAMPLING AND HANDLING

4.2.1  Sample holding times and temperatures must conform to conditions
described in Section 8, Effluent and Receiving Water Sampling and Sample
Handling.

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 cladoceran,  Ceriodaphnia dubia,
and the green alga, Selenastrum capricornutum. The  organisms should  be
disease-free and should be positively identified  to  species.  The fish and
invertebrates should appear healthy, behave normally, feed  well,  and  have low
mortality in cultures and test controls.

4.4  FACILITIES, EQUIPMENT, AND TEST CHAMBERS

4.4.1  Laboratory and bioassay 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 and Equipment).

4.5  ANALYTICAL METHODS

4.5.1  Routine chemical and physical analyses must include  established quality
assurance practices outlined in Agency methods manuals (USEPA,  1979a,b).

4.6  CALIBRATION AND STANDARDIZATION

4.6.1  Instruments used for routine measurements  of  chemical and physical
parameters such as pH, DO, temperature, conductivity, alkalinity, and
hardness, must be calibrated and standardized according to  instrument
manufacturers procedures as indicated in the general section on quality
assurance (see EPA Methods 150.1, 360.1, 170.1, and  120.1,  USEPA, 1979b).
Calibration data are recorded in a permanent log.
^Adapted from: Peltier (1978), Peltier and Weber (1985),  and USEPA (1979a).

                                       15

-------
4.6.2  Wet chemical methods used to measure hardness and alkalinity  must be
standardized according to the procedures for those specific  EPA methods  (see
EPA Methods 130.2 and 310.1, USEPA 1979b).

4.7  DILUTION WATER

4.7.1  The dilution water used in effluent toxicity tests will  depend  on the
objectives of the study and logistical  constraints, as discussed in
Section 7.  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.

4.8  TEST CONDITIONS

4.8.1  Water temperature must be maintained within the limits specified  for
each test.  Dissolved oxygen (DO) concentration and pH in fish  and
invertebrate test chambers should be checked daily throughout the test period,
as prescribed in the methods.

4.9  ACCEPTABILITY OF SHORT-TERM CHRONIC TOXICITY TESTS

4.9.1  To be acceptable, control survival  in fathead minnow  and Ceriodaphnia
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 0.250 mg.
In the controls, the number of young per surviving adult Ceriodaphnia  must be
15 or greater, and at least 60% must have had three broodlTIn algal  toxicity
tests, the mean cell density in the controls after 96 h must equal or  exceed
2 X 10* cells/mL.

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  best
professional judgment and experience of the analyst and regulatory authority.
The deviation from test specifications must be noted when reporting  data from
the test.

4.10  TEST PRECISION

4.10.1 The ability of the laboratory personnel to obtain consistent, precise
results must be demonstrated with reference toxicants before they attempt to
measure effluent toxicity.  The single laboratory precision  of  each  type of
test to be used in a laboratory should be determined by performing at  least
five or more chronic tests with a reference toxicant.  In cases where  the test
data are used to obtain point estimates, such as LCs, ECs, or ICs (see
Section 2),  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 2), precision can
only be described by listing the NOEC-LOEC interval for each test.  In this
case, it is not possible to express precision in terms of a  commonly used
statistic.
                                       16

-------
For instance, when all tests of the same toxicant yield the same  NOEC-LOEC
interval, maximum precision has been attained.   However, the "true"  no  effect
concentration could fall anywhere within the interval,  NOEC + (NOEC-LOEC).

4.10.2  It should be noted here that the dilution factor selected for a test
determines the width of the NOEC-LOEC interval  and the  inherent maximum
precision of the test.  As the absolute value of the dilution factor
decreases, the width of the NOEC-LOEC interval  increases, and the inherent
maximum precision of the test decreases.  When a dilution factor  of  0.3 is
used, the NOEC could be considered to have a relative variability as high as
+_ 300&.  With a dilution factor of 0.5, the NOEC could  be considered to have a
relative variability of +_ 100%.  Other factors which can affect test precision
include test organism age, condition, and sensitivity,  and temperature  control
and feeding.

4.11 REPLICATION AND TEST SENSITIVITY

4.11.1  The sensitivity of the tests will depend in part on the number  of
replicates, the probability level selected, and the type of statistical
analysis.  The minimum recommended number of replicates varies with  the test
and the statistical method used, and is discussed in Section 2 and in each
method.  If the variability remains constant, the sensitivity of  the test will
increase as the number of replicates is increased.

4.12  QUALITY OF TEST ORGANISMS

4.12.1  If the laboratory does not have an ongoing test organism  culturing
program and obtains the test organisms from an outside  source, the sensitivity
(quality) of test organisms will be assumed to be acceptable if a reference
toxicant test is conducted side-by-side with the effluent toxicity test.  If
the laboratory maintains breeding cultures, the sensitivity of the offspring
should be determined in a chronic toxicity test performed with a  reference
toxicant at least once each month.  If preferred, this  reference  toxicant test
may be performed concurrently with an effluent toxicity test.

4.13  FOOD QUALITY

4.13.1  The quality of the food for fish and invertebrates is an  important
factor in toxicity tests.  Suitable trout chow, Artemia, and other foods must
be obtained as described in the manual.  Limited quantities of reference
Artemia cysts, information on commerical sources of good quality  Artemia
cysts, and procedures for determining cyst suitability  as food are available
from the Quality Assurance Research Division, Environmental Monitoring  Systems
Laboratory - Cincinnati.  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.

4.14  DOCUMENTING LABORATORY PERFORMANCE

4.14.1  Satisfactory laboratory performance is demonstrated by performing at
least one acceptable test per month for each of the toxicity test methods
                                       17

-------
commonly used in the laboratory, employing the same reference toxicant,  at the
same concentrations, in the same dilution water.

4.14.2  A control chart is prepared for each reference-toxicant-organism
combination, and successive toxicity values are plotted and examined to
determine if the results are within prescribed limits (Figure 2).   In this
technique, a running plot is maintained for the toxicity values (X-,*) from
successive tests with a given reference toxicant.  The type of control chart
illustrated (USEPA, 1979a) is used to evaluate the cumulative trend of the
statistics from a series of tests.  For point estimation techniques, the
mean (X) and upper and lower control limits (+_ 2S) are re-calculated with each
successive point, until the statistics stabilize.  Outliers,  which are values
which fall outside the upper and lower control limits, and trends  of
increasing or decreasing sensitivity are readily identified.   At the Pg.05
probability level, one in 20 tests would be expected to fall  outside of'the
control limits by chance alone.  For hypothesis testing results, it is assumed
that the same concentrations of reference toxicants are used for each toxicity
test.  The NOEC from each successive test is entered on the control  chart, and
the values should fall within one concentration interval above or  below  the
central tendency.

4.14.3  If the toxicity value from a given test with the reference toxicant
does not fall  in 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.

4.14.4  Four reference toxicants are available from EMSL-Cincinnati  to
establish the precision and validity of toxicity data generated by
biomonitoring laboratories: sodium dodecylsulfate (SDS), copper sulfate
(CuS04), sodium chloride (NaCI), and cadmium chloride (CdCl2).  The
reference toxicants may be obtained by contacting the Quality Assurance
Research Division, EMSL-Cincinnati, FTS 684-7325, commercial  513-569-7325.
Instructions for the use and the toxicity values for the reference toxicants
are provided with the samples.  Note: To assure comparability of QA data on a
national scale, all laboratories should use the same source of reference
toxicant (EMSL-Cincinnati), and periodically (such as quarterly) use the same
formulation of dilution water — moderately hard dilution water described in
Section 7, for fathead minnows and Ceriodaphnia,  and algal growth  medium
described in Tables 1 and 2, Section 13, for Selenastrum.

4.15  RECORD KEEPING

4.15.1  Proper record keeping is required.  Bound notebooks should be used to
maintain detailed records of the test organisms such as species, source, age,
date of receipt, and other pertinent information relating to their history and
health, and information on the calibration of equipment and instruments, test
conditions employed, and test results.  Annotations should be made on a
real-time basis to prevent the loss of information.
                                       18

-------
     o
     LLJ
     O
              UPPER CONTROL LIMIT
                   CENTRALTENDENCY
              LOWER CONTROL LIMIT
         I  I  f !  I
!  LI III! I  I  I 1  _1_
                          . I
                       10
                                        20
     o
     LU
              UPPER CONTROL LIMIT(X+2S]
                   CENTRALTENDENCY
 LOWER CONTROL LI MIT (X - 2S)
                              i  1
                    11!!
       0       5       10       IS       20
       TOXICITY TEST WITH REFERENCE TOXICANTS

   Figure 2. Control charts. (A) hypothesis testing results;
            (B)  point estimates (EC, LC, or 1C).
                   n
Where:
              n- 1

=  Successive toxicity values from toxicity tests
Number of  tests.
Mean toxicity value.
Standard deviation.
                          19

-------
                                   SECTION 5

                            FACILITIES AND EQUIPMENT1
5.1  GENERAL REQUIREMENTS
5.1.1  Effluent toxicity tests may be performed in a fixed or mobile
laboratory.  Facilities should include equipment for rearing  and holding
organisms.

5.1.2  Culturing and testing areas should be separated.

5.1.3  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,  surface  water,
dechlorinated tap water, or synthetic water.  Dechlorination  can be
accomplished by aerating for 24 h, carbon filtration, or the  use of sodium
thiosulfate.  Use of 1.0 mg (anhydrous) sodium thiosulfate/L  will  reduce 1.5
mg chlorine/L.  After dechlorination, total  residual  chlorine should  be
non-detectable.  Air used for aeration must be free of oil  and fumes.
Oil-free air pumps should be used where possible.   If air pumps  are not
oil-free, at a minimum the air should be filtered through cotton.
Particulates can be removed from the air using BALSTONR Grade BX or
equivalent filters (Balston, Inc., Lexington, Massachusetts), and oil and
other organic vapors can be removed using activated carbon filters
(BALSTONR, C-l filter, or equivalent).  The facilities must be well
ventilated and free of toxic fumes.  During rearing,  holding, and testing,
test organisms should be shielded from external disturbances.

5.1.4  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.   Plastics such as
polyethylene, polypropylene, polyvinyl chloride, TYGONR,  etc., may be used
to store and transfer effluents,  but they should not be reused unless
absolutely necessary, because they could carry over toxicants from one test to
another if reused.  The use of glass carboys is discouraged for  safety
reasons.  Glass or disposable polystyrene containers are used for test
chambers.

5.1.5  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 which cannot be discarded after each
use bee.  .^e of cost, must be decontaminated according to the  cleaning
procedures listed below.  Fiberglass, in addition to the previously mentioned
materials, can be used for holding, acclimating, and dilution water storage
tanks, and in the water delivery system.  All material should be flushed or
rinsed thoroughly with the test media before using in the test.   Copper,
^Adapted from: Peltier and Weber (1985).

                                       20

-------
galvanized material, rubber,  brass,  and lead must not come  in contact with
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 test.

5.3  CLEANING

5.3.1  New plasticware used for sample collection or organism test chambers
does not require cleaning.  It is sufficient to rinse new sample containers
once with sample before use.   New, disposable,  plastic test chambers
normally do not have to be rinsed before use.   New glassware, however,
should be soaked overnight in acid (see below).

5.3.2  It is recommended that all sample containers, test vessels, tanks,
and other equipment that has come in contact with effluent  be washed after
use in the manner described below to remove surface contaminants.  Special
cleaning requirements for glassware  used in algal toxicity  tests are
described in Section 13.

    1. Soak 15 min, and scrub with detergent in tap water,  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 well with deionized water.

5.3.3  All test chambers and equipment must be  thoroughly rinsed with the
dilution water immediately prior to  use in  each test.
                                     21

-------
                                   SECTION 6

                                 TEST ORGANISMS
6.1  SPECIES
6.1.1  The organisms used In the chronic tests described in  this manual  are
the fathead minnow, Pimephales promelas, the cladoceran,  Ceriodaphm'a  dubia
(Berner, 1985), and the green alga, Selenastrum caprlcornutum.

6.2  SOURCE

6.2.1  The test organisms are easily cultured in the laboratory.   Culturing,
care, and handling procedures for Ceriodaphm'a and Selenastrum  are described
in the respective test methods sections.  A fathead minnow culturing procedure
using laboratory water is described in Peltier and Weber (1985).

6.2.2  Starter cultures of Selenastrum capricornutum are available from  the
following sources:

    1. Aquatic Biology Branch, Quality Assurance Research Division,
       Environmental Monitoring Systems Laboratory, USEPA, Cincinnati,
       Ohio  45268.

    2. Environmental Research Laboratory, USEPA, 200 SW 35th Street,
       Corvallis, Oregon  97330.

    3. American Type Culture Collection {Culture No. ATCC 22662),  12301
       Parklawn Drive, Rockvilie, Maryland  10852.

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

6.2.3  Starter cultures of fathead minnows and Ceriodaphm'a  can be obtained
from the Aquatic Biology Branch, Quality Assurance Research  Division,
EMSL-Cincinnati Newtown Facility, Environmental Monitoring Systems
Laboratory, USEPA, 3411 Church Street, Newtown, Ohio  45244   (Phone: FTS
684-8114; commercial 513-533-8114).

6.2.4  If there is any uncertainty concerning the identity of the  test
organisms, it is advisable to have them examined by a second party to
confirm their identification.
                                     22

-------
6.3  SHIPMENT

6.3.1  Many states have strict regulations regarding the importation  and
disposal  of non-native fishes.  Required clearances  should  be  obtained  from
state fisheries agencies before arrangements are made for the  interstate
shipment of fathead minnows.

6.4  DISPOSAL

6.4.1  Test organisms must be destroyed after use.
                                       23

-------
                                    SECTION  7

                                 DILUTION WATER

7.1  The source of dilution water used in effluent toxicity tests will  depend
largely on the objectives of the study:

    1. If the objective of the test is to estimate the inherent chronic
       toxicity of the effluent, which is the primary objective of NPDES
       permit-related toxicity testing, a standard dilution water
       (moderately hard water) is used.
    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 upstream and outside the
       influence of the the outfall, or with other uncontaminated surface
       water or standard dilution water having approximately the same
       characteristics (pH, hardness, alkalinity, conductivity, and total
       suspended solids) as the receiving water.  Seasonal variations in
       the quality of surface waters may affect effluent toxicity.
       Therefore, the pHs alkalinity, hardness, and conductivity of
       receiving water samples should be determined before each use.
    3. If the objective of the test is to determine the additive effects of
       the discharge on already contaminated receiving water, the test  is
       performed using dilution water consisting of receiving water
       collected upstream from the outfall.

7.2  When the dilution water is to be taken  from the receiving water
"upstream" from the outfall, it should be collected at a point as close as
possible to the outfall, but upstream from or outside of the zone influenced
by the effluent.  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.

7.3  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.

7.4  Dechlorinated water may be used as a source of dilution water if properly
treated.   Dechlorination can be accomplished by aerating for 24 h, carbon
filtration, or the use of sodium thiosulfate.   Use of 1.0 mg (anhydrous)
sodium thiosulfate/L will reduce 1.5 mg chlorine/L.

7.5  Deionized water may be obtained from a  MILLIPORE MILLI-QR System,  or
equivalent.  It is advisable to provide a preconditioned (deionized) feed
water by using a Culligan, Continental, or equivalent system in front of the
MILLI-QR System to extend the life of the MILLI-QR cartridges.  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.

                                       24

-------
7.6  Synthetic, moderately hard dilution water can  be prepared using reagent
grade chemicals (Table 1) or mineral  water (Table 2).

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

    1. Place 19 L of MILLI-QR, or equivalent,  water in  a  properly cleaned
       plastic carboy.
    2. Add 1.20 g of Mg$04,  1.92 g NaHCOa,  and 0.080g KC1  to the carboy.
    3. Aerate overnight.
    4. Add 1.20 g of CaS04'2H20 to 1  L  of MILLI-QR  or equivalent
       water in a separate flask.  Stir on magnetic stirrer until calcium
       sulfate is dissolved and add to  the 19  L above and mix well.
    5. Aerate vigorously for 24 h to  dissolve  the added chemicals and
       stabilize the medium.
    6. The measured pH, hardness, etc., will be as  listed in Table 1.

7.6.2  To prepare 20 L of synthetic,  moderately hard water using mineral water
(Table 2), follow the instructions below.  Note: These  instructions are
specific for PERRIERR Water.  The properties of other commercially available
mineral waters are not well  enough known at this time to  permit inclusion of
recommendations for their use.

    1. Place 16 L of MI1_LI-QR or equivalent water in a  properly cleaned
       plastic carboy.
    2. Add 4 L of PERRIERR Water.
    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 2.
    5. The synthetic water prepared with PERRIERR Water is referred to as
       20% diluted mineral water (20£ DMW) in  the toxicity test methods.

7.7  A given batch of dilution water should not be  used for more than 14 days
following preparation because of the  possible  build-up  of slime growth and the
problems associated with it.  The container should  be kept covered and the
water should be protected from light.
                                       25

-------
TABLE 1.  PREPARATION OF SYNTHETIC FRESH WATER USING REAGENT GRADE  CHEMICALS3
Reagent Added
Water
Type
Very soft
Soft
Moderately
Hard
Very hard


(mg/L)t>

NaHC03 CaS04-2H20 MgS04
12.0
48.0
Hard 96.0
192.0
384.0
7.5
30.0
60.0
120.0
240.0
7.5
30.0
60.0
120.0
240.0


KC1
0.5
2.0
4.0
8.0
16.0
Final

pHC
6.4-6.8
7.2-7.6
7.4-7.8
7.6-8.0
8.0-8.4
Water Quality

Alka-
Hardness^ linityd
10-13
40-48
80-1 00
160-180
280-320
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.
^Expressed as mg CaC03/L .
TABLE 2.  PREPARATION OF SYNTHETIC FRESH WATER USING MINERAL WATER*
Final Water Quality

Water
Type
Very soft
Soft
Moderately
Hard
Very hard6
Volume of
Mineral Water
Added (nt/L)b>
50
100
Hard 200
400
— -
Proportion
of Mineral
Water (%)
2.5
10.0
20.0
40.0
—


pHc
7.2-8.1
7.9-8.3
7.9-8.3
7.9-8.3
—


Hardness^
10-13
40-48
80-1 00
160-180
—

Alka-
linityd
10-13
30-35
60-70
110-120
— — -
aFrom Mount et al., 1987, and data provided by Philip Lewis,  EMSL-Cincinnati.
&Add mineral water to Mil1i-QR water or equivalent to prepare DMW (Diluted
 Mineral Water).
cApproximate equilibrium pH after 24 h of aeration.
dExpressed as mg CaC03/L.
Dilutions of PERRIERR Water form a precipitate when concentrations  equivalent
 to "very hard water" are aerated.
                                       26

-------
                                   SECTION 8

           EFFLUENT AND RECEIVING WATER SAMPLING AND SAMPLE HANDLING

8.1  EFFLUENT SAMPLING

8.1.1  The effluent sampling point usually should be the same  as  that
specified in the NPDES discharge permit (USEPA,  1979c).   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 to the receiving waters, 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 prohibitive  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.

8.1.3  Sample Type

8.1.3.1  The advantages and disadvantages of effluent grab and composite
samples are listed below:

8.1.3.2. Grab Samples

8.1.3.2.1  Advantages:

    1. Easy to collect; require a minimum of equipment and on-site time.
    2. Provide a measure of instantaneous toxicity.  Toxicity  spikes are
       not masked by dilution.

8.1.3.2.2  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.
                                     27

-------
8.1.3.3.  Composite Samples:

8.1.3.3.1.  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  and  contains
       all toxicity spikes.

8.1.3.3.2.  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.

8.1.4 SAMPLING RECOMMENDATIONS

8.1.4.1.  When tests are conducted on-site,  samples are collected daily,
except for the algal tests.

8.1.4.2  When tests are conducted off-site,  a minimum of three  samples are
collected.  It is recommended that these  samples not be collected on  a more
frequent schedule than one every other day.   This collection schedule would
provide fresh sample on Test Days 1, 3, and  5.  The first sample  would be used
for test initiation, Day 1, 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.

8.1.4.3  The following effluent sampling  methods are recommended:

8.1.4.3.1.  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,  composite samples are used.

    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 in chemical  composition or toxicity regardless
       of holding time, grab samples are  used.

    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.
                                       28

-------
8.1.4.3.2.   Intermittent Discharges

8.1.4.3.2.1.  If the facility discharge is intermittent,  composite  samples  are
collected during the discharge period.   Examples  of intermittent  discharges
are:

    1. When the effluent is continuously discharged during  a  single 8-h
       work shift or two successive 8-h work shifts.

    2. When the facility retains the wastewater during an 8-h work  shift, and
       then treats and releases the wastewater as a batch discharge.

    3. When the facility discharges wastewater to an estuary  only during an
       outgoing tide (usually during the 4 h following slack  high tide).

    4. At the end of the shift, clean up activities may result in the
       discharge of a slug of toxic waste.

8.1.5  Aeration during collection and transfer of effluents should  be
minimized to reduce the loss of volatile chemicals.

8.2  RECEIVING WATER SAMPLING

8.2.1  It is common practice to collect grab samples for receiving  water
toxicity studies.

8.2.2  When non-toxic receiving water is required for a test, it  may be
possible to obtain it upstream from the outfall or from another surface water
which is known to be uncontaminated and has properties similar to the
receiving water (see Section 7).  If the objective of the test is to determine
the additive effects of the discharge on receiving water which may  already  be
contaminated, the test is performed using dilution water consisting of
receiving water collected daily upstream from the outfall.

8.2.3  Dilution water to be taken from the receiving water  "upstream" from  the
outfall is collected at a point as close as possible to the outfall, but
upstream from or outside of the zone influenced by the effluent.

8.2.4  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 test organisms as used in one effluent toxicity  test with
five effluent dilutions.
                                       29

-------
8.3  SAMPLE HANDLING, PRESERVATION, AND SHIPPING

8.3.1  If the data from the samples are to be acceptable for use in the NPDES
Program, the lapsed time from collection of a grab or composite sample and its
first use for initiation of a test, or for test solution renewal,  should not
exceed 36 h.  Composite samples should be chilled during collection,  where
possible, and maintained at 4°C until  used.

8.3.2 Samples Used in On-Site Tests

8.3.2.1  Samples collected for on-site tests should be used within 24 h.

8.3.3  Samples Shipped to Off-Site Facilities

8.3.3.1  Samples collected for off-site toxicity testing are to be chilled to
4°C when collected, shipped iced to the central  laboratory, and there
transferred to a refrigerator (4°C) until used.   Every effort must be made
to initiate the test with an effluent sample on the day of arrival in the
laboratory.

8.3.3.2  Samples may be shipped in 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,  CUBITAINERSR and plastic jugs are
punctured to prevent reuse.

8.3.3.3  Several sample shipping options are available, including Express
Mail, air express, bus, and courier service.  Express Mail  is delivered seven
days a week.  Shipping and receiving schedules of private carriers on weekends
vary with the carrier.

8.4  SAMPLE PREPARATION

8.4.1  With the Ceriodaphm'a and fathead minnow tests, effluents and surface
waters must be filtered through a 60-um plankton net to remove indigenous
organims that may attack or be confused with the test organisms (see
Ceriodaphm'a test method for details).  Surface waters used in algal  toxicity
tests must be filtered through a 0.45-um pore diameter filter before use.  It
may be necessary to first coarse-filter the dilution and/or waste water
through a nylon sieve having 2- to 4-mm holes to remove debris and/or break up
large floating or suspended solids.  Caution: filtration may remove toxicity.

8.4.2  The DO concentration in the dilution water 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.

8.4.3  If the dilution water and effluent must be wanned to bring them to the
prescribed test temperature, supersaturation of the dissolved gases may become
a problem.  To prevent this problem, the effluent and dilution water are
checked for dissolved oxygen (DO) with a probe after heating to 25°C.  If
the DO is greater than 100% saturation or lower than 40£ saturation,  the
solutions are aerated moderately with a pipet tip for a few minutes until the
DO is within the prescribed range.

                                       30

-------
                                   SECTION 9

                              REPORT PREPARATION!

    The following general  format and content  are recommended  for  the  report:

9.1  INTRODUCTION

     1. Permit number
     2. Toxicity testing requirements of permit
     3. Plant location
     4. Name of receiving water body
     5. Contractor (if contracted)
        a. Name of firm
        b. Phone number
        c. Address
9.2  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 (MGD9 CFS9 GPM)
     8. Design flow of treatment facility at time of sampling

9.3  SOURCE OF EFFLUENT (AMBIENT) AND DILUTION WATER

     1. Effluent Samples
        a. Sampling point
        b. Collection dates and times
        c. Sample collection method
        d. Physical and chemical data

     2. Surface 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)
^Adapted from: Peltier and Weber (1985)
                                     31

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

9.4  TEST METHODS

      1. Toxicity test method  used
      2. End point(s) of test
      3. Deviations  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 (mean and range)
     11. Test temperature (mean and range)
9.5  TEST ORGANISMS

      1.  Scientific name
      2.  Age
      3.  Life stage
      4.  Mean length and weight (where applicable)
      5.  Source
      6.  Diseases and treatment (where applicable)
9.6  QUALITY ASSURANCE

      1.  Standard toxicant used and source
      2.  Date and time of most recent test
      3.  Dilution water used in test
      4.  Results (LC50 or, where applicable,  NOEC and/or EC1)
      5.  Physical and chemical  methods used
9.7  RESULTS

      1.  Provide raw biological  data in  tabular form,  including daily
         records of affected organisms in  each  concentration  (including
         controls)
      2.  Provide table of LC50s, NOECs,  etc.
      3.  Indicate statistical  methods to calculate  endpoints
      4.  Provide summary table of physical  and  chemical  data
      5.  Tabulate QA data
                                    32

-------
                                   SECTION 10

                                  TEST METHOD

      FATHEAD MINNOW, PIMEPHALES PROMELAS, LARVAL SURVIVAL AND GROWTH TEST
                                 METHOD 1000.0


1.  SCOPE AND APPLICATION

1.1  This method estimates the chronic toxicity  of whole effluents  and
receiving water to the fathead minnow, Pimephales promelas, 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, the
concentrations of highly degradable or highly volatile toxicants, such as
chlorine, present in the source may fall  below detectable levels before the
samples are used in a test.

1.5  This test is commonly used in one of two forms: (1) a definitive test,
consisting of a minimum of five effluent concentrations  and a control, and  (2)
an abbreviated test, consisting of only one concentration such as 100%
effluent or the in-stream waste concentration and a control.  Abbreviated
tests are used for toxicity screening or a pass/fail permit condition.
Failure of the screening test usually results in a followup definitive test.

1.6  This method should be restricted to use by  or under the supervision of
professionals experienced in aquatic toxicity testing.

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 (increase in weight) of  the larvae.
                                       33

-------
3.  INTERFERENCES

3.1  Toxic substances may be introduced by contaminants in dilution water,
glassware, sample hardware, and testing equipment (see Section 5,  Facilities
and Equipment).

3.2  Adverse effects of low dissolved oxygen (DO) concentrations,  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 handling may adversely affect  test results
(see Section 8, Effluent and Receiving Water Sampling and Sample Handling).

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 Peltier  and Weber
(1985),  This test requires 180-360 larvae.  It is preferable  to obtain larvae
from an inhouse fathead minnow culture unit.  If it is not feasible to culture
fish inhouse, embryos or newly hatched larvae can be shipped in well
oxygenated water in insulated containers.

5.2  Samplers — automatic sampler, preferrably 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 and Sample Handling).

5.4  Environmental chamber or equivalent facility with temperature control
(25+ IOC).

5.5  Water purification system -- MILLIPORE MILLI-QR or equivalent.

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.
                                       34

-------
5.8  Test chambers — four (minimum of three)  borosilicate glass or non-toxic
disposable plastic test chambers are required  for each concentration and
control.  Test chambers may be 1L,  500 mL,  or  250 ml  beakers, 220 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, 1/4  in thick).

5.9  Volumetric flasks and graduated cylinders — Class A, borosilicate glass
or non-toxic plastic labware,  10-1000 ml for making test solutions.

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, 4mm ID — for
transferring larvae.

5.14  Wash bottles — for washing embryos from substrates and containers and
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  pH, DO, and specific conductivity meters -- for routine physical and
chemical measurements.  Unless the test is  being conducted to specifically
measure the effect of one of the above parameters, a  portable, field-grade
instrument is acceptable.

6.  REAGENTS AND CONSUMABLE MATERIALS

6.1  Reagent water — defined as MILLIPORE  MILLI-QR or equivalent water  (see
paragraph 5.5 above).

6.2  Effluent, surface water,  and dilution  water  — see Section 7, Dilution
Water, and Section 8, Effluent and Surface  Water Sampling and Sample Handling.

6.3  Reagents for hardness and alkalinity tests (see  USEPA Methods 130.2 and
310.1, USEPA 1979b).

6.4  Standard pH buffers 4, 7, and 10 (or as per instructions of instrument
manufacturer) for instrument calibration (see  USEPA Method 150.1, USEPA 1979b)
                                       35

-------
6.5  Specific conductivity standards (see USEPA Method 120.1,  USEPA 1979b).

6.6  Laboratory quality assurance samples and standards for the above methods.

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

6.8  Ethanol (70%) for use as a preservative for the fish larvae.

6.9  Membranes and filling solutions for dissolved oxygen probe (see USEPA
Method 360.1, USEPA 1979b), or reagents for modified Winkler analysis.

6.10  Brine Shrimp (Artemia) Cysts — see Peltier and Weber (1985).

6.10.1  Although there are many commercial  sources of brine shrimp eggs,  the
Brazilian or Colombian strains are preferred because the supplies  examined
have had low concentrations of chemical residues.   (A source of brine shrimp
eggs that has been found to be satisfactory is Aquarium Products,  180 L Penrod
Ct., Glen Burnie, MD, 21061).  Each new batch of Artemia cysts should be
evaluated for nutritional  suitability against known suitable reference cysts
by performing a larval growth test.  It is recommended that a sample of
newly-hatched Artemia nauplii from each new batch of cysts be chemically
analyzed to determine that the concentration of total organic chlorine does
not exceed 0.15 ug/g wet weight or the total concentration of organochlorine
pesticides plus PCBs does not exceed 0.3 ug/g wet weight.  If those values are
exceeded, the Artemia should not be used.

6.10.2  Limited quantities of reference Artemia cysts, information on
commerical sources of good quality 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.

7.  TEST ORGANISMS

7.1  Fathead minnow larvae are used for the test (for fathead minnow culturing
methods, see Peltier and Weber, 1985).

8.  SAMPLE COLLECTION, PRESERVATION AND STORAGE

8.1  See Section 8, Effluent and Receiving Water Sampling and Sample Handling.

9.  CALIBRATION AND STANDARDIZATION

9.1  See Section 4, Quality Assurance.

10.  QUALITY CONTROL

10.1  See Section 4, Quality Assurance.

11.  TEST PROCEDURES

11.1  TEST SOLUTIONS

                                       36

-------
11.1.1  Surface Waters

11.1.1.1  Surface water toxicity is  determined with  samples 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, depending on the test volumes  selected.

11.1.2  Effluents

11.1.2.1  The selection of the effluent test concentrations should be based on
the objectives of the study.  One of two dilution  factors, approximately 0.3
or 0.5, is commonly used.  A dilution factor of  approximately  0.3 allows
testing between 100% and 1% effluent using  only  five effluent  concentrations
(100%, 30%, 10%, 3%, and 1%).   This  series  of dilutions minimizes the level of
effort, but because of the wide interval  between test concentrations provides
poor test precision (+ 300%).   A dilution factor of  0.5 provides greater
precision (+ 100%), but requires several  additional  dilutions  to span the same
range of effluent concentrations.  Improvements  in precision decline rapidly
as the dilution factor is increased  beyond  0.5

11.1.2.2  If the effluent is known or suspected  to be highly toxic, a lower
range of effluent concentrations should be  used, beginning at  10%.  If a high
rate of mortality is observed during the first 1 to  2 h of the test,
additional dilutions at the lower range of  effluent  concentrations can be
added.

11.1.2.3  Based on a 0.3 dilution factor, the volume of effluent required for
daily renewal of four replicates per concentration,  each  containing 250 ml of
test solution, would be approximately 1500  ml for  a  screening  test with 100%
effluent and a control, and 2.5 L for a definitive test with five
concentrations of effluent and a control.  Sufficient test solution
(approximately 400 rnL) 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.

11.2  START OF THE TEST

11.2.1  On-site tests should be initiated within 24  h of  sample collection,
and off-site tests should be initiated within 36 h of sample collection.  Just
prior to testing, the temperature of the sample  should be adjusted to (25 +_
1°C) and maintained at that temperature until portions are added to the
dilution water.

11.2.2  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.

                                       37

-------
11.2.3  Randomize the position of test chambers at the beginning  of  the  test.

11.2.4  The larvae are pooled and placed one to four at a  time  into  each test
chamber in sequential order, until  each chamber contains 15 (minimum of  10)
larvae, for a total of 60 larvae (minimum of 30) for each  concentration.   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.

11.3  LIGHT, PHOTOPERIOD AND TEMPERATURE

11.3.1  The light quality and intensity should be at ambient laboratory  levels,
which is approximately 10-20 uE/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 ± IOC.

11.4  DISSOLVED OXYGEN (DO)

11.4.1  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 not fall below 40% saturation.  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.

11.5  FEEDING

11.5.1  The fish in each test chamber are fed 0.1 mL (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 mL are fed
twice daily at an interval of 6 h.

11.5.2  The feeding schedule will depend on when the test  solutions  are
renewed.  If the test is initiated after 1200 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.

11.5.3  The nauplii should be rinsed with freshwater before use.   The amount of
food provided in each feeding should be sufficient to ensure the  presence  of a
small amount of uneaten food at the next feeding.

11.6  DAILY CLEANING OF TEST CHAMBERS

11.6.1  At the time of 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
                                        38

-------
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.  A note of this should be made in the log.

11.7  TEST SOLUTION RENEWAL

11.7.1  For on-site tests, test solutions are renewed daily with freshly
collected samples.  For off-site tests, test solutions are also renewed daily,
using the most recently collected sample.  A minimum of three  samples are
collected, preferrably for use beginning on  Days  1,  3, 5.  The first sample is
used for test initiation on Day 1 and test solution  renewal on Day 2.  The
second sample is used for test solution renewal on Days 3 and  4, and the third
sample is used for test solution renewals on Days 5, 6, and 7.  Samples first
used on Days 1, 3, and 5, are held over in the  refrigerator for use on the
following day(s).

11.7.2  Several sample shipping options are  available, including Express Mail,
air express, bus, and courier service.  Express Mail is delivered seven days a
week.  For private carriers, shipping and receiving  schedules  on weekends vary
with the carrier.

11.7.3  The test solutions are renewed immediately after cleaning the test
chambers.  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 subjecting
the larvae to excessive turbulence.

11.8  ROUTINE CHEMICAL AND PHYSICAL ANALYSIS

11.8.1  At a minimum, the following measurements  are made:

11.8.1.1  DO and pH are measured at the beginning and end of each 24-h
exposure period in one test chamber at the high,  medium, and low test
concentrations, and in the control.

11.8.1.2  Temperature should be monitored continously or observed and recorded
daily for at least two locations in the environmental control  system or the
samples.

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

11.8.1.4  Record the data (as shown in Figure 1).

11.9  OBSERVATIONS DURING THE TEST

11.9.1  The number of live and dead larvae in each test chamber are recorded
daily (see Figure 2 of this Section), and the dead larvae are  discarded.

                                       39

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

11.10  TERMINATION OF THE TEST

11.10.1  The test is terminated after seven  days  of exposure.  At test
termination, the surviving larvae in  each  test chamber (replicate) are counted
and prepared as a group for dry weight determination,  or are  preserved in 70%
ethanol for later analysis.   Inmediately prior to the  dry weight  analysis,
each group of larvae is rinsed with distilled water to remove food particles,
transferred to a tared weighing boat, and  dried at  100°C for  a minimum of
2 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 (see
Figure 3).  If the larvae are preserved, they must  be  dried and weighed within
two weeks.

11.10.2  Prepare a summary table as illustrated in  Figure 4.

11.11   ACCEPTABILITY OF TEST RESULTS

11.11.1  For the test results to be acceptable, survival  in the controls must
be at least 80%.  In tests initiated  with  larvae  less  than 24-h old, the
average dry weight of control larvae  surviving at the  end of  the  test should
equal  or exceed 0.25 mg.

11.12  SUMMARY OF TEST CONDITIONS

11.12.1  A summary of test conditions is listed in  Table 1.

12.  DATA ANALYSIS

12.1  GENERAL

12.1.1   Tabulate and summarize the data.   A  sample  set of survival and growth
response data is listed in Table 2.

12.1.2  The endpoints of toxicity tests using the fathead minnow  larvae are
based on the adverse effects on survival and growth.   Point estimates, such as
LCs and ICs, are calculated using point estimation  techniques (see
Section 2).  LOEC and NOEC values, for survival and growth, are obtained using
a hypothesis test approach such as Dunnett's Procedure (Dunnett,  1955) or
Steel's Many-one Rank Test (Steel, 1959; Miller,  1981).   See  the  Appendices
for examples of the manual computations and  data  input and program output for
the computer programs.

12.1.3  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 the  Appendices.   The  assistance of a
statistician is recommended for analysts who are  not proficient in statistics.

                                       40

-------
     TABLE 1.  SUMMARY OF RECOMMENDED EFFLUENT TOXICITY  TEST  CONDITIONS
               FOR THE FATHEAD MINNOW (PIMEPHALES  PROMELAS) LARVAL SURVIVAL
               AND GROWTH TEST
 1. Test type:
 2. Temperature (OC):
 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:
10. No. larvae per test chamber:
11. No. replicate chambers
     per concentration:
12. No. larvae per concentration
13. Feeding regime:
14. Cleaning:

15. Aeration:
Static renewal
25 + 1oc
Ambient laboratory illumination
10-20 uE/m2/s (50-100 ft-c)(ambient
laboratory levels)
16 h light, 8 h darkness
500 mL
250 mL/replicate

Daily
Newly hatched larvae less than 24 h old.
15 (minimum of 10)

4 (minimum of 3)
60 (minimum of 30)
Feed 0.1 mL newly hatched (less than 24-h
old) brine shrimp nauplii three times
daily at 4-h intervals or, as a minimum,
0.15 mL twice daily, 6 h between feedings
(at the begining of the work day prior to
renewal, and at the end of the work day
following renewal).  Sufficient larvae are
added to provide an excess.   Larvae are
not fed during the final 12 h of the test
Siphon daily, immediately before test
solution renewal
None, unless DO concentration falls below
40% saturation.  Rate should not exceed
100 bubbles/min
                                       41

-------
     TABLE 1.  SUMMARY OF RECOMMENDED EFFLUENT TOXICITY TEST  CONDITIONS
               FOR FATHEAD MINNOW (PIMEPHALES  PROMELAS) LARVAL SURVIVAL
               AND GROWTH TEST (CONTINUED)
16. Dilution water:


17. Effluent concentrations
18. Dilution factor:1
19. Test duration:
20f Endpoints:
21. Test acceptability

22. Sampling requirement:
23. Sample volume required:
Moderately hard synthetic water is prepared
using MILLIPORE MILLI-QR or equivalent
deionized water and reagent grade chemicals
or 20% DMW (see Section 7)
Minimum of 5 and a control
Approximately 0.3 or 0.5
7 days
Survival and growth (weight)
80% or greater survival in controls; Average
dry weight of surviving controls equals or
exceeds 0.25 mg
For on-site tests, samples are collected
daily, and used within 24 h of the time they
are removed from the sampling device.  For
off-site tests, a minimum of three samples
are collected, and used as described in
Paragraph 11.7.1
2.5 L/day
^Surface water test samples are used as  collected  (undiluted).
                                       42

-------
       Figure 1.  Data form for the fathead minnow larval  survival
                 and growth test.   Routine chemical  and physical
                 determinations.
Discharger:
Location:
Test Dates:
Analyst:

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

Day
1










2










3










4










5










6











7











Remarks











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

Daj
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










                                       43

-------
       Figure 1.   Data form for the fathead minnow larval  survival  and growth
                   test.  Routine chemical  and physical  determinations.
                   (Continued).
Discharger:
Location:
Test Dates:
Analyst:

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

Day
1










2










3










4










5










6










7











Remarks











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

Day
1










2










3










4










5










6










7











Remarks











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

Day
1










2










3










4










5










6










7











Remarks










                                       44

-------
  Figure 2. Survival  data for fathead minnow larval  survival  and growth test.
Discharger:
Location:
Test Dates:
Analyst:

Cone: Rep.
No.
Control



Cone:



Cone:



Cone:



Cone:



Cone:



No. Survivors
Day
1
























2
























3
























4
























5

























6
























7


























Remarks
























Comments:
                                       45

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
























A
Wgt. of
boat
(mg}
























B
Dry wgt:
foil and
larvae
(mg)
























B-A
Total dry
wgt of
larvae
(mg)
























C
No. of
larvae
























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
























Remarks
























1 Adapted from Hughes,  et al.,  1987.

-------
       Figure 4. Summary data for fathead minnow larval  survival
                 and growth test."1
Discharger:
Location:
Test Dates:
Analyst:
Treatment
No. live
1 arvae
Survival
I3L\
\%J
Mean dry wgt
of larvae (mg)
±SD
Temperature
range (°C)
Dissolved
oxygen range
(mg/L)
Hardness
Conductivity
Control















































Comments:
^Adapted from Hughes et al., 1987.
                                       47

-------
12.2 EXAMPLE OF ANALYSIS OF FATHEAD MINNOW SURVIVAL DATA

12.2.1   Formal statistical  analysis of the survival data is outlined 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 LCI, LC5, LC10 and LC50 endpoints. Concentrations at which there is no
survival in any of the test chambers are excluded from statistical  analysis of
the NOEC and LOEC, but included in the estimation of the LC endpoints.

12.3.2  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.
      TABLE 2. SUMMARY OF SURVIVAL AND GROWTH DATA FOR FATHEAD MINNOW
               LARVAE EXPOSED TO A REFERENCE TOXICANT FOR SEVEN DAYS'!
NaPCP
Cone.
(ug/L)
0
32
64
128
256
512
Proportion of
Survival in Repl
Chambers
A
1.0
0.8
0.9
0.9
0.7
0.4
B
1.0
0.8
1.0
0.9
0.9
0.3
C
0.9
1.0
1.0
0.8
1.0
0.4
icate
D
0.9
0.8
1.0
1.0
0.5
0.2
Mean
Prop.
Surv
0.95
0.85
0.975
0.90
0.775
0.325
Ave Dry Wgt (mg) In
Replicate Chambers
A
0.711
0.646
0.669
0.629
0.650
0.358
B
0.662
0.626
0.669
0.680
0.558
0.543
C
0.718
0.723
0.694
0.513
0.606
0.488
D
0.767
0.700
0.676
0.672
0.508
0.495
Mean
Dry Wgt
(mg)
0.
0.
0.
0.
0.
0.
714
674
677
624
580
471
^Four replicates of 10 larvae each.

12.3.3  If unequal  numbers of replicates occur among the concentration  levels
tested, there are parametric and nonparametric alternative analyses.   The
parametric analysis is the Bonferroni T-test (see Appendix D).   The Wilcoxon
Rank Sum Test with the Bonferroni adjustment is the nonparametric alternative
(see Appendix F).
                                       48

-------
STATISTICA
+
PROBIT
ANALYSIS
1
ENDPOINT ESTIMATE
LCI. LC5, LCIO. LC50
NORMAL
HOMOGENEOUS VARIANCE
^. 	 EQUAL
REP
L. ANALYSIS OF FATHEAD MINNOW LARVAL
SURVIVAL AND GROWTH TEST
SURVIVAL
SURVIVAL DATA
PROPORTION SURVIVING
1
I

ARCSIN
TRANSFORMATION
1

	 * 	 NUN-PtU
QUAPTQCI-UTI k* Q TFTT

DISTRIBUTION 1
rBARTLETT

NUMBER OF
LICATES?
1 -1
T-TEST HITH mN
BONFERRONI ***"•
ADJUSTMENT





NETT'S STEEL'S
•EST RANK

1



-
EQUAL NUMBER
REPLICATES?
1 YES
MANY-ONE WIL
TEST BONFE



ENDPOINT ESTIMATES
NOEC. LOEC

RMAL DISTRIBUTION
HETEROGENEOUS
VARIANCE
f
3F ^°
V
COXON RANK SUM
TEST HITH
RRONI ADJUSTMENT




Figure 5.  Flow chart for statistical  analysis of fathead
          minnow larval  survival  data.
                        49

-------
12.3.4  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, use the
graphical or other appropriate method.

12.3.5  Example of Analysis of Survival Data

12.3.5.1  This example uses the survival data from the Fathead Minnow Larval
Survival and Growth Test.  The proportion surviving in each replicate must
first be transformed by the arc sine square root transformation procedure
described in Appendix B.  The raw and transformed data, means and standard
deviations of the transformed observations at each toxicant concentration and
control are listed in Table 3.  A plot of the survival proportions is provided
in Figure 6.
                     TABLE 3.  FATHEAD MINNOW SURVIVAL DATA
                                        NaPCP Concentration (ug/L)
          Replicate
Control
32
64
128
256
512

RAW


ARC SINE
TRANS-
FORMED

MEAN(Y.j)
Si2
1
A
B
C
D
A
B
C
D



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
8
8
0
8
107
107
412
107
183
0232

0.9
1.0
1.0
1.0
1.249
1.412
1.412
1.412
1.371
0. 0066
3
0.
0.
0.
1.
1.
1.
1.
1.
1.
0.
4
9
9
8
0
249
249
107
412
254
01 bb

0.
0.
1.
0.
0.
1.
1.
0.
1.
0.
b
7
9
0
5
991
249
412
785
109
U/68

0.4
0.3
0.4
0.2
0.685
0.580
0.685
0.464
0.604
0.0111
6
12.2.6  Test for Normality

12.2.6.1  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.
                                       50

-------
                                                 L9
                                         SURVIVAL PROPORTION
           o  -
to
 c
 Cl



 •o
 fj

 o
n>


3

(/»
C
-s
<


<
Q>
•a

o
-o
o

ft
 a.
 cu
 c*
cr


fD
       o
       g

       i  2
       T)
       5
       m

       1

      "c   oo

      §
           K)

           U)
           O>
           o
-D
a»
mm
                                                                                           — CD— Z
                                                                                           •nmai
                                                                                           -m-— <
                                                                                           r»iQ-i>-
                                                                                           »*— r-
                                                                                           Zr-co
                                                                                              zm
                                                                                              zz

-------
         TABLE 4.  CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S  EXAMPLE
                                   NaPCP Concentration (ug/L)
Replicate Control
A 0.
B 0.
C -0.
D -0.
082
082
081
081
32
-0.
-0.
0.
-0.
076
076
229
076
64
-0.
0.
0.
0.
122
041
041
041
128
-0.
-0.
-0.
0.
005
005
147
158
256
-0.
0.
0.
-0.
118
140
303
324
51
0.
-0.
0.
-0.
2
081
024
081
140
12.2.6.2  Calculate the denominator,  D,  of the  statistic:


                              T  -  X)2
                          n
                      D = 2
    Where   ^j = the ith centered observation
            X  = the overall  mean of the centered observations
            n  = the total  number of centered observations
12.2.6.3  For this set of data:
                                   n  =  24

                                  "X  =   1   (0.000) = 0.000
                                       24

                                   D =  0.4265

12.2.6.4  Order the centered observations from smallest to  largest

                           - ...  - x ...  a^ where k  is approximately  n/2. For  the
data in this example, n = 24 and k = 12.   The aj  values are listed in
Table 6.
                                       52

-------
  TABLE 5.   ORDERED  CENTERED OBSERVATIONS FOR THE SHAPIRO-WILK'S EXAMPLE
                         x(D
                              x(D
1
2
3
4
5
6
7
8
9
10
n
12
-0.324
-0.147
-0.140
-0. 1 22
-0.118
-0.081
-0. 081
-0. 076
-0.076
-0.076
-0. 024
-0. 005
13
14
15
16
17
18
19
20
21
22
23
24
-0.005
0.041
0.041
0.041
0.081
0.081
0.082
0.082
0.140
0.158
0.229
0.303
12.2.6.6  Compute the test  statistic, W, as follows:
                       k
               W =1 [2  a*  (X(n-i+l) . x(D) ]2
                   D  i=l 1


The differences x(n-1+l)  -  x(D are listed in Table 6.
in this example,

                    1
                             For the data
               W =
                   0.4265
(0.6444)2  =  0.974
TABLE 6.   COEFFICIENTS AND  DIFFERENCES FOR SHAPIRO-WILK'S EXAMPLE
                                      - x(D
1
2
3
4
5
6
7
8
9
10
n
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)
X(13)
- x(D
- X(2)
- X(3)

-X(5)
- X(6)
- X<7>
- X(8)
- X^9)
- x'10)
- x^D
- X(12)
                                    53

-------
12.2.6.7  The decision rule for this test is to compare W as calculated  in
12.2.6.6 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.

12.2.7 Test for Homogeneity of Variance

12.2.7.1  The test used to examine whether the  variation  in mean proportion
surviving is the same across all toxicant concentrations  including  the
control, is Bartlett's Test (Snedecor and Cochran, 1980).  The test statistic
is as follows:
                   P             P
               [ (  2 Vi) In S2 - 5 V,-  In Sf2 ]
           B =
    Where V-j =   degrees of freedom for each toxicant concen-
                 tration and control,  V-,* = (n-f  - 1)
          n-j = the number of replicates for concentration i.
          In = loge
          i  = 1, 2, ..., p where p is the number of concentrations
               including the control
                  ( S Vi Si2)
          $2 =     1=1	

                      P
                      T» U.

                     1=1



          C  = 1  + ( 3(p-l)H  [ 2 1/Vi  - (  S Vi)-1
12.2.7.2 For the data in this example, (See Table 3) all  toxicant
concentrations including the control  have the same number of replicates
(n-j =4 for all i).  Thus, Vi = 3 for all i.
                                       54

-------
12.2.7.3  Bartlett's statistic is therefore:
       B =  [{18)ln(0.0236) - 3 S IntSf  2)]/1.1296
         =  [18(-3.7465) - 3(-24.7516)]/1.1296

         =  6.8178/1.1296

         =  6.036

12.2.7.4  B is approximately distributed as chi  square with  p  -  1  degrees  of
freedom, when the variances are in fact the same.   Therefore,  the  appropriate
critical value for this test, at a significance  level  of 0.01  with 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.

12.2.8  Dunnett's Procedure

12.2.8.1  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
df Sum of Squares
(SS)

p - 1 SSB

N .- p SSW
N - 1 SST
Mean Square(MS)
(SS/df)
2
SB = SSB/(p-l)
2
SW = SSW/(N-p)

Where:      p  = number toxicant concentrations including  the  control
            N  = total number of observations n-|  + r\2 ...  +np
            n-  = number of observations in concentration i
           SSB = S Tf2/nf - G2/N
                               Between Sum of Squares
SST =
                            - G2/N
Total Sum of Squares
           SSW = SST - SSB
                               Within Sum of Squares
                                    55

-------
            G  = the grand total of all sample observations,  G = 2 TJ
                                                                i=l
            TJ = the total of the replicate measurements for
                 concentration "i"
           Y-JJ = the jth observation for concentration "i"  (represents
                 the proportion surviving for toxicant concentration
                 i in test chamber j)

12.2.8.2  For the data in this example:

    n  = n  = n3 = n4 = ns = ng = 4
N  =
T! -
T2 =

T4 =
T5 •
T6 =

G  =


SSB =
         24
             + Y12 + Y13 + Y14 = 5.322
             + Y22 + Yga + Y24 = 4.733
             + Y32 + Y33 + Y34 = 5.485
             + Y42 + Y43 + Y44 = 5.017
         Y51 + Y52 + Y53 + Y54 = 4.437
             + Y62 + Y63 + Y64 = 2.414

            + T2 + T3 + T4 + 15 + T 6 = 27.408
          S Tj2/ni _ Q2/N
         1=1

          _1_(131.495) - (27.408)2  =1.574
           4                 24
          P   "i
    SST = S   S
         1=1 j=l
                 - G2/N
        = 33.300 - (27.408)2  = 2.000
                      24

    SSW = SST - SSB = 2.000 - 1.574 = 0.426
    SB2 = SSB/p-1  = 1.574/6-1 = 0.315
    Sw2 = SSW/N-p = 0.426/24-6 = 0.024

12.2.8.3 Summarize these calculations in the ANOVA table  (Table  8).
                                    56

-------
             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

12.2.8.4  To perform the individual  comparisons,  calculate  the  t  statistic for
each concentration, and control  combination as follows:
                         t-i —
                                      Y1  - Yl
                                Swx/ (1/nD  +
Where Yi  = mean proportion surviving for concentration i
      YI  = mean proportion surviving for the control
      $W  = square root of within mean sqaure
      n-|  = number of replicates for control
      n-j  = number of replicates for concentration i.

12.2.8.5  Table 9 includes the calculated t values for each  concentration  and
control combination.  In this example, comparing the 32 ug/L concentration
with the control the calculation is as follows:
                            {  1.330 - 1.183 )
                                                 = 1.341
                       [ 0.155V U/4J  + (1/4)   ]
                       TABLE 9.   CALCULATED T VALUES
NaPCP Concentration ug/L)
32
64
128
256
512
i
2
3
4
5
6
ti
1.341
-0.374
0.693
2.016
6.624
                                    57

-------
12.2.8.6 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-j is greater than the critical
value.  Since tg is greater than 2.41,  the 512 ug/L concentration has
significantly lower survival than the control.   Hence the  NOEC and  the LOEC
for survival  are 256 ug/L and 512 ug/L, respectively.

12.2.8.7  To quantify the sensitivity of the  test, the  minimum significant
difference (MSD) that can be detected statistically may be calculated.
Where: d
       sv
       n
                   MSD = d SW V (1/ni)  + (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
            (this assumes equal replication  at  each  concentration)
            the number of replicates in the  control.
12.2.8.8  In this example:
                   MSD = 2.41  (0.155) >/ (1/4)  + (1/4)
                       = 2.41  (0.155M0.707)
                       = 0.264

12.2.8.9  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 (MSDy)  is  determined by subtracting the
       untransformed values from 2.

                          MSDU = 0.943  - 0.766 = 0.177

12.2.8.10  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.
                                       58

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

12.2.9  Probit Analysis

12.2.9.1  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.

12.2.9.2  For this example, since there is 100%  survival in the controls,
there is no need to adjust for control  mortality.   The test for heterogeneity
was not significant, thus Probit Analysis appears  appropriate for this data.


                       TABLE 10.  DATA FOR  PROBIT ANALYSIS
NaPCP Concentration (ug/L)

Number Dead
Number Exposed
Control
2
40
32 64
6 1
40 40
1 28 256
4 9
40 40
512
27
40
                                       59

-------
TABLE 11.   OUTPUT FROM  EPA  PROBIT ANALYSIS  PROGRAM,  VERSION 1.4,
                     USED FOR CALCULATING  EC VALUES
                   EEA PRDBTT ANALYSIS PROGRM1
                 USED FOR OSCULATING EC VALUES
                          Versicn 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
          Observed
Member   Proportion
Resp.    Responding
    2
    6
    1
    4
    9
   27
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
  0.6452
 Chi - Square Heterogeneity
                    4.522
Mu       =     2.626029
Sigma     =     0.223555

Parameter       Estimate    Std.  Err.
                                 95% Confidence Limits
 Intercept
 Slope

 Spontaneous
 Response Rate
   -6.746692
    4.473178

    0.078182
    3.112017
    1.196026
     (  -12.846246,
     (    2.128967,
    0.022541    (    0.034002,
           -0.647139)
            6.817389)

            0.122363)
      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.

      127.6359
      181.2605
      218.5347
      247.9341
      422.6964
      720.6437
      817.5905
      985.7196
     1399,8575
              lower      Upper
            95% Conf idence  Limits
            34.5885
            71.4914
           104.8182
           135.2143
           345.7290
           562.5553
           616.4054
           702.9269
           893.3054
                195.
                248.
                284.
                311.
                531.
               1420.
               1836.
               2696.
               5581.
        4335
        7074
        0806
        8864
        0254
        7512
        3506
        8005
        7588
                                       60

-------
Frctoit Analysis of Fathead Minnow larval Survival Data

       PLOT OP ADJUSTED PROBITS  AND H3EDICTED REGRESSION UNE

Ercbit
    9-f
    8+
    7+
    St-
    4+
    -o

    -o

    3+
   2+
    1+
                       o....
    (HO
     EC01
BdO    EC25     EC50     ECT75     EC90
EC99
Figure  7.  Plot of adjusted probits  and predicted  regression  line
                .         from EPA Probit Program
                                     61

-------
12.3  EXAMPLE OF ANALYSIS OF FATHEAD MINNOW GROWTH DATA

12.3.1  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 2).  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.

12.3.2  The statistical analysis using hypothesis tests consists of a
parametric test, Dunnett's Procedure, and a non-parametric 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 non-parametric
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.

12.3.3  Additionally, if unequal numbers of replicates occur among the
concentration levels tested there are parametric and non-parametric
alternative analyses.  The parametric analysis is the Bonferroni T-test (see
Appendix D). The Wilcoxon Rank Sum Test with the Bonferroni adjustment is the
non-parametric alternative (see Appendix F).

12.3.5  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 ug/L concentration, its effect on growth is
not considered.
                   TABLE 12.  FATHEAD MINNOW GROWTH DATA
                                   NaPCP Concentration (ug/L)
Replicate    Control
32     64     128    256
512




A
B
C
D
Mean(Y1->
Si2
i
•

0.
0.
0.
0.
0.
0.
1
711
662
718
767
714
0018

0
0
0
0
0
0

.646
.626
.723
.700
.674
.0020
2
0.
0.
0.
0.
0.
0.
3
669
669
694
676
677
0001

0.
0.
0.
0.
0.
0.
4
629
680
513
672
624
0059

0.650
0.558
0.606
0.508
0.580
0. 0037
5

-
-
—

-
6
                                    62

-------
                               m
  POINT ESTIMATION

   HYPOTHESIS TESTIN6
(EXCLUDING CONCENTRATIONS
ABOVE NOEC FOR SUHVIVAU
  ENDPOINT ESTIMATE
     IC25.  IC50
                        SHAPIRO-MILK'S TEST
                     NON-NORMAL DISTRIBUTION
             NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
                           BARTLETT'S TEST
                           HETEROGENEOUS
                              VARIANCE
               EQUAL NUMBER OF
                 REPLICATES?
             EQUAL NUMBER OF
               REPLICATES?
                 YES
                YES
                 DUNNETT'S
                   TEST
    STEEL'S MANY-ONE
       RANK TEST
  HILCOXON RANK SUM
      TEST WITH
BONFERRONI ADJUSTMENT
                          ENDPOINT ESTIMATES
                               NOEC. LOEC
  Figure 8. Flow chart for statistical  analysis  of fathead minnow
                        larval growth data.
                                 63

-------
CTi
            0.77-,
                                                        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 larval  survival and growth test,

-------
12.3.6  Test for Normality

12.3.6.1  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 (ug/L)
Replicate

12.3

A
B
C
D
.6.2 Calculate

Control
-0.003
-0.052
0.004
0.053
32
-0.028
-0. 048
0.049
0.026
the denominator, D,
n
D = S (Xi
I -X)2
64
-0.008
-0.008
0.017
-0. 001
of the test

128
0.005
0.056
-0.111
0.048
statistic:

256
0.070
-0. 022
0.026
-0.072


    Where X-j = the ith centered observation
          IT  - the overall mean of the centered observations
          n  = the total number of centered observations.

For this set of data,            n = 20

                                 J = J_(0.000) = 0.000
                                     20
                                 D = 0.0412

12.3.6.3 Order the centered observations from smallest to  largest

                       - X<2) - ... - X(n)
Where X    is the ith ordered observation.  These ordered observations  are
listed in Table 14.

12.3.6.4  From Table 4, Appendix B, for the number of observations,  n,  obtain
the coefficients ai , 32* ...» a^ where k is approximately n/2.   For  the
data in this example, n = 20, k = 10.   The a-,- values are listed in Table  15.
                                       65

-------
  TABLE 14.  ORDERED CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S EXAMPLE
                                                        X(D
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
12.3.6.5 Compute the  test  statistic, W, as follows

                   1    k       ,   - nV    ,-V   o
               W = -  [  S a,-  (X
-------
12.3.6.6  The decision rule for this  test  is  to compare W with the critical
value found in Table 6, Appendix B.   If  the computed W is less than the
critical value, conclude that the data are not normally distributed.  For this
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.

12.3.7 Test for Homogeneity of Variance

12.3.7.1  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:
[ ( S
                     In S2 - S
                                      In
Where Vi =   degrees of freedom for each  toxicant concen-
             tration and control,  V-j  =  (ni  -  1)
           the number of replicates for concentration i.
           logg
           1, 2, ..., p where p is the  number of concentrations
           including the control
          "1 =
          In =
          1  **
          S2 =
                  ( 2 Vi Si 2)
                      P
                      2
         = i  + (  3(p-i)H  [ s
                            1=1
                                            P
                                            s
                                           1=1
12.3.7.2 For the data in this example, (See Table 12)  all  toxicant
concentrations including the control  have the same number  of  replicates
(nj = 4 for all i).  Thus, Vi - 3 for all i.
                                       67

-------
12.3.7.3  Bartlett's statistic is therefore:

                                p
       B =  [(15)1n(0.0027) - 3 2 1 n(Sf2)]/i.133


         =  [15(-5,9145) - 3{-32.4771 ]/l.133

         =  8.7138/1.133

         =  7.691

12.3.7.4  B is approximately distributed as chi  square with  p  -  1  degrees of
freedom, when the variances are in fact the same.   Therefore,  the  appropriate
critical value for this test, at a significance  level  of  0.01  with 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.

12.3.8  Dunnett's Procedure

12.3.8.1  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
       df
Sum of Squares
     (SS)
Mean Square(MS)
    (SS/df)
    Between
    Within
      p - 1
      N - p
     SSB
     SSW
SB = SSB/(p-l)
 2
SW = SSW/(N-p)
    Total
      N - 1
     SST
Where:      p  = number toxicant concentrations including  the  control
            N  = total  number of observations n-|  + r\2 •-•  +"n
            n-  = number of observations in  concentration i
SSB = 2
     1=1
                        j  - G2/N
              Between Sum of Squares
SST = 2
                            - Q2/N
              Total  Sum of Squares
           SSW = SST - SSB
                               Within Sum of Squares
                                    68

-------
            G  = the grand total  of all  sample observations, G = 2 Tj
                                                               i=0
            Tj = the total of the replicate measurements for
                 concentration "i"
           Y-jj = the jth observation for concentration  "i"  (represents
                 the mean dry weight of  the fish  for toxicant
                 concentration i  in test chamber  j)

12.3.8.2  For the data in this example:
ni  = n2 = n3 = n4 =
N  = 20
Tl  = YH + Yi2 + Y13
T2 = Y21 + Y22 + Y23
T3 =
T4 = Y41
T5 =

G  =
                           = 4

                           Y14 = 2.858
                               = 2.695
               Y32 + Y33 + Y34 = 2.708
             + Y42 + Y43 + Y44 = 2.494
             + Y52 + Y53 + Y54 = 2.322

            + T2 + T3 + T4 + TS = 13.077
    SSB = 2 T1-2/ni - ^
        = J_(34.376) - (13.077)2  = 0.044
           4                20
SST = 2   S
     1-1 j=l
                     - G2/N
        =  8.635 - (13.077)2  = 0.085
                      20

    SSW = SST - SSB = 0.085 - 0.044 = 0.041
    SB2 = SSB/p-1 = 0.044/5-1 = 0.011
    SW2 = SSW/N-p = 0.041/20-5 = 0.0027

12.3.8.3 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)
o.on
0.0027
    Total
              19
0.085
                                      69

-------
12.3.8.4  To perform the individual  comparisons,  calculate  the t  statistic for
each concentration, and control  combination as  follows:
                                SWN/ (1/ni)  + (1/nj)
Where Yj  = mean dry weight for toxicant concentration  i
      Y-j  = mean dry weight for the control
      SW  = square root of within mean sqaure
      n-j  = number of replicates for control
      n-j  = number of replicates for concentration  i.

12.3.8.5  Table 18 includes the calculated t  values for each concentration
and control combination.  In this example, comparing the  32 ug/L concentration
with the control the calculation is as follows:
                            {  0.714 - 0.674}
                                                 =  1.081
                       [ 0.052/(1/4)  + (1/4)   ]
                      TABLE 18.  CALCULATED T VALUES
                     NaPCP
                 Concentration
                     (ug/L)
32
64
128
256
2
3
4
5
1.081
1.000
2.432
3.622
12.3.8.6 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-j  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 the LOEC for growth are 64 ug/L and  128  ug/L,
respectively.
                                       70

-------
12.3.8.7  To quantify the sensitivity of the test, the minimum significant
difference (MSD) that can be  statistically detected may be calculated.
MSD = d Sw  /
                                        +  (l/n)
Where  d  = the critical  value for the  Dunnett's procedure
       $W = 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-j = the number of replicates in the  control.

12.3.8.8  In this example:
                   MSD = 2.36 (0.052)  /  (1/4)  +
                       = 2.36 (0.052M0.707)
                       = 0.087
                            TT7T)
12.3.8.9  Therefore, for this set of data,  the minimum difference that can be
detected as statistically significant is  0.087 mg.

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

13.  PRECISION AND ACCURACY

13.1  PRECISION

13.1.1  Information on the single laboratory precision of the fathead minnow
larval survival and growth test is presented in  Table 19.   The range of NOECs
was only two concentration intervals, indicating good precision.

13.1.2  An interlaboratory study of Method  1000.0 described in the first
edition of this manual (Horning and Weber,  1985), 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 20).  For the growth (weight)
NOECs, an average of 3Z& were at the median, and 84% were within one
concentration interval of the median (Table 21).  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 is  now included in  this revised
edition of the method, had been applied to  the results of the interlaboratory
study, 40 of the 135 completed tests would  have  been considered unacceptable
(Norberg-King, 1988).

13.2  ACCURACY

13.2.1  The accuracy of toxicity tests can  not be determined.

                                       71

-------
TABLE 19.  PRECISION OF THE FATHEAD MINNOW LARVAL SURVIVAL
          AND GROWTH TEST, USING NAPCP AS A REFERENCE TOXICANT*,b
NOEC
Test (ug/L)
1 256
2 128
3 256
4 128
5 128
LOEC
(ug/L)
512
256
512
256
256
Chronic
Value
(ug/L)
362
181
362
181
181
aFrom Pickering, 1988.
 For a discussion of the precision of data from chronic toxicity
 tests see Section 4, Quality Assurance.
                                    72

-------
       TABLE 20.  COMBINED FREQUENCY  DISTRIBUTION FOR SURVIVAL NOECs
                 FOR ALL LABORATORIES9
NOEC Frequency
Tests with Two
Sample
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
Median
35
42
47
41
26
37
56
±1b
53
42
47
41
68
53
33
Reps
>2C
12
16
6
18
6
10
n
(%) Distribution
Tests with Four Reps
Median
57
56
75
50
78
56
56
-Mb
29
44
25
50
22
44
33
>2C
14
0
0
0
0
0
n
aprom DeGraeve et. al., 1988.
bPercent of values within one  concentration  intervals  of the median.
°Percent of values within two  or more concentrations intervals of the median.
                                       73

-------
       TABLE 21. COMBINED FREQUENCY DISTRIBUTION FOR  WEIGHT  NOECs
                 FOR ALL LABORATORIES^
NOEC Frequency

1,
2.
3.
4.
5.
6.
7.
Sampl e
Sodium Pentachlorophenate (A)
Sodium Pentachlorophenate (B)
Potassium Dichromate (A)
Potassium Dichromate (B)
Refinery Effluent 301
Refinery Effluent 401
Utility Waste 501
Tests
Median
59
37
35
12
35
37
11
with Two
±1b
41
63
47
47
53
47
61
Reps
>2c
0
0
18
41
12
16
28
(%) Distribution
Tests with
Median +_
57
22
88
63
75
33
33
Four Reps
lb
43
45
0
25
25
56
56
>2C
0
33
12
12
0
11
11
aFrom DeGraeve et. al., 1988.
^Percent of values within one  concentration intervals  of the median.
cPercent of values within two  or more concentrations intervals of the median.
                                        74

-------
                                   SECTION  11

                                  TEST METHOD

                      FATHEAD MINNOW,  PIMEPHALES  PROMELAS,
                 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 and
larvae in an seven-day, static renewal test.   Th"e 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
degradeable and highly volatile toxicants,  such as  chlorine, in the source may
not be detected in the test.

1.4  This method should be restricted to use by or  under  the supervision of
professionals experienced in aquatic toxicity testing.

1.5  This test is commonly used in one of two forms: (Da definitive test,
consisting of a minimum of five effluent concentrations and  a  control, and  (2)
an abbreviated test, consisting of only one test  concentration, such as 100&
effluent or the instream waste concentration, and a control.  Abbreviated
tests are used for toxicity screening or a  pass/fail permit  condition.
Failure of the screening test usually results in  a  followup  definitive test.

2.  SUMMARY OF METHOD

2.1  Fathead minnow embryos and larvae 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
and Equipment).

                                       75

-------
7.  Test Organisms

7.1   Fathead minnow embryos,  less than 36-h  old,  are  used  for  the test  (for
fathead minnow culturing methods (see Peltier and  Weber,  1985).

7.2  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.  It is recommended that when possible,  the embryos be obtained
from local  sources.  Receipt of embryos via Express Mail, air express, or
other carrier, from a reliable outside source, is  an acceptable alternative.

7.3  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 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.

7.4  The test is conducted with four (minimum of three) test chambers at each
toxicant concentration and control.  Fifteen  (minimum  of  10) embryos are
placed in each replicate test  chamber.  Thus, 60  (minumum of 30) embryos are
exposed per test concentration.

8.  SAMPLE COLLECTION, PRESERVATION AND HANDLING

8.1  See Section 8, Effluent and Receiving Water Sampling and Sample Handling.

9.  CALIBRATION AND STANDARDIZATION

9.1  See Section 4, Quality Assurance.

10.  QUALITY CONTROL

10.1  See Section 4, Quality Assurance.

11.  TEST PROCEDURES

11.1  TEST SOLUTIONS
                                       78

-------
11.1.1   Surface Waters

11.1.1.1   Surface water toxicity  is  determined with  samples used directly as
collected.  Four (minimum of three)  replicate test chambers are used for each
surface water sample.

11.1.2  Effluents

11.1.2.1   The selection of the effluent test concentrations should be based on
the objectives of the  study.  One of two dilution factors, approximately 0.3
or 0.5, is commonly used.  A dilution factor of  approximately 0.3 allows
testing between 100% and 1% effluent using  only  five effluent concentrations
(100%,  30%, 10%, 3%, and 1%).  This  series  of dilutions minimizes the level of
effort, but because of the wide interval  between test concentrations provides
poor test precision (+_ 300%).  A dilution factor of  0.5 provides greater
precision (_+ 100%), but requires  several  additional  dilutions to span the same
range of effluent concentrations.  Improvements  in precision decline rapidly
as the dilution factor is increased  beyond  0.5.

11.1.2.2  If the effluent is known or suspected  to be highly toxic, a lower
range of effluent concentrations  should be  used  (such as  10%, 3%, 1%, 0.3%,
and 0.1%).

11.1.2.3  Based on a 0.3 dilution factor, the volume of effluent required for
daily renewal of four replicates  per concentration,  each  containing 200 ml of
test solution, is 1200 ml for a screening test with  100%  effluent and a
control, and 1800 ml for a definitive test  with  five effluent concentrations
and a control.  Sufficient test solution (approximately 1200 ml) at each
effluent concentration is prepared 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.

11.1.2.4 The hardness of the test solutions must exceed 25 mg/L (CaCOs) to
insure hatching success.  If the hardness of the effluent is less than 25 mg
CaC03/L, adjust the hardness by adding reagents  for  synthetic softwater
listed in Table 1, Section 7.

11.2  START OF THE TEST

11.2.1   On-site tests should be initiated within 24  h of  sample collection,
and off-site tests should be initiated within 36 h of sample collection.  Just
prior to testing, the temperature of the sample  should be adjusted to (25 +_
1°C) and maintained at that temperature until portions are added to the
dilution water.

11.2.2  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.

11.2.3  Add 10-15 embryos to each test chamber using a small bore (2mm) glass
tube calibrated to contain approximately the desired number of embryos.
Repeat the process until the required number of  embryos have been added to
each chamber.

                                       79

-------
11.2.4  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.

11.2.5  Randomize the position of the test chambers at the  beginning of
the test.

11.3  LIGHT, PHOTOPERIOD AND TEMPERATURE

11.3.1  The light quality and intensity  should  be at ambient laboratory
levels, approximately 10-20 uE/m2/s, or  50 to 100 foot candles  (ft-c),
with a photoperiod of 16 h of light and  8 h of  darkness.  The test
solution temperature should be maintained at 25 _+ 1°C.

11.4  DISSOLVED OXYGEN (DO)

11.4.1  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 not fall below 40% saturation. If it is  necessary
to aerate, all concentrations and the control should be  aerated.  The rate
should not exceed TOO bubbles/min, using a pipet with a  1-2 mm  orifice,
such as a 1-mL Kimax Serological  Pipet No. 37033, or equivalent.  Care
should be taken to ensure that turbulence resulting from the aeration does
not cause undue physical stress to the fish.

11.5  FEEDING

11.5.1  Feeding is not required.

11.6  DAILY CLEANING OF  TEST CHAMBERS

11.6.1  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.

11.7  TEST SOLUTION RENEWAL

11.7.1  For on-site tests, test solutions are renewed daily with freshly
collected samples.  For off-site tests,  test solutions are  also renewed
daily, using the most recently collected sample. A minimum of  three
samples are collected, preferrably for use beginning on  Days 1, 3, 5.  The
first sample is used for test initiation on Day 1 and test  solution
renewal on Day 2.  The second sample is  used for test solution  renewal on
Days 3 and 4, and the third sample is used for  test solution renewals on
Days 5, 6, and 7.  Samples first used on Days ls 3, and  5,  are  held over
in the refrigerator for use on the following day(s).

11.7.2  Several sample shipping options  are available, including Express
Mail, air express, bus,  and courier service.  Express Mail  is delivered
seven days a week.  For  private carriers, shipping  and receiving schedules
on weekends vary with the carrier.
                                     80

-------
11.7.3  The test solutions are renewed immediately  after  removing dead embryos
and/or larvae.  During the daily renewal  process, a small amount of water is
left in the chamber to ensure that the embryos and  larvae remain submerged
during the renewal  process.   New test solution should  be  added slowly to avoid
subjecting the embryos and larvae to excessive turbulence.

11.8  ROUTINE CHEMICAL AND PHYSICAL DETERMINATIONS

11.8.1  At a minimum, the following measurements are made:

11.8.1.1  DO and pH are measured at the beginning and  end of each 24-h
exposure period in one test chamber at the high, medium,  and low test
concentrations, and in the control.

11.8.1.2  Temperature should be monitored continously  or  observed and recorded
daily for at least two locations in the environmental  control system or the
samples.

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

11.8.1.4  Record the data (as shown in Figure 1).

11.9  OBSERVATIONS DURING THE TEST

11.9.1  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 non-concentration-related-mortality occurs,  terminate the test and
start a new test with new embryos.

11.9.2  At 25<>C, hatching begins on about 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
(non-distinct 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.

11.9.3  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.
                                       81

-------
11.10  TERMINATION OF THE TEST

11.10.1  The test is terminated after seven days of exposure.   Count  the
number of surviving, dead, and deformed larvae,  and record  the  numbers  of
each.  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.

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

11.11 ACCEPTABILITY OF TEST RESULTS

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

11.12 SUMMARY OF TEST CONDITIONS

11.12.1  A summary of test conditions is listed in Table  1.

12.  DATA ANALYSIS

12.1  GENERAL

12.1.1  Tabulate and summarize the data.

12.1.2  The endpoints of this toxicity test are based on  total  mortality,
combined number of dead embryos, and dead and deformed larvae.   Point
estimates, such as LCI, LC5, LC10 and LC50, are  calculated  using Probit
Analysis (Finney, 1971).  LOEC and NOEC values,  for total mortality,  are
obtained using a hypothesis test approach such as Dunnett's Procedure
(Dunnett, 1955} or Steel's Many-one Rank Test (Steel, 1959; Miller,
1981).  See the Appendices for examples of the manual computations and
examples of data input and output for the computer programs.

12.1.3  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.

12.2  EXAMPLE OF ANALYSIS OF FATHEAD MINNOW EMBRYO-LARVAL SURVIVAL AND
      TERATOGENICITY DATA

12.2.1  Formal statistical analysis of the total  mortality  data is
outlined 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 LCI, LC5, LC10 and LC50 endpoints.  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 LC endpoints.
                                     82

-------
        TABLE 1.  SUMMARY OF RECOMMENDED  EFFLUENT TOXICITY TEST
                 CONDITIONS FOR THE  FATHEAD MINNOW  (PIMEPHALES PROMELAS)
                 EMBRYO-LARVAL SURVIVAL  AND TERATOGENICITY TEST
 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:
16. Effluent test concentrations:
17. Dilution factor:^
Static renewal
25+_ loc
Ambient laboratory illumination
10-20 u£/m2/s or 50-100 ft-c (ambient
laboratory levels)
16 h light, 8 h dark
150-500 mL
70-200 mL
Daily
Less than 36-h old embryos
15 (minimum of 10}
4 (minimum of 3)

60 (minimum of 30)
Feeding not required
None unless DO falls below 40% saturation
Moderately hard synthetic water is
prepared using MILLIPORE MILLI-qR or
equivalent deionized water and reagent
grade chemicals or 20% DMW (see Section
7).  The hardness of the test solutions
must equal or exceed 25 mg/L (CaC03) to
ensure hatching.
5 and a control
Approximately 0.3 or 0.5
^Surface water test samples are used as  collected (undiluted).
                                       83

-------
     TABLE 1.  SUMMARY OF RECOMMENDED EFFLUENT TOXICITY TEST  CONDITIONS
              FOR FATHEAD MINNOW (PIMEPHALES PRQMELAS) EMBRYO-LARVAL
               SURVIVAL AND TERATOGENICITY TEST (CONTINUED)
18. Test duration:

19. Endpolnt:

20. Test acceptability

21. Sampling requirement:
7 days

Combined mortality (dead and deformed organisms)

80% or greater survival  in controls

For on-site tests, samples are collected dally
and used within 24 h of the time they are
removed from the sampling device.  For off-site
tests a minimum of three samples are collected
and used as described in Paragraph 11.7.1.
22.  Sample volume required:     2.5 L/day

-------
       Figure 1.  Data form for the fathead minnow embryo/larval
                 survival  and teratogenicity test.  Routine
                 chemical  and physical  determinations.
Discharger:
Location:
Test Dates:
Analyst:


Control :
Temp.
D.O. Initial
Final
pH Initial
Final
Alkalinity
Hardness
Conductivity
Chlorine
%
Day
1











2











3











4











5











6











7












Remarks













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

Day
1











2











3











4











5











6











7












Remarks













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

Day
1











2











3











4











5











6











7












Remarks











                                       85

-------
       Figure 1. Data form for the fathead minnow embryo/larval
                 survival  and teratogenicity test.  Routine
                 chemical  and physical  determinations.   (Continued)
Discharger:
Location:
Test Dates
Analyst:


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

Day
1











2











3











4











5











6











7












Remarks










*


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

Day
1











2











3











4











5











6











7












Remarks













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

Day
1











2











3











4











5











6











7












Remarks











                                       86

-------
     Figure 2.  Data form for the fathead minnow embryo/larval
                survival and teratogenicity test.  Survival  and
                terata data.
Discharger:
Location:
Test Dates
Analyst:
Condition of
Rep Embryos/Larvae
Control: 1 Live/dead
Terata
2 Live/dead
Terata
3 Live/dead
Terata
4 Live/dead
Terata
Treat: 1 Live/dead
Terata
2 Live/dead
Terata
3 Live/dead
Terata
4 Live/dead
Terata
Treat: 1 Live/dead
Terata
2 Live /dead
Terata
3 Live/dead
Terata
4 Live/dead
Terata
Treat: 1 Live/dead
Terata
2 Live/dead
Terata
3 Live /dead
Terata
4 Live/dead
Terata
Day
1
































2
































3
































4
































5
































6
































7
































                                       87

-------
     Figure 2.
Discharger:
Location:
Data form for the fathead minnow embryo/larval
survival and teratogenicity test.  Survival  and
terata data. (Continued)
                          Test Dates
                          Analyst:

Rep Embryos -Larvae
Treat: 1 Live/dead
Terata
2 Live/dead
Terata
3 Live/dead
Terata
4 Live/dead
Terata
Treat: 1 Live/dead
Terata
2 Live/dead
Terata
3 Live/dead
Terata
4 Live/dead
Terata
Day
1
















2
















3
















4
















5
















6
















7
















                                       88

-------
       Figure 3.     Data form for the fathead minnow embryo/larval
                    survival  and teratogencity test.  Summary data
Discharger:
Location:
Test Dates
Analyst:
Treatment
No. dead embryos
and larvae
No. terata
Total mortality
(dead and deformed
organisms)
Total mortality (%}
Terata (%}
Hatch (%)
Control









































Comments:
                                       89

-------
12.2.2  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.

12.2.3  If unequal numbers of replicates occur among the concentration levels
tested, there are parametric and nonparametric alternative analyses.   The
parametric analysis is the Bonferroni T-test (see Appendix D).   The Wilcoxon
Rank Sum Test with the Bonferroni adjustment is the nonparametric alternative
(see Appendix F).

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

12.2.5  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.OX concentration, it  is
not included in this statistical analysis and is considered a qualitative
mortality effect.

12.2.6  Test for Normality

12.2.6.1  Since only two replicates were run at each concentration level, the
test for normality is invalid.  Additionally, a non-parametric  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.

12.2.7  Dunnett's Procedure

12.2.7.1  To obtain an estimate of the pooled variance for the  Dunnett's
Procedure, construct an ANOVA table as described in Table 4.
                                       90

-------
           TABLE 2.  DATA FROM FATHEAD MINNOW EMBRYO-LARVAL  TOXICITY
                    TEST WITH TRICKLING FILTER WASTE
             A.  REPLICATES A AND B (USED IN DUNNETT'S PROCEDURE)


Repl.  Effl.   No.     Dead at   Dead + Deform.   Dead at Test  Dead + Deform.
       Cone. Eggs at  Hatching   at hatching    Termination    at  termination
        (%)   Start   No.   (%}    No.  (%)      No.   (%)       No.   (%)
A





B





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
6
5
6
8
25
39
9
6
10
10
37
34
12
10
12
16
50
80
18
12
20
20
76
68
6
5
5
9
17
49
10
9
10
16
33
45
12
10
10
18
34
100
20
18
20
32
66
90
7
5
6
15
32
49
10
9
10
20
40
50
14
10
12
30
64
100
20
18
20
40
82
TOO
     B. COMBINED DATA FROM REPLICATES A AND B (USED IN PROBIT ANALYSIS)
Repl.


A&B





Effl.
Cone.
c%,\
\K>)
Cont.
3
5
7
11
16
No.
Eggs at
Start
100
100
102
100
99
99
Dead
at
Hatching
No.
14
11
15
8
40
68
f<£\
\® )
14
11
15
8
40
69
Dead +
Deform.
at hatching
No.
15
11
16
18
62
73
(%)
15
11
16
18
62
74
Dead
at Test
Termination
No.
16
14
15
25
50
94
(%)
16
14
15
25
50
95
Dead +
Deform

at termination
No.
17
14
16
35
72
99
{«)
17
14
16
35
73
100







                                      91

-------
    STATISTICAL ANALYSIS  OF FATHEAD MINNOW EMBRYO-LARVAL
               SURVIVAL  AND TERATOGENECITY TEST
                           TOTAL MORTALITY
                    TOTAL NUMBER OF DEAD EMBRYOS.
                   DEAD LARVAE AND DEFORMED LARVAE
                              ARCSIN
                          TRANSFORMATION
 ENDPOINT ESTIMATE
 EC1, ECS, EC 10. EC50
SHAPIRO-WILK'S  TEST
                                            NON-NORMAL DISTRIBUTION
            NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
       NO
                          BARTLETT'S TEST
                           HETEROGENEOUS
                             VARIANCE
              EQUAL NUMBER OF
                REPLICATES?
             EQUAL NUMBER OF
               REPLICATES?
                YES
                YES
T-TEST WITH
BONFERRONI
ADJUSTMENT



DUNNETT'S
TEST

STEEL'S MANY-ONE
RANK TEST



HILCOXON RANK SUM
TEST WITH
BONFERRONI ADJUSTMENT


                          ENDPOINT ESTIMATES
                               NOEC, LOEC
      Figure 4.  Flow chart for statistical analysis of fathead
                minnow embryo-larval data.
                              92

-------
to
CO
        0.9-
         0.0
CONNECTS THE MEAN VALUE FOR EACH CONCENTRATION
REPRESENTS THE CRITICAL VALUE  FOR DUNNETT'S  TEST
(ANY MEAN OF TOTAL MORTALITY ABOVE THIS  VALUE WOULD
 BE SIGNIFICANTLY DIFFERENT FROM THE  CONTROL)
                                                                                                     11
                                              EFFLUENT CONCENTRATION («)
          Figure  5.   Plot of  fathead minnow total  mortality  data from the  embryo-larval  test.

-------
TABLE 3.   FATHEAD MINNOW EMBRYO-LARVAL TOTAL MORTALITY DATA
Effluent Concentration (%)
Replicate
A
RAW B
ARC SINE
TRANS- A
FORMED B
MEAN(Y,-)
Si2
i

Source df
Between p - 1
Within N - p
Total N - 1
Control 3.0 5.
0.14 0.10 0.
0.20 0.18 0.
0.384 0.322 0.
0.464 0.438 0.
0.424 0.380 0.
0.003 0.007 0.
1 2 3
TABLE 4. ANOVA
Sum of Squares
(SS)
SSB
ssw
SST
0 7.
12 0.
20 0.
354 0.
464 0.
409 0.
006 0.
4
TABLE
0 11.0 16.0
30 0.64 1.0
40 0.82 1.0
580 0.927
685 1.133
632 1.030
006 0. 021
5

Mean Square(MS)
(SS/df)
2
SB
2
Sy

= SSB/(p-l)
= SSW/(N-p)

                            94

-------
Where
        p  = number of effluent concentration levels including the
             control
        N  = total  number of observations n-j  + n2 ••-  +np
        n-j  = number of observations in concentration i
           SSB = 2 Tj2/ni- - G2/N
                1=1
                                      Between Sum of Squares
           SST = 2   2 Yij
                1=1 j=1

           SSW = SST - SSB
                                      Total  Sum of Squares


                                      Within Sum of Squares
            G  = the grand total of all sample observations, G = 2 T-j
                                                                1=1
            T.J = 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)

12.2.7.2  For the data in this example:
"I = n2
N  = 10
                           = 2
    Tl =
    T2 * Y21 + Y22 = 0-760
    T3 = Y31 -*- Y32 = 0-818
    14 = ¥41 + Y42 = 1.265
    T5 = Y51 + Y52 = 2.060
    G  = TI + T2 + T3 + 14 + T5 = 5.751
    SSB = 2 T1-2/ni - G2/N
         i=l

        = JJ 7.810) - (5.751)2  = o.598
           2              10
              n^
    SST - 2   2 Y,-^ _ G2/N
         i=l J-l  J

        = 3.948 - (5.751)2  a 0>640
                     10
    SSW = SST - SSB = 0.640 - 0.598 = 0.042
    SB 2 = SSB/p-1 = 0.598/5-1 = 0.1495

    SW 2 = SSW/N-p = 0.042/10-5 = 0.008
                                     95

-------
12.2.7.3  Summarize these calculations in an ANOVA table (Table 5)

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

12.2.7.4  To perform the individual comparisons,  calculate the t
statistic for each concentration, and control  combination as follows:
Where _Yi  = mean proportion of total mortality for concentration i
      Y]  = mean proportion of total mortality for the control
      $W  = square root of within mean sqaure
      H]  = number of replicates for control
      ni  = number of replicates for concentration i.

Since we are lookfng for an increased response in percent of total
mortality over control, the control mean is subtracted from the mean at a
concentration.

12.2.7.5  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:
                  t2 =
                            (  0.380 - 0.424 )
                       [ 0.0897 U/H) + (1/2) ]
= - 0.494
                                    96

-------
                       TABLE 6.   CALCULATED  T VALUES
           Effluent Concentration (%)
3.0
5.0
7.0
11.0
2
3
4
5
-0.494
-0.168
2.337
6.809
12.2.7.6  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 tj 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%.

12.2.7.7  To quantify the sensitivity of the test, the minimum
significant difference (MSD) that can be detected statistically  may be
calculated.
                   MSD = d SW \/ (l/n-|) + (1/n)
Where: d

       n

       "1
                                               s procedure
                                               square
the critical value for the Dunnett
the square root of the within mean
the common number of replicates at each concentration
(this assumes equal replication at each concentration
the number of replicates in the control.
12.2.7.8  In this example:
                   MSD = 2.85 (0.089) \/ (1/2)  + (1/2)
                       = 2.85 (0.089)0.0)
                       = 0.254

12.2.7.9  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
                                    97

-------
    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 (MSDU) is determined by subtracting the
       untransformed values from 2.

                       MSOU =  0.393 -  0.169 = 0.224

12.2.7.10  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.

12.2.8  Probit Analysis

12.2.8.1   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
                                    98

-------
  TABLE 8.  OUTPUT FROM EPA PROBIT ANALYSIS PROGRAM, VERSION 1.4,
           USED FOR CALCULATING EC VALUES.

    Cone.

   Control
    3.0000
    5.0000
    7.0000
   11.0000
   16.0000
 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
 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
Mi
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
        6.3314
        6.7892
        9.1192
       12.2489
       13.1345
       14.5657
       17.6840
              Lower       Upper
            95% Confidence Limits
             3.6073
             4.6408
             5.3031
             5.7994
             8.3614
            11.4157
            12.1697
            13.3302
            15.7134
                    5.5567
                    6.5196
                    7.1058
                    7.5354
                    9.7763
                   13.3942
                   14.5708
                   16.5676
                   21.2145
                              99

-------
     Probit
       10+
        9+
        8+
        7+
        6+
                          o	
        4+
        3+
       2+
       1+
       0+0  O
         _l	,	1	1
         BC01          EC10    BC25     EE30     EC75    EE90         EC99
Figure  6.   Plot of  adjusted probits and  predicted regression line,
                                    100

-------
13.  PRECISION AND ACCURACY

13.1 PRECISION

13.1.1  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 (Table 9), and 41% for Diquat (Table 10).

13.1.2  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.

13.2 ACCURACY

13.2.1  The accuracy of toxicity tests cannot be determined.
                                    101

-------
  TABLE 9. PRECISION OF THE FATHEAD MINNOW EMBRYO-LARVAL
           SURVIVAL AND TERATOGENICITY TEST, USING CADMIUM
           AS A REFERENCE TOXICANT*»b
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

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's Ohio  (Cornelius I.  Weber,
 Project Officer).

^Cadmium 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.

cDetermined by Probit Analysis.

^Highest no-observed-effect concentration determined
 by independent statistical analysis (2x2 Chi-square Fisher's
 Exact Test).
                             102

-------
TABLE 10.  PRECISION OF THE FATHEAD MINNOW,  EMBRYO-LARVAL,
          SURVIVAL AND TERATOGENICITY TOXICITY TEST,  USING
          DIQUAT AS A REFERENCE TOXICANTa>b
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 (Cornelius I.  Weber,
   Project Officer).
       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.

  dNot calculatable.
                           103

-------
TABLE 11. PRECISION OF FATHEAD MINNOW 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
    aData 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 1

    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.
                       104

-------
                                  SECTION 12

                                  TEST METHOD

         CLADOCERAN, CERIODAPHNIA DUBIA,  SURVIVAL AND REPRODUCTION  TEST
                                 METHOD 1002.0
1.  SCOPE AND APPLICATION

1.1  This method measures the chronic toxicity of whole effluents  and
receiving water to the cladoceran, Ceriodaphnia dubia,  during a  three-brood
(seven-day), static renewal exposure^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-hs 48-h, and 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, such as chlorine, in the  source may
not be detected in the test.

1.5  This method should be restricted to use by or under the supervision of
professionals experienced in aquatic toxicity testing.

1.6  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)
an abbreviated test, consisting of only one concentration such as  100%
effluent or the instream waste concentration and a control.   Abbreviated tests
are used for toxicity screening or a pass/fail  permit condition.   Failure of
the screening test is usually followed by a definitive  test.

2.  SUMMARY OF METHOD

2.1  Ceriodaphnia 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.   Test results  are based on
survival and reproduction.  If the test is conducted as described,  the control
organisms should produce three broods of young during a seven-day  period.
                                      105

-------
3.  INTERFERENCES

3.1  Toxic substances may be introduced by contaminants in dilution water,
glassware, sample hardware, and testing equipment (see Section 5,  Facilities
and Equipment).

3.2  Improper effluent sampling and handling may adversely affect test results
(see Section 8, Effluent and Receiving Water Sampling and Sample Handling).

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 AMD EQUIPMENT

5.1  Ceriodaphm'a and algal culture units — See culturing methods below.

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

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

5.4  Environmental  chambers, incubators, or equivalent facilities  with
temperature control (25+^ l^C; Fisher # 11-679-66 or equivalent).

5.5  Water purification system — MILLIPORE MILLI-QR system or equivalent.

5.6  Balance — Analytical, capable of accurately weighing 0.0001  g.

5.7  Reference weights, Class S -- for checking performance of balance.

5.8  Test Chambers — 10 test chambers are required for each concentration and
control.  Test chambers such as 30-mL borosilicate glass beakers or disposable
polystyrene {salad dressing) cups are recommended because they will  fit in the
viewing field of most stereoscopic microscopes.  Glass beakers are rinsed
thoroughly with dilution water before use.  Plastic cups do not require
rinsing.

5.9  Mechanical shaker or magnetic stir plates — for algal cultures.
                                      106

-------
5.10  Light meter — with a  range  of  0-200 uE/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).

5.16  Counting chamber — Sedgwick-Rafter, Palmer-Maloney, or hemocytometer.

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

5.18  Centrifuge  tubes — 15-100 ml,  screw-cap.

5.19  Filtering apparatus ~ for membrane  and/or glass fiber filters.

5.20  Racks (boards) for test chambers  —  Racks 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).

5.21  Dissecting  microscope  with substage  lighting  —  for examining
Ceriodaphm'a in the test chambers.

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 -- 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  Pi pet bulbs and fillers ~ PropipetR, or equivalent.
                                      107

-------
5.28  Disposable polyethylene pipets, droppers, and glass tubing  with
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  pH, DO, and specific conductivity  meters — for routine physical  and
chemical measurements.  Unless the test  is being conducted to specifically
measure the effect of one of the above parameters, a portable, field-grade
instrument is acceptable.

6.  REAGENTS AND CONSUMABLE MATERIALS

6.1  Reagent water -- defined as MILLIPORE MILLI-QR water, or equivalent;
carbon-filtered, deionized water which does not contain substances  which are
toxic to the test organisms (see paragraph 5.5 above).

6.2  Effluent, surface water, and dilution water — see Section 7,  Dilution
Water, and Section 8, Effluent and Surface Water Sampling and Sample
Handling.  Dilution water that contains  undesirable organisms, that may attack
the test organisms should be filtered through a fine mesh net (30-um or
smaller openings).

6.3  Reagents for hardness and alkalinity tests (see EPA Methods  130.2  and
310.1, USEPA 1979b).

6.4  Standard pH buffers 4,  7, and 10 (or as per instructions of  instrument
manufacturer) for instrument calibration (see USEPA Method 150.1, USEPA 1979b).

6.5  Specific conductivity standards (see EPA Method 120.1,  USEPA 1979b).

6.6  Laboratory quality assurance samples and standards for  the above methods.

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

6.8  Membranes and filling solutions for dissolved oxygen probe (see USEPA
Method 360.1, USEPA 1979b),  or reagents  for modified Winkler analysis.

7.  TEST ORGANISMS

7.1  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.

                                      108

-------
7.2  Neonates used for toxicity tests should be obtained from  individually
cultured organisms.  Mass cultures may be maintained,  however,  to  serve as a
reserve source of organisms for use in case of loss of individual  cultures.

7.3  Starter animals may be obtained from an outside source  by  shipping 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.   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.

7.4  It is best to start the cultures with one animal, which is sacrificed
after producing young, embedded, and retained as a  permanent microscope 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  Ceriodaphm'a
(Beckett and Lewis, 1982):

     1. Pipet the animal onto a watch glass.
     2. Reduce the water volume by withdrawing excess water  with  the
        pipet.
     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-9/9AF Medium"!, prepared by mixing two parts
        of CMCP-9 with one part of CMCP-9AF.  For more viscosity  and
        faster drying, CMC-10 stained with acid fuchsin may  be used.
     5. Using a forceps or a pipet, transfer the animal to the drop
        of mounting medium on the microscope slide.
     6. Cover with a 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 CMC-10 around the edges of
        the covers!ip.
     9. Identify to species (see Pennak, 1978, and  Berner, 1985).
    10. Label with waterproof ink or diamond pencil.
    11. Store for permanent record.
1CMCP-9 and 9AF are available from Polysciences, Inc.,  Paul  Valley
Industrial Park, Warrington, Pennsylvania, 18976 (215-343-6484).
                                    109

-------
7.5  MASS CULTURE

7.5.1  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
Paragraph 12.2.3 below).

7.5.2  One-liter or 21 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.

7.5.4  Mass cultures which will  serve as a source of brood organisms for
individual culture should be maintained in good condition by frequent renewal
of the medium and brood organisms.  Each culture  is started by adding 40-50
neonates per liter of medium.  The stocked organisms should be transferred to
new culture medium at least twice a week for two  weeks.  At each nenewal, 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 each week.

7.5.3  Reserve cultures also may be maintained  in large  (80-L) aquaria or
other large tanks.

7.6  INDIVIDUAL CULTURE

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

7.6.2  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).

7.6.3  Organisms are fed daily 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.

7.6.4  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.

7.6.5  Cultures which are properly maintained should produce at least 15 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.
                                      110

-------
7.6.6  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 15 young per adult on  a board during a one-week period
would indicate problems, such as poor quality  of culture media or food.
Organisms on that board should not be used as  a source  of test organisms.

7.7  CULTURE MEDIUM

7.7.1  Moderately hard synthetic water  prepared using MILLIPORE MILLI-QR or
equivalent deionized water and reagent  grade chemicals  or 20% DWM is
recommended as a standard culture medium (see  Section 7, Dilution Water).

7.8 CULTURE CONDITIONS

7.8.1  Ceriodaphnia should be cultured  at a temperature of  25 +_ 1°C.

7.8.2  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.

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

7.8.4  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.

8.  FOOD PREPARATION AND FEEDING

8.1  Feeding the proper amount of the right food is extremely important in
Ceriodaphnia 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.

8.2  The YCT and algae are prepared as  follows:

8.2.1  Digested trout chow:

  1. Preparation of trout chow 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.
     0. 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).
                                      Ill

-------
     Add 5.0 g of trout chow pellets  to  1  L  of MILLI-QR water.  Mix well in
     a blender and pour into a 2-L 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.
     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,
8.2.2  Yeast:
of
  1. Add 5.0 g of dry yeast,  such as FLEISCHMANN'SR to 1  L
     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.

8.2.3  CEROPHYLLR (Dried, Powdered, Cereal  Leaves):

  1. Place 5.0 g of dried, powdered,  cereal  leaves 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
     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 .

8.2.4  Combined YCT Food:

  1. 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 two weeks.
                                      112

-------
  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.

8.Z.5  Algal (Selenastrum) Food

8.2.5.1  Algal Culture Medium

  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-QR  water. Mix  well after the addition of
     each solution.  Dilute to 1 L, mix well, and  adjust the the pH to 7.5 j^
     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 0.45um 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.

8.2.5.2  Algal Cultures

8.2.5.2.1  See Section 6, Test Organisms,  for information on sources of
"starter" cultures of Selenastrum capricornutum.

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

8.2.5.2.2.1  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-nt 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 Ceriodaphnia toxicity tests.   The  volume of stock culture
     maintained at any one time will depend  on the amount of algal food
     required for the Ceriodaphnia 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

                                      113

-------
     approximately 86 +_ 8.6 uE/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 four
     to six months.

8.2.5.2.2.2  Establishing and Maintaining  "Food" Cultures  of Algae

  1.  "Food" cultures are started seven  days prior to use for Ceriodaphm'a
     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 the excess
     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
     approximately 86 + 8.6 uE/m2/Sj  Or 400 ft-c).
  3.  Cultures are mixecT 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.

8.2.5.2.3  Preparing Algal Concentrate  for Use as Ceriodaphm'a Food

  1.  An algal concentrate containing 3.0 to 3.5 X 10? 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 approximately two-to-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 13), and used to determine the

                                       114

-------
   TABLE 1. NUTRIENT STOCK SOLUTIONS FOR MAINTAINING  ALGAL STOCK CULTURES
                         AND TEST CONTROL CULTURES.
Nutrient
Stock
Solution
1









2
3
±
1
Compound
MgCl2-6H20
CaCl2*2H20
H3B03
MnCl2-4H20
ZnCl2
FeCl3-6H20
CoCl2-6H20
Na2Mo04-2H20
CuCl2-2H20
Na2EDTA-2H20
NaN03
MgS04-7H20
K2HP04
NaKCQ3
Amount dissolved in
500 mL MILLI-OR Water
6.08 g
2.20 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
12.75 g
7.35 g
0. 522 g
7.50 g
aZnd2 - Weigh out 164 mg and  dilute to  100 mL.  Add 1 mL of this
 solution to Stock #1.
bCoC!2 -6H20 - Weigh out 71.4  mg  and dilute to 100 mL.  Add 1 mL of
 this solution to Stock #1.
GNa2Mo04 *2H2° - Weigh out 36.6 mg and dilute to 10 mL.  Add 1 mL
 of this solution to Stock #1.
dCud2 -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.
                                     115

-------
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
Concentration
(mg/L)
25.5
12.2
4.41
14.7
1.04
15.0


Concentration
(ug/U
185
416
3.27
1.43
0.012
7.26
160
300
Element
N
Mg
Ca
S
P
Na
K
C
Element
B
Mn
Zn
Co
Cu
Mo
Fe
—
Concentration
(mg/L)
4.20
2.90
1.20
1.91
0.186
11.0
0.469
2.14
Concentration
(ug/L)
32.5
115
1.57
0.354
0.004
2.88
33.1
	
                               116

-------
                0
           £>
           to
           £>
           :o
           :o
           Cl
           £>
           
-------
12.1.1  Surface Waters

12.1.1.1  Surface water ti
collected.  Approximately
assuming 10 replicates of
analysis,

12.1.2  Effluents

12.1.2.1  The selection o:
the objectives of the stui
or 0.5, is commonly used.
testing between 100% and
(100%, 30%a 10£, 3%, and '
effort, but because of th<
poor test precision (_+ 301
precision (;* 100%), but n
range of effluent concenti
as the dilution factor is

12.1.2.2  If the effluent
range of effluent concenti
and 0.1%).  If a high rat<
of the test, additional d-
concentrations can be add)

12.1.2.3  A volume of 15 r
and will provide a depth •
stereomicroscope with a nr
for each effluent dilutioi
required for daily renewa'
15 ml of test solution, w
Prepare enough test solut
concentration to provide *

12.2  START OF THE TEST

12.2.1  On-site tests shoi
and off-site tests should
prior to testing, the tern]
1°C and maintained at tha-
dilution water.

12.2.2  The test solution:
treatments and a control,
(Figure 1 of this Section
randomized block design, •
beginning of the test.  A
that the same template is

12.2.3  Neonates less thai
required to begin the tes1
cultures using brood boards, as described above.   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).

12.2.4  Ten brood cups, each with 8 or more young, are 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).  A second brood cup  is selected, and  one
neonate from this 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.

12.2.5  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.

12.3  LIGHT, PHOTOPERIOD, AND TEMPERATURE

12.3.1  The light quality and intensity should be at ambient  laboratory
levels, approximately 10-20 uE/m2/s, or 50 to 100 foot candles  (ft-c),  with
a photoperiod of 16 h of light and 8 h of darkness.   It is  critical that the
test water temperature be maintained at 25 +_ 1°C  to obtain  three  broods in
seven days.

12.4  DISSOLVED OXYGEN (DO)

12.4.1  Low DO concentrations may be important when running effluent  toxicity
tests.  However, aeration is not practical  for the Ceriodaphnia test.   If  the
DO in the effluent and/or dilution water is low,  aerate before  preparing the
test solutions.

12.5  FEEDING

11.5.1  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/15 mL test solution and 0.1 mL Selenastrum concentrate/15 mL  test solution
(0.1 mL of algal concentrate containing 3.0-3.5 X 107 cells/mL  will provide
2-2.3 XI05 cells/mL in the test chamber).

12.5.2  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.

12.6  TEST SOLUTION RENEWAL

12.6.1  For on-site tests, test solutions are renewed daily with  freshly
collected samples.  For off-site tests, test solutions are  also renewed daily,
                                                                         120

-------
using the most recently collected sample.   A minimum of three samples are
collected, preferrably for use beginning on Days  1, 3, 5.   The first sample is
used for test initiation on Day 1  and test solution renewal on Day 2.  The
second sample is used for test solution renewal on Days 3 and 4, and the third
sample is used for test solution renewals  on Days 5, 6, and 7.  Samples first
used on Days 1, 3, and 5, are held over in the  refrigerator at 4°C for use
on the following day(s).

12.6.2  Several sample shipping options are available, including Express Mail,
air express, bus, and courier service.  Express Mail is delivered seven days a
week.  For private carriers, shipping and  receiving schedules on weekends vary
with the carrier.

12.6.3  New test solutions are prepared daily,  and the test organisms are
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.

12.7  ROUTINE CHEMICAL AND PHYSICAL DETERMINATIONS

12.7.1  At a minimum, the following measurements  are made:

12.7.1.1  DO and pH are measured at the beginning and end of each 24-h
exposure period in the high, medium, and low test concentrations, and in the
control.

12.7.1.2  Temperature should be monitored  continuously or observed and
recorded daily for at least two locations  in the  environmental control system
or the samples.

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

12.7.1.4  Record the data (as shown in Figure  1).

12.8  OBSERVATIONS DURING THE TEST

12.8.1  Three broods are usually obtained  in the  controls in a seven-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,  soon after the adults are
transferred to fresh test solutions.  Successive  broods are released every 36
to 48 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.

12.8.2  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
                                      121

-------
brood of five might be released on Day 4,  just prior to  test solution  renewal,
and the fifth released just after renewal, and counted on  Day 5.   Partial
broods, released over a two-day period, should be counted  as one  brood.

12.8.3  Each day, the live adults are transferred to fresh test solutions, and
the numbers of live young are recorded (see data form, Figure 2).   If
difficultly is encountered in counting the live young because of  their erratic
motion, two drops of IN HC1 can be added to the chamber  (except the chambers
used for DO and pH measurements) after the adult has been  transferred.   Upon
addition of acid, the young die quickly and settle to the  bottom  of the test
chamber where they may be counted with a minimum of effort and error.  The
young are discarded after counting.

12.8.4  The young are best counted with the aid of a stereomicroscope  with
substage lighting.  If counts are made without the aid of  a stereomicroscope,
it is helpful to place the test chambers on a black strip  of tape on a light
box.

12.8.5  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.

12.9  TERMINATION OF THE TEST

12.9.1  Tests should be terminated when 60% or more of the surviving females
in the controls have produced their third brood.  Because  of the  rapid rate of
development of Ceriodaphnia, 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.

12.9.2  The data recorded in Figure 2 is summarized as illustrated in  Figure 3,

12.10. ACCEPTABILITY OF TEST RESULTS

12.10.1  For the test results to be acceptable, survival in the controls must
be at least 80%, and reproduction in the controls must average 15 or more
young per surviving female.

12.11  SUMMARY OF TEST CONDITONS

12.11.1  A summary of test conditions is listed in Table 3.

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.

                                       122

-------
      TABLE 3.   SUMMARY OF RECOMMENDED  EFFLUENT TOXICITY TEST CONDITIONS
                FOR THE CERIODAPHNIA SURVIVAL AND REPRODUCTION TEST
 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. Aeration:
15. Dilution water:
Static renewal
25+ loc
Ambient laboratory illumination
10-20 uE/m2/s,  or 50-100 ft-c
(ambient laboratory levels)
16 h light, 8 h dark
30 mL
15 mL
Daily
Less than 24 h; and all released
within a 8-h period
16.  Effluent concentrations;
10
10
Feed 0.1 mL each of YCT and algal
suspension per test chamber daily.
None
Moderately hard synthetic water is
prepared using MILLIPORE MILLI-QR
or equivalent deionized water and
reagent grade chemicals or 20% DMW
(see Section 7).
Minimum of 5 effluent concentrations
and a control.
                                   123

-------
        TABLE 3.  SUMMARY OF RECOMMENDED EFFLUENT TOXICITY TEST  CONDITIONS
                  FOR THE CERIODAPHNIA SURVIVAL AND  REPRODUCTION TEST
                  (CONTINUED!
  17. Dilution factor;!

  18. Test duration:



  19. Endpoints:

  20. Test acceptability
  21. Sampling requirements
  22. Sample volume required
Approximately 0.3 or 0.5

Until 60% of control females have three
broods (may require more or less than 7
days).

Survival  and reproduction

8(K or greater survival  and an average of
15 or more young/surviving female in the
control solutions.  At least 60% of
surviving females in controls should have
produced their third brood.

For on-site tests, samples are collected
daily, and used within 24 h of the time
they are removed from the sampling
device.  For off-site tests, a minimum of
three samples are collected, and used as
described in Paragraph 12.6.1.
1 L
^Surface water test samples are used undiluted.
                                     124

-------
Figure 2. Data form for the Ceriodaphnia survival  and reproduction test.
          Daily record.  See Figure 1  for key to positions on randomized
          test board.  (The chart on the right was reduced to save space)
Discharger:
Location:
Date Sample Collected:
Analyst:
Test Dates:
Template No. :
Dilution Water:
Test Chambers (glass/plastic):
Food :
Test Temp:
Test Organisms (age):
Comments:








%/ - Test organism alive
x = Test organism dead
0 = Number of live young
(-0) = Number of dead young
M - Lost or missing
y - Male


















































































8
7
10








9








a







7












s








4








3








2








1







20








19








18







17












15








14








13








12








!I







M








29








28







27












25








2*








23








22








Zf







40








39








38







37 _,












3S








34








33








32








31







50








49








48







S7












45








44








43








42








41







EO








59








58







57












55








54








53








5?








51







                                   125

-------
   Figure 3. Data form for the Ceriodaphnia survival and reproduction test.
             Summary of data from form in Figure 2.
Discharger:
Location:
Date Sample Collected:
Analyst:	
Test Start-Date/Time:
Test Stop -DateAime:"
Cone.


Day
1
2
3
4
5
6
7
8
Total
Replicate
1









2









3









4









5









6









7









8









9









10









No. of
Young









No. of
Adults









Young per
Adult









Cone.


Day
1
2
3
4
5
6
7
8
Total
Replicate
1









2









3









4









b









b









1









«









9









10









No. of
Young









No. of
Adults









Young per
Adult









Cone.


Day
1
2
3
4
5
6
7
8
Total
Replicate
1









2









3









4









5









b









7









8









9









!U









No. Of
Young









No. of
Adults









Young per
Adult









                                      126

-------
Figure 3. Data form for the Ceriodaphm'a survival  and reproduction  test.
          Summary of data from form in Figure 2.  (Continued)
Cone.


Day
1
2
3
4
5
6
7
y
Total
1









2









Replicate
3









4









b









6









/









b









9









iU









No. of
Young









No. of
Adul ts









Young per
Adult









Cone.

Replicate
Day
1
2
3
4
5
6
7
8
Total
I









2









3









4









b









b









/









ti









9









IU









No. of
Young









No. of
'Adults









Young per
Adult









Cone.

Replicate
Day
1
2
3
4
5
6
7
8
Total
1









2









3









4









b









b









;









ii









9









10









No. Of
Young









140. Of
'Adults









Young per
Adult









                                    127

-------
13.1.2  The endpoints of toxicity tests using Ceriodaphnia are based on  the
adverse effects on survival  and reproduction.  Point estimates, such as  LCs
and ICs, are calculated using point estimation techniques {see Section 2).
LOEC and NOEC values, for survival  and growth, 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 the Appendices for examples of the manual
computations and data input and output 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.   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 CERIODAPHNIA SURVIVAL DATA

13.2.1  Formal statistical analysis of the survival  data  is outlined 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  LCI,
LC5, LC10 and LC50 endpoints.  Concentrations at which there is no survival
are excluded from statistical analysis of the NOEC and LOEC, but included in
the estimation of the LC 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
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 Ceriodaphnia 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 with the control and the effluent  concentration you
wish to compare, you can determine statistical significance by looking up a
value in the table provided in the Appendix (Table G.5).   However, to use this
table the contingency table must be arranged in the  format illustrated in
Table 5.

                                       128

-------
    STATISTICAL  ANALYSIS  OF CERIODAPHNIA
        SURVIVAL  AND REPRODUCTION  TEST
                  SURVIVAL
                   SURVIVAL DATA
               PROPORTION SURVIVING
                                FISHER'S EXACT
                                     TEST
 ENDPOINT ESTIMATE
 LCI, LC5. LCIO, LC50
                                     i
ENDPOINT ESTIMATES
     NOEC. LOEC
Figure 4.  Flow chart for statistical  analysis of
          Ceriodaphnia survival data.
                       129

-------
      TABLE  4. SUMMARY OF SURVIVAL AND REPRODUCTION DATA FOR CERIODAPHNIA
                    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
Replicate
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
b
16
18
36
31
7
0
6
15
29
33
27
10
0
/
18
27
33
33
10
0
8
17
16
27
31
16
0
<)
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

Failures
A -
B -
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 SB).
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/AS
b/B).  For this data, a success may be 'alive'  or 'dead1 whichever  causes
a/AS 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.
                                    130

-------
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
                                 Number of
           Total
13
                            Number of

Control
2S% Effluent
Alive
10
3
Dead
0
7
Observations
10
10
20
13.2.5.4  Since 10/10 23/10,  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.4 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  For this example there is only one  partial mortality, and  Probit
analysis is not appropriate.

                       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
                                      131

-------
13.3  EXAMPLE OF ANALYSIS OF CERIQDAPHNIA REPRODUCTION DATA

13.3.1  Formal statistical analysis of the reproduction data is  outlined  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 2).  Hypothesis testing can be used to obtain a 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 non-parametric 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 non-parametric 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 non-parametric
alternative analyses.  The parametric analysis is the Bonferrom" T-test (see
Appendix D). The Wilcoxon Rank Sum Test with the Bonferrom*  adjustment is the
non-parametric alternative (see Appendix F).

13.3.5  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
Figure 6.  Since there is significant mortality in the 25% effluent
concentration, its effect on reproduction is not considered.
                                      132

-------
                        REPRODUCTION  DATA
                       NO.  OF YOUNG PRODUCED
       1
 POINT ESTIMATION

                     HYPOTHESIS TESTING
                  (EXCLUDING CONCENTRATIONS
                  ABOVE NOEC FOR  SURVIVAL)
 ENDPOINT ESTIMATE
    IC25.  IC50
                             I
                    SHAPIRO-MILK'S TEST
          NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
                                1
                            NON-NORMAL DISTRIBUTION
                        BARTLETT'S TEST
                    i
                                                   HETEROGENEOUS
                                                      VARIANCE
                                               t
               EQUAL NUMBER OF
                 REPLICATES?
                 YES
T-TEST WITH
BONFERRONI
ADJUSTMENT
    [
                                  EQUAL NUMBER (
                                    REPLICATES?
DUNNETT'S
   TEST

                                     YES
                            STEEL'S MANY-ONE
                               RANK THST
  HILCOXON RANK SUM
      TEST WITH
BONFERRONI ADJUSTMENT
                                   I
                          ENDPOINT ESTIMATES
                               NOEC. LOEC
    Figure 5. Flow chart for statistical  analysis  of Ceriodaphnia
              reproduction data.
                                  133

-------
CO
        501
                      CONNECTS THE MEAN VALUE  FOR  EACH  CONCENTRATION
                      REPRESENTS THE  CRITICAL  VALUE  FOR DUNNETT'S TEST
                      (ANY MEAN NO OF OFFSPRING  BELOW THIS VALUE WOULD
                       BE SIGNIFICANTLY DIFFERENT  FROM  THE CONTROL)
          0.00
1.56
         1.00

EFFLUENT CONCENTRATION (%)
                                                                            6.25
                                                                  12.50
         Figure 6.  Plot of number  of young per  adult female from a Ceriodaphm'a survival
                    and reproduction test.

-------
                 TABLE 8.  CERIODAPHNIA REPRODUCTION DATA
Replicate    Control
                                   Effluent  Concentration  (%)
                               1.56
                                         3.12
6.25
12.5










Mear
Si2
i
1
2
3
4
5
6
7
8
9
10
i(Yi)


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
13.3.6  Test for Normality

13.3.6.1  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.

13.3.6.2  Calculate the denominator, D,  of the  test statistic:
                         n
                     D = 2
                             i  - X)2
    Where X-\ - the ith centered observation
          X  = the overall  mean of the centered observations
          n  = the total  number of centered observations.
For this set of data,
                                 n = 50

                                 I =  1  (0.0)  =  0.0
                                     ~HT
                                 D = 1433.4
                                    135

-------
         TABLE 9.  CENTERED OBSERVATIONS FOR SHAPIRO-WILK'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
13.3.6.3 Order the centered observations from smallest to  largest

                       - x(2) - ...  - x
-------
TABLE 10.   ORDERED CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S  EXAMPLE
i
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
x(D
-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(D
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
                                  137

-------
TABLE 11.   COEFFICIENTS  AND  DIFFERENCES FOR SHAPIRO-WILK'S EXAMPLE
1 a, x(n-HD-x(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
0.3751
0.2574
0.2260
0.2032
0.1847
0.1691
0. 1 554
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. 031 4
0. 0244
0. 01 74
0. 01 04
0. 0035
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(50)
X(49)
x(48)
X<47)
X(46)
x(45)
X(44)
X(43)
x(42)
X(41 )
X(40)
X(39)
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)

I X(5)
- x(6)
- X<7)
- X(8)
- x(9)
- xdo)
- xdi)
- X(12)
- X(13)
- Xtl4)
- X(15)
- X(16)
- X(17)
- X(18)
- xH9)
- X(20)
- X(21)
- X(22)
- X(23)
- X(24)
- X(25)
                                138

-------
13.3.7 Test for Homogeneity of Variance

13.3.7.1  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
               [ ( S v-j) In S2 - 2 VT  In Sf2 ]
           R =    i=l           1=1
    Where V-j =   degrees of freedom for each effluent concen-
                 tration and control, V-j = (n-j - 1)

          p  =   number of levels of effluent concentration, including control
          n-j = the number of replicates for concentration i
          In = loge
          i  = 1, 2, ..., p where p is the number of concentrations
               including the control
                  ( 2 Vi Si2)
          S2 =     1=1
          C  = 1 + ( 3(p-l)H [ S 1/Vj - { S Vf
13.3.7.2 For the data in this example, (See Table 8) all effluent
concentrations including the control have the same number of replicates (nj
= 10 for all i).  Thus, Vj = 9 for all i.

13.3.7.3  Bartlett's statistic is therefore:
       B =  [(45)ln(31.8) - 9 2 ln(Sf2)]/i.04
                             1=1

         =  [45(3.5) - 9(16.!)]/!.04

         =  12.6/1.04

         *  12.1
                                       139

-------
13.3.7.4  B is approximately distributed as chi  square  with  p  -  1 degrees of
freedom, when the variances are in fact the same.   Therefore,  the appropriate
critical value for this test, at a significance  level of  0.01  with  four
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.

13.3.8  Dunnett's Procedure

13.3.8.1  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 df


Between p - 1

Within N - p
Total N - 1
Sum of Squares
(SS)

SSB

SSW
SST
Mean Square(MS)
(SS/df)
2
SB = SSB/(P-D
2
SM = sswy(N-p)

Where:      p  = number effluent concentrations  including  the control
            N  = total  number of observations n-]  + r\% ...  +np
            r\\ = number of observations in  concentration i
           SSB = 2 Tt2/ni - G2/N
                1=1
                               Between Sum of Squares
SST = Z
                            _ G2/N
Total Sum of Squares
           SSW = SST - SSB
                               Within Sum of Squares
            G  = the grand total  of all  sample  observations,  G  = 2 Tj
                                                               i=l
            Tj = the total of the replicate measurements  for
                 concentration "i"
           YJJ = the jth observation for concentration  "i"  (represents
                 the number of young produced by  female j  in
                 effluent concentration  i)
                                    140

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

    n-|  = n2 = n3 = n4 = n5 = 10
    N = 50
    TT  = YH  + Y12 +  . .  . + Yno" 224
    T2 = Y2l  + Y22 +  • •  • + Y21Q = 263
    T3 = Y3]  + Y32 +  . .  . + Y310 = 346
    T4 = Y41  + Y42 +  . .  . + Y410 = 317
    T5 " Y51  + Y52 +  . .  . + Y510=  94
    G  = TI  + T2  +  T3 + T4 + Tg = 1244

          P    o
    SSB = S  T1-2/ni  - G2/N
         1=1

        = JJ 348,386) - (1244)2  = 3887.88
          10              50
SST = 2   S
                     -  Q2/N
        = 36,272  -  (1244)2  = 5321.28
                      50

    SSW = SST -  SSB  =  5321.28 - 3887.88 = 1433,40
    SB2 = SSB/p-1  =  3887.88/5-1 = 971.97
    SW2 = SSW/N-p  =  1433.40/50-5 = 31.85
13.3.8.3 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
                                      141

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

                                    ( Yi  - T*  )
                                SWV tl/n-j)  +
Where _Yi  = mean number of young produced for effluent concentration  i
      Y-J  = mean number of young produced for the  control
      $W  = square root of within mean sqaure
      n-|  = number of replicates for control
      n-j  = 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,


13.3.8.5  Table 14 includes the calculated t  values for each  concentration
and control combination.  In this examples comparing the 1.56%  concentration
with the control the calculation is as follows:
                              (  22.4 - 26.3   )
                                                   = -1.55
                       [ 5.64 \/ 11/1UJ  + U/1U)   1
                      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
13.3.8.6 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.
                                      142

-------
The mean reproduction for concentration  i   is considered  significantly  less
than the mean reproduction for the control  if tj  is  greater  than the
critical value.  Since ts is greater than 2.23, the  12.5%  concentration  has
significantly lower reproduction than the control.   Hence  the  NOEC and the
LOEC for reproduction are 6.25% and 12.5%,respectively.

13.3.8.7  To quantify the sensitivity of the test, the minimum significant
difference (MSD) that can be statistically detected  may  be calculated.
Where  d

       n

       "i
                   MSD = d SW  V (1/ni)  + (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
            (this assumes equal replication at each concentration
            the number of replicates in the control.
13.3.8.8  In this example:
                   MSD - 2.23 (5.64)  >/ (1/10)  + (1/10)
                       = 2.23 (5.64)(0.45)
                       = 5.66

13.3.8.9  Therefore, for this set of data, the  minimum difference  that can  be
detected as statistically significant is 5.66.
13.3.8.10
control.
           This represents a 25% decrease in mean reproduction from the
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  Ceriodaphnia
reproduction test based on the NOEC and LOEC values from nine tests with  the
reference toxicant NaPCP is provided in Table 15.   The NOECs and LOECs  of all
tests fell in the same concentration range, indicating maximum possible
precision.

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. a!., 1988a).   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 pre-publication draft of Method 1002.
Some deviations from the standard protocol were reported by the  participating
laboratories.
                                      143

-------
14.1.2.2  Ten sets of data from six laboratories met the acceptability
criteria, and were statistically analyzed using non-parametric  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 two-fold range in concentration (Table 16).

14.1.2.3  An second interlaboratory stucty of Method 1002.0 (using the first
edition of this manual; Horning and Weber, 1985), was coordinated by Battelle,
Columbus Division, and involved 11  participating laboratories  (DeGraeve et
al., 1989).   All  participants used 10% DMW (lOfc PERRIER& Water)  as the
culture and dilution water, and used their own  formulation of food for
culturing and testing the Ceriodaphnia.  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).  The  overall
precision (CV) of the test was 27% for the survival  data (7-day LC50s) and 40%
for the reproduction data (IC50s).

14.2  ACCURACY

14.2.1  The accuracy of toxicity tests cannot be determined.
                                      144

-------
 TABLE 15.   SINGLE LABORATORY PRECISION OF  THE CERIQDAPHMIA SURVIVAL AND
            REPRODUCTION TEST,  USING NAPCP  AS A  REFERENCE TOXICANTS.b
NOEC
Test (mg/L)
lc 0.25
2d 0.20
3 0.20
46 0.30
5 0.30
6 0.30
7 0.30
8 0.30
9 0.30
LOEC
(mg/L)
0.50
0.60
0.60
0.60
0.60
0.60
0.60
0.60
0.60
Chronic
Value
(mg/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 CaCOs/L; pH = 8.1).
cConcentrations 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.
                                 145

-------
  TABLE  16.   INTERLABORATORY PRECISION OF CERIODAPHNIA SURVIVAL
AND REPRODUCTION TEST1



Endpoints
Reproductl on
Analyst
3
4
4
5
5
6
6
10
10
11
Test
1
1
2
1
2
1
2
1
2
1
NOEC
12
6
6
6
12
12
6
6
6
12
LOEC
25
12
12
12
25
25
12
12
12
25
(% Effluent)
Survival
NOEC
25
12
25
12
12
25
25
12
12
25


LOEC
50
25
50
25
25
50
50
25
25
50
Neiheisel et al., T988a,
                              146

-------
                                   SECTION 13

                                  TEST METHOD

                 ALGAL, SELENASTRUM CAPRICORNUTUM, GROWTH TEST
                                 METHOD 1003.0

1.  SCOPE AND APPLICATION

1.1  This method measures the chronic  toxicity of whole effluents  and
receiving water to the fresh water alga,  Selenastriim capricornutum,  during  a
four-day, static exposure.  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  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
degradeable and volatile toxicants,  such as chlorine, in the source  may not be
detected in the test.

1.4  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.

1.5  This method is restricted to use  by or under the supervision  of
professionals experienced in aquatic toxicity testing.

2.  SUMMARY OF METHOD

2.1  A Selenastrum 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
and Equipment).

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.
                                      147

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

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 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 capable  of collecting a 24-h composite
sample of 1 L.

5.3  Sample containers -- for sample shipment and storage  see Section 8,
Effluent and Receiving Water Sampling and Sample Handling.

5.4  Environmental  chamber, incubator, or equivalent facility —  with
"cool-white" fluorescent illumination (86 +. 8.6 uE/m2/s, or 400 + 40 ft-c)
and temperature control (25 _+ 1°C, for compatibility with  other tests).

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  uE/m2/s (0-1000 ft-c).

5.7  Water purification system — MILLIPORE MILLI-QR or equivalent.

5.8  Balance — Analytical, capable of accurately weighing 0.0001  g.

5.9  Reference weights, Class S — for checking performance of balance.

5.10  Glass or electronic thermometers -- for measuring water temperatures.

5.11  Bulb-thermograph or electronic-chart  type thermometers -- for
continuously recording temperature.

5.12  National Bureau of Standards Certified thermometer  (see EPA Method
170.1, USEPA 1979b).

5.13  Meters: pH and specific conductivity  -- for routine  physical and
chemical measurements.  Unless the test is  being conducted to specifically
measure the effect of one of the above parameters, a portable, field-grade
instrument is acceptable.

                                      148

-------
5.14  Tissue grinder — for chlorophyll  extraction.
5.15  Fluorometer (Optional) — Equipped with chlorophyll  detection light
source, filters, and photomultiplier tube (Turner Model  110 or equivalent).
5.16  UV-VIS spectrophotometer — capable of accommodating 1-5 cm cuvettes.
5.17  Cuvettes for spectrophotometer —  1-5 cm light path.
5.18  Electronic particle counter (Optional) —  Coulter Counter, Model ZBI,
or equivalent, with mean cell  (particle) volume determination.
5.19  Microscope — with 10X,  45X, and 100X objective lenses, 10X ocular
lenses, mechanical stage, substage condenser, and light source (inverted or
conventional microscope).
5.20  Counting chamber — Sedgwick-Rafter, Palmer-Maloney,  or hemocytometer.
5.21  Centrifuge — with swing-out buckets having a  capacity of  15-100 ml.
5.22  Centrifuge tubes — 15-100 ml, screw-cap.
5.23  Filtering apparatus — for membrane and/or glass fiber filters.
5.24  Volumetric flasks and graduated cylinders — Class A, 10-1000 ml,
borosilicate glass, for culture work and preparation of test solutions.
5.25  Volumetric pi pets— Class A, 1-100 ml.
5.26  Serological pipets— 1-10 ml, graduated.
5.27  Pipet bulbs and fillers  — PropipetR, or equivalent.
5.28  Wash bottles — for rinsing small  glassware, instrument electrodes, and
probes.
5.29  Culture chambers ~ 1-4  L borosilicate, Erlenmeyer flasks.
5.30  Test chambers — 125 or  250 ml borosilicate, Erlenmeyer flasks, with
stainless steel closures.
5.31  Preparation of glassware — prepare all graduated cylinders, test
flasks, bottles, volumetric flasks, centrifuge tubes and vials used in algal
bioassays as follows:
5.31.1  Wash with non-phosphate 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.
5.31.2  Rinse with tap water.
                                      149

-------
5.31.3  Test flasks should be thoroughly rinsed with 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.

5.31.4  Rinse twice with MILLI-QR water.

5.31.5  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 5.31.7.  If acetone is used,  go to 5.31.6.

5.31.6  Rinse thoroughly with MILLI-QR water, and dry in an 105°C oven.

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

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

6. REAGENTS AND CONSUMABLE MATERIALS

6.1  Reagent water — defined as  MILLIPORE MILLI-QR or equivalent water  (see
paragraph 5.7 above).

6.2  Effluent, surface water, and dilution water -- see Section 7,  Dilution
Water, and Section 8, Effluent and Receiving  Water Sampling and Sample
Handling.

6.3  Reagents for hardness and alkalinity tests (see EPA Methods 130.2 and
310.1, USEPA 1979b).

6.4  Standard particles — polymer microspheres,  5.0 + 0.03 urn diameter,
65.4 um3 volume, for calibration  of electronic particle counters (available
from Duke Scientific Co., 1135D,  San Antonio  Road,  Palo Alto,  California,
94303).

6.5  Standard pH buffers 4, 7, 8  and 10 (or as per instructions of instrument
manufacturer) for instrument calibration (see USEPA Method 150.1, USEPA  1979b),

6.6  Specific conductivity standards (see EPA Method 120.1, USEPA 1979b).

6.7  Laboratory quality assurance samples and standards for the above methods.

6.8  Reference toxicant solutions (see Section 4,  Quality Assurance).
                                      150

-------
6.9  Acetone — pesticide-grade or equivalent.

6.10  Dilute (10%) hydrochloric acid  — carefully add 10  ml  of concentrated
HC1 to 90 ml of MILLI-QR water.

7.  TEST ORGANISMS

7.1  Log-phase-growth Selenastrum capricornutum are used for  the  test.

7.2  CULTURE MEDIUM

7.2.1  The culture medium is used to maintain stock cultures  of the  test
organisms.

7.2.2  Prepare five stock nutrient solutions using reagent grade  chemicals  as
described in Table 1.

7.2.3  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 Ls mix well, and  adjust the pH to 7.5  +_ 0.1,  using 0.1N
sodium hydroxide or hydrochloric acid, as appropriate.  The final
concentration of macronutrients and micronutrients in the  culture medium is
given in Table 2.

7.2.4  Immediately filter the pH-adjusted medium through a 0.45um 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
prior to use by passing 500 mL of distilled water through  it.

7.2.5  If the filtration is carried out  with sterile apparatus, filtered
medium can be placed immediately into sterile culture flasks, and no further
sterilization steps are required before  the inoculation of the medium.  The
medium can also be sterilized by autoclaving before placing in the culture
flasks.  However, the pH should be checked after autoclaving  to determine if
it was changed.

7.2.6  Unused sterile medium should not  be stored in the (250 mL) test culture
flasks more than one week prior to use,  because there may  be  substantial loss
of water by evaporation.

7.3  ALGAL CULTURES

7.3.1  Test organisms —  Selenastrum capricornutum, a unicellular coccoid
green alga.  See Section 6, Test Organisms, for information on sources of
"starter" cultures.

7.3.2   Stock algal cultures

7.3.2.1 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

                                      151

-------
    TABLE  1. NUTRIENT STOCK SOLUTIONS FOR MAINTAINING ALGAL  STOCK CULTURES
                         AND TEST CONTROL CULTURES.
Nutrient
Stock
Solution
1









2
_3
4-
i
Compound
MgCl2-6H20
CaCl2-2H20
H3B03
MnCl2-4H20
ZnCl2
FeCl3-6H20
CoCl2-6H20
Na2Mo04-2H20
CuCl2-2H20
Na2EDTA-2H20
NaN03
MgS04.7H20
K2HP04
NaHC03
Amount dissolved in
500 mL Distilled Water
6.08 g
2.20 g
92.8 mg
208.0 mg
1 . 64 mga
79.9 mg
0. 71 4 mgb
3 . 63 mgc
0. 006 mgd
1 50. 0 mg
12.750 g
7.350 g
0.522 g
7.50 g
aZnd2 - Weigh out 164 mg  and dilute to TOO mL.  Add 1 mL of this
 solution to Stock #1.
bCoCl2 -6H20 - Weigh out 71.4 mg and dilute to 100 mL.  Add 1 mL of
 this solution to Stock #1.
GNa2Mo04 -2H20 - Weigh out 36.6 mg and dilute to 10 mL.  Add 1 mL
 of this solution to Stock #1.
dCud2 -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.
                                    152

-------
TABLE 2.   FINAL CONCENTRATION OF MACRONUTRIENTS AND MICRONUTRIENTS
          IN THE CULTURE MEDIUM
Macronutrient
NaN03
MgC12-6H20
CaCl2-2H20
MgS04-7H20
K2HP04
NaHC03


Micronutrient
H3B03
MnC72-4H20
ZnC12
CoCl2-6H20
CuC12-2H20
Na2Mo04.2H20
FeCl3-6H20
Na?EDTA-2H20
Concentration
(mg/L)
25.5
12.2
4.41
14.7
1.04
15.0


Concentration
(ug/L)
185
416
3.27
1.43
0.012
7.26
160
300
Element
N
Mg
Ca
S
P
Na
K
C
Element
B
Mn
Zn
Co
Cu
Mo
Fe
__
Concentration
(mg/L)
4.20
2,90
K20
1.91
0.186
1T.O
0.469
2.14
Concentration
(ug/L)
32.5
115
1.57
0.354
0.004
2.88
33.1
____
                               153

-------
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.

7.3.2.2  Maintain the stock cultures at 25 +_ 1°C, under continuous
"Cool-White" fluorescent lighting of 86 +_ 8.6 uE/m2/s  (400  + 40 ft-c).
Shake continuously at 100 cpm or twice daily by hand.

7.3.2.3  Transfer 1  to 2 ml of stock culture weekly to 50 - 100 nL 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.

7.3.2.4  Viable unlalgal culture material may be maintained long periods of
time if placed in a refrigerator at 4°C.

8.  SAMPLE COLLECTION, PRESERVATION AMD HANDLING

8.1  See Section 8,  Effluent and Receiving Water Sampling and Sample Handling.

9.  CALIBRATION AND STANDARDIZATION

9.1  See Section 4,  Quality Assurance.

10.  QUALITY CONTROL

10.1  See Section 4, Quality Assurance.

11.  TEST PROCEDURES

11,1  TEST SOLUTIONS

11.1.1  Surface Waters

11.1.1.1  Surface water toxicity is determined with samples used directly  as
collected.

11.1.2  Effluents

11.1.2.1  The selection of the effluent test concentrations should be based on
the objectives of the study.  One of two dilution factors,  approximately 0.3
or 0.5, is commonly used.  A dilution factor of approximately 0.3  allows
testing between 100% and 1% effluent using only five effluent concentrations
(100%, 30%, 10%, 3%, and 1%).  This series of dilutions minimizes  the level of
effort, but because of the wide interval between test  concentrations provides
poor test precision (+_ 300%).  A dilution factor of 0.5 provides greater
precision (+_100fc),  but requires several additional dilutions to span the  same
range of effluent concentrations.  Improvements in precision decline rapidly
as the dilution factor is increased beyond 0.5,
                                      154

-------
11.1.2.2  If the effluent is known  or suspected  to be highly toxic, a
range of effluent concentrations should  be  used  (such as
and 0.
lower
0.3%,
11.1.2.3  The volume of effluent required  for the test is 600 - 1000 ml.
Prepare enough test solution at each  effluent concentration  (approximately
700 ml) to provide 50 - 100 ml of test solution  for each of  three replicate
test flasks and 400 ml for chemical analyses.

11.1.3  Dilution water may consist of stock  culture medium without EDTA, or
other water such as surface water, depending on  the objectives of the test.
However, if water other than the stock culture medium is used for dilution
water, 1 ml of each stock nutrient solution  (except for EDTA) should be added
per liter of dilution water.  Surface 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 urn particle size  retention.

11.1.4  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 (except EDTA) is added
per liter of effluent prior to use in preparing  the test dilutions.  Thus, all
test treatments and controls will contain  as a minimum the concentration of
nutrients in stock culture medium.

11.1.5  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 urn
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.

11.1.6  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.

11.2  PREPARATION OF INOCULUM

11.2.1  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 miililiter  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.  Estimate the volume of stock culture  required to  prepare the
inoculum as described in the following example:

    If the seven-to 10-day stock culture used as the source  of the
    inoculum has a cell density of 2,000,000 cells/ml, a test
    employing 18 flasks, each containing 100 ml  of test medium and
    inoculated with a total of 1,000,000 cells,  would require
    18,000,000 cells or 12.5 ml of stock solution

                                  155

-------
    (18,000,000/2,000,000) 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.

    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
       distilled or deionized water.
    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
Stock Culture
Requi red
                    Number of flasks X Volume of Test X 10,000  cells/ml
                      to be used	Solution/Flask	
                     Cell  density (cells/ml)  in  the stock  culture

                =       18 flasks X 100 ml/flask X 10,000  cells/ml
                                 2,000,000 cells/mL

                     9.0 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.

11.3  START OF THE TEST

11.3.1  On-site tests should be initiated within 24 h of sample collection,
and off-site tests should be initiated within 36 h of sample  collection.   Just
prior to testing, the temperature of the sample  should be  adjusted to  (25  +_
1°C) and maintained at that temperature until portions are added to  the
dilution water.

11.3.2  The test begins when the algae are added to the test  flasks.

    1. Mix the inoculum well, and add 1  mL to the test solution in each  flask.
    2. Make a final check of the cell density in three of  the test solutions
       at time "zero "  (within 2 h of the inoculation).

11.4  LIGHT, PHOTOPERIOD,  AND TEMPERATURE

11.4.1  Test flasks are incubated under continuous illumination at
86 + 8.6 uE/m2/s (400 ^ 40 ft-c), at 25 +_ IOC, and should  be  shaken
continously at 100 cpm on a mechanical shaker or twice daily  by hand.

                                      156

-------
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.

11.5  ROUTINE CHEMICAL AND PHYSICAL DETERMINATIONS

11.5.1  At a minimum, the following measurements are made:

11.5.1.1  Temperature should be monitored continously  or observed and recorded
daily for at least two locations in the environmental  control  system or the
samples.

11.5.1.2  pH, 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 (see Figure  1).

11.6  OBSERVATIONS DURING THE TEST

11.6.1  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-to-two days in the test.  It may be desirable, therefore, to
determine the algal growth response daily.

11.7  TERMINATION OF THE TEST

11.7.1  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).

11.7.2  Cell counts

11.7.2.1  Automatic Particle Counters

11,7.2.1.1  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 um^/cell.  The Coulter Counter is
widely used and is discussed in detail by Miller et  al., 1978.

11.7.2.1.2  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.

11.7.2.1.3  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 Isoton^, to facilitate counting. The resulting dilution is
counted using an aperture tube with a 100-um diameter  aperature.  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.
                                      157

-------
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).

11.7.2.2  Manual microscope counting methods

11.7.2.2.1  Cell counts may be determined using a Sedgwick-Rafter,
Palmer-Maloney, hemocytometer, inverted microscope,  or similar  methods.  For
details on microscope counting methods, see APHA, 1985,  and Weber, 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.

11.7.3  Chlorophyll Content

11.7.3.1   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).

11.7.3.2  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 urn 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.

11.7.3.3  For chlorophyll measurement methods, see APHA, 1985.
                                      158

-------
11.7.4  Turbidity (Absorbance)

11.7.4.1  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.

11.7.4.2  The algal growth  measurements are made as  follows:
       A blank is prepared as described for the fluorometric analysis.
       The culture is thoroughly mixed as described  above.
       Sufficient sample is withdrawn from the test  flask with a sterile pipet
       and transferred to a 1- to 5-cm cuvette.
       The absorbance is read at 750 nm and divided  by  the  light path length
       of the cuvette, to obtain an "absorbance-per-centimeter" value.
       The 1-cm absorbance values are used in  the same  manner as the cell
       counts.
11.7.5.  Record the data as indicated in Figure 2.

11.8  SUMMARY OF TEST CONDITIONS

11.8.1  A summary of test conditions is listed in Table 3.

11.9  ACCEPTABILITY OF TEST RESULTS

11.9.1  The test results are acceptable if the algal  cell  density  in the
control flasks (without EDTA) exceeds 2 X 105 cells/mL  at  the  end  of the
test, and does not vary more than 20% among replicates.

12.  DATA ANALYSIS
12.1  GENERAL

12.1.1  Tabulate and summarize the data.
response data is listed in Table 4.
A sample set of algal  growth
12.1.2  The endpoints of toxicity tests using Selenastrum capricornutum are
based on the adverse effect on cell  growth (see Section 2).   LOEC and  NOEC
values, for growth, are obtained using a hypothesis test approach such as
Dunnett's Procedure (Dunnett, 1955)  or Steel's Many-one Rank  Test (Steel,
1959; Miller, 1981).  Point estimates, such as EC!, EC5, EC10 and EC50, would
also be appropriate in analyzing algal growth response data.   See the
Appendices for examples of the manual  computations  and examples  of computer
program data input and output.
                                      159

-------
12.1.3  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.

12.2  EXAMPLE OF ANALYSIS OF ALGAL GROWTH DATA

12.2.1  Formal statistical  analysis of the growth data is outlined in
Figure 3.  The response used in the statistical  analysis is the number of
cells per milliliter per replicate.

12.2.2  The statistical analysis consists of a parametric test, Dunnett's
Procedure, and a non-parametric 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-Milk's Test, and Bartlett's Test is used to test for
homogeneity of variance.  If either of these tests fail, the
non-parametric 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.

12.2.3  Additionally, if unequal numbers of replicates occur among the
concentration levels tested there are parametric and non-parametric
alternative analyses.  The  parametric analysis is the Bonferroni T-test
(see Appendix D).  The Wilcoxon Rank Sum Test with the Bonferroni
adjustment is the non-parametric alternative (see Appendix F).

12.2.4  Data from an algal  growth test with cadmium chloride will  be used
to illustrate the statistical analysis.  The cell counts were log-|Q
transformed in an effort to stabilize the variance for the ANOVA
analysis.  The raw data, log-jo transformed data, mean and standard
deviation of the observations at each concentration including the  control
are listed in Table 4.  A plot of the log-jo transformed cell counts for
each treatment is provided  in Figure 4.
                                    160

-------
 TABLE 3.  SUMMARY OF RECOMMENDED EFFLUENT TOXICITY TEST  CONDITIONS FOR
          THE ALGAL (SELENASTRUM CAPRICORNUTUM)  GROWTH TEST
     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/concentration:
    11.  Shaking rate:

    12.  Dilution water:

    13.  Effluent concentrations:
    14.  Dilution factor^:
    15.  Test duration:
    16.  Endpoint:

    17.  Test acceptability:

    18.  Sample volume required:
Static
25
"Cool white" fluorescent lighting
86 + 8.6 uE/m2/s (400 +_ 40 ft-c)
Continuous illumination
125 mL or 250 mL
50 mL or 100 mL
None
4 to 7 days

10,000 cells/mL
3

100 cpm continuous, or twice daily
by hand
Algal stock culture medium without
EDTA or enriched surface water
Minimum of 5 and a control
Approximately 0.3 or 0.5
96 h
Growth (cell counts, chlorophyll
fluorescence, absorbance, biomass)
2 X 105 cells/mL in the controls;
Variability of controls should not
exceed 20%
1 L (one sample for test initation)
"•Surface water samples for toxicity tests are used undiluted.
                                      161

-------
      Figure 1.  Data form for algal  growth test.  Routine chemical  and
                 physical determinations.
Discharger:
Location:
 Test Dates:
 Analyst:
Treatment
Temp.
PH
Alkalinity
Hardness
Salinity
Conductivity
Chlorine

Contr








Effluent Concentration
•












































Remarks








       Figure 2. Data form for algal  growth test,
                 determinations.
          Cell  density
Discharger:
Location:
Test Dates:
Analyst:
Cone:
Control
Cone:
Cone:
Cone:
Cone:
Cone:
Cell Density Measurement
Replicate
1






2






3






Treatment
Mean






Comments






Comments
                                       162

-------
                      TABLE 4.   ALGAL GROWTH  RESPONSE DATA
                                 Toxicant Concentration  fug Cd/L)
     Replicate  Control
10
20
40
80
A
B
C
Logi o A
Trans- B
formed C
Mean(Ti)
Si2
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. 001 3
5
49.3
40.0
44.0
1.693
1.602
1.643
1.646
0. 0021
6
12.2.5  Test for Normality

12.2.5.1  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 (ug  Cd/L)
     Replicate    Control
10
20
40
80
A -0.012
B -0.022
C 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
                                    163

-------
          STATISTICAL ANALYSIS OF AL6AL GROWTH TEST
                       GROWTH RESPONSE DATA
                              (BIS/ML
                      | SHAPIRO-MILK'S TEST
                                            NON-NORMAL DISTRIBUTION
            NORMAL DISTRIBUTION
HOMOGENEOUS VARIANCE
       NO
                          BARTLETT'S TEST
                                 HETEROGENEOUS
                                    VARIANCE
              EQUAL NUMBER OF
                REPLICATES?
                    EQUAL NUIfflER OF
                      REPLICATES?
YES

                                       YES
T-TEST WITH
BONFEHRONI
ADJUSTMENT



»
DUNNETT'S
TEST

»
STEEL'S MANY-ONE
RANK TEST



MILCOXQN RANK SUM
TEST WITH
BONFEHRONI ADJUSTMENT


                          ENDPOINT ESTIMATES
                               NOEC, LOEC
        Figure 3.  Flow chart for statistical  analysis  of algal
                         growth response  data.

                                  164

-------
en
                                                    _  CONNECTS THE MEAN VALUE FOR EACH CONCENTRATION
                                                    -  REPRESENTS THE CRITICAL VALUE FOR DUNNETT'S TEST
                                                       (ANY MEAN GROWTH BELOW THIS VALUE WOULD BE
                                                        SIGNIFICANTLY DIFFERENT FROM THE CONTROL)
                                             TOXICANT CONCENTRATION (UG CD/L)
      Figure 4, Plot  of log-jo transformed cell count data from algal  growth response test (Table  4)

-------
12,2.5.2  Calculate the denominator, D,  of the test statistic:

                         n       _
                     D = 2 (XT  - X)2
                        1=1

    Where Xj = the ith centered observation
          X  = the overall mean of the centered observations
          n  = the total number of centered observations,

For this set of data,            n = 18

                                 I =  1  (0.000) = 0.000
                                     W
                                 D = Q.Q214

12.2.5.3  Order the centered observations from smallest to largest:

                  X(l) - X(2) - ... - X) is the ith ordered observation.  These ordered  observations
are listed in Table 6.

12.2.5.4  From Table 4, Appendix B, for the number of observations,  n,
obtain the coefficients a-|, 32, .... ak where k is approximately
n/2.  For the data in this example, n =  18, k = 9.  The a-j  values  are
listed in Table 7.
  TABLE 6.  ORDERED CENTERED OBSERVATIONS FOR SHAPIRO-WILK'S EXAMPLE
i
1
2
3
4
5
6
7
8
9
XH)
-0, 078
-0.044
-0.042
-0.027
-0. 024
-0.022
-0. 01 2
-0.006
-0.003
i
10
11
12
13
14
15
16
17
18
xin
0.001
0.004
0.021
0.021
0.033
0.036
0.041
0.047
0.050
                                    166

-------
12.2.5.5 Compute the test statistic,  W,  as  follows
                       k
               W =1 [ Sa*  (X(n-i+l)  -  x(i))  ]2
                   D  1=1 '

the differences X  are listed  in  Table  7.

For this set of data:
                     W =	5	 (0.1436)2  = 0.964
                          0.0214


   TABLE 7.  COEFFICIENTS AND  DIFFERENCES FOR  SHAPIRO-WILK'S EXAMPLE


       i       a*         x(n-i+l)  - x^')
1 0.4886
2 0.3253
3 0.2553
4 0.2027
5 0.1587
6 0.1197
7 0.0837
8 0.0496
9 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)
xH5)
X(14)
X<13)
X(12)
xdi)
xOO)
- X 1)
- X 2)
- X 3)
- x(^)
- X(5)
- X(6)
- X(7)
- x(8)
- XO)
12.2.5.6  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.

12.2.6 Test for Homogeneity of  Variance

12.2.6.1  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:
                                    167

-------
                   p             p
               [ ( 2 Vj) In S2 - 2 Vi In S-,-2 ]
           B =__!=]	1=]	
    Where V-j =   degrees of freedom for each toxicant concen-
                 tration and control, Vj = (n-f - 1)

          p  =   number of levels of toxicant concentration
                 including the control
          n-j = the number of replicates for concentration i
          In = log*
          i  = 1, 2, ..., p where p is the number of concentrations
               including the control

                    p
                  ( 2 VT Si2)
                      p
                     1=1
                                 p          p
          C  = 1 + { 3(p-l))-l  [ 2 1/V;  - (  S VfJ-
12.2.6.2  For the data in this example, (See Table 4) all  toxicant
concentrations including the control  have the same number  of replicates
(HI = 3 for all  i).   Thus, Vj = 2 for all  i.

12.2.6.3  Bartlett's statistic is therefore:

                                P
       B =  [(12)ln(0.0018) -22 1 n(Si2)]/! .194
                               i=l

         =  [12(-6.3200) - 2( -41. 9082) ]/!.! 94

         =  7.9764/1.194

         =  6. 6804

12.2.6.4  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.
                                      168

-------
12.2.7  Dunnett's Procedure

12.2.7.1  To obtain an estimate of the  pooled  variance  for  the Dunnett's
Procedure, construct an ANOVA table as  described in  Table 8.

                           TABLE 8.   ANOVA TABLE
Source

Between

Within
Total
df Sum of Squares
(SS)

p - 1 SSB

N - p SSW
N - 1 SST
Mean
2
SB =
2
Sw =

Square(MS)
(SS/df)

SSB/(p-l)

SSW/(N-p)

Where:
            p  = number of toxicant concentrations including the  control
 N  = total number of observations
                                                   n2
                                                          +n
            n-j = number of observations in concentration i
           SSB = S T^/nj - G2/N
                                          Between Sum of Squares
SST = 2
                            - 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-J = the total of the replicate measurements for
                 concentration "i"
            JJ = the jth observation for concentration "i"  (represents
                 the cell count for toxicant concentration  i  in test
                 chamber j)
                                    169

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

    HI  = n2 = n3 = n4 = n5 = ng = 3
    N  = 18
    Tl  = YH + Y12 + Y13 = 9.281
    T2  = Y21 + Y22 + Y23 = 9.239
    T3  = Y31 + Y32 + Y33 = 8.627
    14  = Y41 + Y42 + Y43 = 7.928
    Tb  = Y5i + Y52 + Y53 = 6.438
    T6  = Y6l + Y62 + Y63 = 4.938
    G  = TT  + T2 + T3 + T4 + T5 + T6 =  46.451

          p
    SSB = £T1-2/ni - Q2/N
         1 = 1

        = JJ374.606) - (46.451)2  = 4.997
           3                 TS~~

          p    n-i
    SST = S   S Yjj2 - G2/N
         i=l  j=l

        = 124.890  - (46.451)2  = 5.018
    SSW = SST - SSB = 5.018 - 4.997 = 0.021
    SB2 = SSB/p-1  = 4.996/6-1 = 0.999
    SW2 = SSW/N-p = 0.021/18-6 = 0.0018

12.2.7.3 Summarize these calculations in  the  ANOVA table  (Table  9).


             TABLE 9.  ANOVA TABLE FOR DUNNETT'S PROCEDURE EXAMPLE
Source
Between
Within
df
5
12
Sum of Squares
(SS)
4.997
0.021
Mean Square(MS)
(SS/df)
0.999
0.0018
    Total          17            5.017
                                      170

-------
12.2.7.4  To perform the individual  comparisons, calculate the t
statistic for each concentration,  and control  combination as follows:
                                     (1/ni)  + (1AM)
Where Yi  = mean cell  count for toxicant concentration  i
      Y|  = mean cell  count for the control
      $W  = square root of within mean sqaure
      n-|  = number of replicates for control
      n-j  = number of replicates for concentration i.

12.2.7.5  Table 10 includes the calculated t values for each
concentration and control combination.  In this  example,  comparing  the
5 ug/L concentration with the control  the calculation is  as follows:
                            (  3.094 - 3.080 )
                                                  = 0.405
                       [ 0.0424V (1/3)  + (1/3)  ]
                       TABLE 10.  CALCULATED T VALUES
           Toxicant Concentration
                  (ug Cd/L)
5
10
20
40
80
2
3
4
5
6
0.405
6.300
13.035
27.399
41.850
12.2.7.6 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-j is greater than the critical
                                    171

-------
value.  Since t3, t4, ts and tg are greater than 2.50,  the 10,
20, 40 and 80 ug/L concentrations have significantly lower mean  cell
counts than the control.  Hence the NOEC and the LOEC for the test are
5 ug/L and 10 ug/L,respectively.

12.2.7.7  To quantify the sensitivity of the test,  the minimum
significant difference (MSD) that can be statistically detected  may be
calculated.
                   MSD = d Sw  >/ (l/n-|) + (1/n)

Where  d  = the critical value for the Dunnett's procedure
       SN = 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-j = the number of replicates in the control.

12.2.7.8  In this example:
                   MSD = 2.50 (0.0424) V (1/3)  + (1/3J
                       = 2.50 (0.0424K0.8165)
                       = 0.086

12.2.7.9  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 untransfonned values for the control mean  and  the
       difference calculated in 1.

          10(3.094) « 1241.6
          10(3.008) s -,018.6

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

        MSUU = 1241.6 - 1018.6 = 223

12.2.7.10  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.

12.2.7.11  This represents a decrease in  growth  of  18% from  the  control.
                                    172

-------
12.3  BIOSTIMULATION

12.3.1  Where the growth response in effluent (or surface  water)  exceeds
growth in the control  flasks, the percent stimulation,  $(%),  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 . , 1980b):
                                     _      x 100


13.  TEST PRECISION AND ACCURACY

13.1  PRECISION

13.1.1  Data from repetitive 96-h toxicity tests conducted with three
reference toxicants, using medium containing EDTA, are shown in Table 11
The relative standard deviation (coefficient of variation) of the LCI s
ranged from 47% to 83%.

13.2  ACCURACY

3.2.1  The accuracy of toxicity tests cannot be determined.
                                    173

-------
              TABLE 11.   PRECISION  OF  THE  SELENASTRUM  CAPRICORNUTUM, 96-H
TOXICITY TEST
, USING
REFERENCE TOXICANTS
Toxicant
Test
No.

1
2
3
4
5
6
7
8
9
10
11
N
Mean
SO
CV
Cadmium
Chloride
GC1 a NOECb
(ug Cd/L)
0.201
0.647
.372
.242
.638
2.37
2.27
1.23
0.347
.608
1.72
n
0.968
0.806
83%
UL
0.272
1.33
5.45
0.446
0.972
3.27
2.98
1.78
0.652
1.01
2,38




LL
0.181
0.198
0.220
0,0981
0,352
1.54
1.59
0.748
0.137
0.296
1.11




tug
LT
LT

LT
LT
LT


LT
LT





Cd/L)
0.49C
10.0
1.0
2.0
2.0
8.0
5.0
5.0
5.0
5.0
5.0




(ug/L)
20.7
NC
16.7
41.3
40.2
47.0
43.4
84.3
40.5
33.4

9
40.8
19.3
47%
Sodium
Pentachlorophenate
EC1
UL
27.6
—
21.6
46.8
45.5
53.1
48.7
90.0
48.4
40.6






LL
13.9
—
11.9
34.8
33.9
39.8
37.1
76.3
30.8
25.5





NOEC
(ug/L)
62.5
80.0
40.0
66.0
LT 66.0
82.0
LT 66.0
102
LT 66.0
82.0






(mg/L)
2.57
1.32
5.57
6.41
1.26
2.85





6
3.33
1.98
60%
Sodium
Uodecyl Sulfate
EC1
UL
3.13
1.77
6.60
7.52
1.81
2.98










LL
1.97
0.890
4.30
4.98
0.766
2.72









NOEC
(mg/L)
5.0
2.5
10.0
7.5
LT 5.0
5.0









a EC1  (threshold concentration)  and upper (UL) and lower (LL) confidence limits  determined by Probit Analysis,
bNOEC determined with Dunnett's  Test.
CLT = NOEC less than the lowest  concentration tested.
Reference  toxicant concentrations
  Cadmium  Chloride (ug Cd/L):
  1: 0.49,  0.95, 1.88, 3.77,  7.27
  2: 10.0,  20.0, 40.0, 80.0
  3: 1.0,  2.0, 4.0, 8.0, 16.0
  4: 2.0,  4.0, 8.0, 16.0, 32.0
  5: 2.0,  4.0, 8.0, 16.0, 32.0
  6: 8.0,  16.0, 32.0, 64.0, 128
  7: 5.0,  10.0, 20.0, 40.0, 80.0
  8: 5.0,  10.0, 20.0, 40.0, 80.0
  9: 5.0,  10.0, 20.0, 40.0, 80.0
 10: 5.0,  10.0, 20.0, 40.0, 80.0
 11: 5.0,  10.0, 20.0, 40.0, 80.0
                                 used in the toxidty tests are listed below:
                                     Sodium Pentachlorophenate Jug/LJ:
1:
2:
3:
4:
5:
6:
7:
8:
9:
10:
62.
40.
40.
66.
66.
66.
66.
66.
66.
66.
5,
0,
0,
0,
0,
0,
o,
0,
0,
0,
125
80.
80.
82.
82.
82.
82.
82.
82.
82.
o,
0,
0,
0,
0,
o,
o,
o,
o,
250,
160
160
102
102
102
102
102
102
102
500,
, 320,
, 320,
, 128,
, 128,
, 128,
, 128,
, 128,
, 128,
, 128,
1000
640
640
160
160
160
160
160
160
320
Sodium Dodecyl Sulfate (mg/L}:
 1: 2.5,  5.0,  7.5, 10.0, 12.5,  15.0
 2: 2.5,  5.0,  10.0, 12.5, 15.0,  20
 3: 2.5,  10.0, 12.5, 15.0, 20.0
 4: 5.0,  7.5,  12.5, 15.0, 20.0
 5: 5.0,  12.5, 20.0, 40.0, 80.0
 6: 2.5,  5.0,  12.5, 15.0, 20.0,  40.0

-------
                              SELECTED REFERENCES
Adams, N., et al.   1985.   Toxicity of eight water-soluble  organic chemicals
   to Selenastrum capricornutum: A study of methods for calculating toxic
   values using different growth parameters.   Arch.  Environ.  Contain. Toxicol.
   14:333.

Adams, N., et al.   1986.   Effect of acetone on the toxicity  of  four chemicals
   to Selenastrum capricornutum.  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.   Hydrobiol.  5(5):19-26.

APHA.  1985.  Standard Methods  for the Examination of Water  and
   Wastewater.  16th Ed.  American Public Health Association,  Washington, D.C.

ASTM.  1988.  Standard practice for conducting acute toxicity tests with
   fishes, macroinvertebrates,  and amphibians. ASTM E-729-88,  American
   Society for Testing and Materials, Philadelphia, Pennsylvania.

ASTM.  1988.  Standard guide for conducting acute toxicity tests on aqueous
   effluents with fishes, macroinvertebrates,  and amphibians.   ASTM
   E-1192-88,  American Society for Testing and Materials, Philadelphia,
   Pennsylvania.

ASTM.  1988.  Standard guide for conducting early life-stage toxicity tests
   with fishes.  ASTM E-1241-88, American Society for Testing and Materials,
   Philadelphia,  Pennsylvania.

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.

Barron, M. G., and I. R.  Adelman.  1985. Temporal characterization of growth
   of fathead minnow (Pimephales promelas)  larvae during sublethal hydrogen
   cyanide exposure.  Comp. Biochem. Physiol.  C.  81:341.

Bartlett, M.S.  1937.  Some examples of statistical  methods  of  research
   in agriculture and applied biology.  J.  Royal  Statist.  Soc.  Suppl.
   4:137-183.

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.  1982.  User's guide for conducting life-cycle chronic
   toxicity tests with fathead  minnows (Pimephales promelas).   Environmental
   Research Laboratory, U. S. Environmental Protection  Agency,  Duluth,
   Minnesota, EPA-600/8-81-011.

                                      175

-------
Benoit, D. A., F. A. Puglisi, and D.  L.  Olson.   1982.   A fathead minnow,
   Plmephales promelas, early life stage toxicity test method evaluation  and
   exposure to four organic chemicals.   Environ.  Pollut.  (Series A)  28:189-197,

Berner, D. B.  1985.  The taxonomy of Ceriodaphm'a (Crustacea:Cladocera)
   in U. S. Environmental Agency Cultures.   Environmental  Monitoring and
   Support Laboratory, U. S. Environmental  Protection  Agency, Cincinnati,
   Ohio, 45268, EPA/600/4-86/032.

Birge, W. J., and J. A. Black.   1981.  In situ  acute/chronic  toxicological
   monitoring of industrial effluents for the NPOES biomonitoring  program
   using fish and amphibian embryo/larval stages as test organisms.   Office of
   Water Enforcement and Permits, U.  S.  Environmental  Protection Agency,
   Washington, DC, OWEP-82-001.

Birge, W. J., J. A. Black, and B. A.  Ramey.   1981.  The reproductive
   toxicology of aquatic contaminants.   Hazard  assessments of chemicals,
   current developments, Vol. 1, Academic Press,  Inc., p.  59-114.

Birge, W. J., J. A. Black, 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): In Press.

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, D.C., p.  108-118.

Birge, W. J., J. A. Black, and A. G.  Westerman.  1985.   Short-term
   fish and amphibian embryo-larval tests for determining the effects of
   toxicant stress on early life stages  and  estimating chronic values for
   single compounds and complex effluents.   Environ. Tox.  Chem. 4:807-821.

Birge, W. J., J.A., Black, A. G., Westerman, and B.  A.  Ramey.  1983.
   Fish and amphibian embryos - A model  system  for evaluating teratogenicity.
   Fundam. Appl. Toxicol. 3: 237-242.

Birge, W. J., and R. A. Cassidy.  1983.   Importance of structure-activity
   relationships in aquatic toxicology.   Fundam.  Appl.  Toxicol. 3:359-368.

Black, J. A., W. J. Birge, A. G. Westerman,  and P. C.  Francis.  1983.
   Comparative aquatic toxicology of  aromatic hydrocarbons.  Fundam.  Appl.
   Toxicol. 3: 353-358.

Blaise, C. R.  1986.  Micromethod for acute  aquatic toxicity  assessment using
   the green alga Selenatrum capricornutum.   Toxic.  Assess. 1:377-385.

Burton, G. A., Jr., D. Nimmo, D. Murphey, and F.  Payne.   1987.  Stream profile
   determinations using microbial activity assays and  Ceriodaphm'a.   Environ.
   Toxicol. Chem. 6:505-513.

Burton, G. A., Jr., A. Drotar,  J. Lazorchak, and L.  Bahls. 1987.
   Relationship of microbial activity and Ceriodphnia  responses to mining

                                      176

-------
   impacts on the Clark Fork River, Montana.  Arch.  Environ.  Contam.  Toxicol.
   16:523-530.

Chengalath, R.  1982.  A faunistic and ecological  survey of the littoral
   Cladocera of Canada.  Can. J. Zoo!. 60:2668-2682.

Chiaudani, G., and M. Vighi.  1974.  The N:P ratio and tests with
   Selenastrum to predict eutrophication in lakes.  Water Res.  8:1063-1069.

Conit, R. 0.  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. p. 466-467.

Cooney, 0. D., G. M. DeGraeve, E. L. Moore, T. 1.  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, Ohio.

Cowgill, U. M.  1987.  Critical analysis of factors affecting the sensitivity
   of zooplankton and the reproducibility of toxicity test results.  Wat. Res.
   21(12):1453-1462.

Cowgill, U. M., K. I. Keating, and I.  T. Tokahashi.   1985.  Fecundity and
   longevity of Ceriodaphnia dubia/affinis in relation to diet at two
   different temperatures.  J. Crust.  Biol. 5(31:420-429.

Cowgill, U. M., D. P. Milazzo, and C.  K. Meagher.   1988.  New diet for
   Ceriodaphnia dubia.  Bull. Environ. Contam. 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 Daphnia magna and Ceriodaphnia
   dubia-affinis tested at two different temperatures.  Environ. Toxicol.
   rn"em7  4:415-422.

Czeczuga, B., and E. Bobiatynskaksok.   1972.  The  extent of consumption of the
   energy contained in the food suspension by Ceriodaphnia reticulata
   (Jurine).  In: Z. Kajak, and A. Hillbricht-Ilkowska, eds., Productivity
   Problems of Freshwaters, Proc. IBP-UNESCO Symposium, Kazimierz Dolny,
   Poland, May 6-12, 1970. Krakow, Poland.

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, 0. H.  Dean,
   M. D. Marcus, and D. 0. Mclntyre. 1988. Fathead minnow 7-day test: round
   robin study.  Intra- and inter!aboratory study  to determine the
   reproducibility of the seven-day fathead minnow larval survival and  growth
   test.  Columbus, Battelle, Columbus, Ohio.  Available from:  American
   Petroleum Institute, 1220 L Street, NW, Washington, DC  20005, Rept.
   No. 4486.

                                       177

-------
DeGraeve, G. M., J. D. Cooney, B.  H.  Marsh,  T.  L.  Pollock,  and  N. G.
   Reichenbach.  1989.  Intra- and Interlaboratory study  to determine the
   reproducibility of the seven-day Ceriodaphnia dubia survival  and
   reproduction tests.  Battelle,  Columbus Division,  Columbus,  Ohio  (In
   preparation).

DeWoskin, R. S.  1984.  Good laboratory practice regulations:
   a comparison.  Research Triangle Institute,  Research Triangle Park, North
   Carolina. 63 pp.

Dixon, W. J., and F. J.  Massey, Jr.   1983.  Introduction  to statistical
   analysis.  Fourth edition.   McGraw Hill,  New York.

Draper, N. R., and J. A. Oohn.  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.  1988.  Comparison
   of measured in-stream biological  responses with responses predicted by
   Ceriodaphnia chronic  toxicity test.   North Carolina Division of
   Environmental Management, Raleigh, North  Carolina  (In  Preparation).

Elnabarawy, M. T., A. N. Welter, and R. R. Tobide.au.   1986.  Relative
   sensitivity of three  daphnid species to selected organic and inorganic
   chemicals.  Environ.  Toxicol. Chem.   5:393-398.

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.

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.

Finney, D. J.  1971.  Probit analysis.  Third Edition.   Cambridge Press,
   New York.  668 pp.
                                      178

-------
Fitzgerald, G. P.  1972.  Bioassay analysis of nutrient availability.
   In: Allen, H.E., and J.R. Kramer, eds.   Nutrients in natural  waters.  John
   Wiley & Sons, Inc., New York.  pp. 147-165.

Forsberg, C. G.  1972.  Algal assay procedure.  JWPCF 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.,  Proc.  of the
   Eutroph. Bioassess. Workshop,  19-21  June, 1969.  U.S. Pacific Northwest
   Water Laboratory, Corvallis, Oregon, p. 158-170.

Gophen, M.  1976.  Temperature dependence on food intake,  ammonia
   excretion and respiration in Ceriodaphnia reticulata (Jurine)  (Lake
   Kinneret, Israel).  Freshw. Biol. 6(5):451-455.

Gophen, M.  1977.  Food and feeding habits of Mesocylops leuckarti  (Glaus)
   in Lake Kinneret (Israel).  Freshwat.  Biol. 7:513-518.

Gophen, M.  1979.  Bathymetrical  distribution and diurnal  migrations of
   zooplankton in Lake Kinneret (Israel)  with particular emphasis on
   Hesocyclop leuckarti (Glaus).   Hydrobiologia 64(3):199-208.

Gophen, M., B. Z. Cavari, and T.  Berman.   1974.   Zooplankton  feeding on
   differentially labelled algae and bacteria.  Nature 247:393-394.

Gotham, I. J., and G. Rhee.  1982.  Effects of a hexachlorobiphenyl  and
   pentachlorophenol on growth and photosynthesis of phytoplankton.  J.  Great
   Lakes Res. 8(2):328-335.

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 capricornutum.
   Environ. Sci. Techn. 16(4):225-229.

Green, J.  1976.  Changes in the zooplankton of Lakes Mutanda,  Bunyonyi
   and Mulehe (Uganda).  Freshw.  Biol.  6:433-436.

Greene, J. C., C. L. Bartels, W.  J. Warren-Hicks, B. R. Parkhurst,
   G. L. Linder, S. A. Peterson,  and W. E. Miller.  1988.   Protocols for
   short-term toxicity screening of hazardous waste sites. Environmental
   Research Laboratory, U. S. Environmental Protection Agency,  Corvallis,
   Oregon, EPA/600/3-88/029.

Greene, J. C., W. E. Miller, T. Shiroyama, and E. Merwin.   1975.   Toxicity
   of zinc to the green alga Selenastrum capricornutum as  a function of
   phosphorus or ionic strength.   In: Middlebrooks,  E.  J., T.  E.  Maloney,

                                       179

-------
   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, Oregon. EPA-660/3-75-034, p.  28-43.

Greene, J. C., R. A. Soltero, W. E. Miller, A. F. Gasperino,  and
   T. Shiroyama.  1976.  The relationship of laboratory algal  assays to
   measurements of indigenous phytoplankton in Long Lake,  Washington.
   In: Middlebrooks, E. J., E. H. Falkenborg, and T. E. Maloney, eds.,
   Biostimulation and nutrient assessment.   Ann Arbor Science Publ.,
   Ann Arbor, Mich. p. 93-126.

Hall, W. S., R. L. Paulson, L W. Hall, Jr., and D.  T. Burton.  1986.   Acute
   toxicity of cadmium and sodium pentachlorophenate to daphnids and fish.
   Bull. Environ. Contam. Toxicol. 37:308-316.

Hamilton, M. A.  1986.  Statistical analysis of cladoceran reproductivity
   test. Environ. Toxicol. Chem. 5:205-212.

Hart, W. B., Doudordoff, P. and Greenbank,  J.  1945.  The  Evaluation of
   the toxicity of industrial wastes, chemicals and other  substances to
   fresh-water fishes.  The Atlantic Refining Company, Philadelphia,
   Pennsylvania.

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.

Hodgkiss, I. J., and L. T. H. Chan.  1976.   Studies on Plover Cove
   Reservoir, Hong Kong.  Part 4.  The composition  and spatial  distribution of
   the crustacean zooplankton.  Freshw. Biol. 6(41:301-31 5.

Horning, W. B., II, and C. I. Weber.  1985   Short-term methods for  estimating
   the chronic toxicity of effluents and receiving  waters  to  freshwater
   organisms.  Environmental Monitoring and Support Laboratory - Cincinnati,
   U. S. Environmental Protection Agency, Cincinnati, Ohio,  EPA-600/4-85/014.

Hughes, M. M., M. A. Heber, S. C. Schimmel, and W.  J. Berry.   1987.   Guidance
   manual for conducting complex effluent and receiving water larval fish
   growth-survival studies with the sheepshead minnow (Cyprinodon
   variegatus).  Contribution No. X104.  In: Schimmel, S.  C.,  ed.,  Users guide
   to the conduct and interpretation of complex effluent toxicity tests at
   estuarine/marine sites.  Environmental Research  Laboratory,  U. S.
   Environmental Protection Agency, Narragansett, Rhode Island, Contribution
   796, 265 pp.

Keller, A. E., R. J. Dutton, G. Bitton, and T. L. Crisman.  1988.  Chronic
   toxicity of hydrothol-191 to Cerigdaphnia dubia  at 25 and  15°C.   Bull.
   Environ. Contam. Toxicol. 41:233-240.

                                       180

-------
Kititsyna, L. A., and 0. A. Sergeeva.  1976.   Effect of temperature
   increases on the size and weight of some invertebrate populations  in  the
   cooling reservoir of the Kurakhovsk State Regional Electric Power  Plant.
   Ekologiya 5:99-102.

Knight, J. T.s and W. T. Waller.  1987.  Incorporating Daphnia magna  into the
   seven-day Ceriodaphm'a effluent toxicity test method.  Environ! Toxicol.
   Chem. 6:635-645.

Kopp, J. F.  1983.  Guidelines and format for EMSL-Cincinnati  methods.
   Environmental Monitoring and Support Laboratory, U. S. Environmental
   Protection Agency, Cincinnati, Ohio, EPA-600/8-83-020.

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.

Lynch, M.  1979.  Predation, competition, and zooplankton community
   structure: An experimental study.  Limnol. Oceanogr. 24(2):253-272.

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:  F.  L. Mayer and J. L. Hamelink,
   eds., Aquatic Toxicology and Hazard Evaluation, ASTM STP 634,  American
   Society for Testing and Materials, Philadelphias Pennsylvania, p.  137-146.

Marking, L. L.5 and V. K. Dawson.  1973.  Toxicity of quinaldine sulfate to
   fish.  Invest. Fish Contr. No. 48, U. S. Fish and Wildlife Service,
   Washington, D.C.  8 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.   1986.
   Acute and chronic toxicity of ammonia to freshwater fish: A site-specific
   study.  Environ. Toxicol. Chem. 5:437-442.

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. Bahner and D. J. Hansen,  eds.
   Aquatic Toxicology and Hazard Assessment: Eighth Symposium. ASTM  STP  891.
   American Society for Testing and Materials, Philadelphia, Pennsylvania,
   p. 375-381.

McPhee, C.  1961.  Bioassay of algal production in chemically altered
   waters.  Limnol. Oceanogr. 6:416-422.
                                       181

-------
Michaels, R. A., R. G. Rowland, and C.  F.  Wurster.   1982.   Polychlorinated
   biphenyls (PCB) inhibit photosynthesis  per cell  in the  marine  diatom
   Thalassiosi'ra pseudonana. Environ.  Pollut. (Ser.  A.)  27:9-14.

Miller, R. G.  1981.  Simultaneous statistical  inference.   Springer-Verlag,
   New York. 299 pp.

Miller, W. E., J. C. Greene, and T. Shiroyama.   1978.  The Selenastrum
   capricornutum Printz algal  assay bottle test.   Environmental Research
   Laboratory, U. S. Environmental Protection Agency, Corvallis,  Oregon,
   EPA-600/9-78-018.

Miller, W. E., and T. E. Maloney.  1971.   Effects of secondary and tertiary
   wastewater effluents on algal  growth in a lake-river system.   JWPCF
   43(12):2361-2365.

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 T. J. Norberg-King,  eds.  1985.   Validity of effluent and
   ambient toxicity tests for predicting biological  impact, Scippo Creek,
   Circleville, Ohio.  Environmental Research Laboratory,  U.  S. Environmental
   Protection Agency, Duluth,  Minnesota, EPA/600/3-85/044.

Mount, D. I., and T. J. Norberg-King.   1986.   Validity of  effluent and ambient
   toxicity tests for predicting biological  impact,  Kanawha River, Charleston,
   West Virginia.  Environmental  Research  Laboratory, U. S. Environmental
   Protection Agency, Duluth,  Minnesota, EPA/600/3-86/006.

Mount, D. I., T. J. Norberg-King, R. Keen, and J.  T. Taraldsen.   1987.
   A reference tests water for cladocerans.   Abstract, llth Annual Symposium,
   Aquatic Toxicology and Hazard Assessment, American Society for Testing and
   Materials, May 10-12, 1987, Cincinnati, Ohio.

Mount, D. I., T. J. Norberg-King, and A. E.  Steen.   1986a.  Validity  of
   effluent and ambient toxicity tests for predicting biological  impact,
   Naugatuck River, Connecticut.   Environmental  Research Laboratory,  U. S.
   Environmental Protection Agency, Duluth,  Minnesota, EPA/600/8-86/005.

Mount, D. I., A. E. Steen, and T. J. Norberg-King,  eds.  1985b.   Validity of
   effluent and ambient toxicity testing for predicting biological impact on
   Five Mile Creek, Birmingham, Alabama.   Environmental  Research  Laboratory,
   U. S. Environmental Protection Agency,  Duluth,  Minnesota,  EPA/600/8-85/015.

Mount, D. I., A. E. Steen, and T. J. Norberg-King,  eds.  1985b.   Validity
   of effluent and ambient toxicity tests  for predicting biological  impact,
   Ohio River, near Wheeling,  West Virginia.  Environmental Research
   Laboratory, U. S. Environmental Protection Agency, Duluth, Minnesota,
   EPA/600/3-85/071.

Mount, D. I., A. E. Steen, and T. J. Norberg-King.   1986b.  Validity  of
   effluent and ambient toxicity tests for predicting biological  impact, Back
   River, Baltimore Harbor,  Maryland.   Environmental Research Laboratory,
   U. S. Environmental Protection Agency,  Duluth,  Minnesota,  EPA/600/8-86/001.
                                       182

-------
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.

Mount, D. I., N. A. Thomas, T. J. Norberg, M.  T.  Barbour, T.  H.  Roush,
   and W. F. Brandes.  1984.  Effluent and ambient toxicity testing and
   instream community response on the Ottawa River, Lima, Ohio.   Environmental
   Research Laboratory, U. S. Environmental Protection Agency,  Duluth,
   Minnesota, EPA/600/3-84/080.

Neiheisel, T. W., J. R. Menkendick, F. A. Kessler,  W. B.  Horning, and C.  I.
   Weber.  1988a.  Multi-laboratory study of Ceriodaphnia chronic toxicity
   test.  In Press.

Neiheisel, T. M., W. B. Horning,  II, B.  M. Austern, D. F. Bishop, T.  L.  Reed,
   and J. F. Estenik.  1988b.  Toxicity  reduction at municipal  wastewater
   treatment plants.  OWPCF 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, Virginia.

Norberg, T. J., and D. I. Mount.   1985.   A new fathead minnow (Pimephales
   promelas) subchronic toxicity  test.  Environ.  Toxicol. Chem.  4(5):711-718.

Norberg, T. J., and D. I. Mount.   1985.   Diets for Ceriodaphnia reticulata
   life-cycle tests.  In: R. 0. Cardwell, R. Purdy, and R.  C.  Bahner, eds.,
   Aquatic Toxicology and Hazard Assessment, Seventh Symposium,  ASTM STP 854,
   American Society for Testing and Materials, Philadelphia,  p.  42-52.

Norberg-King, T.  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,  Montana.

Norberg-King, I.  1989.  An evaluation of the fathead minnow  seven-day
   subchronic test for estimating chronic toxicity.  Environ.  Toxicol. Chem.
   In Press.

Norberg-King, T. J., and D. I. Mount, eds.  1986   Validity  of effluent and
   ambient toxicity tests for predicting biological impact, Skeleton Creek,
   Enid, Oklahoma.  Environmental Research Laboratory, U. S.  Environmental
   Protection Agency, Duluth, Minnesota, EPA/600/8-86/002.

O'Brien, W. J.  1972.  Limiting factors  in phytoplankton algae:  their
   meaning and measurement.  Science 178:616.

O'Brien, W. J.  1974.  Filtering  rate of Ceriodaphnia reticulata in pond
   waters of varying phytoplankton concentration.flier. Midi.  Nat.
   91(2):509-512.

Owsley, J. A., and D. E. McCauley.  1986.  Effect of extended sublethal
   exposure to sodium selenite on Ceriodaphnia affinis. Bull.  Environ.
   Contain. Toxicol. 36:876-880.
                                      183

-------
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.O. Hartley.   1962.  Biometrika tables  for
   statisticians. Vol. 1.  Cambridge Univ.  Press, England, p. 65-70.

Peltier, W.  1978.  Methods for measuring the acute toxicity of  effluents
   to aquatic organisms.  Second  edition.   Environmental  Monitoring and
   Support Laboratory - Cincinnati,  U. S.  Environmental  Protection Agency,
   Cincinnati, Ohio, EPA-600/4-78-012.

Peltier, W., and C. I. Weber, eds.   1985.   Methods for measuring the acute
   toxicity of effluents to freshwater and marine organisms.  Third edition.
   Environmental Monitoring and Support Laboratory, U. S. Environmental
   Protection Agency, Cincinnati, Ohio, EPA 600/4-85-013, 230 pp.

Pennak, R. W.  1978.  Freshwater  invertebrates of the United States.
   Second edition.  Ronald Press, New York, 769 pp.

Pickering, Q. H.  1988.  Evaluation and comparison of two short-term fathead
   minnow tests for estimating chronic toxicity.  Wat. 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.  JWPCF 54(2):181-186.

Sachdev, D. R., and N. L.  Clesceri.   1978.  Effects of organic fractions
   from secondary effluent on Selenastrum capricornutum  (Kutz.). OWPCF
   52:1810-1820.

Scheffe, H.  1959.  The analysis  of variance.  John Wiley and Sons,
   New York, 477 pp.

Schoenberg, S. A., and A.  E. Maccubbin.  1985. Relative  feeding  rates of free
   and particule-bound bacteria by  freshwater macrozooplankton.   Limnol.
   Oceanogr. 30{5) :1084-1090.

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 Ceriodaphnia
   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, p. 262-299.

Snedecor, G. W., and W. G. Cochran.   1980.  Statistical  Methods.  Seventh
   edition.  Iowa State University  Press,  Ames, 593 pp.
                                      184

-------
Spencer, D. F., and L. H. Nichols.   1983.   Free nickel  ion  inhibits  growth
   of two species of green algae.  Environ. Pollut.  (Ser. A.)  31:97-104.

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.s New York.

Takahashi, I. T., U. M. Cowgill, and P. li. Murphy.  1987.   Comparison of
   ethanol toxicity to Daphnia magna and Ceriodaphnia 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,  25-29
   October, 1972.  Nordforsk. Secretariat of Environm.  Sci.,  Helsinki.   Publ.
   1973-2, p. 5-17.

Tarzwell, C. M.  1971.  Bioassays to determine allowable waste
   concentrations in the aquatic environment.  In:  H.  A.  Cole (Organizer),
   A discussion on biological effects of pollution in the  sea.  Proc. Roy.
   Soc. Lond. B 177(1048):279-285.

Thomas, R. E., and R. L. Smith.  1975.  Assessing treatment process
   efficiency with the algal assay test.   In: E. J.  Middlebrooks,
   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, Oregon, EPA-660/3-75-034,  p.  244-248.

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.  National
   Eutrophication Program. U. S. Environmental Protection Agency,
   Environmental Research Laboratory, Corvallis, Oregon, 82 pp.

USEPA.  1975.  Methods for acute toxicity tests with fish,
   macroinvertebrates, and amphibians.  Environmental Research Laboratory,
   U. S. Environmental Protection Agency, Duluth, Minnesota,  EPA-660/3-75-009,

USEPA.  1977.  Occupational health and safety manual.  Office of Planning
   and Management, U. S. Environmental Protection Agency, Washington, DC.

USEPA.  1979a.  Handbook for analytical quality control in  water and
   wastewater laboratories.  U. S.  Environmental Protection Agency,
   Environmental Monitoring and Support Laboratory,  Cincinnati, Ohio,
   EPA-600/4-79-019.
                                       185

-------
USEPA.  1979b.  Methods for chemical  analysis of  water and wastes.
   Environmental Monitoring and Support Laboratory,  U.  S. Environmental
   Protection Agency, Cincinnati, Ohio, EPA-600/4-79-020.

USEPA. 1979c.  Interim NPDES compliance biomonitoring  inspection manual.
   Office of Water Enforcement, U.  S.  Environmental  Protection Agency,
   Washington, DC, MCD-62.

USEPA.  1979d.  Good laboratory practice standards for health effects.
   Paragraph 772.110-1, Part 772 - Standards for  development of test data.
   Fed. Reg. 44:27362-27375, May 9, 1979.

USEPA.  1980a.  Appendix B  - Guidelines for Deriving Water Quality
   Criteria for the Protection of Aquatic  Life and Its Uses.  Federal
   Register, Vol. 45, No. 231, Friday, November 28,  1980.

USEPA.  1980b.  Proposed good laboratory practice guidelines for toxicity
   testing.  Paragraph 163.60-6.  Fed. Reg. 45:26377-26382, April 18, 1980.

USEPA.  1980c.  Physical, chemical, persistence,  and ecological effects                   I
   testing; good laboratory practice standards (proposed rule).  40 CFR  772,
   Fed. Reg. 45:77353-77365, November 21,  1980.

USEPA.  1985.  Technical Support Document for Water Quality-based
   Control.  Office of Water, U. S. Environmental Protection Agency,
   Washington, DC.

USEPA.  1986. 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.  1987a.  Ambient water quality criteria for zinc.  Criteria
   and Standards Division,  Office of Water Regulations and Standards, U. S.
   Environmental Protection Agency, Washington,  UC,  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.

Vigerstad, T. J., and L. J. Tilly.   1977.   Hyperthermal  effluent effects
   on heleoplanktonic Cladocera and the influence of submerged macrophytes.
   Hydrobiol. 55(l):81-86.

Walsh, G. E., and S. V. Alexander.   1980.   A marine algal bioassay method:
   Results with pesticides  and industrial  wastes. Water,  Air, and 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.
                                      186

-------
Walters, D. B., and C. W,  Jameson.   1984.   Health and safety for toxiclty
   testing.  Butterworth Pub!., Woburn,  Massachusetts.

Ward, G. S.s and P. R. Parrish.  1980.   Evaluation of early life stage
   toxicity tests with embryos and juveniles of sheepshead minnows.   In:
   0. G. Eaton, P. R.  Parrish and A. C.  Hendricks, Aquatic Toxicology,  ASTM
   STP 707, American Society for Testing and Materials,  Philadelphia,
   Pennsylvania, p. 243-247.

Weber, C. I., ed.  1973.  Biological field and laboratory methods for
   measuring the quality of surface waters and wastes.   Office of Research and
   Development, U, S.  Environmental Protection Agency,  Cincinnati, Ohio,
   EPA 670/4-73-001, 200 pp.

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, Michigan, p. 93-126.

Weiss, C. M.  1976.  Evaluation of the algal assay procedure.
   U. S. Environmental Protection Agency,  Corvallis, Oregon, EPA-660/3-76-064.

Weiss, C. M., and R. W. Helms.  1971.  Inter!aboratory precision test:
   An eight-laboratory evaluation of the Provisional Algal Assay Procedure
   Bottle Test.  National Eutrophication Research Program.  U. S.
   Environmental Protection Agency, Corvallis, Oregon.

Winner, R. W.  1988.  Evaluation of the relative sensitivities of 7-D
   Daphnia magna and Cerlodaphnia dubia toxicity tests for cadmium and sodium
   pentachlorophenate.  Environ. Toxicol.  Chem. 7:153-159.

Weltering, 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.  0. Fish. Res. Board.  Can. 35:479-481.

Wong, C. K.  1981.  Predatory feeding behavior of Epischura lacustris
   (Copepoda:Calanoida) and prey defense.   Can. J. Fish Aquat. Sci.
   38(3):275-279.

Wright, J. Jr., F. A. Camp, and J. Cairns, Jr.  1974.  Preliminary algal
   bioassays to determine nutrients limiting algal productivity in Mountain
   Lake, Virginia.  Assoc. SE Biol. Bull.  21(2):92.

Yoshioka, Y., et al.  1986.  Correlation of the five test methods to
   assess chemical toxicity and relation to physical properties.  Ecotoxicol.
   Environ. Saf. 12:15.

Young, T. C., J. V. DePinto, S. E. Flint,  M. S. Switzenbaum, and
   J. K. Edzwald.  1982.  Algal availability of phosphorus in municipal
   wastewater.  JWPCF 54(11):1505-1516.
                                       187

-------
                                   APPENDICES


A.  Independence, Randomization, and Outliers  	   189

    1. Statistical Independence  	   189

    2. Randomization	189

    3. Outliers	190

B.  Validating Normality and Homogeneity of Variance
      Assumptions	192

    1. Introduction	192

    2. Test for Normal  Distribution of Data	192

    3. Test for Homogeneity of Variance	200

    4. Transformations  of Data	201

C.  Dunnett's Procedure  	   204

    1. Manual Calculations 	   204

    2. Computer Calculations 	   211

D.  Bonferroni's T-test  	   216

E.  Steel's Many-one Rank Test	221

F.  Wilcoxon Rank Sum Test	225

G.  Fisher's Exact Test	231

H.  Toxicity Screening  Test - Comparison of Control  with
      100% Effluent or  Instream Waste Concentration   	   240

I.  Probit Analysis	244
                                      188

-------
                                   APPENDIX A

                   INDEPENDENCE, RANDOMIZATION,  AND  OUTLIERS1


1.  STATISTICAL INDEPENDENCE

1.1  Dunnett's Procedure and Bonferroni's T-test are parametric  procedures
based on the assumptions that (1) the observations within  treatments  are
independent and normally distributed, and (2)  that the variance  of  the
observations is homogeneous across all toxicant concentrations and  the
control.  Of the three possible departures from the  assumptions,
non-normality, heterogeneity of variance, and  lack of independence, those
caused by lack of independence are the most difficult to deal with  (see
Scheffe, 1959).  For toxicity data, statistical  independence means  that given
knowledge of the true mean for a given concentration or control,  knowledge  of
the error in any one actual observation would  provide no information  about  the
error in any other observation.  Lack of independence is difficult  to assess
and difficult to test for statistically.  It may also have serious  effects  on
the true alpha or beta level.  Therefore, it is of utmost  importance  to be
aware of the need for statistical independence between observations and to  be
constantly vigilant in avoiding any patterned  experimental procedure  that
might compromise independence.  One of the best ways to help insure
independence is to follow proper randomization procedures  throughout  the test.

2.  RANDOMIZATION

2.1  Randomization of the distribution of test organisms among test vessels,
and the arrangement of treatments and replicate vessels is an important part
of conducting a valid test.  The purpose of randomization  is to  avoid
situations where test organisms are placed serially  by level of  concentration
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 is described  using  the Fathead  Minnow Larval
Survival and Growth test.  For a test design with five treatments,  a  control,
and four replicates at each treatment, there would be 24 experimental units,
i.e., 24 positions to be randomized.  There are several ways to  randomly
assign the positions.  Random numbers may be selected from a random numbers
table or may be generated by computer software.

2.3  In this example, the first four random numbers  selected would  be used  for
the four control replicates.  The selection of random numbers would continue,
four at a time, each group being assigned to the four replicates of a given
test concentration, progressing from the lowest concentration to the  highest.
The rank ordering of these random numbers would determine  the relative
positioning for the controls and concentration levels.
                                      189

-------
2.4  The result of this randomization procedure is presented in Table A.I,
using an effluent concentration series of 1.0%, 3,2%,  10.0%, 32.0%,  and 100%,


 TABLE A.I.  RANDOMIZATION OF THE POSITIONS OF EXPERIMENTAL UNITS USING A
             DESIGN OF FOUR ROWS AND SIX COLUMNS
1002
3.2%
10.0%
3.2%
10.0%
3.2%
Control
Control
1.0%
100%
1.0%
10.0%
100%
32. t^
19 n^
•JL., \Jto
19 W/
O£. \f/o
i y%.
O. £a
Control
10.0%
100%
1.0%
Control
32.0%
1.0%
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.

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).
                                    190

-------
    TABLE A.2.   TABLE OF RANDOM NUMBERS1
10 oe
37 54
08 42
99 01
12 80
06 06
31 06
85 26
03 57
73 79
98 52
11 80
83 45
88 08
90 59
e& 48
80 12
74 35
m 91
09 89
91 49
80 33
44 10
12 55
03 00
01 19
IS 47
94 55
42 48
23 52
04 49
00 54
35 96
£9 80
46 05
32 17
09 23
19 50
45 15
94 86
98 08
33 18
80 95
79 75
18 63
74 02
54 17
11 66
48 32
09 07
73
20
26
9O
76
57
01
97
33
04
01
SO
29
64
46
11
43
09
02
32
91
69
48
07
64
09
44
72
11
37
35
99
31
80
88
W
46
54
51
43
02
51
10
24
33
94
84
44
47
49
25 33
48 05
89 S3
25 29
99 70
47 17
08 05
76 02
21 35
57 63
77 67
54 31
96 34
02 00
73 48
76 74
56 35
98 17
S8 03
05 05
45 23
45 98
19 49
37 42
93 29
04 40
52 66
85 73
02 13
S3 17
24 94
76 54
53 07
83 91
52 30
05 97
14 06
14 30
49 38
19 94
48 26
02 32
04 06
91 40
25 37
39 02
56 11
98 83
79 38
41 38
76
«4
19
09
80
34
45
02
05
03
14
39
OS
8Q
87
17
17
77
06
14
08
26
85
11
16
26
95
07
97
73
75
64
26
45
01
87
20
01
19
36
45
41
96
71
58
77
SO
52
31
87
52 01
89 47
64 50
37 67
IS 73
07 27
57 18
05 16
32 54
62 96
90 66
80 82
28 89
SO 75
51 76
46 85
72 70
40 27
25 22
22 56
47 92
94 03
IS 74
10 00
50 53
45 74
27 07
89 75
34 4O
20 88
24 63
05 18
89 80
42 72
39 09
37 92
11 74
75 87
47 60
10 81
24 02
94 15
38 27
96 12
14 50
55 73
9» 33
07 98
24 96
63 79
35 86
42 96
93 03
07 IS
01 47
68 50
24 06
fie 92
70 48
47 78
86 07
77 32
80 83
84 01
49 68
09 6O
80 15
72 14
91 48
85 14
70 86
68 58
79 54
20 40
44 84
77 74
99 53
43 87
87 21
98 37
38 24
81 59
93 54
08 42
22 86
52 41
52 04
53 79
72 46
08 51
84 04
09 49
07 74
82 96
66 71
22 70
71 43
48 27
47 10
19 76
84 07
24 80
23 20
38 31
04 03
30 69
35 30
08 06
90 S5
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
69 36
54 62
16 86
68 93
45 80
96 11
33 35
83 60
77 28
05 56
15 95
40 41
43 00
34 88
44 99
89 43
20 15
69 86
31 01
97 79
05 33
59 38
02 29
35 58
35 48 76
52 « 37
90 25 60
13 11 05
23 66 53
73 61 70
34 26 14
57 48 IS
35 75 48
S3 42 82
94 05 58
56 82 48
67 00 78
66 79 51
60 89 28
77 69 74
82 23 74
60 02 10
68 72 03
75 67 88
28 35 54
73 41 35
92 65 75
07 46 97
95 25 63
43 37 29
78 38 48
24 44 31
84 87 67
59 14 16
25 10 25
96 38 96
13 54 62
94 97 00
14 40 77
70 70 07
06 00 00
92 15 85
79 45 43
88 15 53
90 88 96
54 85 81
12 33 87
10 25 91
02 46 74
01 71 19
51 29 69
17 15 39
53 68 70
40 44 01
80 95 90
20 63 01
15 95 33
88 67 67
98 95 11
65 81 33
86 79 90
73 05 38
28 40 82
00 93 52
00 97 09
29 40 52
18 47 54
90 26 47
93 78 56
73 03 95
21 11 57
45 62 16
76 62 11
96 29 77
94 75 08
53 14 03
67 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
IS 74 39
06 67 43
59 04 79
01 54 03
99 09 47
88 69 54
25 01 62
74 85 22
OS 45 56
52 52 75
66 12 71
09 97 33
32 30 75
10 51 82
91 17
04 02
47 64
43 97
08 77
98 85
74 39
62 47
87 09
03 44
34 33
42 01
06 10
ft* 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
62 98
05 39
14 27
80 21
B2 ?5
34 40
75 46
16 15
39 29
00 82
35 08
04 43
12 17
It 19
23 40
18 62
83 49
35 27
50 50
62 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 81
27 49 45
29 16 66
03 36 08
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 28 55
05 W 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 S3 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
40 11 71
75 95 79
8» » 36
45 17 «
OS 03 24
12 33 M
00 M «
87 69 38
Dixon and Massey,  1983.
                     191

-------
                                   APPENDIX B

          VALIDATING NORMALITY AND  HOMOGENEITY OF VARIANCE ASSUMPTIONS!

1.  INTRODUCTION

1.1  Dunnett's Procedure and Bonferroni's T-test 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 non-parametric procedure such as Steel's Many-one
Rank Test may be more appropriate.   However, the decision on whether to use
parametric or non-parametric 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  A 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 Kolomogorov "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:

                          n      _ ?
                     D =  S {X.- Xr
                         1=1  ]

    Where: Xj =    the centered  observations and X"  is the overall  mean of
                   the centered  observations.  For  this set  of data, 1-0,
                   and D = 0.0412.
                                     192

-------
2.5  Order the centered observations  from  smallest to largest.
                               _  x(2)    _        x(n)
    Where x(i)  denotes the ith ordered observation.   The  ordered
observations are listed in Table B.3.

2.6  From Table B.4,  for the number of observations,  n, obtain  the
coefficients a-,, a2,  	, a., where k is  approximately n/2.  For the

data in this example, n = 20, k = 10.   The a-j  values  are  listed in
Table B.5.
2.7  Compute the test statistic,  W,  as follows:
                        1    [z  a  (X<"-1+1>   -  X»>)]2
    The differences, X
-------
        TABLE B.I.   FATHEAD LARVAL GROWTH  DATA (WEIGHT  IN NG)
                    FOR THE SHAPIRO-WILK'S TEST
Replicate
Control
                                   NaPCP  Concentration  (ug/L)
32
64
128
256
A
B
C
D
Mean(Yi)
Si2
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  (ug/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
                                     194

-------
TABLE B.3.  EXAMPLE OF THE SHAPIRO-WILK'S  TEST: ORDERED OBSERVATIONS
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
                                  195

-------
TABLE B.4. COEFFICIENTS FOR THE SHAPIRO-WILK'S  TEST!
V
\
I \
1
2
3
4
5

2
\
0.7071
_
—
_
—

3

0.7071
0.0000
—
—
—

4

0.6872
0.1667
—
—
—

5

0.6646
0.2413
0.0000
—
—

e

0.6431
0.2806
0.0875
—
—

T

0.6233
0.3031
0.1401
0.0000
—

8

0.6052
0.3164
0.1743
0.0561
—

9

0.5888
0.3244
0.1976
0.0947
0.0000

10

0.5739
0.3291
0.2141
0.1224
0.0399









\ n

i\
1
2
3
4
S
6
7
8
9
10

n
\
0.5601
0.3315
0.2260
0.1429
0.0695
0.0000
—
—
—
—

12

0.5475
0.3325
0.2347
0.1586
0.0922
0.0303
—
—
—
—

13

0.5359
0.3325
0.2412
0.1707
0.1099
0.0539
0.0000
—
—
—

14

0.5251
0.3318
0.2460
0.1802
0.1240
0.0727
0.0240
—
—
—

15

0.5150
0.3306
0.2495
0.1878
0.1353
0.0880
0.0433
0.0000
—
—

16

0.5056
0.3290
0.2521
0.1939
0.1447
0.1005
0.0593
0.0196
—
—

17

0.4968
0.3273
0.2540
0,1988
0.1524
0.1109
0.0725
0.0359
0.0000
—

18

0.4886
0.3253
0.2553
0.2027
0.1587
0.1197
0.0837
0.0496
0.0163
—

19

0.4808
0.3232
0.2561
0.2059
0.1641
0.1271
0.0932
0.0612
0.0303
0.0000

20

0.4734
0.3211
0.2565
0.2085
0.1686
0.1334
0.1013
0.0711
0.0422 '
0.0140

\ n
\
t\
I
2
3
4
5
6
7
8
9
10
11
12
13
14
15

21
\
0.4643
0.3185
0.2578
0.2119
0.1736
0.1399
0.1092
0.0804
0.0530
0.0263
0.0000
—
—
—
—

22

0.4590
0.3156
0.2571
0.2131
0.1764
0.1443
0.1150
0.0878
0.0618
0.0368
0.0122
—
—
—
—

23

0.4542
0.3126
0.2563
0.2139
0.1787
0.1480
0.1201
0.0941
0.0696
0.0459
0.0228
0.0000
—
—
—

24

0.4493
0.3098
0.2554
0.2145
0.1807
0.1512
0.1245
0.0997
0.0764
0.0539
0.0321
0.0107
—
—
—

25

0.4450
0.3069
0.2543
0.2148
0.1822
0.1539
0.1283
0.1046
0.0823
0.0610
0.0403
0.0200
0.0000
—
—

26

0.4407
0.3043
0.2533
0.2151
0.1836
0.1563
0.1316
0.1089
0.0876
0.0672
0.0476
0.0284
0.0094
—
—

27

0.4366
0.3018
0.2522
0.2152
0.1848
0.1584
0.1346
0.1128
0.0923
0.0728
0.0540
0.0358
0.0178
0.0000
—

28

0.4328
0.2992
0.2510
0.2151
0.1857
0.1601
0.1372
0.1162
0.0965
0.0778
0.0598
0.0424
0.0253
0.0084
—

29

0.4291
0.2968
0.2499
0.2150
0.1864
0.1616
0.1395
0.1192
0.1002
0.0822
0.0650
0.0483
0.0320
0.0159
0.0000

30

0.4254
0.2944
0.2487
0.2148
0.1870
0.1630
0.1415
0.1219
0.1036
0.0862
0.0697
0.0537
0.0381
0.0227
0.0076
             from: Conover, 1980.
                        196

-------
TABLE B.4  COEFFICIENTS FOR  THE  SHAPIRO-WILK'S  TEST  (Continued)
V
i\
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
31
\
0.4220
0.2921
0.2475
0.2145
0.1874
0.1641
0.1433
0.1243
0.1066
0.0899
0.0739
0.0585
0.0435
0.0289
0.0144
0.0000
—
—
_
—
32

0.4188
0.2898
0.2462
0.2141
0.1878
0.1651
0.1449
0.1265
0.1093
0.0931
0.0777
0.0629
0.0485
0.0344
0.0206
0.0068
—
—
—
—
33

0.4156
0.2876
0.2451
0.2137
0.1880
0.1660
0.1463
0.1284
0.1118
0.0961
0.0812
0.0669
0.0530
0.0395
0.0262
0.0131
0.0000
—
— _
—
34

0.4127
0.2854
0.2439
0.2132
0.1882
0.1667
0.1475
0.1301
0.1140
0.0988
0.0844
0.0706
0.0572
0.0441
0.0314
0.0187
0.0062
—
—
—
35

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
—
—
36

0.4068
0.2813
0.2415
0.2121
0.1883
0.1678
0.1496
0.1331
0.1179
0.1036
0.0900
0.0770
0.0645
0.0523
0.0404
0.0287
0.0172
0.0057
—
—
37

0.4040
0.2794
0.2403
0.2116
0.1883
0.1683
0.1505
0.1344
0.1196
0.1056
0.0924
0.0798
0.0677
0.0559
0.0444
0.0331
0.0220
0.0110
0.0000
—
38

0.4015
0.2774
0.2391
0.2110
0.1881
0.1686
0.1513
0.1356
0.1211
0.1075
0.0947
0.0824
0.0706
0.0592
0.0481
0.0372
0.0264
0.0158
0.0053
—
39

0.3989
0.2755
0.2380
0.2104
0.1880
0.1689
0.1520
0.1366
0.1225
0.1092
0.0967
0.0848
0.0733
0.0622
0.0515
0.0409
0.0305
0.0203
0.0101
0.0000
40

0.3964
0.2737
0.2368
0.2098
0.1878
0.1691
0.1526
0.1376
0.1237
0.1108
0.0986
0.0870
0.0759
0.0651
0.0546
0.0444
0.0343
0.0244
0.0146
0.0049

\ n
\
i\
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

41
\
0.3940
0.2719
0.2357
0.2091
0.1876
0.1693
0.1531
0.1384
0.1249
0.1123
0.1004
0.0891
0.0782
0.0677
0.0575
0.0476
0.0379
0.0283
0.0188
0.0094
0.0000

—
—
—

42

0.3917
0.2701
0.2345
0.2085
0.1874
0.1694
0.1535
0.1392
0.1259
0.1136
0.1020
0.0909
0.0804
0.0701
0.0602
0.0506
0.0411
0.0318
0.0227
0.0136
0.0045
—
—
—
	

43

0.3894'
0.2684
0.2334
0.2078
0.1871
0.1695
0.1539
0.1398
0.1269
0.1149
0.1035
0.0927
0.0824
0.0724
0.0628
0.0534
0.0442
0.0352
0.0263
0.0175
0.0087
0.0000
—
—
—

44

0.3872
0^2667
0.2323
0.2072
0.1868
0.1695
0.1542
0.1405
0.1278
0.1160
0.1049
0.0943
0.0842
0.0745
0.0651
0.0560
0.0471
0.0383
0.0296
0.021 1
0.0126
0.0042
—
—
—

45

0.3850
0.2651
0.2313
0.2065
0.1865
0.1695
0.1545
0.1410
0.1286
0.1170
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
—
	

46

0.3830
0.2635
0.2302
0.2058
0.1862
0.1695
0.1548
0.1415
0.1293
0.1180
0.1073
0.0972
0.0876
0.0783
0.0694
0.0607
0.0522
0.0439
0.0357
0.0277
0.0197
0.0118
0.0039
—
—

47

0.3808
0.2620
0.2291
0.2052
0.1859
0.1695
0.1550
0.1420
0.1300
0.1189
0.1085
0.0986
0.0892
0.0801
0.0713
0.0628
0.0546
0.0465
0.0385
0.0307
0.0229
0.0153
0.0076
0.0000
—

48

0.3789
0.2604
0.2281
0.2045
0.1855
0.1693
0.1551
0.1423
0.1306
0.1197
0.1095
0.0998
0.0906
0.0817
0.0731
0.0648
0.0568
0.0489
0.0411
0.0335
0.0259
0.0185
0.0111
0.0037
—

49

0.3770
0.2589
0.2271
0.2038
0.1851
0.1692
0.1553
0.1427
0.1312
0.1205
0.1105
0.1010
0.0919
0.0832
0.0748
0.0667
0.0588
0.0511
0.0436
0.0361
0.0288
0.0215
0.0143
0.0071
0.0000

50

0.3751
0.2574
0.2260
0.2032
0.1847
0.1691
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
                              197

-------
TABLE B.5. EXAMPLE OF THE SHAPIRO-WILK'S TEST:
           TABLE OF COEFFICIENTS AND DIFFERENCES
ai
1
2
3
4
5
6
7
8
9
10
0.4734
0.3211
0.2565
0.2085
0.1686
0.1334
0.1013
0.0711
0.0422
0.0140
0.181
0.128
0.105
0.097
0.076
0.048
0.034
0.025
0.008
0.005
X(20)
X(19)
X 18)
xH7)
X(16)
X(15)
Xp4)
X(13)
x(1 2}
xdi)
x(D
- X<2)

- x^4)
-XJ5)
- x(6)
- x(7J
- x(8)
- x(9)
- xH°)
                     198

-------
TABLE B.6 QUANTILES OF THE  SHAPIRO-WILK'S  TEST STATISTIC*
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
       ^Taken from Conover, 1980.
                           199

-------
3.  TEST FOR HOMOGENEITY OF VARIANCE

3.1  For Dunnett's Procedure and Bonferroni's  T-test,  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.7,  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:
                P       _2   p        2
              [{S V.)  In S  - S V.  In  S-]
Where: V
       C
       In
            = Degrees of freedom for each  toxicant  concentration and control
            = Number of levels  of toxicant concentration  including the
              control
            = The average of the individual  variances.
                             P          P
            = 1  + [l/3(p-l)][Sl/V.  - 1/2 V.]

            = Loge           "          '
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.
                                                  __2
3.5  For the data in this example,  v-f = 3, p = 5,  S  =  0.0027,  and
C = 1.133.  The calculated B value  is:
                 B =
                      (15)[ln(0.0027)j - 3 S In(sf)
                              	      i=l
                                1.133

                      15(- 5.9145)  - 3(- 32.4771)]
                              1.133
                   =  7.691
                                    200

-------
3.5  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.7.  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(Ti)
Si2
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 a
non-parametric 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 Transformation!

4.2.1  For data consisting of proportions from a binomial  (response/no
response; live/dead) response variable, the variance within the i-th
treatment is proportional to Pj (1 - P-j), where P-f is the expected
proportion for the treatment.  This clearly violates the  homogeneity of
variance assumption required by parametric procedures such as Dunnett's
iFrom: Peltier and Weber (1985).
                                    201

-------
or Bonferroni's, since the existence of a treatment effect implies different
values of P-j for different treatments, i.  Also, when the observed
proportions are based on small samples, or when Pj is close to zero or one,
the normality assumption may be invalid.  The arc sine square root
(arc sine  %/P) transformation is commonly used for such data to stabilize the
variance and satisfy the normality requirement.

4.2.2  Arc sine transformation consists of determining the angle (in radians)
represented by a sine value.  In the case of arc sine square root
transformation of mortality data, the proportion of dead (or affected)
organisms is taken as the sine value, the square root of the sine value is
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)°'5.

         Example:  If RP = 0.40:

                  Angle = arc sine (0.40)°-5

                        = arc sine 0.6325

                        = 0.6847 radians
                                      202

-------
4.2.4.2  Modification of the arc  sine when RP = 0.
         Angle (in radians)  = arc sine  (1/4N)0'5
         Where:  N - Number of animals/treatment
         Example: If 20 animals are used:
                  Angle = arc sine (1/80)0-5
                        = arc sine 0.1118
                        = 0.1120  radians

4.2.4.3  Modification of the arc  sine when RP ~ 1.0.
         Angle = 1.5708 radians - (radians for RP  = 0)
         Example: Using above value:
                  Angle = 1.5708  - 0.1120
                        - 1.4588   radians
                                     203

-------
                                   APPENDIX C

                              DUNNETT'S PROCEDURE
i.  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,  Bonferroni's
T-test 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.  One way to  obtain an estimate of
the pooled variance is to construct an ANOVA table including  all  sums of
squares, using the following formulas:
        TABLE C.I.  FATHEAD LARVAL GROWTH DATA (WEIGHT  IN MG)
                   USED FOR DUNNETT'S PROCEDURE
                                   NaPCP Concentration  (ug/L)
Replicate
Control
32
64
128
256
A
B
C
D
MeanfYj)
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
0.606
0.508
0.580
2.322
                                    204

-------
1.3  One way to obtain an estimate of the pooled variance is to construct an
ANOVA table including all sums of squares, using the following formulas:

Total Sum of Squares:   SST = S  y? - -  G2/N
                              ij   J

Between Sum of Squares: SSB = S Tj/n. -  G2/N

Within Sum of Squares:  SSW = SST - SSB
    Where: G = The grand total of all sample observations; G = ST.
           N = The total sample size; N = 2 n.

          n. = The number of replicates for concentration "i".
          T. = The total of the replicate measurements for concentration "i"
         Y..- • = The jth observation for concentration "i".
          ' j
1.4  Calculations:
Total Sum of Squares:    SST = S  Y?.  -  G2/N
                               ij   J
                             = 8.635 - {13.077)2/20
                             = 0.085

Between Sum of Squares:  SSB = S  T?/n. -  G2/N
                               i
                             = 8.594 - (13.077)2/20
                             = 0.044

Within Sum of Squares:   SSW = SST - SSB
                             = 0.085 - 0.044
                             = 0.041
                                      205

-------
1.5  Prepare the ANOVA table as follows:





                      TABLE C.2  GENERALIZED ANOVA TABLE
Source DF
*
Between p - 1
Within N - p
Total N - 1
Sum of
Squares (SS)
SSB
SSW
SST
Mean Square (MS)
(SS/DF)
S* = SSB/(p-l)
S* = SSW/(N-p)

*n =
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             0.011



Within    20 - 5 = 15      0.041              0.0027





Total        19            0.085
                                      206

-------
1.7  To perform the individual  comparisons,  calculate  the  t  statistic for
each concentration and control  combination,  as follows:
                             [Su /(1/n,)  + (l/n.)3
                               W      I         1

    Where: 7f   =  Mean for each concentration

           7|   =  Mean for the control

           Sw   -  Square root of the within mean square

           n*|   =  Number of replicates  in the control.

           n^   =  Number of replicates  for concentration "i".
1.8  Table C.4  includes the calculated t values for each concentration  and
control combination.
                       TABLE C.4.  CALCULATED T VALUES
              NaPCP             i                   t1
          Concentration
             (ug/L)
32
64
128
256
2
3
4
5
1.081
1.000
2.432
3.622
                                     207

-------
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-j 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 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 Sw>/{l/n1) + (1/n)
    Where: d   = Critical value for the Dunnett's Procedure

           Sw  = The square root of the within mean square

           n   = The number of replicates at each concentration,
                   assuming an equal number of replicates at all
                   treatment concentrations
               = Number of replicates in the control
    For example:
        MSD = 2.36 (0.052)[s/(l/4) + (1/4)] = 2.36 (0.052H /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.
                                     208

-------
1.11.2.1  Subtract the MSD from the transformed control  mean.   Call  this
difference D.  Next, obtain untransformed values for the control  mean  and  the
difference, D.

            MSDU  =  Controlu - Du

Where:

          MSDU = The minimum significant difference for untransformed  data

      Controlu = The untransformed control  mean

            Du = The untransformed difference

1.11.2.2  Calculate the percent reduction from the control  that MSDU
represents as:
                                  u        v  i nn
         Percent Reduction =   Contro1u


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 (MSDU) 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 [(0.085/0.429H100)].
                                      209

-------
1.12   Table of Dunnett's  "t"  values.
                       TABLE C.5.  DUNNETT'S "T" VALUES!
                                     (One-tailed) d

'X
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
24
30
40
60
120
«

1
2.02
l.M
1.89
1.86
1.83
1.81
1.80
1.78
1.77
1.76
1.75
1.75
1.74
1.73
1.73
1.72
1.71
1.70
1.68
1.67
1.66
1.64

2
2.44
2.34
2.27
2.22
2.18
2.15
2.13
2.11
2.00
2.08
2.07
2.06
3.05
2.04
2.03
2.03
2.01
1.99
1.97
1.95
1.93
1.92

a
2.68
2.56
2.46.,
2.42
2.37
2.34
2.31
2.29
2.27
2.25
2.34
2.23
2.23
2.21
2.20
2.19
2.17
2.15
2.13
2.10
2.08
2.06

4
2.85
2.71
2.62
2.55
2.50
2.47
2.44
2.41
2.39
2.37
3.36
2.34
2,33
2.32
2.31
2.30
2.28
2.2S
2.23
2.21
2.18
2.16
at = .05
5
2.98
2.83
2.73
2.66
2.60
2.56
2.53
2.50
2.48
2.46
2.44
2.43
2.42
2.41.
2.40
2.39
3.36
2.33
2.31
2.28
2.26
2.23

6
3.08
2.92
2.82
2.74
2.68
2.64
2.60
2.58
2.55
3.53
2.51
3.50
3.4»
3.48
2.47
3.46
3.43
3.40
2.37
2.35
2.32
3.39

7
3.16
3.00
2.89
2.81
2. 75
2.70
2.6T
2.64
2.61
2.59
2.57
2.56
2.S4
2.53
2.53
2.51
2.48
2.45
2.42
3.39
3.37
3.34

8
3.24
3.07
3.95
3.87
3.81
1.76
2.72
2.69
2.66
2.64
2.62
2.61
2.59
2.58
2.57
2.56
2.53
2.50
2.47
2.44
2.41
2.38

9
3*. 30
3.12
3.01
2.92
2.86
2.81
Z.77
2.74
2.71
2.69
3.67
3.65
2.64
2.63
2.61
3.60
3.57
2.54
2.51
2.48
2.45
2.42

1
3.37
3.14
3.00
2.90
2.82
2.76
2.72
2.68
2.85
2.62
2.60
2.58
2,57
2.55
3.54
2.53
3.49
2.46
2.42
2.39
2.36
2.33

2
3.90
3.61
3.42
3.29
3.19
3.11
3.06
3.01
2.97
3.94
2.91
2.88
2.86
2.84
2.83
2.81
2.77
2.72
2.68
2.64
2.60
2.M

3
4.21
3.88
3.66
3.51
3.40
3.31
3.23
3.19
3.15
3.11
.08
.05
.03
.01
.99
2.97
2.92
3.87
3.82
3.78
3.73
3.68

4
4.43
4.07
3.83
3.67
3.55
3.45
S.3S
3.33
3.37
3.23
3.20
3.17
3.14
3.13
3.10
3.08
3.03
2.97
2.92
2.87
2.82
2.77
a> .01
5
4.60
4.21
3.96
9.79
3.66
3.56
3.48
3.42
3.37
3.32
3.29
3.26
3.23
3.21
3.18
3.17
3.11
3.05
2.99
2.94
2.89
2.84

6
4.73
4.33
4.07
3.88
3.75
3.64
J.M
3.50
3.44
3.40
3.36
3.33
3.30
3.27
3.25
3.33
3.17
3.11
3.05
3.00
3.94
3.89

7
4. as
4.43
4.15
3.M
3.82
3.71
$.«
3.56
3.51
3.46
3.42
3.39
3.36
3.93
3.31
3.29
3.22
3.16
3.10
3.04
2.99
3.99

8
4.94
4.51
4.33
4.03
3.89
3.78
3.69
3.63
3.56
3.51
3.47
3.44
3.41
3.38
3.36
3.34
3.27
3.31
3.14
3.08
3.03
2.97

9
5.03
4.59
4.30
4.09
9.94
3.83
3.74
3.67
3.61
3.56
3.52
3.48
3.45
3.42
3.40
3.38
3.31
3.24
3.18
3.12
g.ofl
3.00
                    Vrom:  Miller,  1981
                                      210

-------
2. COMPUTER CALCULATIONS

2.1  This computer program incorporates two analyses:  an  analysis  of  variance
(ANOVA), and a multiple comparison of treatment means  with  the  control mean
(Dunnett's Procedure).  The ANOVA Is used to obtain the error value.
Dunnett's Procedure indicates which toxicant concentration  means  (if  any)  are
statistically different from the control  mean at the 5% level of
significance.  The program also provides  the minimum difference between  the
control and treatment means that could be detected as  statistically
significant, and tests the validity of the homogeneity of variance assumption
by Bartlett's Test.  The multiple comparison is based  on  Dunnett,  C.  W., 1955,
"Multiple Comparison Procedure for Comparing Several Treatments with  a
Control," J. Amer. Statist. Assoc. 50:1096-1121.

2.2  The source code for the Dunnett's program is structured into  a series of
subroutines, controlled by a driver routine.  Each subroutine has  a specific
function in the Dunnett's Procedure, such as data input,  transforming the
data, testing for equality of variances,  computing p values, and calculating
the one-way analysis of variance.

2.3  The program compares up to seven toxicant concentrations against the
control, and can accommodate up to 50 replicates per concentration.

2.4  If the number of replicates at each  toxicant concentration and control
are not equal, Bonferroni's T-test is performed instead of  Dunnett's  Procedure
(see Appendix D).

2.5  The program was written in IBM-PC FORTRAN (XT and AT)  by D.  L. Weiner,
Computer Sciences Corporation, 26 W. Martin Luther King Drive,  Cincinnati,
Ohio 45268.  A complete listing of the program is contained in
EPA/600/4-87/028.  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 Ceriodaphm'a 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
                                      211

-------
                TABLE C.6.   SAMPLE DATA FOR DUNNETT'S  PROGRAM.
                            CERIODAPHNIA REPRODUCTION  DATA
                                   Effluent Concentration  (%)
Replicate    Control            1.56      3.12       6.25      12.5
1
2
3
4
5
6
7
8
9
10
27
30
29
31
16
15
18
17
14
27
32
35
32
26
18
29
27
16
35
13
39
30
33
33
36
33
33
27
38
44
27
34
36
34
31
27
33
31
33
31
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 below.
                                      212

-------
                     MAIN MENU AND DATA  INPUT
 1) Create a data file
 2) Edit a data file
 3) Perform ANOVA on existing  data set
 4) Stop

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
                               213

-------
2.6.3  Program Output

2.6.3.1  When Option 3  (Perform ANOVA on existing data set) is selected from
the main menu, the  user  is asked to select the  transformation desired,  and
indicate whether they expect the means of the test groups to be less or
greater than the mean for the control group (see below).
  1)  Create a  data  file
  2)  Edit a data file
  3)  Perform AWOVA  on existing data set
  4)  Stop

Your choice ? 3


File name ? cerio
 Available Transformations
     1)  no transform
     2)  square root
     3)  loglO
     4)  arcsine square root

Your choice ? l
 Durmett's test as implemented in this program is
 a one-sided test. You must specify the direction
 the test is to be run; that is, do you expect the
 means for the test groups to be less than or
 greater than the mean for the control group mean.

Direction for Dunnetts test : L=less than, Ogreater than ?  1
                                    214

-------
r
              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.
                               Summary Statistics and

                           Transformation =       None

                Group      n        Msan           s. d.             cv%
1 =




control 10
2 10
3 10
4 10
5* 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
              Minumum 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
                                                  215

-------
                                APPENDIX D

                            BONFERRONI'S T-TEST
1.  Bonferroni's T-test 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.  Bonferroni's T-test 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 Bonferroni's T-test is provided below.  The
data used in the example are the same as in Appendix C, except that the
third replicate from the 256 ug/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 LARVAL GROWTH DATA (WEIGHT IN MG)
                   USED FOR BONFERRONI'S T-TEST
Replicate
Control
                                   NaPCP Concentration (ug/L)
32
64
128
256
A
B
C
D
MeantYf)
Total (\i)
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
                                    216

-------
3.1  One way to obtain an estimate of the  pooled  variance  is  to construct an
ANOVA table including all sums of squares,  using  the  following formulas:

Total Sum of Squares:   SST = S  Y?.  -  G2/N
                              "i "i

Between Sum of Squares: SSB = S T?/n. - G2/N

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 =  S n.                 1

          n- = The number of replicates for concentration  "i
          T. = The total of the replicate  measurements for concentration "i
         Y.. = The jth observation for concentration  "i".
          • j
3.2  Calculations:
Total Sum of Squares:    SST = S  Y2   -  G2/N
                               i i
                             = 8.268 - (12.471)2/19
                             = 0.082

Between Sum of Squares:  SSB = Z  T?/n- -   G2/N

                             = 8.228 - (12.471)2/19
                             = 0.042

Within Sum of Squares:   SSW = SST - SSB
                             = 0.082 - 0.042
                             = 0.040
                                       217

-------
3.3 Prepare the ANOVA table as follows:
                     TABLE D.2.  GENERALIZED ANOVA TABLE
Source DF
*
Between p - 1
Within N - p
Total N - 1
Sum of
Squares (SS)
SSB
SSW
SST
Mean Square (MS)
(SS/DF)
S* = SSB/(p-l)
S^ = SSW/(N-p)

*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 BONFERRONI'S  T-TEST
Source
 DF
  SS
Mean Square
Between    5 - 1  =  4
Within    19 - 5 = 14
Total
18
0.042
0.040

0.082
                                0.0105
                                0.0028
                                      218

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



                                  
-------
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 tj 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.
                TABLE D.5. CRITICAL VALUES FOR BONFERRONI'S "T"
                      P = 0.05 CRITICAL LEVEL, ONE TAILED
O.F.
1
2
3
4
5
6
7
8
9
10
11
1Z
13
14
15
16
17
la
19
20
21
22
23
2*
25
26
27
28
?9
30
31
32
33
34
35
3*
ii
38
39
40
SO
60
70
ao
•90
100
110
120
INF.
D.F. =
K =
K " 1
6.314
2.920
2.354
2.132
2.016
1.944
1.395
1.860
1.334
1.S13
1.796
1.783
1.771
1.762
1.754
1.746
1.740
1.735
1.730
1.725
1.721
1.719
1.714
1.711
1.709
1.706
1.704
1.702
1,700
l.trtB
1.696
1.694
1.693
1.691
1.690
1.689
1.688
1.686
1.695
1.684
1.676
1.671
1.667
1.665
1.662
1.661
1.659
1.658
1.64S
Degrees
Number
X - 2
12.707
4.303
3.183
2.777
2.571
2.447
2.365
2.307
2.263
2.229
2.201
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
2.06O
2.056
2.052
2.049
2.046
2.043
2.040
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
1.997
1.984
1.982
1.980
1.960
K - 3
19.002
5.340
3.741
3.107
2.912
2.7SO
2.642
2.567
2.510
2.466
2.432
2.404
2.300
2.360
2.343
2.329
2.316
2.305
2.2SS
2.206
2.278
2.271
2.264
2.258
2.253
2.248
2.243
2.239
2.235
2.231
2.228
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
2.159
2.156
2.153
2.129
of freedom
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
2.352
2.349
2.346
3.3*2
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
for MSE
of concentrations to
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
2.5(19
2.500
2.493
2.486
2.479
2.473
2.468
2.463
2.458
2.453
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
38. 199
7.649
4.857
3.961
3.535
3.288
3.128
3.016
2.93*
2.871
2.821
2.780
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
2.5*1
2.536
2.531
2.527
2.523
2.519
5.515
2.512
2.508
2.505
2.502
2.499
2.479
2.463
2.453
2.446
2.440
2.435
2.432
2.429
2.394
K - 7
44.556
8.277
S.138
4.148
3.681
3.412
3.239
3.118
3.029
2.961
2.907
2.863
2.827
2.797
2-7T1
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
2.592
2.587
2.583
2.579
2.575
2.572
2.568
2.565
2.562
2.539
2.52*
2.513
2.505
2.499
2.494
2.490
2.487
2.450
X - 8
50.924
8.861
5.392
4.315
a. an
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
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
X - 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
2.696
Z.691
2.696
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
63
9
5
4
4
3
3
3
3
3
3
3
3
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
- 10
.657
.925
.841
.605
.033
.703
.500
.356
.250
.170
.106
.055
.013
.977
.947
.921
.894
.879
.861
.8^6
.832
.819
.808
.797
.788
.779
.771
.764
.757
.750
.745
.739
.734
.729
.724
.720
.716
.712
.708
.705
.679
.661
.648
.639
.632
.626
.622
.618
.576
(Mean Square Error) from ANOVA.
be come
tared to
the control .
                                       220

-------
                                   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
Procedures 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 Ceriodaphnia 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.
                                      221

-------
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
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 CERIODAPHNIA 7-DAY CHRONIC TEST
Effluent
Concentration 1
Replicate
234567
No.
Live
8 9 10 Adults
Cont            20   26   26   23   24   27   26   23   27   24     10
                                                                     9
                                                                    10
                                                                    10
 25%             9    0    9    7    6   10   12   14    9   13      8
 50%             0000000000      0
20
13
18
14
9
0
26
15
22
22
0
0
26
14
13
20
9
0
23
13
13
23
7
0
24
23
23
20
6
0
27
26
22
23
10
0
26
0
20
25
12
0
23
25
22
24
14
0
27
26
23
25
9
0
24
27
22
21
13
0
                                    222

-------
     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%
TPL
•Jto
7<£
OA
yy.
J/o
Control
Control
Control
T£
•3/0
Control
Control
"V9
•J/O
Control
Control
Control
3%
3%
Control
Control
3%
                            TABLE  E.3   TABLE  OF  RANKS
Replicate Control3
(Organism)
1
2
3
4
5
6
7
8
9
10
20
26
26
23
24
27
26
23
27
24
(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)
Effluent Concentration (%}

13
15
14
13
23
26
0
25
26
27
3
(2.5)
(5)
(4)
(2.5)
(8)
(15)
(1)
(12)
(15)
(19)

18
22
13
13
23
22
20
22
23
22
6
(3)
(7.
(1.
(1.
(11
(7.
(4.
(7.
(11
(7.


5)
5)
5)
.5)
5)
5)
5)
.5)
5)
1
14
22
20
23
20
23
25
24
25
21
12
(1)
(6)
(3)
(8.5)
(3)
(8.5)
(14.5)
(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)
aControl ranks are given in the order of the concentration with which they
 were ranked.
                                       223

-------
                       TABLE  E.4.  RANK  SUMS
              Effluent
           Concentration
                (X)
Rank Sum
                3
                6
               12
               25
   84
   63.5
   76
   55
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

It -
2
11
18
15
27
23
37
32
49
43
63
56
79
71
97
87
116
105
138
125
161
147
186
170
213
196
241
223
272
252
304
282
339
315
number
3
10
17
-
26
22
36
31
48
42
62
55
77
69
95
85
114
103
135
123
158
144
182
167
209
192
237
219
267
248
299
278
333
310
of treatments
4 5
10
17
-
25
21
35
30
47
41
61
54
76
68
93
84
112
102
133
121
155
142
180
165
206
190
234
217
264
245
296
275
330
307
10
16
-
25
21
35
30
46
40
60
53
75
67
92
83
111
100
132
120
154
141
178
164
204
188
232
215
262
243
294
273
327
305
(excluding
6 7
10
16
-
24
-
34
29
46
40
59
52
74
66
91
82
110
99
130
119
153
140
177
162
203
187
231
213
260
241
292
271
325
303
-
16
-
24
-
34
29
45
40
59
52
74
66
90
81
109
99
129
118
152
139
176
161
201
186
229
212
259
240
290
270
323
301
control)
a 9
*
16
-
24
-
33
29
45
39
58
51
73
65
90
81
108
98
129
117
151
138
175
160
200
185
228
211
257
239
288
268
322
300
-
IS
—
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.
                             224

-------
                                   APPENDIX  F

                             WILCOXON RANK SUM  TEST


1. Wilcoxon's Rank Sum Test is a non-parametric 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 vs.  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 12% 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 vs. 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.
                                      225

-------
   TABLE F.I. EXAMPLE OF WILCOXON'S RANK SUM TEST:
              DATA FOR CERIODAPHNIA 7-DAY CHRONIC TEST

Effluent
Concentration
Cont
3%
6%
m
25%
50%










Replicate
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
b
24
23
23
M
6
0
6
27
26
22
23
10
0
/
26
0
20
25
12
0
8
23
25
22
24
14
0
9
27
26
23
25
9
0
lu
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 TU THE CONTROL AND EFFLUENT CONCENTRATIONS
Rank
1
2.5
2.5
4
5
7
7
7
9.5
9.5
11
14
14
14
14
14
18
18
18
Number of Young
Produced
0
13
13
14
15
23
23
23
24
24
25
26
26
26
26
26
27
27
27
Control or % Effluent
3£
3%
3%
3%
TfZ.
Oib
Control
Control
"WL
•ja
Control
Control
^
o«
Control
Control
Control
3%
3%
Control
Control
3%
                                226

-------
                           TABLE F.3  TABLE OF RANKS
Replicate
Control3
(Organism)
1
2
3
4
5
6
7
8
9
10
M
26
26
23
24
27
26
23
27
24

(14,16,15,16)
(14,16,15,16)
(7,10.5,6.5,11
(9.5,13.5,10,1
(18,18.5,17.5,
(14,16,15,16)
(7,10.5,6.5,11
(18,18.5,17.5,
(9.5,13.5,10,1



.5)
3.5)
18.5)

.5)
18.5)
3.5)
13
15
14
13
23
26
0
25
26
27
Effluent Concentration (%)
3
(2.5)
(5)
(4)
(2.5)
(7)
(14)
(1)
(11)
(14)
(18)

18
22
13
13
23
22
20
22
23
22
6
(3)
(6.5)
(1.5)
(1.5)
(10.5)
(6.5)
(4)
(6.5)
(10.5)
(6.5)
1
14
22
20
23
M
23
25
24
25
21
2
(1)
(4)
(2)
(6.5)

(6.5)
(12.5)
(10)
(12.5)
(3)
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
                                     227

-------
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. Replicates No. of Replicates 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
                                 228

-------
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 Repl
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
n
n
12
13
13
„
—
10
11
n
12
12
13
_„
—
—
10
n
n
12
13
icates Per Effluent

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

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
                                  229

-------
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 Replicates Per Effluent
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
n
n
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

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
Concentration

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
                                 230

-------
                                   APPENDIX G

                              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  Ceriodaphnia 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 G.I.  FORMAT FOR CONTINGENCY  TABLE
                                  Number of	
                                                          Number of
                            Successes     Failures       Observations
Row 1
Row 2
a
b
A - a
B - b
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 Sb/B).  For the ceriodaphnia
survival data, a success may be 'alive' or 'dead1 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.
                                       231

-------
4.  To illustrate Fisher's Exact Test, a set of survival  data  (Table  G.2)  from
the Ceriodaphnla 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 MORTALITY DATA

Effluent
Concentration (%)
Control
1
3
6
12
25

No. Dead
1
0
0
0
0
10

Total ]
9
10
10
10
10
10
^Total  number of live adults at the beginning of  the  test.
         TABLE G.3.   2X2 CONTINGENCY TABLE FOR  CONTROL  AND  1%  EFFLUENT
Number
Alive
H Effluent 10
Control 8
of
Dead
0
1
Number of
Observations
10
9
           Total              18             1               19
                                      232

-------
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
                               Dead         Alive        Observations
            25% Effluent        10            0              10

             Control              1             8               9


             Total               11             8              19
                                      233

-------
TABLE G.5.
SIGNIFICANT LEVELS OF B:  VALUES OF  B  (LARGE  TYPE)
AND CORRESPONDING PROBABILITIES (SMALL TYPE)!

A-3 B=3


A*4 B*4
3

A*S 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

5
4
5
4

5
5


6
5
4

6
5
4
6
5
6
5

*


7
6
5
4

7
6
5
4
7
6
5
7
6
5
7
6
7
Probability
005
0 4)0


0 414
0 429

1 424
0 424
1 441
0 440

0 411
0 441


2 4V
1 •OtO
0 410

1 -013+
0 413
0 443+
1 411
0 414
0 413
0 44*

0 436


3 413-
1 413-
0 410+
0 433-

2 431
1 42)+
0 41*
0 44*
2 44J+
1 44)+
0 427
1 424
0 413+
0 443+
0 40C
0 4»
0 421
0025



0 414
_

1 424
0 424
0 40t
_

0 411
—


1 401
0 -oo*


0 41)+
0 411
	
0 403-
0 424
0 412


~*


2 410+
1 413-
0 410+


2 431
0 404
0 416
^_
1 410+
0 401
	
1 424
0 413+
0 401
—
—
0*1



„_
,^

0 404
_.
0 401
^_

—
—


1 401
0 401
.M.

0 403
—
	
0 403-
—
	
	

*— •


1 401
0 402
^_
^—

I 403-
0 404
	
^_
0 401
0 40*
—
0 403
—
0 401
—
— *
0005



_
	

0 404
-_
_
^^

—
—


0 401
_
__

0 402
—
	
0 40J-
—
	
	

	


I 402
0 402
^_
__

1 403-
0 404
	
.
0 401
	
—
0 401
^~
	
	
—

A=8 B=8




7


6



5




4



3

2


A=» B=9






8





7



6



0
8
7
6
5
4
8
7
6
5
8
7
6
5
8
7
6
5

8
7
6

8
7
8


9
8
7
6
5
4

9
8
7
6
5

9
8
7
6
5
9
8
7
6
5
Probability
OO5
4 431
2 420
1 420
0 41)
0 431
3 426
2 433"
1 4)2
0 419
2 413-
1 416
0 409
0 421
2 41)-
1 412
0 41«
0 444

1 411
0 -010+
0 430

0 4W
0 414
0 -022


5 441
3 -KS-
2 428
1 423-
0 413-
0 441

4 429
3 441
2 444
1 4M
0 420

3 419
2 424
1 -020
0 410-
0 429
3 444
2 447
1 -03)-
0 -017
0 443
O025
3 411
A jyfc—
1 420
0 413
—
2 407
1 409
0 406
0 41*
2 413-
1 -016
0 409

I 40T
0 403-
0 -OK


1 4IS
0 410*


0 -006
0 424
0 422


4 413-
3 -02J-
1 401
1 423-
0 41)-
__

3 -009
2 -013
1 412
0 407
0 420

3 419
2 424
1 420
0 410+
— . -
2 411
1 411
0 406
0 417
•™
O01
2 401
1 40)+
0 403
	
	
2 407
1*WM
409
0 406
	
1 403
0 402
0 409

1 407
0 -oo)-

	

0 403
^__
	

0 -006
_
—


3 403-
2 40!
1 -ON
0 403-
^_
	

3 409
1 403
0 402
0 407
	

2 403-
1 -006
0 403
	
	
1 402
0 401
0 406
—
—
o-ooi
2 401
0 401
0 403
—
—
1 401
0 401
—
1 40)
0 402
^_
__
0 401
0 403-

_

0 402
_
	

—
—
—


3 403-
1 -002
0 -ooi
0 403-
	
	

2 403
I 403
0 402
	
	

2 403-
0 401
0 403
	
	
1 402
0 401
—
	
^—
       table shows:(1) In bold type, for given a,  A and B,  the
   value of b ([a) which is just significant at the probability
   level quoted (one-tailed test); and (2) In small type, for
   given A, B and r = a + b, the exact probability (if there is
   independence) that b is equal to or less than the integer shown
   in bold type.  From Pearson and Hartley, 1962.
                             234

-------
TABLE G.5.   SIGNIFICANT LEVELS OF B:  VALUES OF B  (LARGE  TYPE)
            AND CORRESPONDING PROBABILITIES (SMALL TYPE)
            (CONTINUED)

A*9 B=5



4



3


2

A* 10 B=10






9





8





7





6




5




a
9
8
7
6
9
8
7
6
9
8
7
9

10
9
8
7
6
5
4
10
9
8
7
6
5
10
9
8
7
6
5
10
9
8
7
6
5
10
9
8
7
6
10
9
8
7
6
Probability
045
2 417
1 433
0 41 0+
0 431
2 414
0 407
0 431
0 449
1 443+
0 411
0 443+
0 411

6 443
4 429
3 433-
2 433-
1 429
0 41*
0 441
5 43)
4 4)0-
2 419

1 440
0 422
4 413
3 432
2 431
1 42)
0 411
0 429
3 413-
2 411
1 411
1 436
0 417
0 441
3 436
2 4M
1 434
0 410+
0 426
2 422
1 417
1 447
0 419
0 442
0-025
1 403-
1 431
0 410+
	
1 414
0 407
0 421
_
0 403-
0 411
__
0 411

5 416
3 410-
2 412
1 410-
0 40)-
0 416
—
4 411
3 417
2 419
I 413-
0 40S
0 422
4 423
2 409
1 401
1 423
0 411
—
3 413-
2 411
1 413
0 406
0 417
	
2 401
1 401
1 424
0 410*
—
2 422
1 417
0 407
0 419
—
041
1 40)-
0 403
—
_
0 401
0 407
_
—
0 40)-
—
—
	

4 403+
3 410-
1 403
1 410-
0 403*
—
—
3 403
2 405-
1 404
ft 402
0 -oot
—
3 407
2 409
1 401
0 404
	
—
2 401
1 404
0 402
0 4W
	
	 ,
2 401
1 401
0 403
__
	
1 404
0 401
0 407
—
—
0-005
1 403-
0 403
—
—
0 401
_
—
—
0 403-
_ ;
—
»

3 402
2 403
1 401
0 402
_
—
_
3 401
2 403-
1 404
0 402
__
—
2 402
1 402
0 401
0 404
^
—
2 401
1 404
0 402
—
	
—
1 401
0 401
0 403
— -
_
1 404
0 402
—
_
—

A=IO B=4



3


t



A=ll B=ll






10






9






8






7





6


a
10
9
8
7
10
9
8
10
9


11
10
9
8
7
6
5
4
11
10
9
8
7
6
5
It
10
9
8
7
6
5
11
10
9
8
7
6
5
n
10
9
8
7
6
11
10
9
Probability
MS
1 411
1 441
0 413-
0 4))-
1 411
0 414
0 413-
0 41)*
0 4*3+


7 443-
5 432
4 440
3 443
2 440
1 4)2
0 411
0 443*
6 433-
4 421
3 424
2 4U
1 4IT
1 443
0 413
5 416
4 431
3 440
2 43)-
1 423-
0 412
0 430
4 -on
3 424
2 422
1 413-
1 -037
0 417
0 410
4 441
3 447
2 4)»
1 423-
0 410*
0 433-
3 43
2 421
1 411
0-025
1 411
0 403-
0 413-
—
0 40)
0 414
«
0 413*
__


6 411
4 413
3 41)-
2 413-
1 412
0 40*
0 411
—
5 413
4 421
3 42<
2 423
1 417
0 409
0 423
4 401
3 412
2 412
1 409
I 413-
0 412
—
4 411
3 424
2 433
1 413-
0 407
0 417
—
3 411
2 41)
1 409
1 423-
0 410*
0 423-
2 406
1 403+
1 411
041
0 401
0 403-
—
—
0 40)
_
_
—
__


5 406
3 404
2 404
1 404
0 402
0 406
—
—
4 404
3 407
2 407
1 -006
0 403
0 409
—
4 401
2 403
I 403
1 409
0 404
	
—
3 403-
2 406
1 -003-
0 402
0 407
	
	
2 402
1 003
1 409
0 404
	
	
2 40*
1 403+
0 403
0405
0 401
0 40)-
._
—
0 401
_
—
_
_


4 403
3 404
2 404
1 404
0 402
—
—
—
4 404
2 402
1 403
0 40)
0 403
—
—
3 402
2 401
1 401
0 401
0 404
—
	
3 40J-
1 401
1 403*
0 401
	
—
—
2 403
1 402
0 401
0 404
	
—
1 401
0 401
0 402
                             235

-------
TABLE G.5.   SIGNIFICANT LEVELS OF B: VALUES OF B  (LARGE TYPE)
            AND CORRESPONDING PROBABILITIES (SMALL TYPE)
            (CONTINUED)

A-1I B=6


5




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
II
10
9
8
7
6
5
12
11
10
9
8
7
6
5
12
11
10
9
8
Probability
0-05
1 443
0 417
0 437
2 411
1 413 .
1 4M
0 413
0 429
1 40»
1 413
0 411
0 42*
1 4)3
0 411
0 427

0 413
0 431

«44T
6 434

4 430-

3 430-
2 44S-
1 41*
0 419
0 447
7 437
5 424
4 419
3 430
2 42*
I '419
I 443-
0 434
6 42*
5443
4 441
* 44*
2 431
1 4»
0 411
0 430
5 421
449
3 42»
2 434
1 41*
0-025
0 4O7
0 417
_
2 4tl
• 1 413
0 403-
0 413
	
I 40V
0 404
0 411
	
0 403
0 411
	

0 41)
^_

7 41*
5 414
4 411
3 430

2 411
1 414
0 407
0 419
__
6 414
5 -024
3 410*
2 40*
1 407
1 419
0 4B*
0 414
5 410-
4 41)*
3 417
2 41)-
1 410*
0 403-
0 411
—
5 421
3 40.
2 4M
2 434
1 41*
0-01
0 407
—
—
1 403
0 401
0 403-
	
_
1 40*
0 404
—
—
0 4W
	
	

—
_»

6 -007
4 403-
3 406
2 40*

I 40J-
0 401
0 407
	
_
5 403-
4 401
2 40)
2 4»
1 407
0 40)
0 40*
—
5 410-
3 403-
2 40J-
1 404
0 401
0 40)-
—
—
440*
3 40*
2 401
1 40*
0 401
0-005
_
—
—
1 403
0 401
0 403-
—
—
0 401
0 404
	
	
0 403
	
_

—
.^

5 401
4 403-
2 402
I 401

1 403-
0 403
	
__
__
5 403-
3 4oi
2 403
1 401
0 401
0 403
—
—
4 403
3 403-
2 403-
1 404
0 403
0 403-
—
—
3 401
2 401
I 401
0 401
0 403

A=12 B=9


8






7







6





5





4




3



2



A= 13 B= 13





«
7
6
5
12
11
10
9
8
7
6
12
II
10
9
8
7
6

12
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
8
Probability
(M)5
1 437
0 41T
0 439
5 449
3 411
2 41 )*
2 440
1 425-
0 410-
0 424
4 436
3 431
2 429
I 417
1 440
0 416
0 434

3 423-
2 422
1 411
1 432
0 -on
0 411-
0 430-

2 413-
1 410-
1 411
0 40*
0 420
0 441
2 430
1 427
0 -cot
0 419
0 41!
1 429
0 40*
0 422
0 444
0 411
0 431


9 441
7 437
6 441
4 424
3 414
2 -eii
0-025
0 407
0 417
	
4 414
3 411
2 413*
1 410-
1 -02J-
0 410-
0 42'
3 409
2 4io-
1 406
0-01
0 40T
_
—
3 4M
2 03.
1 -OOJ
1 410-
0 404
—
—
3 409
2 410-
1 4M
1 417 i 0 oo:
0 -00? 0 40T
0 416 i —
i
1
3 423-
1 413
0 403-
0 411
0 4M-


2 413-
1 410"
0 -003
0 409
0 420
—
1 407
0 403
0 40*
0 419
	
0 402
0 409
0 -022
—
0 411
__


8 400
6 41)*
5 421
4 424
3 424
2 421
2 40)-
0 402
0 403-
	
—


1 402
1 410-
0 403
0 409
—
—
1 407
0 -003
0 401
—
—
0 -oo:
0 409
—
—
_
—


7 407
5 40*
4 401
3 40t
2 401
1 -OM
0-005
-
—
—
3 404
2 404
1 40)
0 401
0 4M
—
—
2 402
1 40J
0 401
0 402
	
^
	

2 40S-
0 402
0 40)-
—
—


1 402
0 401
0 403
—
—
—
0 -001
0 40)
—
—
—
0 401
—
—
—
—
—


6 403
4 401
3 402
2 401
1 402
0 401
                             236

-------
TABLE G.5.  SIGNIFICANT LEVELS OF B:  VALUES OF B  (LARGE  TYPE)
            AND CORRESPONDING PROBABILITIES (SMALL TYPE)
            (CONTINUED)

* = 13 B=13



12








n








10








9






8







7

a
7
6
S
4
13
12
11
10
9
8
7
6
5
13
12
11
10
9
8
7
6
5
13
12
11
10
9
8
7
6
5
13
12
11
10
9
8
7
6
5
13
12
11
10
9
8
7
6
13
12
Probability
0-05
2 *4i
1 437
0 420
0 441
8 -039
6 -017
5 433
4 -oic
3 4w
2 -019
1 -010
1 -044
0 >014
7 -osi
644.
4 -on
3 -on
3 450-
2 440
1 -017
0 -on
0 430
6 424
5 43S-
4 -037
3 43)
2 42*
1 -on
I -03J
0 -on
0 4«
5 417
4 423
3 -012
2 -OIT
2 -040
I 42S-
0*410"
0 -023
0 -049
5 -042
4 -047
3 -MI
2 429
1 -017
1 4JT
0 413-
0 -032
4 -031
3 -031
0-025
1 4is+
0 -007
0 -020
—
7 415-
5 4to-
4 413
3 -013
2 -on
1 -001
1 -010
0 4io-
0 -02*
6 -011
5 -on
4 -021
3 -021
2 -017
1 -Oil
0 403-
0 -013
—
6 424
4 -on
3 -012
2 -oto*
1 406
1 -Oil
0 -007
0 -017
—
5 -017
4 -013
3 -012
2 -on
1 -oio*
1 -023-
0 410*
0 -023
—
4 -012
3 on
2 -on
1 -007
1 -017
0 -006
0 4is-
—
3 -tor
2 -007
0-01
0 -001
0 -007
—
—
6 -oos*
5 4io-
3 -ecu
2 -OM
1 -001
I 4M
0 -004
0 4io-
_
5 -003
4 4os
3 -007
2 -o«
I -004
0 -002
0 -003-
—
—
5 -OOT
3 -003
2 -003'
1 -002
1 -006
0 -003
0 -007
	
—
4 403-
3 -007
2 -006
1 -004
0 -001
0 -OM
—
—
3 -003
2 -003
1 -002
1 -007
0 -001
0 -006
—
—
3 -007
2 -007
0-005
0 -003
—
—
—
5 -ooj
4 -003
3 404
2 404
1 -003
0 -001
0 -OH
—
—
5 -003
3 -ooj
2 -002
1 -001
1 40*
0 -001
0 -ooj-
—
—
4 -ooi
3 -003
2 -003
1 -002
0 -ooi
0 -001
	
	
—
4 403-
2 -ooi
1 431
1 -004
0 4ot
0 -004
—
—
3 -003
2 -003
1 -002
0 -ooi
0 -ooi
—
—
—
2 -ooi
1 -001

A=13 B=7





fi






5





4




3




2

A = 14 B=14









13






a
11
10
9
8
7
6
13
12
11
10
9
8
7
13
12
11
10
9
S
13
12
11
10
9
13
12
11
10

13
12
14
13
12
11
10
9
8
7
6
5
4
14
13
12
11
10
9
8
Probability
0-05
2 -012
1 -012
1 4»
0 -010+
0 -012
0 -044
3 -031
2 -017
2 -o««
I "014
1 4»-
0 -Ot7
0 414
2 -012
2 4*4
1 -on
1 -047
0 415-
0 -019
2 -044
1 -012
0 -006
0 -013"
0 4»
I 415
0 -007
0 -on
0 -ox

0 -oio-
0 •«»
10 *w
8 -OH
6 -023
5 -«7
4 -021
3 -017
2 -013
1 -016
1 -o»
0 -oio
0 -ow
9 -041
7 -o»
6 -017
S 441
4 «*t
3 -oil
2 -on
0425
2 -on
1 -Oil
0 -O04
0 -010*
0 -022
—
3 -021
2 -017
I -oio-
1 -024
0 -ooi
0 -017
—
2 -an
1 -oo»
1 -021
0 -OOT
0 -013-
	
1 -006
1 -022
0 -oot
0 -015-
_
1 -OM
0 -007
0 -on


0 -oio-
—
9 420
7 -01*
6 -021
4 -oil
3 -on
2 -009
2 -021
1 -OIS
0 -DOt
0 420
—
8) 4I«
6 411
5 4M+
4 417
3 416
2 413
I 409
OOI
1 404
0 401
0 -004
—
—
—
2 404
1 403
1 -oio-
0 403
0 401
—
—
1 402
I -001
0 402
0 -007
	
	
1 406
0 402
0 406
__
	
0 402
0 407



0 410-
	
8 40t
6 406
5 409
3 404
2 403
2 409
1 406
0 -003
0 401
—
—
7 40<
5 404
4 405+
3 «w
2 40J-
1 -001
1 40V
ooo;
1 -004
0 402
0 404
_
_
—
2 404
1 -003
0 401
0 -003
—
—
—
1 402
0 401
0 -003
—
—
—
0 400
0 402
—
—
	
0 402
	
	


—
_~_
7 403
5 -002
4 403
3 404
2 403
1 402
0 401
0 403
—
—
—
6 401
5 4M
3 -003
2 401
2 40J-
1 403
0 401
                             237

-------
TABLE G.5.   SIGNIFICANT LEVELS OF B:  VALUES OF  B  (LARGE TYPE)
            AND CORRESPONDING PROBABILITIES (SMALL TYPE)
            (CONTINUED)

A = 14 B=13


12









n









10










9








S








a
7
6
5
14
13
12
11
10
9
8
7
6
5
14
13
12
11
10
9
8
7
6
5
14
13
12
11
10
9

8
7
6
5
14
13
12
11
10
9
S
7
6
14
13
12
II
10
9
8
7
6
Probability
0-05
1 421
1 441
0 423-
8 411
6 421
5 423-
4 -026
3 424
2 419
2 441
1 421
0 413
0 410
7 426
6 4»
5 443
4 442
3 43«
2 -027
1 -017
1 431
0 417
0 431
6 420
5 421
4 42t
3 424
2 -oil
2 440

I -014
0 410-
0 -032
0 447
6 -047
4 411
3 417
3 4*1
2 429
1 .417
1 43*
0 414
0 430
5 43*
4 43*
3 -on
2 422
2 *41
1 42*
0 40*
0 420
0 440
M*
1 421
0 -010*
0 423-
7 412
6 421
4 409
'J-009
3 424
2 419
1 412
0 403*
0 411
	
6 409
5 414
4 41*
3 4»-
2 411
1 407
1 417
0 407
0 417
—
6 4lo
4 409
3 -ow
3 424
2 411
1 411

1 424
0 410-
0 422
	
S 414
4 411
3 417
2 411
1 407
1 417
0 40C
0 4(4
—
4 -oio-
3 411
2 401
2 ^22
1 -Oil
0 404
0 -oo*
0 420
~~
041
0 404
,—
—
6 404
3 407
4 409
3 409
2 407
1 40)-
0 402
0 403*
—
—
6 409
4 404
3 403-
2 404
1 403
1 407
0 403
0 407
—
—
5 -006
4 409
3 409
2 407
1 404
0 402

0 404
0 410-
_
	
4 404
3 403-
2 404
1 402
1 407
0 401
0 406
—
—
4 -010-
2 402
2 401
1 403-
0 402
0 404
0 40*
	
~~
0-005
0 404
—
—
6 404
4 402
3 40)
2 402
1 402
1 403-
0 402
	
—
—
S 401
4 404
3 4M-
2 -004
I 401
0 401
0 401
—
—
—
4 402
3 -ooi
2 -oo:
1 OOI
1 404
0 -002

0 404
	
	
	
4 404
3 403-
2 4O*
1 oo:
0 -001
0 002
—
—
	
3 -co:
2 -002
1 40 1
1 403-
0 402
0 404
—
—
—

A=!4 B=7







6







5



1


4







3



2




A=15 B=15











a
14
13
12
11
10
9
8
7
14
13
12
11
10
9
8
7
14
13
12
11
10
9
S
14
13
12
11
10

9

14
13
12
11
14
13
12


15
14
13
12
11
10
9
8
7
6
5
4
Probability
0-05
4 42S
3 423
2 417
2 441
1 411
1 -043
0 413-
0 410
3 411
2 414
2 -037
1 411
1 431
0 412
0 42*
0 444
2 410-
2 437
1 417
1 431
0 411
0 422
0 440
2 439
1 419
1 444
0 411
0 423
A
0 441

1 412
0 406
0 4(3-
0 -029
0 401
0 4:1
0 430


11 430-
9 440
7 423*
O 410
S 431
4 411
3 430
2 42S."
1 411
1 440
0 421
0 4»-
0-025
3 406
2 40fi
2 417
1 409
1 421
0 407
0 415-
—
3 411
2 414
1 407
1 411
0 40J-
0 412
0 -02'
—
2 410-
1 406
1 417
0 -oo;-
0 -on
0 -022
	
1 403-
1 419
0 403-
0 411
0 413



1 422
0 406
0 41)-
—
0 401
0 423
	


10 421
8 -on
6 410*
S 413
4 411
3 41)
2 410*
I 407
1 411
0 401
0 421
—
0-01
3 406
2 40*
1 403
1 -009
0 403
0 407
	
—
2 401
1 402
0-005
2 401
1 401
1 401
0 401
0 403
—
	
	
2 40J
1 402
1-007 ! 0 401
0 402
0 40)-
	
—
—
1 401
1 40*
0 401
0 -003-

	
	
1 40)-
0 -ooi
0 40)-





0 401
0 -oo*
—
—
0 401
	
_


9 401
7 407
5 404
4 -003-
3 403-
2 404
1 401
1 407
0 401
0 401
_
—
0 402
	
—
—
	
1 401
0 -ooi
0 402
0 4M-
	
	
—
1 403-
0 402
0 40«-

	 .



0 401
—
—
—
—
	
—


8 401
6 4oi
5 404
4 40S-
3 40J-
2 404
1 401
0 401
0 401
—
_
—
                             238

-------
TABLE G.5.   SIGNIFICANT LEVELS OF B: VALUES OF B (LARGE TYPE)
            AND CORRESPONDING PROBABILITIES (SMALL TYPE)
            (CONTINUED)

A- 15 2=14










13










12










11










10









9

a
15
14
13
12
11
10
9
8
7
6
S
15
14
13
12
11
10
9
8
7
6
5
15
14
13
12
11
10
9
8
7
6
5
15
14
13
12
11
10
9
8
7
6
5
15
14
13
12
11
10
9
8
7
6
15
14
Probability
5
10 041
8 011
7 041
6 046
5 041
4 04*
3 041
2 O33
1 on
1 O49
0 O13+
9 ois-
7 -023
6 -029
5 O3I
4 O3Q
3 004
2 030
2 04)
1 O29
0 OI3
0 031
8 on
7 O43
6 O49
S O49
4 043-
3 oji
2 021
1 on
1 ou
Q 017
0 O37
7 O32
6 031
5 O34
4 031
3 -026
2 «1»
2 O40
1 O24
1 -049
0 on
0 O46
6 on
5 ou
4 O23
3 -01*
3 041
2 019
1 OU
1 O34
0 ou
0 on
.6 -043
5 04T
0-025
9 017
7 OI3
6 017
5 010
4 020
3 4.11
2 0|4
1 009
1 032
0 Oil
—
8 013
7 023
5 on
4 -013
3 on
2 oot
2 030
1 ou
0 003+
0 ou
—
7 010-
6 OI6
5 019
4 019
3 017
2 -on
1 007
1 Oil
0 007
0 017
_
7 032
5 on
4 013
3 010+
2 on
2 O|9
1 Oil
1 -014
0 oio-
0 021
—
6 017
S 023
4 OZZ
3 oil
2 ou
1 007
1 0|6
0 006
0 ou
—
5 -013
4 -013-
0-01
8 006
6 403-
5 007
4 O07
3 O07
2 004
1 O04
1 009
0 -004
—
—
7 ooj-
6 O09
4 O04
3 004
2 O03
2 ooi
1 003+
0 O02
0 O03+
—
_
7 oio-
5 O06
4 007
3 -006
2 405-
1 lOOJ
1 -007
0 O03
0 O07
—
	
6 OOT
4 003
3 -003
2 O03
2 oot
1 -004
0 -002
0 oo4
0 010-
—
—
5 005-
4 007
3 007
2 005-
1 O03
1 007
0 O02
0 OM
—
—
4 O03
3 -oo*
0005
7 ooi
6 oo)-
4 ooi
3 002
2 O03
1 OOI
1 O04
0 ooi
0 O04
—
—
7 00)-
5 003
4 004
3 004
2 O03
1 003
0 ooi
0 O03
—
—
_
6 003
4 ooi
3 -002
2 ooi
2 003-
1 003
0 ooi
0 O03
—
__
—
5 002
4 O03
3 O03
2 -003
1 001
1 -004
0 ooz
0 004
—
—
—
5 ooj-
3 002
2 ooi
2 ooj-
1 00)
0 ooi
0 4oj
—
_.
—
4 -403
3 004

A = 15 B = 9







8









7








6







5






4





3




2


a
13
12
11
10
9
8
7
6
15
14
13
12
11
10
9
8
7
6
15
14
13
12
II
10
9
S
7
15
14
13
12
11
10
9
8
15
14
13
12
11
10
9
15
14
13
12
11
10
15
14
13
12
11
15
14
13
Probability
0-05
4 04i
3 o»
2 -021
2 045-
1 034
1 O4I
0 OI9
0 -037
5 032
4 O31
3 42S
2 on
2 O37
1 O19
1 031
0 OI3
0 O16
0 O50-
4 41)
3 431
2 414
2 432
1 015+
1 -031
0 OIO+
0 O20
0 4)1
3 015+
2 on
2 O31
1 OI4
1 O29
0 -009
0 on
0 OJ2
2 009
2 431
1 O14
1 0)1
0 oot
0 O16
0 O30
2 O3J-
1 416
1 437
0 409
0 Oil
0 033
1 030
0 003-
0 012
0 013-
0 043
0 007
0 412
0 444
0-025
3 013
2 009
2 421
1 on
I O34
0 409
0 019
—
4 oot
3 409
2006
2 417
1 oot
1 OI9
0 006
0 ou
—
—
4 013
3 0:1
2 414
I 007
1 413+
0 OM-
0 410+
0 410
—
3 015+
2 411
1 -00*
1 OM
0 404
0 009
0 017
— "•
2 409
1 405-
1 OI4
0 404
0 401
0 016
_
1 404
1 416
0 404
0 409
0 Olt
—
1 O20
0 405-
0 oiz
0 013-
_
0 407
0 4Z2
—
0-01
2 403
2 on
1 403-
0 403
0 -*M
0 on
—
—
4 401
3 409
2 006
1 403
1 OM
0 OOJ
0 O06
—
—
—
3 ooj-
2 404
1 00!
1 O07
0 401
0 oos-
—
—
—
2 403
1 402
1 406
0 401
0 OM
0 409
—
—
2 009
1 40)-
0 401
0 OM
0 oot
—
—
1 404
0 ODI
0 404
0 409
—
— -
0 401
0 oos-
—
_
—
0 407
—
—
0-005
2 003
1 401
1 oos-
0 401
0 4M
—
—
—
3 ooz
2 003
1 OOI
1 403
0 401
0 003
—
—
—
—
3 405-
2 004
1 403
0 OOI
0 402
0 403-
—
—
—
2 001
1 002
0 401
0 402
0 404
— -
—
—
1 OOI
1 403-
0 401
0 404
—
—
—
I O04
0 401
0 404
—
—
—
0 oot
0 005-
—
—
—
—
—
—
                              239

-------
                                 APPENDIX H

              TOXICITY  SCREENING  TEST  - COMPARISON OF CONTROL
            WITH  100%  EFFLUENT OR  INSTREAM WASTE CONCENTRATION
1.  To statistically compare a control with one concentration, such as
100% effluent or the instream waste concentration,  a t test is the
recommended analysis.  The t test is based on the assumptions that the
observations are independent and normally distributed and that the
variances of the observations are equal between the two groups.

2.  Shapiro-Wilk's test may be used to test the normality assumption (See
Appendix B for details).  If the data do not meet the normality
assumption, the non-parametric 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
5.  Compare F with the 0.005 level of a tabled F value with n-|  - 1  and
r\2 - 1 degrees of freedom, where n-|  and n2 are the number of replicates
for each of the two groups.

6.  A set of Ceriodaphnia 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.   CERIQDAPHNIA REPRODUCTION DATA
                               FROM AN EFFLUENT SCREENING  TEST
Replicate

Control
100% Effluent
1
36
23
2
38
14
3
35
21
4
35
7
5
28
12
6
41
17
7
37
23
8
33
8
9
*
18
10 J
. 35.4
. 15.9
S2
14.5
36.6
                                      240

-------
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.

                               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  =
                                     Al^T
                                    \l  ni     n?
     Where:
                _
                Yj  =  Mean for the control
                V2  =  Mean for the effluent concentration
                         (n1 - 1)  S   +  {n2 - 1) S

                               R   +  n   -  2
                    =  Estimate of the variance for the control

                    =  Estimate of the variance for the effluent
                       concentration
                n-,   =  Number of replicates for the control

                n«   =  Number of replicates for the effluent
                       concentration
                                    241

-------
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:
                                    35.4 - 15.9
                             t  =  	   =  7.82
5.13/1+1
            9
                                      3/T
                                      J  8
     Where:
                         (8 - 1) 14.5  +  (9-1)  36.6

                Sp  -J  	8  +  9  -  2	   =  5-13

9.3  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
follows:
                                      Y  - Y
                             t  -      1     2
                                         +  S2
                                            II
                                            n2

     Where:      _
                Y_l   =  Mean for the control
                Y2   =  Mean for the effluent concentration
                 2
                S,   =  Estimate of the variance for the control

                 2
                Sp   =  Estimate of the variance for the effluent
                       concentration
                n,   =  Number of replicates  for the control

                                    242

-------
                n2  =  Number of replicates for the effluent
                       concentration
10.2  Additionally, the degrees of freedom for the test are adjusted
using the following formula:
               d
         Where:                  9
                                sf
(n, - l)(n 0 - 1)
1 c.
(« i\ r*- j.
n« - 1)C +
(1 - C)2(n1
- 1)
                                "
                c  =
10.3  The modified degrees of freedom is usually not an integer.  Common
practice is to round down to the nearest integer.

10.4  The t test is then conducted as the equal variance t test.  The
calculated t is compared to the critical t at the 0.05 significance level
with the modified degrees of freedom.  If the calculated t exceeds the
critical t» the mean responses are found to be statistically different.
                                    243

-------
                                APPENDIX  I

                              PROBIT ANALYSIS

1.1  This program calculates the EC50, EC15, EC10, ECS,  and EC!  values,
and 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 full listing of the program is
contained in EPA/600/4-87/028.  A compiled version of the program can be
obtained from EMSL-Cincinnati by sending a diskette with a written
request.

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 begins with a request for the following information:


1. Output designation (P = printer, D = disk file).
2. Title for the output.
3. A selection of model  fitting options (see sample output
   for a detailed description of options).  If Option 2 is
   selected, the theoretical lower threshold needs to be entered.
   If option 3 is selected, the program requests the number of
   animals responding in the control group and the total number
   of original animals in the control  group be entered.
4. The number of test concentrations.

2.2.  The program then requests information on the results at each
concentration, beginning with the lowest concentration.

1. Concentration.
2. Number of organisms responding.
3. Total number of exposed organisms.

2.2.1.  See sample data input on the next page.
                                    244

-------
2.2.1  Sample Data Input.
 yuuuuuuuuuuuuuuuuuuuuuuuuuuTJUuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu
 ii                                                            tr
 U                  EPA PROBIT ANALYSIS  PROGRAM               tJ
 U               USED  FOR CALCULATING EC VALUES               U
 U                         Version 1.4                        U
 uuuuuuuuutnJuuutJtJuumJuuuuutJuuuuuuuuuuuuuuuuuuuuuuu
Output to printer or  disk  file  (P  /  D)?  p
Title ? Example  for 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 (l, 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
                                245

-------
2.2.1 Sample Data Input (Continued).
Input data starting with the  lowest concentration

Concentration = ? 3.0
Number responding = ? 14
Number exposed = ? 100

Concentration = ? 5.0
Number responding = ? 16
Number exposed = ? 102

Concentration = ? 7.0
Number responding = ? 35
Number exposed = ? 100

Concentration = ? 11.0
Number responding = ? 72
Number exposed = ? 99

Concentration = ? 16.0
Number responding = ? 99
Number exposed - ? 99
Number
1
2
3
4
5
Cone.
3.0000
5.0000
7.0000
11.0000
16.0000
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
                                246

-------
2.3  Sample Data Output

2.3.1  The program output includes  the  following:

2.3.1.1  Statistical  table (Table  I.I.)

    1.  The observed, adjusted (using Abbott's  formula)
        and predicted proportions  responding  at each concentration.
    2.  Chi-square statistic for heterogeneity.   This  test  is one
        indicator of how well  the  data  fit the  model.
    3.  Estimates of the mean (mu)  and  standard deviation  (sigma)
        of the underlying tolerance distribution.
    4.  Estimates and standard errors of the  intercept and  slope of
        the fitted probit regression line.
    5.  Estimate and standard error of  the lower threshold  (if
        requested - requires control data on  input).
    6.  A list of estimated EC values and 95% confidence limits.
        Please note that EC, effective  concentration,  is a  broad term
        and applies to any response, such as  fertilization,  death or
        immobilization.  If mortality data is entered  in the program
        as the response, the EC estimates are equivalent to  LC
        (lethal  concentration) estimates.

2.3.1.2  Plot (Figure I.I.)

    1.  A plot of the fitted probit regression  line with observed
        data overlaid on the plot.
                                     247

-------
                   TABLE I.I.  OUTPUT FROM PROBIT PROGRAM
    Cone.

   Control
    3.0000
    5.0000
    7.0000
   11.0000
   16.0000
 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
  -.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
-2.764127
8.088003
1.002530
0.990954
( -4.729086,
( 6.145732,
-0.799168)
10.030273)
Spontaneous
Response Rate
    0.156014
                            0.022593
                     0.111732,
                          0.200296)
      Estimated EC Values and confidence Limits
Point

EC l.OO
EC 5.00
EC10.00
EC15.00
EC50.00
EC85.00
EC90.00
EC95.00
EC99.00
       Cone.

        4.7025
        5.7093
        6.3314
        6.7892
        9.1192
       12.2489
       13.1345
       14.5657
       17.6840
                                      Lower       Upper
                                    95% Confidence Limits
             3.6073
             4.6408
             5.3031
             5.7994
             8.3614
            11.4157
            12.1697
            13.3302
            15.7134
                    5.5567
                    6.5196
                    7.1058
                    7.5354
                    9.7763
                   13.3942
                   14.5708
                   16.5676
                   21.2145
                                      248

-------
Probit
   10+
    8+
    4+
O+o  o
  — ^•^
-------

-------

-------
Environmental  Protection
Agency
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use,  5300
                                                                   Please make all necessary changes on the above label,
                                                                   detach or copy, and return lo the address in the upper
                                                                   left-hand corner

                                                                   If you do no! wish to receive these reports CHECK HERE D;
                                                                   deiach, or copy this cover, and return to the address in the
                                                                   upper lefl-hand corner
                                                                 EPA/600/4-89/001

-------