&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
------- |