US Army Corps
• of Engineers *
Engineer Research and
Development Center
United States Office of Research and Office of Water Department of the Army EPA/600/R-01/020
Environmental Protection Development Washington, DC 20460 U.S. Army Corps of March 2001
Agency Washington, DC 20460 Engineers
__ Vicksburg, MS 39180
&EPA Methods for Assessing the
Chronic Toxicity of
Marine and Estuarine
Sediment-associated
Contaminants with the
Amphipod
Leptocheirus plumulosus
First Edition
-------
-------
EPA 600/R-01/020
March 2001
Method for Assessing the Chronic Toxicity
of Marine and Estuarine Sediment-associated
Contaminants with the Amphipod
Leptocheirus plumulosus
First Edition
Office of Research and Development
Western Ecology Division
U.S. Environmental Protection Agency
Newport, OR 97365
Office of Science and Technology
Office of Water
U.S. Environmental Protection Agency
Washington, D.C. 20460
Engineer Research and Development Center
Waterways Experiment Station
U.S. Army Corps of Engineers
Vicksburg,MS39180
-------
-------
Foreword
Sediment contamination is a widespread environmental problem that can potentially pose a threat to a variety
of aquatic ecosystems. Sediment functions as a reservoir for common chemicals such as pesticides,
herbicides, polychlorinated biphenyls (RGBs), polycyclic aromatic hydrocarbons (PAHs), and metals, such
as lead, mercury, and arsenic. In-place contaminated sediment can result in depauperate benthic
communities. Because relationships between bioavailability and concentrations of chemicals in sediment
are not fully understood, determination of contaminated sediment effects on aquatic organisms requires the
use of controlled toxicity and bioaccumulation tests.
As part of USEPA's Contaminated Sediment Management Strategy, Agency programs have agreed to use
consistent methods to determine whether sediments have the potential to affect aquatic ecosystems. More
than ten federal statutes provide authority to many USEPA program offices to address the problem of
sediment contamination. The use of this uniform sediment testing procedure is expected to increase data
accuracy and precision, facilitate test replication, and increase the comparative value of test results. The
sediment test method in this manual may be useful in assessing sediment contamination, registration of
pesticides, assessment of new and existing industrial chemicals, Superfund site assessment, and assessment
and cleanup of hazardous waste treatment, storage, and disposal facilities. Each EPA Program will, however,
retain the flexibility of deciding when and how to use this test and whether identified risks would trigger actions.
A chronic sediment toxicity test (which is used to study the effects of continuous, long-term exposure of a
toxicant on an organism) using the estuarine benthic amphipod, Leptocheirus plumulosus, was developed by
DeWitt et al. (1992a) for the USEPA. McGee et al. (1993) and Emery et ai. (1997) independently developed
chronic test methods with L. plumulosus that measured similar endpoints. Subsequent to these method
development efforts, the USEPA and the U.S. Army Corps of Engineers (USAGE) have funded further
research to refine this chronic method. Findings from studies at both organizations have been incorporated
into the chronic testing method described in this document. The protocol for the L. plumulosus 28-d sediment
toxicity test will be revised periodically, as such, users of this manual are encouraged to contribute to this effort
by sending to the USEPA the results of experiments that could bring to light any deficiencies or improvements
to the L. plumulosus 28-d sediment toxicity test. Send these results and all supporting information (i.e.,
experimental conditions and procedures) to the U. S. Environmental Protection Agency, Office of Science and
Technology/Standards and Health Protection Division (mail code 4305), ATTN: Contaminated Sediment
Program, 1200 Pennsylvania Avenue, NW., Washington, D.C. 20460. Contributors to the improvement of the
methodology will be acknowledged in future revisions to this manual.
This document is supplementary to USEPA (1994d), but does not replace it. The approaches described in
this manual were developed from DeWitt et al. (1992; 1997a; 1997b), McGee et al. (1993), Emery et al.
(1997), Scott and Redmond (1989), DeWitt etal. (1989), Schlekatetal. (1992), American Society for Testing
and Materials (ASTM, 2000a; 2000f), U.S. Army Corps of Engineers (Emery and Moore, 1996), and U.S.
Environmental Protection Agency (USEPA 1994d, 2000).
For additional guidance on the technical considerations in the manual, please contact Ted DeWitt, USEPA,
Newport, OR (541/867-4029, fax -4049, email dewitt.ted@epamail.epa.gov) or Todd Bridges, USAGE,
Vicksburg, MS (601/634-3626, fax-3713, email Todd.S.Bridges@erdc.usace.army.mil).
The cover art is an illustration of Leptocheirus plumulosus, by E.L. Bousfield, reproduced with permission of
the Canadian Museum of Nature, Ottawa, Canada.
-------
Disclaimer
This technical manual describes recommended procedures for testing an estuarine organism in the laboratory
to evaluate the potential toxicity of contaminants in whole sediments. This manual has no immediate or direct
regulatory consequence. It does not impose legally binding requirements on the U. S. Environmental
Protection Agency (EPA), the U.S. Army Corps of Engineers (USAGE), states, tribes, other regulatory
authorities, or the regulated community, and may not apply to a particular situation based upon the
circumstances. EPA, USAGE, state, tribal, and other decision makers retain the discretion to adopt
approaches on a case-by-case basis that differ from those in this manual where appropriate. EPA or USAGE
may change this manual in the future.
The information in this document has been funded in part by EPA and USAGE. It has been subjected to
review by EPA's National Health and Environmental Effects Research Laboratory and Office of Science and
Technology and approved for publication. Mention of trade names or commercial products does not constitute
endorsement by either Agency or recommendation for use.
IV
-------
Abstract
A laboratory method is described for determining the chronic toxicity of contaminants associated with whole
sediments. Sediments may be collected from estuarine or marine environments or spiked with compounds
in the laboratory. The toxicity method outlined uses an estuarine crustacean, the amphipod Leptocheirus
plumulosus. The toxicity test is conducted for 28 d in 1 -L glass chambers containing 175 mL of sediment and
about 725 ml of overlying water. Test temperature is 25° ±2°C, and the recommended overlying water salinity
is 5%o ±2%o (for test sediment with pore water at 1 %o to 10%0) or 20%o ±2%o (for test sediment with pore water
>10%o). Four hundred milliliters of overlying water is renewed three times per week, at which times test
organisms are fed. The endpoints in the toxicity test are survival, growth, and reproduction of amphipods.
Performance criteria established for this test include the average survival of amphipods in negative control
treatment must be greater than or equal to 80% and there must measurable growth and reproduction in all
replicates of the negative control treatment. This test is applicable for use with sediments from oligohaline
to fully marine environments, with a silt content greater than 5% and a clay content less than 85%.
-------
Acknowledgments
This document is a general purpose toxicity testing manual for estuarine and marine sediment. The
approaches described in this manual were developed primarily from DeWitt et al. (1992, 1997a, 1997b),
McGee et al. (1993), and Emery et al. (1997). That work and the impetus for this manual were derived from
previous key papers and reports, including Swartz et al. (1985), Scott and Redmond (1989), DeWitt et al.
(1989), Schlekatetal. (1992), American Society for Testing and Materials (ASTM, 1998b; 1998e), U.S. Army
Corps of Engineers (Emery and Moore, 1996), and U.S. Environmental Protection Agency (USEPA, 1994d;
2000). This manual incorporates general guidelines that reflect the consensus of the USEPA Program Offices
and the USAGE.
The principal authors of this manual are Theodore H. DeWitt (USEPA, Office of Research and Development,
National Health and Environmental Effects Research Laboratory, Western Ecology Division), Todd S. Bridges
(USAGE, Engineer Research and Development Center), D. Scott Ireland and Leanne L. Stahl (USEPA, Office
of Water, Office of Science and Technology), and Margaret R. Pinza and Liam D. Antrim (Battelle Marine
Science Laboratory).
Some of the material in this manual was excerpted from USEPA (1994a, 2000) for the purpose of consistency
among sediment toxicity test methods manuals. Contributors to specific sections of the manual, including
contributors to relevant sections of USEPA (1994a and 2000), are:
1. Section 1-9; General Guidelines
G.T. Ankley, USEPA, Duluth, MN
T.S. Bridges, USAGE, Vicksburg, MS
G.A. Burton, Wright State University, Dayton, OH
T.D. Dawson, ILS, Duluth, MN
T.H. DeWitt, USEPA, Newport, OR
F.J. Dwyer, USGS, Columbia, MO
R.A. Hoke, DuPont, Newark, DE
C.G. Ingersoll, USGS, Columbia, MO
D.S. Ireland, USEPA, Washington, DC
N.E. Kemble, USGS, Columbia, MO
J.O. Lamberson, USEPA, Newport, OR
D.R. Mount, USEPA, Duluth, MN
T.J. Norberg-King, USEPA, Duluth, MN
M.S. Redmond, Northwestern Aquatic Sciences, Newport, OR
C.E. Schlekat, USGS, Menlo Park, CA
K.J. Scott, SAIC, Narragansett, Rl
L. Stahl, USEPA, Washington, DC
R.C. Swartz, USEPA, Newport, OR(retired)
2. Sections 10-11; Culture and Test Methods
L.D. Antrim, NOAA Olympic Marine Sanctuary, Port Angeles, WA
T.S. Bridges, USAGE, Vicksburg, MS
T.H. DeWitt, USEPA, Newport, OR
V.L. Emery, USAGE, Vicksburg, MS
B.D. Gruendell, Battelle, Sequim, WA
vi
-------
B.L.. McGee, USFWS, Annapolis, MD
D.W. Moore, MEC Analytical Systems, Carlsbad, CA
L.A. Niewolny, Battelle, Sequim, WA
M.R. Pinza, Battelle, Sequim, WA
3. Sections 12; Statistical Analysis
V.I. Cullinan, Battelle, Sequim, WA
J.D. Farrar, AScI, Vicksburg, MS
B.R. Gray, AScI, Vicksburg, MS
J. Heltshe, SAIC, Narragansett, Rl
R.A. Hoke, DuPont, Newark, DE
H. Lee, USEPA, Newport, OR
T.J. Norberg-King, USEPA, Duluth, MN
C.E. Schlekat, USGS, Menlo Park, CA
4. Section 13; Precision and Accuracy
V.I. Cullinan, Battelle, Sequim, WA
T.H. DeWitt, USEPA, Newport, OR
C.G. Ingersoll, USGS, Columbia, MO
T.J. Norberg-King, USEPA, Duluth. MN
M.R. Pinza, Battelle, Sequim, WA
Review comments from the following individuals led to substantial improvements in the manual for which we
are grateful: B.A. Barbo, Battelle, Sequim, WA; Walter Berry, USEPA, Narragansett, Rl; G.A. Burton, Wright
State Univ., Dayton, OH; P. Crocker, USEPA, Dallas, TX; V.L. Emery, Vicksburg, MS; Daniel Farrar, AScI,
Vicksburg, MS; C.G. Ingersoll, USGS, Columbia, MO; Laura Johnson, USEPA, Washington, DC; N.P. Kohn,
Battelle, Sequim. WA; J. Lazorchak, USEPA, Cincinnati, OH; M. Lewis, USEPA, Gulf Breeze, FL; B.L. McGee,
USFWS, Annapolis, MD; D.W. Moore, MEC, Carlsbad, CA; C.E. Schlekat, USGS, Menlo Park, CA; J. Serbst,
USEPA, Narragansett, Rl; J. Smrchek, USEPA, Washington, DC; and Jeff Stevens, USAGE, Vicksburg, MS.
Participation by the following laboratories in the round-robin testing is greatly appreciated: Battelle Marine
Sciences Laboratory, Sequim, WA; Environment Canada, Dartmouth, Nova Scotia, Canada; EVS Consultants,
Inc., Vancouver, British Columbia, Canada; Science Applications International Corporation (SAIC),
Narragansett, Rl; University of California at Santa Cruz, Marine Pollution Studies Laboratory, Monterey, CA;
University of Maryland, Wye Research and Education Center, Queenstown, MD; University of South Carolina,
Environmental Health Sciences Department, Columbia, SC; USAGE Waterways Experiment Station,
Vicksburg, MS; USEPA Atlantic Ecology Division Laboratory, Narragansett, Rl; and USEPA Gulf of Mexico
Ecology Division Laboratory, Gulf Breeze, FL.
We are very grateful to USEPA's Office of Water, Office of Science and Technology, USEPA's Office of
Research and Development, and USAGE'S Long-term Effects of Dredging Operations research program for
supporting the development of this manual and for supporting much of the underlying research.
VII
-------
Contents
Foreword Hi
Disclaimer iv
Abstract * v
Acknowledgments vi
1 Introduction 1
1.1 Significance of Use 1
1.2 Program Applicability 3
1.3 Scope and Application 5
1.4 Performance-based Criteria 10
2 Summary of Method 11
2.1 Method Description and Experimental Design 11
2.2 Types of Tests 13
2.3 Test Endpoints 13
3 Definitions 14
3.1 Terms 14
4 Interferences 16
4.1 General Introduction 16
4.2 Noncontaminant Factors 17
4.3 Changes in Bioavailability 19
4.4 Presence of Indigenous Organisms 20
5 Health, Safety, and Waste Management 22
5.1 General Precautions . 22
5.2 Safety Equipment > • 22
5.3 General Laboratory and Field Operations 22
5.4 Disease Prevention 23
5.5 Safety Manuals 23
5.6 Pollution Prevention, Waste Management, and Sample Disposal 23
6 Facilities, Equipment, and Supplies 24
6.1 General 24
6.2 Facilities '. 24
viii
-------
6.3 Equipment and Supplies 24
7 Water, Reagents, and Standards '. 28
7.1 Water 28
7.2 Reagents 29
7.3 Standards ., 29
8 Sample Collection, Storage, Manipulation, and Characterization 30
8.1 Collection . . : 30
8.2 Storage 30
8.3 Manipulation 31
8.4 Characterization : 33
9 Quality Assurance and Quality Control ; 36
9.1 Introduction - 36
9.2 Performance-based Criteria 36
9.3 Facilities, Equipment, and Test Chambers 36
9.4 Test Organisms 37
9.5 Water 37
9.6 Sample Collection and Storage 37
9.7 Test Conditions 37
9.8 Quality of Test Organisms 37
9.9 Quality of Food , 38
9.10 Test Acceptability 38
9.11 Analytical Methods 38
9.12 Calibration and Standardization 38,
9.13 Replication and Test Sensitivity 38
9.14 Demonstrating Acceptable Performance 39
9.15 Documenting Ongoing Laboratory Performance 39
9.16 Reference Toxicants 39
9.17 Record Keeping 40
10 Collection, Culture, and Maintaining of Test Organisms 42
10.1 Life History I 42
10.2 General Culturing Procedures 43
10.3 Culturing Procedure for Leptocheirus plumulosus . 44
10.4 Field Collection ....:...'. ." ; 47
ix
-------
10.5 Obtaining Amphipods for a Test 49
10.6 Minimization of Risk of Release of Nonindigenous Organisms 49
11 Leptocheirus plumulosus 28-d Chronic Test for Sediment .: , 50
11.1 Introduction 50
11.2 Procedure for Conducting a Leptocheirus plumulosus 28-d Test for Measuring
Subiethal Effects of Sediment-associated Contaminants 51
11.3 General Procedures 51
11.4 Interpretation of Results 60
12 Data Recording, Data Analysis and Calculations, and Reporting 63
12.1 Data .Recording 63
12.2 Data Analysis 63
12.3 Data Interpretation 81
12.4 Reporting '. 82
13 Precision and Accuracy 83
13.1 Determining Precision and Accuracy 83
13.2 Accuracy , -.: .84
13.3 Replication and Test Sensitivity ; 84
13.4 Demonstrating Acceptable Laboratory Performance '....' 84
13.5 Precision of the 28-d Chronic Sediment Toxicity Test Method 86
14 References '. 91
Tables
1.1 Sediment Quality Assessment Procedures 4
1.2 Statutory Needs for Sediment Quality Assessment 6
4.1 Advantages and Disadvantages for Use of Sediment Tests 16
6.1 Equipment and Supplies for Culturing and Testing Leptocheirus plumulosus 26
9.1 Recommended Test Conditions for Conducting Reference-toxicity Tests 41
11.1 Test Conditions for Conducting a 28-d Chronic Sediment Toxicity Test with Leptocheirus
plumulosus 52
11.2 General Activity Schedule for Conducting a 28-d Chronic Sediment Toxicity Test with
Leptocheirus plumulosus : 54
11.3 Test Acceptability Requirements for a 28-d Chronic Sediment Toxicity Test with Leptocheirus
plumulosus 55
x
-------
12.1 Suggested alpha Levels to Use for Tests of Assumptions 73
13.1 Intralaboratory Precision Distribution of the Coefficient of Variation for Each Test Endpoint .... 86
13.2 Ranges of the Black Rock Harbor Sediment Dilution Series Chemical Concentrations 87
13.3 Interlaboratory Precision for Each Endpoint for Leptocheirus plumulosus in a 28-d Long-term
Sediment Toxicity Test using Black Rock Harbor Spiked Sediments 88
13.4 Summary of Intralaboratory Precision at Five Laboratories for the 28-d Leptocheirus plumulosus
Chronic Test Using Five Dilutions of Black Rock Harbor Sediment 89
Figures
10.1 Leptocheirus plumulosus morphology (A) and characteristics used to determine the gender (B-C)
of the amphipod 42
12.1 Treatment response for a Type I and Type II error 66
12.2 Power of the test vs. percentage reduction in treatment response relative to the control mean at
various CVs (8 replicates, a = 0.05 [one-tailed]) 68
12.3 Power of test vs. percentage reduction in treatment response relative to the control mean at
various CVs (5 replicates, a = 0.05 [one-tailed]) 68
12.4 Power of the test vs. percentage reduction in treatment response relative to the control mean at
various CVs (5 replicates, a = 0.10 [one-tailed]) ... 69
12.5 Effect of CV and number of replicates on the power to detect a 20% decrease in treatment
response relative to the control mean (a = 0.05 [one-tailed]) 69
12.6 Effect of alpha and beta on the number of replicates at various CVs 71
12.7 Decision tree for analysis of survival, growth rate, and reproduction data subjected to hypothesis
testing 71
12.8 Decision tree for analysis of point estimate data 78
13.1 Control charts: (A) hypothesis testing results; and (B) point estimates (LC, EC, or 1C) 85
Appendix A Example Data Forms for Use with the 28-d Chronic Test 104
XI
-------
-------
Section 1
Introduction
1.1 Significance of Use
1.1.1 Sediment provides habitat for many
estuarine and marine organisms and is a major
repository for many of the more persistent
chemicals that are introduced into surface waters.
In the aquatic environment, most anthropogenic
chemicals and waste materials, including toxic
organic and inorganic chemicals, eventually
accumulate in sediment. Mounting evidence exists
of environmental degradation in areas where U.S.
Environmental Protection Agency (USEPA, or the
Agency) Water Quality Criteria (WQC; Stephan et
al., 1985) are not exceeded, yet organisms in or
near sediment are adversely affected (Chapman,
1989). The WQC were developed to protect
organisms in the water column and were not
intended to address issues to protect organisms in
sediment. Concentrations of contaminants in
sediment might be several orders of magnitude
higher than in the overlying water; however, bulk
sediment concentrations have not been strongly
correlated to bioavailability (Burton, 1991).
Partitioning or sorption of a compound between
water and sediment may depend on many factors,
including aqueous solubility, pH, redox, affinity for
sediment organic carbon and dissolved organic
carbon, grain size of the sediment, sediment
mineral constituents (oxides of iron, manganese,
and aluminum), and the quantity of acid volatile
sulfides in sediment (Di Toro et al., 1990; 1991).
Although certain chemicals are highly sorbed to
sediment, these compounds may still be available
to the biota. Contaminated sediment may be
directly toxic to aquatic life and can also be a
source of contaminants for bioaccumulation in the
food chain.
1.1.2 Assessments of sediment quality have
commonly included sediment chemical analyses,
and surveys of benthic community structure.
Determination of sediment chemical
concentrations on a dry weight basis alone offers
little insight into predicting adverse biological
effects because bioavailability may be limited by
the intricate partitioning factors mentioned above.
Likewise, benthic community surveys may be
inadequate, because they sometimes fail to
discriminate between effects of contaminants and
those that result from unrelated noncontaminant
factors, including water quality fluctuations,
physical parameters, and biotic interactions. To
obtain a direct measure of sediment toxicity, or
bioaccumulation, laboratory tests have been
developed in which surrogate organisms are
exposed to sediments under controlled conditions.
Sediment toxicity tests have evolved into effective
tools that provide direct, quantifiable evidence of
biological consequences of sediment
contamination that can on|y be inferred from
chemical or benthic community analyses. To
evaluate sediment quality nationwide, USEPA
developed the National Sediment Inventory (NSI),
which is a compilation of existing sediment quality
data and protocols used to evaluate the data. The
NSI was used to produce the first biennial report to
Congress on sediment quality in the United States
as. required under the Water Resources
Development Act of 1992 (USEPA, 1997a;1997b;
1997c). -USEPA's evaluation of the data shows
that sediment contamination exists in every region
and state of the country, and various waters
throughout the United States contain sediment that
is sufficiently contaminated with toxic pollutants to
pose.potential risks to fish and to humans and
wildlife who eat fish. The use of. consistent
sedimenttesting methods described in this manual
will provide high-quality data needed for the NSI,
future reports to Congress, and regulatory
programs to prevent, remediate, and manage
contaminated sediments (USEPA, 1998).
-------
1.1.3 The objective of a sediment test is to
determine whether contaminants in sediment are
harmful to or are bioaccumulated by benthic
organisms. The tests can be used to measure
interactive toxic effects of complex contaminant
mixtures in sediment. Furthermore, knowledge of
specific interactions among sediments and test
organisms is not necessary in order to conduct the
tests (Kemp and Swartz, 1988). However, such
knowledge can be useful to interpret toxicity data.
Sediment tests can be used to (1) determine the
relationship between toxic effects and
bioavailability, (2) investigate interactions among
contaminants, (3) compare the sensitivities of
different organisms, (4) determine spatial and
temporal distribution of contamination, (5) evaluate
dredged material disposal suitability, (6) measure
toxicity as part of product licensing or safety
testing or chemical approval, (7) rank areas for
cleanup, and (8) develop cleanup goals and
estimate the effectiveness of remediation or
management practices for marine or estuarine
environments.
1.1.4 Most standard whole sediment toxicity tests
have been developed to produce a lethality
endpoint (survival/mortality) with potential for a
sublethal endpoint (reburial) in some species.
Methods that measure sublethal effects have not
been available or have not been routinely used to
evaluate sediment toxicity in marine or estuarine
sediments (Scott and Redmond, 1989; Green and
Chandler, 1996; Levin et al., 1996; Ciarelli et al.,
1998; Meador and Rice, 2001). Most
assessments of contaminated sediment rely on
short-term lethality tests (e.g., <;10 d; USEPA-
USACE, 1991; 1998). Short-term lethality tests
are useful in identifying "hot spots" of sediment
contamination, but might not be sensitive enough
to evaluate moderately contaminated areas.
However, sediment quality assessments using
sublethal responses of benthic organisms, such as
effects on growth and reproduction, have been
used to successfully evaluate moderately
contaminated areas (Ingersoll et al., 1998; Kemble
etal., 1994; McGeeetal., 1995; Scott, 1989). The
28-d toxicity test with Leptocheirus plumulosus has
two sublethal endpoints: growth and reproduction.
These sublethal endpoints have potential to exhibit
a toxic response from chemicals that otherwise
might not cause acute effects or significant
mortality in a test. Sublethal response to chronic
exposure is also valuable for population modeling
of contaminant effects. This data can be used for
population-level risk assessments of benthic
pollutant effects.
1.1.5 Results of toxicity tests on sediments spiked
at different concentrations of chemicals can be
used to establish cause-and-effect relationships
between chemicals and biological responses.
Results of toxicity tests with test materials spiked
into sediments at different concentrations may be
reported in terms of a median lethal
concentration(LC50), a median effect
concentration (EC50), an inhibition concentration
(IC50), or as a no observed effect concentration
(NOEC) or lowest observed effect concentration
(LOEC). However, spiked sediment might not be
representative of contaminated sediment in the
field. Mixing time (Stemmer et al., 1990a) and
aging (Word etal., 1987; Landrum, 1989; Landrum
and Faust, 1992) of spiked sediment can influence
bioavailability of contaminants.
1.1.6 Evaluating effect concentrations for
chemicals in sediment requires knowledge of
factors controlling their biqavailability. Similar
concentrations of a chemical in units of mass of
chemical per mass of sediment dry weight often
exhibit a range in toxicity in different sediments (Di
Toro et al., 1990; 1991). Effect concentrations of
chemicals in sediment have been correlated to
interstitial water concentrations, and effect
concentrations in interstitial water are often similar
to effect concentrations in water-only exposures.
The bioavailability of nonionic organic compounds
in sediment is better correlated with the organic
carbon normalized concentration. Whatever the
route of exposure, these correlations of effect
concentrations to interstitial water concentrations
indicate that predicted or measured concentrations
in interstitial water can be used to quantify the
concentration to which an qrganism is exposed.
Therefore, information on partitioning of chemicals
between solid and liquid phases of sediment is
useful for establishing effect concentrations (Di
Toro etal., 1991).
1.1.7 Field surveys can be designed to provide
either a qualitative reconnaissance of the
distribution of sediment contamination or a
quantitative statistical comparison of contamination
-------
among sites. Surveys of sediment toxicity are
usually part of more comprehensive analyses of
biological, chemical, geological, and hydrographic
data. Statistical correlations may be improved and
sampling costs may be reduced if subsamples are
taken simultaneously for sediment tests, chemical
analyses, and benthic community structure.
1.1.8 Table 1.1 lists several approaches the
USEPA has considered for the assessment of
sediment quality (USEPA, 1992c). These
approaches include (1) equilibrium partitioning,
(2) tissue residues, (3) interstitial water toxicity,
4) benthic community structure, (5) whole-
sediment toxicity and sediment-spiking tests,
(6) Sediment Quality Triad, and (7) sediment
quality guidelines (see Chapman, 1989; USEPA,
1989a; 1990a; 1990b; 1992b for a critique of these
methods). The sediment assessment approaches
listed in Table 1.1 can be classified as numeric
(e.g., equilibrium partitioning), descriptive (e.g.,
-whole-sediment toxicity tests), or a combination of
numeric and descriptive approaches (e.g., Effects
Range Median; USEPA, 1992c). Numeric
methods can be used to derive chemical-specific
equilibrium partitioning sediment guidelines
(ESGs) or other sediment quality guidelines
(SQGs). Descriptive methods, such as toxicity
tests with field-collected sediment, cannot be used
alone to develop numerical ESGs or other SQGs
for individual chemicals. Although each approach
can be used to make site-specific decisions, no
single approach can adequately address sediment
quality. Overall, an integration of several methods
using the weight of evidence is the most desirable
approach for assessing the effects of
contaminants associated with sediment (Long and
Morgan, 1990; MacDonald et al., 1996; Ingersoll et
a!., 1996; 1997). Hazard evaluations integrating
data from laboratory exposures, chemical
analyses, and benthic community assessments
provide strong complementary evidence of the
degree of pollution-induced degradation in aquatic
communities (Chapman etal., 1992; 1997; Burton,
1991).
1.2 Program Applicability
1.2.1 The USEPA has authority under a variety of
statutes to manage contaminated sediments
(Table 1.2 and USEPA, 1990c). USEPA's
Contaminated Sediment Management Strategy
(USEPA, 1998) establishes the following four
goals for contaminated sediments and describes
actions that the Agency intends to take to
accomplish these goals: (1) to prevent further
contamination of sediments that may cause
unacceptable ecological or.human health risks;
(2) when practical, to clean up existing sediment
contamination that adversely affects the Nation's
waterbodies or their uses, or that causes other
significant effects on human health or the
environment; (3) to ensure that sediment dredging
and the disposal of dredged material continue to
be managed in an environmentally sound manner;
and (4) to develop and consistently apply
methodologies for analyzing .contaminated
sediments. The Agency plans to employ its
pollution prevention and source control programs
to address the first goal. To accomplish the
second goal, USEPA will consider a range of risk
management alternatives to reduce the volume
and effects of existing contaminated sediments,
including in-situ containment and contaminated
sediment removal. Finally, the Agency is
developing tools for use in pollution prevention,
source control, remediation, and dredged material
management to meet the collective goals. These
tools include national inventories of sediment
quality and environmental releases of
contaminants, numerical assessment guidelines to
evaluate contaminant concentrations, and
standardized bioassays to evaluate the
bioaccumulation and toxicity potential of sediment
samples.
1.2.2 The Clean Water Act (CWA) is the single
most important law dealing with, environmental
quality of surface waters in the United States.
Section 101 of the CWA sets forth provisions
calling for the restoration an maintenance of the
chemical, physical, and biological integrity of the
Nation's waters. Federal and state monitoring
programs traditionally have focused on evaluating
water-column problems caused by point-source
dischargers. Findings in the National Sediment
Quality Survey, Volume I of the first biennial report
to Congress on sediment quality in the United
States, indicate that this focus .needs to be
expanded to include sediment quality impacts
(Section 1.1.2 and USEPA, 1997a).
-------
Table 1.1 Sediment Quality Assessment Procedures1
Type
Method
Numeric Descriptive Combination
Approach
Equilibrium Partitioning X
Tissue Residues X
Interstitial Water X
Toxicity
Benthic Community
Structure
Whole Sediment " X
Toxicity and Sediment
Spiking
X
X
X
Sediment Quality Triad X
Sediment Quality
Guidelines
X
X
A sediment quality value for a given
contaminant is determined by calculating the
sediment concentration of the contaminant that
corresponds to an interstitial water
concentration equivalent to the USEPA water
quality criterion for the contaminant.
Safe sediment concentrations of specific
chemicals are established by determining the
sediment chemical concentration that results in
acceptable tissue residues.
X Toxicity of interstitial water is quantified and
identification evaluation procedures are applied
to identify and quantify chemical components
responsible for sediment toxicity.
Environmental degradation is measured by
evaluating alterations in benthic community
structure.
X Test organisms are exposed to sediment that
may contain known or unknown quantities of
potentially toxic chemicals. At the end of a
specified time period, the response of the test
organisms is examined in relation to a specified
endpoint. Dose-response relationships can be
established by exposing test organisms to
sediments that have been spiked with known
amounts of chemicals or mixtures of chemicals.
X Sediment chemical contamination, sediment
toxicity, and benthic community structure are
measured on the same sediment sample.
Correspondence between sediment chemistry,
toxicity, and field effects is used to determine
sediment concentrations that discriminate
conditions of minimal, uncertain, and major
biological effects.
X The sediment concentration of contaminants
associated with toxic responses measured in
laboratory exposures or in field assessments
(i.e., Apparent Effect Threshold [AET], Effect
Range Median [ERM], Probable Effect Level
[PEL]). - -
1) Modified from USEPA (1992c).
4
-------
1.2.3 The Office of Water (OW), the Office of
Prevention, Pesticides, and Toxic Substances
(OPPTS), the Office of Solid Waste (OSW), and
the Office of Emergency and Remedial Response
(OERR) are all committed to the principle of
consistent tiered testing described in the
Contaminated Sediment Management Strategy
(USEPA, 1998). Consistent testing is desirable,
because the use of uniform testing procedures is
expected to increase data accuracy and precision,
facilitate test replication, and increase the
comparative value of test results. Each USEPA
program will, however, retain the flexibility of
deciding whether identified risks would trigger
actions.
1.2.4 Several programs use a tiered testing
approach. Tiered testing refers to a structured,
hierarchical procedure for determining data needs
relative to decision-making that consists of a
series of tiers, or levels, of investigative intensity.
Typically, increasing tiers in a tiered testing
framework involve increased information and
decreased uncertainty (USEPA, 1998). Each
USEPA program office intends to develop
guidance for interpreting the tests conducted within
the tiered framework and to explain how
information within each tier would trigger
regulatory action. Depending on statutory and
regulatory requirements, the program specific
guidance will describe decisions based on a
weight of evidence approach, a pass-fail
approach, or comparison to a reference site. The
following two approaches are currently being used
by USEPA: (1) OW-U.S. Army Corps of Engineers
(USAGE) dredged material testing framework and
(2) the OPPTS ecological risk assessment tiered
testing framework. USEPA-USACE (1991; 1998)
describes the dredged material testing framework,
and Smrchek and Zeeman (1998) summarizes the
OPPTS testing framework. A tiered testing
framework has not yet been chosen for Agency-
wide use, but some of the components have been
identified to be standardized. These components
include toxicity tests, bioaccumulation tests,
sediment quality guidelines, and other
measurements that may have ecological
significance, including benthic community structure
evaluation, colonization rate, and in situ sediment
testing within a mesocosm (USEPA, 1992a).
1.3 Scope and Application
1.3.1 Procedures are described for laboratory
testing of an estuarine amphipod to evaluate the
sublethal toxicity of contaminants in whole
sediments. Sediments can be collected from the
field or spiked with compounds in the laboratory.
The test species is L plumulosus, an Atlantic
coast estuarine species. The toxicity test is
conducted for 28 d in 1-L glass chambers
containing 175 ml_ of sediment and about 725 mL
of overlying seawater. Four hundred milliliters of
overlying water is renewed three times per week,
at which time test organisms are fed. Tests are
initiated with neonate amphipods that mature and
reproduce during the 28-d test period. The
endpoints in the 28-d toxicity test are survival,
growth rate, and reproduction of amphipods.
Survival is calculated as the percentage of newly
born (neonate) amphipods at test initiation that
survive as adults at test termination. Growth rate
is calculated as the mean dry weight gain per day
per adult amphipod surviving at test termination.
Reproduction is calculated as the number of
offspring per surviving adult. See Section 11.4 for
discussion on relative sensitivity of sublethal test
endpoints. This test is applicable for use with
sediment having pore water salinity ranging from
1%oto35%o.
1.3.2 This 28-d sediment toxicity test method
manual serves as a companion to the marine
acute sediment test methods manual (USEPA,
1994d) and the freshwater sediment test methods
manual (USEPA, 2000).
1.3.3 Procedures described in this manual are
based on method refinements described in DeWitt
et al. (1992a; 1997a), Emery et al. (1997), Emery
and Moore (1996) and USEPA (2000). This
USEPA/USACE manual outlines test methods for
evaluating the chronic toxicity of sediment with
L plumulosus. Although standard procedures are
described in the manual, further investigation of
certain issues could aid in the interpretation of test
results. Some of these issues include further
investigation to evaluate the relative toxicological
-------
Law2
Table 1.2 Statutory Use for Sediment Quality Assessment1
Area nf I lap
CERCLA
Assessment of need for remedial action with contaminated sediments; assessment of
degree of cleanup required, disposition of sediments
CWA
National Pollutant Discharge Elimination System (NPDES) permitting, especially under
Best Available Technology (BAT) in water-quality-limited water
Section 403(c) criteria for ocean discharges; mandatory additional requirements to
protect marine environment
Section 301 (h) waivers for publicly owned treatment works (POTWs) discharging to
marine waters
Section 404 permits for dredge and fill activities (administered by the U.S. Army Corps
of Engineers [USAGE])
FIFRA
MPRSA
Reviews of uses for new and existing chemicals
Pesticide labeling and registration
Permits for ocean dumping of dredged material
NEPA
Preparation of environmental impact statements for projects with surface water
discharges
TSCA
Section 5: Premanufacture notification reviews for new industrial chemicals
Sections 4, 6, and 8: Reviews for existing industrial chemicals
RCRA
Assessment of suitability (and permitting of) on-land disposal or beneficial use of
contaminated sediments considered "hazardous"
1 Modified from Dickson et al. (1987) and Southerland et al. (1992).
2 CERCLA Comprehensive Environmental Response, Compensation and Liability Act (Superfund)
CWA Clean Water Act
FIFRA Federal Insecticide, Fungicide, and Rodenticide Act
MPRSA Marine Protection, Resources and Sanctuary Act
NEPA National Environmental Policy Act
TSCA Toxic Substances Control Act
RCRA Resource Conservation and Recovery Act.
-------
sensitivity of the lethal and sublethal endpoints to
a wide variety of chemicals spiked in sediment and
to mixtures of chemicals in sediments from
pollution gradients in the field. Additional research
is needed to evaluate the ability of the test's lethal
and sublethal endpoints to estimate the responses
of populations and communities of benthic
invertebrates to contaminated sediments.
Research is also needed to link the toxicity test's
endpoints to a field-validated population model of
L plumulosus that would then generate estimates
of population-level responses of the amphipod to
test sediments and thereby provide additional
ecologically relevant interpretive guidance for the
toxicity test.
1.3.4 Additional sediment toxicity research and
methods development are now in progress to
(1) develop standard sediment bioaccumulation
tests (i.e., 28-d exposures with the bivalve
Macoma nasuta, and the polychaete Nereis
virens) (Lee et al., 1989), (2) refine sediment
spiking procedures, (3) refine sediment dilution
procedures, (4) refine sediment Toxicity
Identification Evaluation (TIE) procedures,
(5) produce additional data on confirmation of
responses in laboratory tests with natural
populations of benthic organisms (i.e., field
validation studies), (6) develop sediment toxicity
test methods for additional species (e.g.,
Neanthes), and (7) evaluate relative sensitivity of
endpoints measured in 10- and 28-d toxicity tests
using marine and estuarine amphipods. This
information will be described in future editions of
this manual or in other USEPA or USAGE
manuals.
1.3.5 Altering the procedures described in this
manual might affect contaminant bioavailability or
organism sensitivity and produce results that are
not directly comparable with results of accepted
procedures. Comparison of results obtained using
modified versions of these procedures might
provide useful information concerning new
concepts and procedures for conducting sediment
tests with aquatic organisms. If tests are
conducted with procedures different from those
described in this manual, additional tests are
required to determine comparability of results.
1.3.6 Where states have developed culturing and
testing methods for indigenous species other than
L. plumulosus, data comparing the sensitivity of
the substitute species and L plumulosus must be
obtained with sediments or reference toxicants to
ensure that the species selected are at least as
sensitive and appropriate as the recommended
species.
1.3.7 Selection of Test Organisms
1.3.7.1 The choice of a test organism has a major
influence on the relevance, success, and
interpretation of a test. Test organism selection
should be based on both environmental relevance
and practical concerns (DeWitt et al., 1989;
Swartz, 1989). Ideally, a test organism for use in
sediment tests should (1) have a toxicological
database demonstrating relative sensitivity to a
range of contaminants of interest in sediment;
(2) have a database for interlaboratory
comparisons of procedures (e.g., round-robin
studies); (3) be in direct contact with sediment;
(4) be readily available from culture, commercial
supplier, or through field collection; (5) be easily
maintained in the laboratory; (6) be easily
identified; (7) be ecologically or economically
important; (8) have a broad geographical
distribution, be indigenous (either present or
historical) to the site being evaluated, or have a
niche similar to organisms of concern (e.g., similar
feeding guild or behavior to the indigenous
organisms); (9) be tolerant of a broad range of
sediment physico-chemical characteristics (e.g.,
grain size); and (10) be compatible with selected
exposure methods and endpoints (ASTM, 2000a).
The method should also be (11) peer reviewed
(e.g., journal articles, American Society of Testing
and Materials [ASTM] guides) and (12) confirmed
with responses with natural populations of benthic
organisms.
1.3.7.2 The primary criterion used for selecting
L. plumulosus was that it met the above criteria.
Amphipods have been used extensively to test the
toxicity of marine, estuarine, and freshwater
sediments (Swartz et al., 1985; DeWitt et al, 1989;
Scott and Redmond, 1989; DeWitt et al., 1992a;
Schlekat et al., 1992; ASTM, 2000a).
L. plumulosus is an infaunal amphipod intimately
associated with sediment, due to its burrowing and
sediment ingesting nature. L. plumulosus is found
in both oligohaline and mesohaline regions of
estuaries on the East Coast of the United States
-------
and is tolerant to a wide range, of sediment grain
size distribution. This species is easily cultured in
the laboratory and has a relatively short generation
time (i.e., about 24d at 23°C, DeWitt et al. 1992a)
that makes this species adaptable to chronic
testing (see Section 10). Using a similar set of
criteria, L. plumulosus was selected as one of four
amphipod species recommended for short-term
toxicity testing of whole sediments (USEPA,
1994d).
1.3.7.3 An important consideration in the selection
of species for test method development is the
organism's sensitivity to single chemicals and to
complex mixtures. Studies (Schlekat 1995; DeWitt
et al., 1992a) have evaluated the sensitivities in
acute tests of amphipod species, including
L. plumulosus, either relative to one another, or to
other commonly tested estuarine or marine
species. For example, the sensitivity of marine
amphipods was compared with that of other
species that were used in generating saltwater
WQC. Seven amphipod genera, were among the
test species used to generate saltwater WQC for
12 chemicals. Acute amphipod toxicity data from
4-d water-only tests for each of the 12 chemicals
were compared with data for (1) all other species,
(2) other benthic species, and (3) other infaunal
species. Amphipods were generally of median
sensitivity for each comparison. The average
percentile rank of amphipods among all species
tested was 57.2%; among all benthic species,
55.5%; and, among all infaunal species, 54.3%.
Thus, amphipods are not uniquely sensitive
relative to all species, benthic species, or even
infaunal species (D. Hansen, USEPA,
Narragansett, Rl, personal communication).
Additional research may be warranted to develop
tests using species that are consistently more
sensitive than amphipods, thereby offering
protection to less sensitive groups.
1.3.7.3.1 Several studies of acute tests (10-d)
have compared the sensitivity of L plumulosus to
other commonly used amphipod species.
L plumulosus was as sensitive as the freshwater
amphipod Hyalella azteca to an artificially created
gradient of sediment contamination when the latter
was acclimated to oligohaline salinity (i.e., 6 %o)
(McGee et al., 1993). DeWitt et al. (1992b)
compared the sensitivities of L. plumulosus,
three other amphipod species, two molluscs,
and one polychaete to highly contaminated
sediment collected from Baltimore Harbor, MD,
and serially diluted with clean sediment.
L plumulosus was more sensitive than the
amphipods H. azteca and Lepidactylus dytiscus
and exhibited sensitivity equal to that of
Eohaustorius estuarius. A study using dilutions of
sediment collected from Black Rock Harbor (BRH),
CT, showed that Ampelisca abdita demonstrated
greater sensitivity than L plumulosus when the
latter was tested at 20°C (SAIC, 1993a).
However, the same study showed that
L. plumulosus was more sensitive at 25°C (the
test temperature for both the L. plumulosus 10-
and 28-d toxicity tests) than A. abdita at 20°C
(SAIC, 1993a). ;
1.3.7.3.2 The relative sensitivity and precision of
10-d acute toxicity tests with three marine and
estuarine amphipod species (A. abdita,
E. haustorius, and L. plumulosus) following
USEPA methods (USEPA, 1994d) were evaluated
in a round-robin test (Schlekat et al., 1995). All
three toxicity tests consistently characterized
moderate to highly contaminated sediments as
toxic relative to uncontaminated control
sediments. In addition, significant concordance
was exhibited by all toxicity tests in ranking the
toxicity of different sediments. Although there was
considerable interlaboratory variability
demonstrated in the round-robin, sensitivity of
these three toxicity tests was similar enough to
produce agreement in the categorization of
sediments as toxic or nontoxic.
1.3.7.3.3 Studies have been conducted to evaluate
the comparative sensitivity of the 28-d toxicity test
and the 10-d toxicity test with L. plumulosus
(DeWitt et al, 1992a; 1997b; McGee and Fisher
1999). DeWitt et al. (1992a; 1997b) found that in
general, the reproductive endpoint of the 28-d test
was more sensitive to chemical contaminants than
the survival and growth endpoints of either the
10-d or 28-d toxicity tests. Studies conducted by
the USAGE demonstrated similar sensitivity
amoung the lethal and sublethal endpoints of both
toxicity tests. In contrast, McGee and Fisher
(1999) found the sublethal endpoints less sensitive
than the survival endpoint. It is possible that the
different conclusions about the relative sensitivities
-------
of the 10- and 28-d L. plumulosus tests resulted
from either subtle differences in the testing
procedures used by DeWitt et al. (1992a; 1997b)
and McGee and Fisher (1999), or from response
of the amphipods to different chemical
contaminants in the test sediments used in the
three studies. In any case, the L. plumulosus 28-d
toxicity test provides valuable information on the
impact of contaminated sediments on both lethal
and sublethal endpoints, which the 10-d test does
not provide.
1.3.7.3.4 Limited comparative data are available
for concurrent water-only exposures of different
amphipod species in single-chemical tests.
Studies that have been conducted generally show
that no single amphipod species is consistently the
most sensitive. The relative sensitivity of four
amphipod species to ammonia was determined in
10-d water-only toxicity tests to aid interpretation
of results of tests on sediments in which this
toxicant is present (SAIC, 1993c). These tests
were static exposures that were generally
conducted under conditions (e.g., salinity,
photoperiod) similar to those used for standard
10-d sediment tests. Departures from standard
conditions included the absence of sediment and
a test temperature of 20°C for L plumulosus,
rather than 25°C as dictated in the acute method
(USEPA, 1994d). Sensitivity to total ammonia
increased with increasing pH for all four species.
The rank sensitivity was Rhepoxynius abronius
>A. abdita >E. estuarius >L. plumulosus. In
addition, cadmium chloride has been a common
reference toxicant for all four species in 4-d
exposures. DeWitt et al. (1992a) reports the rank
sensitivity to cadmium as R. abronius >A. abdita
>L. plumulosus >E. estuarius at a common
temperature of 15°C and salinity of 28%o. A series
of 4-d exposures to cadmium that were conducted
at species-specific temperatures and salinity
values showed the following rank sensitivity:
A. abdita >L. plumulosus > R. abronius >
E. estuarius (SAIC, 1993a; 1993b; 1993c).
1.3.7.3.5. Ammonia is a naturally occurring
compound in marine sediment that results from the
degradation of organic debris. Interstitial pore
water ammonia concentrations in test sediment
can range from <1 mg/L to in excess of 400 mg/L
(Word et al., 1997). Some benthic infauna show
toxicity to ammonia at concentrations of
approximately 20 mg/L (Kohn et al., 1994). Based
on water-only and spiked-sediment experiments
with ammonia, threshold limits for test initiation
and termination have been established for the
L. plumulosus chronic test. Smaller (younger)
individuals are more sensitive to ammonia than
larger (older) individuals (DeWitt et al., 1997).
Results of a 28-d test indicated that neonates can
tolerate very high levels of pore water ammonia
(>300 mg/L total ammonia) for short periods of
time with no apparent long-term effects (Moore et
al., 1997). It is not surprising the L. plumulosus
has a high tolerance for ammonia given that these
amphipods are often found in organic rich
sediments in which diagenesis can result in
elevated pore water ammonia concentrations.
Insensitivity to ammonia by L, plumulosus should
not be construed as an indicator of the sensitivity
of the L. plumulosus sediment toxicity test to other
chemicals of concern.
1.3.7.4 The sensitivity of an organism is related to
route of exposure and biochemical response to
contaminants. Sediment-dwelling organisms can
receive exposure from three primary sources:
interstitial water, sediment particles, and overlying
water. Food type, feeding rate, assimilation
efficiency, and clearance rate will control the dose
of contaminants from sediment. Benthic
invertebrates often selectively consume different
particle sizes (Harkey et al., 1994) or particles with
higher organic carbon concentrations, which may
have higher contaminant concentrations. Grazers
and other collector-gatherers that feed on
aufwuchs, or surface films, and detritus might
receive most of their body burden directly from
materials attached to sediment or from actual
sediment ingestion. In amphipods (Landrum,
1989) and clams (Boese et al., 1990), uptake
through the gut can exceed uptake across the gills
for certain hydrophobic compounds. Organisms in
direct contact with sediment can also accumulate
contaminants by direct adsorption to the body wall
or by absorption through the integument
(Knezovich et al., 1987). Particle type and organic
coating may affect uptake of contaminants, such
as metals (Schlekat, 1998).
1.3.7.5 Despite the potential complexities in
estimating the dose that an organism receives
-------
from sediment, the toxicity and bioaccumulation of
many contaminants in sediment such as Kepone®,
fluoranthene, organochlorines, and metals have
been correlated with either the concentration of
these chemicals in interstitial water, or in the case
of nonionic organic chemicals, in sediment on an
organic-carbon normalized basis (Di Toro et al.,
1990; 1991). The relative importance of whole
sediment and interstitial water routes of exposure
depends on the test organism and the specific
contaminant (Knezovich et al., 1987). Because
benthic communities contain a diversity of
organisms, many combinations of exposure routes
may be important. Therefore, behavior and
feeding habits of a test organism can influence its
ability to accumulate contaminants from sediment
and should be considered when selecting test
organisms for sediment testing.
1.3.7.6 Although no laboratory information was
available at the time of publication of this
document, a review of the distribution of
L. plumulosus in Chesapeake Bay indicates that its
distribution is negatively correlated with the degree
of sediment contamination (Pfitzenmeyer, 1975;
Reinharz, 1981). A field validation study of the 10-
d and 28-d L. plumulosus tests by McGee et
al.(1999) in Baltimore Harbor provides evidence
that L plumulosus mortality in 10-d toxicity tests is
negatively correlated with population density of
indigenous L. plumulosus. Protocol used by
McGee et al. (1999) for the 28-d L. plumulosus
test used a different diet than outlined in Section
11.3.6. Field validation studies with the revised
28-d L plumulosus sediment toxicity test have not
been conducted. The McGee and Fisher (1999)
study was a field validation of the 10-d and 28-d
tests; however, the feeding protocols have
changed slightly from what was used in that study.
1.4 Performance-based Criteria
1.4.1 USEPA's Environmental Monitoring
Management Council (EMMC) recommended the
use of performance-based methods in developing
chemical analytical standards (Williams, 1993).
Performance-based methods were defined by
EMMC as a monitoring approach that permits the
use of appropriate methods that meet pre-
established demonstrated performance standards
(Section 9.2). '
1.4.2 The key consideration for methods used
to obtain test organisms, whether they are field-
collected or obtained from culture, is to
procure healthy organisms of known quality. A
performance-based criteria approach, rather
than use of control-based criteria (See Section 3
for definitions), was selected as the preferred
method through which individual laboratories
should evaluate culture methods or the quality of
field-collected organisms. This method was
chosen to allow each laboratory to optimize culture
methods, determine the quality of field-collected
organisms, and minimize effects of test organism
health on the reliability and comparability of test
results. Performance criteria used to assess the
quality of cultured and field-collected amphipods
and to determine the acceptability of 28-d
sediment toxicity tests are listed in Table 11.3.
10
-------
Section 2
Summary of Method
2.1 Method Description and
Experimental Design
2.1.1 Method Description
This manual describes a laboratory method
for determining the sublethal toxicity
of contaminated sediment using an
estuarine crustacean, the amphipod
Leptocheirus plumulosus. Test sediments may be
collected from estuarine or marine environments,
or spiked with compounds in the laboratory. The
toxicity test is conducted for 28 d in 1-L chambers
containing 175 mL of sediment and about 725 mL
of overlying water. Tests are initiated with neonate
amphipods that mature and reproduce during the
28-d test period. Four hundred milliliters of
overlying water is renewed three times per week,
and test organisms are fed after each water
renewal. The endpoints are survival, growth rate,
and reproduction of test organisms. The use or
choice of additional control and reference
sediments depends on the nature of the test
sediments or the application. This test is
applicable for use with sediment from oligohaline
to fully marine environments (1 %o to 35 %o).
2.1.2 Experimental Design
The following section is a general summary of
experimental design. See Sections 11 and 12 for
additional details on actual procedures and data
analysis.
2.1.2.1 Control and Reference Sediment
2.1.2.1.1 Sediment tests include a control
sediment (sometimes called a negative control).
A control sediment is one that is essentially free of
contaminants and is used routinely to assess the
acceptability of a test; it is not necessarily
collected near the site of concern. Any
contaminants in control sediment are thought to
originate from the global spread of pollutants and
do not reflect any substantial input from local or
nonpoint sources (ASTM, 2000d). A control
sediment provides a measure of test acceptability,
evidence of test organism health, and a basis for
interpreting data obtained from the test sediments.
A reference sediment is typically collected near an
area of concern (e.g., a disposal site) and is used
to assess sediment conditions exclusive of
material(s) of interest. Testing a reference
sediment provides a site-specific basis for
evaluating toxicity.
2.1.2.1.1.1 In general, the performance of
organisms in the negative control(s) is used to
judge the acceptability of a test. Either the
negative control or reference sediment may be
used to evaluate performance in the experimental
treatments, depending on the purpose of the
study. Any study in which organisms in the
negative control do not meet performance criteria
must be considered questionable, because it
suggests that unknown adverse factors affected
the test organisms. Key to avoiding this situation
is using only control sediments that have a
demonstrated record of performance using the
same test procedure. This includes testing of new
sediment collections from sources that have
previously provided suitable control sediment.
2.1.2.1.1.2 Because of the uncertainties
introduced by poor performance in the negative
control, such studies should be repeated to insure
accurate results. However, the scope or sampling
associated with some studies may make it difficult
or impossible to repeat a study. Some
researchers have reported cases where
performance in the negative control is poor, but
control performance criteria are met in a reference
sediment included in the study design. In these
cases, it may be possible to infer that other
11
-------
samples that show good performance are probably
not toxic; however, any samples showing poor
performance should not be judged to have shown
toxicity, because it is unknown whether the
adverse factors that caused poor control
performance might have also caused poor
performance in the test treatments.
2.1.2.1.2 Natural geomorphological and physico-
chemical characteristics, such as sediment
texture, may influence the response of test
organisms (DeWitt et al., 1988). The physico-
chemical characteristics of test sediment should
be within the tolerance limits of the test organism.
Ideally, the limits of a test organism should be
determined in advance; however, controls for
factors such as grain size and organic carbon can
be evaluated if the recommended limits are
approached or exceeded in a test sediment.
See Section 10 and Table 11.1 for tolerance
limits of L. plumulosus for physico-chemical
characteristics. If the physico-chemical
characteristic(s) of a test sediment approach
or exceed the tolerance limits of the test organism,
it may be desirable to include an additional control
sedimentthat encompasses those characteristics.
The effects of some sediment characteristics (e.g.,
grain size or total organic carbon [TOG]) on test
results may be addressed with regression
equations (DeWitt et al., 1988; Ankley et al.,
1994a). L. plumulosus is relatively insensitive to a
wide-range of grain sizes in test sediments (95%
sand to 35% clay) (DeWitt et al., 1997a; Emery et
al., 1997).
2.1.2.2 The experimental design depends on the
purpose of the study. Variables that need to be
considered include the number and type of control
sediment(s), the source of reference sediment, the
number of treatments and replicates, and water
quality characteristics.
2.1.2.2.1 The purpose of the study might be to
determine a specific endpoint, such as
reproduction, and may include a negative control
sediment, a positive control or reference toxicant,
a solvent control, and several concentrations of
sediment spiked with a chemical.
2.1.2.2.2 The purpose of the study might be to
determine whether field-collected sediments are
toxic, and may include controls, reference
sediments, and test sediments. Controls are used
to evaluate the acceptability of the test
(Table 11.3) and a test might include one or more
control sediments. Testing a reference sediment
provides a site-specific basis for evaluating toxicity
of the test sediments. Comparisons of test
sediments to multiple reference or control
sediments representative of the physical
characteristics of the test sediment (i.e., grain size,
organic carbon) may be useful in these
evaluations. A summary of field sampling
design is presented by Green (1979). See Section
12 for additional guidance on experimental design
and statistics.
2.1.2.3 If the purpose of the study is to conduct a
reconnaissance field survey to identify
contaminated sites for further investigation, the
experimental design might include only one
sample from each site to allow for maximum
spatial coverage. The lack of replication at a site
usually precludes statistical comparisons (e.g.,
analysis of variance [ANOVA]) among sites), but
these surveys can be- used to identify
contaminated sites for further study or may be
evaluated using regression techniques (Sokal and
Rohlf, 1981; Steel and Torrie, 1980).
2.1.2.4 In other instances, the purpose of the study
might be to conduct a quantitative sediment survey
of chemical contaminants and toxicity to determine
statistically significant differences between effects
among control and test sediments from several
sites. The number of replicates per site should be
based on the need for sensitivity or power (Section
12). In a quantitative survey, replicates (separate
samples from different grabs collected at the same
site) would need to be taken at each site.
Chemical and physical characterizations of each of
these grabs would be required for each of these
replicates used in sediment testing. Separate
subsamples might be used to determine within-
sample variability or to compare test procedures
(e.g., comparative sensitivity among test
organisms), but these subsamples cannot be
considered to be true field replicates for statistical
comparisons among sites (ASTM, 2000b).
2.1.2.5 Sediments often exhibit high spatial and
temporal variability (Stemmer et al., 1990a).
Therefore, replicate samples may need to be
collected to determine variance in sediment
12
-------
characteristics. Sediments should be collected
with as little disruption as possible; however,
subsampling, compositing, or homogenization of
sediment samples may be necessary for some
experimental designs.
2.1.2.6 Site locations might be distributed along a
known pollution gradient, in relation to the
boundary of a disposal site, or at sites identified as
being contaminated in a reconnaissance survey.
Both spatial and temporal comparisons can be
made. In predredging studies, a sampling design
can be prepared to assess the contamination of
samples representative of the project area to be
dredged. Such a design should include
subsampling of cores taken to the project depth.
2.1.2.7 The primary focus of the physical and
experimental test design, and statistical analysis of
the data, is the experimental unit. The
experimental unit is defined as the smallest
physical entity to which treatments can be
independently assigned (Steel and Torrie, 1980)
and to which air and water exchange between test
chambers is kept to a minimum. As the number of
test chambers per treatment increases, the
number of degrees of freedom and the power of a
significance test increase, and therefore, the width
of the confidence interval on a point estimate, such
as an LC50, decreases (Section 12). Because of
factors that might affect test results, all test
chambers should be treated as similarly as
possible. Treatments should be randomly
assigned to individual test chamber locations.
Assignment of test organisms to test chambers
should be nonbiased.
2.2 Types of Tests
2.2.1 A 28-d toxicity method is outlined for the
estuarine amphipod L plumulosus. The manual
describes procedures for testing sediments from
oligohaline to fully marine environments.
2.3 Test Endpoints
2.3.1 In toxicity tests, the method chosen to
evaluate an endpoint has the potential to affect
that bioassay's quality and cost. For example, an
endpoint measure exhibiting high variance will
decrease test power and increase the likelihood of
false negative results (Fairweather, 1991).
Typically, endpoint selection for new bioassays is
generally guided by methodologies for related
bioassays (Gray et al., 1998). Sediment
bioassays using macroinvertebrates often
incorporate standard survival and growth
endpoints (Ingersoll, 1995). Gray et al. (1998)
recommend optimal endpoint measures for the
L. plumulosus bioassay based on four criteria:
relevance of each measure to its respective
endpoint; signal-to-noise ratio (the ratio between
the response to stressor and the normal variation
in the response variable); redundancy to other
measures of the same endpoint; and cost of labor,
training, and equipment. Signal-to-noise ratios are
independent of experiment design considerations
(i.e., Type I and Type II errors, and sample size)
and are positively correlated with power (Gray et
al., 1998).
The recommended endpoint measures for this
species in 28-d tests are survival, calculated as
the percentage of neonates at test initiation that
survive as adults at test termination; growth rate,
calculated as the mean dry weight gain per day
per adult amphipod surviving at test termination;
and reproduction, calculated as the number of
offspring per surviving adult. Behavior of test
organisms (e.g., avoidance of sediment) should be
qualitatively observed three times per week during
the test, before water renewals.
13 '
-------
Section 3
Definitions
3.1 Terms
The following terms were defined in Lee (1980),
National Research Council (NRC, 1989), USEPA
(1989b), USEPA-USACE(1991), USEPA-USACE
(1998), Leeetal. (1994), orASTM (2000b; 2000c;
2000i).
3.1.1 Technical Terms
3.1.1.1 Clean. Denotes a sediment or water that
does not contain concentrations of test materials
which cause apparent stress to the test organisms
or reduce their survival.
3.1.1.2 Concentration. The ratio of weight or
volume of test material(s) to the weight or volume
of sediment or water.
3.1.1.3 Contaminated sediment. Sediment
containing chemical substances at concentrations
that pose a known or suspected threat to
environmental or human health.
3.1.1.4 Control sediment Sediment that is
essentially free of contaminants and is used
routinely to assess the acceptability of a test. Any
contaminants in control sediment may originate
from the global spread of pollutants and do not
reflect any substantial input from local or nonpoint
sources. Comparing test sediments to control
sediment(s) is a measure of the toxicity of test
sediments beyond inevitable background
contamination. Control sediment is also called a
negative control because no toxic effects are
anticipated in this treatment.
3.1.1.5 Effect concentration (EC). The toxicant
concentration that would cause an effect in a given
percentage of the test population. Identical to LC
when the observable adverse effect is death. For
example, the EC50 is the concentration of toxicant
that would cause a specified effect in 50% of the
test population.
3.1.1.6 Inhibition concentration (1C). The
toxicant concentration that would cause a given
percent reduction in a nonquantal measurement
for the test population. For example, the IC25 is
the concentration of toxicant that would cause a
25% reduction in growth for the test population,
and the IC50 is the concentration of toxicant that
would cause a 50% reduction.
3.1.1.7 Interstitial water or pore water. Water
occupying space between sediment or soil
particles.
3.1.1.8 Lethal concentration (LC). The toxicant
concentration that would cause death in a given
percentage of the test population. Identical to EC
when the observable adverse effect is death. For
example, the LC50 is the concentration of toxicant
that would cause death in 50% of the test
population.
3.1.1.19 Lowest observed effect concentration
(LOEC). The lowest concentration of a toxicant to
which organisms are exposed in a test that causes
an adverse effect on the test organisms (i.e.,
where a significant difference exists between the
value for the observed respbnse and that for the
controls).
3.1.1.10 No observed effect concentration
(NOEC). The highest concentration of a toxicant
to which organisms are exposed in a test that
causes no observable adverse effect on the test
organisms (i.e., the highest concentration of a
toxicant in which the value for the observed
response is not statistically significantly different
from the controls).
3.1.1.11 Overlying water. The water placed over
sediment in a test chamber during a test.
3.1.1.12 Reference sediment. A whole sediment
near an area of concern used to assess sediment
conditions exclusive of material(s) of interest. The
14
-------
reference sediment may be used as an indicator of
localized sediment conditions exclusive of the
specific pollutant input of concern. Such sediment
would be collected near the site of concern and
would represent the background conditions
resulting from any localized pollutant inputs as well
as global pollutant input. This is the manner in
which reference sediment is used in dredged
material evaluations.
3.1.1.13 Reference-toxicity test. A test
conducted with reagent-grade reference chemical
to assess the sensitivity of the test organisms
response to a toxicant challenge. Deviations
outside an established normal range may indicate
a change in the sensitivity of the test organism
population. Reference-toxicity tests are most often
performed in the absence of sediment.
3.1.1.14 Sediment. Particulate material that
usually lies below water. Formulated particulate
material that is intended to lie below water in a
test.
3.1.1.15 Spiked sediment. A sediment to which
a material has been added for experimental
purposes.
3.1.1.16 Whole sediment. Sediment and
associated pore water that have had minimal
manipulation. The term bulk sediment has been
used synonymously with whole sediment.
3.1.2 Grammatical Terms
The words "must," "should," "may," "can," and
"might" have very specific meanings in this
manual.
3.1.2.1 "Must" is used to express an absolute
requirement, that is, to state that a test ought to be
designed specifically to satisfy the specified
conditions, unless the purpose of the test requires
a different design. "Must" is only used in
connection with the factors that directly relate to
the acceptability of a test. .
3.1.2.2 "Should" is used to state that the specified
condition is recommended and ought to be met if
possible. Although a violation of one "should" is
rarely a serious matter, violation of several will
often render the results questionable.
3.1.2.3 Terms such as "is desirable," "is often
desirable," and "might be desirable" are used in
connection with less important factors.
3.1.2.4 "May" is used to mean "is (are) allowed
to," "can" is used to mean "is (are) able to," and
"might" is used to mean "could possibly." Thus,
the classic distinction between "may" and "can" is
preserved, and "might" is never used as a
synonym for either "may" or "can."
15
-------
Section 4
Interferences
4.1 General Introduction
4.1.1 Interferences are characteristics of a
sediment or sediment test system, aside from
those related to sediment associated chemicals of
concern, that can potentially affect test organism
survival, growth, or reproduction (Environment
Canada, 1994; ASTM, 2000c; USEPA, 2001).
These interferences can potentially confound test
interpretation in two ways: (1) false-positive
response (i.e., toxicity is observed in the test when
contamination is not present at concentrations
known to elicit a response, or there is more toxicity
than expected); and (2) false-negative response
(i.e., no toxicity is observed when contaminants
are present at concentrations known to elicit a
response, or there is less toxicity than expected).
4.1.2 There are three categories of interfering
factors that can cause false-negative or false-
positive responses: (1) those physical or chemical
characteristics of sediments affecting survival,
growth, or reproduction, independent of chemical
concentration (e.g., sediment grain size), (2)
changes in chemical bioavailability as a function of
sediment manipulation or storage, and (3) the
presence of indigenous organisms. Although test
procedures and test organism selection criteria
were developed to minimize these interferences,
this section describes the nature of these
interferences. Procedures for minimizing the
effects of interfering factors are presented in
Section 11.
4.1.3 Because of the heterogeneity of natural
sediments, extrapolation from laboratory studies to
the field can sometimes be difficult (Table 4.1;
Burton, 1991). Sediment collection, handling, and
storage procedures may alter contaminant
bioavailability and concentration by changing the
physical, chemical, or biological characteristics of
the sediment. Maintaining the integrity of a field-
collected sediment during removal, transport,
storage, mixing, and testing is extremely difficult
and may complicate the interpretation of effects
(Environment Canada, 1994; USEPA, 2001).
Direct comparisons of organisms exposed
Table 4.1 Advantages and Disadvantages of Use of
Sediment Tests1
Advantages
Sediment tests measure bioavailable fraction of contaminant(s).
Sediment tests provide a direct measure of benthic effects,
assuming no field adaptation or amelioration of effects.
Limited special equipment is required for testing.
Ten-day toxicity test methods are rapid and inexpensive.
Legal and scientific precedence exists for use; ASTM standard
guides are available.
• Sediment tests measure unique information relative to chemical
analyses or benthic community analyses.
• Tests with spiked chemicals provide data on cause-effect
relationships.
• Sediment toxicity tests can be applied to all chemicals of
concern.
• Tests applied to field samples reflect cumulative effects of
contaminants and contaminant interactions.
• Toxicity tests are amenable to confirmation with natural benthos
populations.
Disadvantages
• Sediment collection, handling, and storage may alter
bioavailability.
• Spiked sediment may not be representative of field
contaminated sediment.
• Natural geochemical characteristics of sediment may affect the
response of test organisms.
• Indigenous animals may be present in field-collected sediments.
* Route of exposure may be uncertain and data generated in
sediment toxicity tests may be difficult to interpret if factors
controlling the bioavailability of contaminants in sediment are
unknown.
• Tests applied to field samples may not discriminate effects of
individual chemicals.
• Few comparisons have been made of methods or species.
• Only a few chronic methods for measuring sublethal effects
have been developed or extensively evaluated.
• Laboratory tests have inherent limitations in predicting
ecological effects.
'Modified from Swartz (1989).
16
-------
in the laboratory and in the field would be useful to
verify laboratory results. However, spiked
sediment may not be representative of
contaminated sediment in the field. Mixing time
(Stemmer et al., 1990a), aging (Word et al., 1987;
Landrum, 1989; Landrum and Faust, 1992) and
the chemical form of the material can affect
responses of test organisms in spiked sediment
tests. Detailed recommendations for sample
collection and handling are provided in Section 8
and USEPA (2001).
4.1.4 Laboratory testing with field-collected
sediments may be useful in estimating cumulative
effects and interactions of multiple contaminants in
a sample. Tests with field samples usually cannot
discriminate between effects of individual
chemicals. Most sediment samples contain a
complex matrix of inorganic and organic
contaminants with many unidentified compounds.
The use of TIE in conjunction with sediment tests
with spiked chemicals may provide evidence of
causal relationships and can be applied to many
chemicals of concern (Ankley and Thomas, 1992;
Adams et al., 1985; USEPA, 1996): Sediment
spiking can also be used to investigate additive,
antagonistic, or synergistic effects of specific
contaminant mixtures in a sediment sample
(Swartz et al., 1988).
4.1.5 Most assessments of contaminated sediment
rely on short-term lethality testing methods (e.g.,
<10d; USEPA-USACE, 1977; 1991). Short-term
lethality tests are useful in identifying "hot spots" of
sediment contamination, but may not be sensitive
enough to evaluate moderately contaminated
areas. Sediment quality assessments using
sublethal responses of benthic organisms, such as
effects on growth and reproduction, have been
used to successfully evaluate moderately
contaminated areas (Ingersoll etal., 1998; McGee
and Fisher, 1999; Scott et al., 1996).
4.1.6 Despite the interferences discussed in this
section, existing sediment testing methods can be
used to provide a rapid and direct measure of
effects of contaminants on benthic communities
(e.g., Canfield et al., 1996; Niewolny et al. 1997;
DeWitt et al. 1997c). Laboratory tests with field-
collected sediment can also be used to determine
temporal, horizontal, or vertical distribution of
contaminants in sediment. Most tests can be
completed within 2 to 4 weeks. Legal and
scientific precedents exist for use of toxicity and
bioaccumulation tests in regulatory decision-
making (e.g., USEPA, l990c). Furthermore,
sediment tests with complex contaminant mixtures
are important tools for making decisions about the
extent of remedial action for contaminated aquatic
sites and for evaluating the success of remediation
activities.
4.2 Noncontaminant Factors
4.2.1 Noncontaminant characteristics of sediment
are defined as chemical or physical characteristics
that can cause reduced test organism survival,
growth, or reproduction. These interferences
include, but are not limited to, sediment grain size,
interstitial pore water salinity, TOG, dissolved
sulfides, and interstitial pore water ammonia.
L plumulosus is considered to be remarkably
tolerant of these noncontaminant factors; however,
the physico-chemical properties of each test
sediment must be within .the acceptable tolerance
limits to ensure that a toxicological response is
caused by contaminants. Tolerance limits of
L. plumulosus for the factors listed above are
summarized in Table 11.1 and defined below.
4.2.1 Sediment Grain Size
4.2.1.1 L. plumulosus are found in very fine muds
and muddy sands and are tolerant of variable grain
size. Laboratory studies have shown significant
reduction in survival when clay content exceeded
84% (Emery et al., 1997). Emery et al. (1997)
found an increase in growth as sediment
coarseness increased up to 75% sand. However,
DeWitt et al. (1997a) reported enhanced growth in
finer-grained sediment as compared with more
coarse-grained material, but the difference in
growth was not considered to be biologically
important (DeWitt et al., 1997a). L. plumulosus
survival, growth and reproduction were
significantly reduced when exposed to pure sand
(Emery et al., 1997). Therefore, L plumulosus
should be tested with sediment with > 5% silt-clay
(i.e., s95% sand), but <85% clay (Table 11.1). If
sediment characteristics exceed these bounds, an
17
-------
appropriate clean control/reference sediment
should be incorporated into the test to separate
effects of sediment-associated contaminants from
effects of particle size.
4.2.2 Interstitial Pore Water Salinity
4.2.2.1 L plumulosus is an estuarine species
tolerant of a wide range of salinity. No adverse
effects have been observed in laboratory
exposures to pore water salinity values ranging
from 0%o to 35%o, with overlying water salinity at
20 %o (DeWitt et al., 1997a). Furthermore,
laboratory cultures have been successfully
maintained at 5%o (Emery and Moore, 1996) and
20%o (DeWitt et al., 1997a) which demonstrates
that reproduction had occurred in the cultures at a
variety of salinity values. Although not currently
recommended for testing in truly freshwater
sediment, L plumulosus can be used to test
sediment having pore water salinity >1%o (DeWitt
et al., 1997a). Further research is required to
determine whether L. plumulosus can be used to
test sediment having pore water salinity >35%o.
4.2.3 Total Organic Carbon
4.2.3.1 Test sediment TOG content can vary
greatly, ranging from near 0% to >10%. The
amount of TOG can affect test organism survival,
growth, and reproduction. Limited evidence
suggests that the L. plumulosus chronic test is
tolerant to most TOG concentrations; however,
Scott et al. (1996) reported that growth and
reproduction may be lower in uncontaminated field
sediments having <2% TOG concentrations. An
analysis of organism response over a wide range
of sediment TOG was completed by DeWitt et al.
(1997b) using reference sediment data from two
studies. No effect on survival, growth, or
reproduction was detected for sediments with TOG
concentrations ranging from 1% to 7% TOG.
There was some evidence of significantly
decreased survival, growth, and reproduction in
<1% TOG sediments. No data were available for
test sediments with TOG >7%. Therefore until
additional data are generated, if test sediment
TOG concentrations are <1% or >7%, a TOG
control or reference sediment with similar TOG
should be tested concurrently.
4.2.4 Dissolved Sulfides
4.2.4.1 Hydrogen sulfide occurs naturally in
anoxic marine sediments. Sims and Moore (1995)
conducted an extensive review of the literature
that focused on the effects of hydrogen sulfide on
benthic organisms. Sims and Moore (1995)
reported that tube-building amphipods circulate
oxygenated water through their burrows, thus
reducing or eliminating exposure to pore water
hydrogen sulfide. In acute experiments, however,
dissolved sulfides have been shown to be toxic to
marine amphipods R. abronius and E. estuarius
(48-h LOECs of 1.47 and 1.92 mg/L total sulfide
respectively; Knezovich et al., 1992). Currently,
no data exist regarding the sensitivity of
L plumulosus to hydrogen sulfide in 28-d
exposures. Additional information on the tolerance
of aquatic organisms to sulfides can be found in
Bagarinao(1992).
4.2.5 Interstitial Pore Water Ammonia
4.2.5.1 Ammonia is present in sediment as a
result of several independent microbial processes
as well as anthropogenic sources, and ammonia
concentrations may be enhanced in areas that
exhibit organic enrichment. Ammonia
concentrations are sometimes high in
contaminated sediments. Interstitial pore water
ammonia concentrations in test sediment can
range from <1 mg/L to in excess of 400 mg/L
(Word et al., 1997). Some benthic infauna show
toxicity to ammonia at concentrations of
approximately 20 mg/L (Kohn et al., 1994). Based
on water-only and spiked-sediment experiments
with ammonia, threshold limits for test initiation
and termination have been established for the L.
plumulosus chronic test. Smaller (younger)
individuals are more sensitive to ammonia than
larger (older) individuals (DeWitt et al., 1997a).
Results of a 28-d test indicated that neonates can
tolerate very high levels of pore water ammonia
(>300 mg/L total ammonia) for short periods of
time with no apparent long-term effects (Moore et
al., 1997). At test initiation, pore water should not
exceed 60 mg/L total ammonia (Table 11.1; DeWitt
et al., 1997a; USEPA, 1994d). One study
indicated that pore water ammonia levels
>16 mg/L measured at test termination can be
18
-------
associated with lethal and sublethal impacts to
L plumulosus (DeWitt et al. 1997a). Thus, if pore
water ammonia concentrations exceed 16 mg/L at
test termination, toxicity test results could be
affected by ammonia.
4.2.7 If a particular sediment characteristic
exceeds the tolerance of L. plumulosus, several
measures can be taken. Suggested procedures to
account for or reduce the effects of
noncontaminant interferences are presented in
Section 11.4.
4.3 Changes in Contaminant
Bioavailability
4.3.1 Sediment toxicity tests are meant to serve as
an indicator of contaminant-related toxicity that
might be expected under field or natural
conditions. Some studies have indicated
differences between results of laboratory testing
and results of field testing of sediments using in
situ exposures (Sasson-Brickson and Burton,
1991).
4.3.2 Sediment collection, handling, and storage
may alter contaminant bioavailability and
concentration by changing the physical, chemical,
or biological characteristics of the sediment.
Manipulations such as mixing, homogenization,
and sieving are generally thought to increase
availability of organic compounds because of
disruption of the equilibrium with organic carbon in
the pore water/particle system. Similarly, oxidation
of anaerobic sediment increases the availability of
certain metals (Di Toro et al., 1990). Because the
availability of contaminants can be a function of
the degree of manipulation, this manual
recommends that handling, storage, and
preparation of the sediment for testing be as
consistent as possible. Maintaining the integrity of
a field-collected sediment during removal,
transport, storage, mixing, and testing is extremely
difficult. Direct comparison of organisms exposed
in the laboratory and in the field would be useful to
verify laboratory results. Detailed
recommendations for sample collection and
handling are provided in Section 8 and USEPA
(2001).
4.3.2.1 Sediment Sampling. Sediment collection
techniques include moderately disruptive
(sediment coring and grab sampling) to highly
disruptive (dredging) methods. It is impossible to
collect sediment samples and remove them from
samplers without altering conditions to some
degree that control contaminant availability (e.g.,
redox potential, anaerobic environment, spatial
distributions, biological activity). Oxidation,
compaction, volatilization, homogenization, and
exposure to light can all occur and affect
contaminant distribution, speciation, partitioning,
and ultimately bioavailability. It is important to
select sampling techniques that not only achieve
study goals, but also minimize sediment
disturbance.
4.3.2.2 Sediment Storage. Sediment storage
conditions can also affect contaminant availability
and speciation. Type of storage container, storage
time, temperature, exposure to air, and drying
need to be controlled to maintain sample integrity
(USEPA, 2001). It is generally recommended that
sediment should be stored at 4°C, in the dark, in
sealed containers with minimal headspace.
4.3.2.3 Sieving and Homogenization. Test
sediments should be sieved only when there is
compelling concern that indigenous predator or
amphipods from the test site could accidentally be
introduced into the test chamber. However,
because sieving of test sediments disrupts the
physical properties of the sediment, and may also
affect chemical properties of the sediment, sieving
should be avoided whenever possible. Press-
sieving is preferable to wet-sieving because the
use of water in the latter processing will dilute the
pore water (and its chem.ical constituents) of the
test sediment. To press sieve use a clean inert
surface such as teflon, to help push sediment
through either a nytex or stainless steel sieve
(depending on project requirements). When
sediments are sieved, it may be desirable to take
samples before and after sieving to compare the
concentration of contaminants (especially in the
pore water), total organic carbon, dissolved
organic carbon (in pore water), acid volatile
sulfides (AVS), and sediment grain-size
distribution. USEPA does not recommend
unnecessary sieving of test sediments on a routine
basis (see USEPA 1997d, 2000; ASTM 20QOc).
4.3.2.4 Testing Conditions. Conducting
sediment toxicity tests at temperatures different
from those at the collection site might affect
19
-------
contaminant solubility, partitioning coefficients, and
other physical and chemical characteristics.
Interaction between sediment and overlying water
and the ratio of sediment to overlying water can
influence bioavailability (Stemmer et al., 1990b).
Salinity of the overlying water is another factor that
can affect the bioavailability of contaminant,
particularly metals. Some metals (e.g., cadmium)
are more bioavailable at lower salinity values.
Therefore, if a sediment sample from a low salinity
location is tested with overlying waters of high
salinity, there is the potential that metal toxicity
may be reduced. The broad tolerance of L
plumulosus allows tests to be conducted over the
range of pore water salinity values routinely
encountered in field-collected sediments from
North American estuarine and marine
environments. For standardization purposes,
testing should be conducted with overlying water
at either 5%o or 20%o. Sediment samples with pore
water salinity values £lO%o should be tested with
overlying water at 5%o, and test sediment with pore
water salinity values >10%o should be tested with
overlying water at 20%o (DeWitt et al., 1997a).
Photoinduced toxicity caused by ultraviolet (UV)
light may be important for some compounds
associated with sediment (e.g..polycyclic aromatic
hydrocarbons [PAHs]); Davenport and Spacie,
1991; Ankleyetal., 1994b). However, fluorescent
lighting typically used to conduct laboratory tests
does not include the appropriate spectrum of UV
radiation to photoactivate compounds (Oris and
Giesy, 1985). Therefore, laboratory tests might
not account for toxicity expressed by this mode of
action.
4,3.2.5 Additions to Test Chambers. The
addition of food, water, or solvents to the test
chambers might obscure the bioavailability of
contaminants in sediment or might provide a
substrate for bacterial or fungal growth (Harkey et
al., 1997). Without addition of food, the test
organisms may starve during exposures (Ankley et
al., 1994a; DeWitt et al., 1997a). However, the
addition of food may alter the availability of the
contaminants in the sediment (Harkey et al., 1994;
Bridges et al., 1997) depending on the amount of
food added, its composition (e.g., TOG), and the
chemical(s) of interest.
4.3.2.6 Contaminant Uptake. Depletion of
aqueous and sediment-sorbed contaminants
resulting from uptake by an organism or
absorption to test chamber can also influence
contaminant availability. In most cases, the
organism is a minor sink for contaminants relative
to the sediment. However, within the burrow of an
organism, sediment desorption kinetics might limit
uptake rates. Within minutes to hours, a major
portion of the total chemical can be inaccessible to
the organisms because of depletion of available
residues. The desorption of a particular
compound from sediment can range from easily
reversible (labile; within minutes) to irreversible
(nonlabile; within days or months; Karickhoff and
Morris, 1985). Interparticle diffusion or advection
and the quality and quantity of sediment organic
carbon can also affect sorption kinetics.
4.3.3 The route of exposure may be uncertain and
data from sediment tests may be difficult to
interpret if factors controlling the bioavailability of
contaminants in sediment are unknown. Bulk-
sediment chemical concentrations may be
normalized to factors other than dry weight. For
example, concentrations of nonionic organic
compounds might be normalized to sediment
organic-carbon content (USEPA, 1992c) and
certain metals normalized to acid volatile sulfides
(Di Toro et al., 1990). Even with the appropriate
normalizing factors, determination of toxic effects
from ingestion of. sediment or from dissolved
chemicals in the interstitial water can still be
difficult (Lamberson and Swartz, 1988).
4.4 Presence of Indigenous Organisms
4.4.1 Indigenous organisms may be present in
field-collected sediments. An abundance of the
same organism or organisms taxonomically similar
to the test organisms in the sediment sample may
make interpretation of treatment effects difficult.
Competing or predatory organisms can adversely
affect L. plumulosus survival, growth, or
reproduction.
4.4.2 If compelling evidence exists that indigenous
organisms may be introduced" into the test
chamber with the test sediments, the test
sediments can be sieved (see Section 4.3.2.3).
20
-------
However, USEPA does not recommend basis. Alternatively, short-term storage of test
unnecessary sieving of test sediments on a routine sediments may eliminate indigenous organisms in
the test sediments (see Section 8.2.2).
21
-------
Section 5
Health, Safety, and Waste Management
5.1 General Precautions
5.1.1 Development and maintenance of an
effective health and safety program in the
laboratory requires an ongoing commitment by
laboratory management and includes (1) the
appointment of a laboratory health and safety
officer with the responsibility and authority to
develop and maintain a safety program; (2) the
preparation of a formal, written health and safety
plan, which is provided to each laboratory staff
member; (3) an ongoing training program on
laboratory safety; and (4) regular safety
inspections.
5.1.2 This manual addresses procedures that may
involve hazardous materials, operations, and
equipment, but it does not purport to address all of
the safety problems associated with their use. It is
the responsibility of the user to establish
appropriate safety and health practices and
determine the applicability of regulatory limitations
before use. While some safety considerations are
included in the manual, it is beyond the scope of
this manual to encompass all safety requirements
necessary to conduct sediment tests.
5.1.3 Collection and use of sediment may involve
substantial risks to personal safety and health.
Contaminants in field-collected sediment may
include carcinogens, mutagens, and other
potentially toxic compounds. Inasmuch as
sediment testing is often begun before chemical
analyses can be completed, worker contact with
sediment needs to be minimized by (1) using
gloves, laboratory coats, safety glasses, face
shields, and respirators as appropriate;
(2) manipulating sediment under a ventilated hood,
in an enclosed glove box; and (3) enclosing and
ventilating the exposure system. Personnel
collecting sediment samples and conducting tests
should take all safety precautions necessary for
the prevention of bodily injury and illness that
might result from ingestion or invasion of infectious
agents, inhalation or absorption of corrosive or
toxic substances through skin contact, and
asphyxiation because of lack of oxygen or
presence of noxious gases.
5.1.4 Before beginning sample collection and
laboratory work, personnel should determine that
all required safety equipment and materials have
been obtained and are in good condition.
5.2 Safety Equipment
5.2.1 Personal Safety Gear
5.2.1.1 Personnel should use appropriate safety
equipment, such as rubber aprons, laboratory
coats, respirators, gloves, safety glasses, face
shields, hard hats, and safety shoes.
5.2.2 Laboratory Safety Equipment
5.2.2.1 Each laboratory should be provided with
safety equipment such as first aid kits, fire
extinguishers, fire blankets, emergency showers,
and eye wash stations.
5.2.2.2 All laboratories should be equipped with a
telephone to enable personnel to summon help in
case of emergency.
5.3 General Laboratory and Field
Operations
5.3.1 Laboratory personnel should be trained in
proper practices for handling and using chemicals
that are encountered during procedures described
in this manual. Routinely encountered chemicals
include acids, organic solvents, and standard
22
-------
materials for reference-toxicity tests. Special
handling and precautionary guidance in Material
Safety Data Sheets should be followed for
reagents and other chemicals purchased from
supply houses.
5.3.2 Work with some sediment might require
compliance with rules pertaining to the handling of
hazardous materials. Personnel collecting
samples and performing tests should not work
alone.
5.3.3 It is advisable to wash exposed parts of the
body with bactericidal soap and water immediately
after collecting or manipulating sediment samples.
5.3.4 Strong acids and volatile organic solvents
should be used in a fume hood or under an
exhaust canopy over the work area.
5.3.5 An acidic solution should not be mixed with
a hypochlorite solution because hazardous vapors
might be produced.
5.3.6 To prepare dilute acid solutions,
concentrated acid should be added to water, not
vice versa. Opening a bottle of concentrated acid
and adding concentrated acid to water should be
performed only under a fume hood.
5.3.7 Use of ground-fault systems and leak
detectors is strongly recommended to help prevent
electrical shocks. Electrical equipment or
extension cords not bearing the approval of
Underwriter Laboratories should not be used.
Ground-fault interrupters should be installed in all
"wet" laboratories where electrical equipment is
used.
5.3.8 All containers should be adequately labeled
to identify their contents.
5.3.9 Good housekeeping contributes to safety
and reliable test results.
5.4 Disease Prevention
5.4.1 Personnel handling samples that are known
or suspected to contain human wastes should be
given the opportunity to be immunized against
hepatitis B, tetanus, typhoid fever, and polio.
Thorough washing of exposed skin with
bactericidal soap should follow handling these
samples.
5.5 Safety Manuals
5.5.1 For further guidance on safe practices when
handling sediment samples and conducting toxicity
tests, check with the permittee and consult general
industrial safety manuals including USEPA
(1986b) and Walters and Jameson (1984).
5.6 Pollution Prevention, Waste
Management, and Sample Disposal
5.6.1 It is the laboratory's responsibility to comply
with the federal, state, and local regulations
governing the waste management, particularly
hazardous waste identification rules and land
disposal restrictions, and to protect the air, water,
and land by minimizing and controlling all releases
from fume hoods and bench operations. Also,
compliance is required with any sewage discharge
permits and regulations. For further information on
waste management, consult The Waste
Management Manual for Laboratory Personnel,
available from the American Chemical Society's
Department of Government Relations and Science
Policy, 1155 16th Street N.W., Washington, D.C.
20036.
5.6.2 Guidelines for the handling and disposal of
hazardous materials should be strictly followed.
The federal government has published regulations
for the management of hazardous waste and has
given the states the option of either adopting those
regulations or developing their own. If states
develop their own regulations, they are required to
be at least as stringent as the federal regulations.
As a handler of hazardous materials, it is a
laboratory's responsibility to know and comply with
the applicable state regulations. Refer to The
Bureau of National Affairs, Inc. (1986) for the
citations of the federal requirements.
5.6.3 Substitution of nonhazardous chemicals and
reagents should be encouraged and investigated
whenever possible. For example, use of a
nonhazardous compound for a positive control in
reference-toxicity tests is advisable. Reference-
toxicity tests with copper can provide appropriate
toxicity at concentrations below regulated levels.
23
-------
Section 6
Facilities, Equipment, and Supplies
6.1 General
6.1.1 Before a sediment test is conducted in any
test facility, it is desirable to conduct a
"nontoxicant" test with each test species in which
all test chambers contain a control sediment
(sometimes called the negative control) and clean
overlying water. Survival, growth, or reproduction
of the test organisms will demonstrate whether
facilities, water, control sediment, and handling
techniques are adequate to result in acceptable
species-specific control numbers. Evaluations
may also be made on the magnitude of between-
chamber variance in a test. See Section 9.14.
6.2 Facilities
6.2.1 The facility must include separate areas for
culturing test organisms and sediment testing to
reduce the possibility of contamination by test
materials and other substances, especially volatile
compounds. Holding, acclimation, and culture
chambers should not be in a room where sediment
tests are conducted, stock solutions or sediments
are prepared, or equipment is cleaned. Test
chambers may be placed in a temperature-
controlled recirculating water bath, environmental
chamber, or equivalent facility with temperature
control. An enclosed test system is desirable to
provide ventilation during tests to limit exposure of
laboratory personnel to volatile substances.
6.2.2 Light of the quality and illuminance normally
obtained in the laboratory (about 500 to 1000 lux
using wide-spectrum fluorescent lights; e.g., cool-
white or daylight) is adequate to culture L
plumulosus and to conduct the chronic toxicity
test. Lux is the unit selected for measuring
luminance in this manual, and should be measured
at the surface of the water in test or culture
chambers. A uniform photoperiod of 16 h light
and 8 h dark shall be maintained for cultures and
during the tests, and can be achieved in the
laboratory or in an environmental chamber using
automatic timers.
6.2.3 During phases of rearing, holding, and
testing, test organisms should be shielded from
external disturbances such as rapidly changing
light or pedestrian traffic.
6.2.4 The test facility should be well ventilated and
free of fumes. Laboratory ventilation systems
should be checked to ensure that return air from
chemistry laboratories or sample handling areas is
not circulated to culture or testing rooms, or that
air from testing rooms does not contaminate
culture rooms. Air pressure differentials between
rooms should not result in a net flow of potentially
contaminated air to sensitive areas through open
or loose-fitting doors. Air used for aeration must
be free of oil and fumes. Oil-free air pumps should
be used where possible. Filters to remove oil,
water, and bacteria are desirable. Particles can be
removed from the air using filters such as
BALSTON® Grade BX (Balston, Inc., Lexington,
MA) or equivalent, and oil and other organic
vapors can be removed using activated carbon
filters (e.g., BALSTON® C-1 filter), or equivalent.
6.3 Equipment and Supplies
6.3.1 Equipment and supplies that contact stock
solutions, sediment, or overlying water should not
contain substances that can be leached or
dissolved in amounts that adversely affect the test
organisms. In addition, equipment and supplies
that contact sediment or water should be chosen
to minimize sorption of test materials from water.
Glass, type 316 stainless steel, nylon, and high-
density polyethylene, polypropylene, poly-
24
-------
carbonate, and fluorocarbon plastics should be
used whenever possible to minimize leaching,
dissolution, and sorption. Concrete and high-
density plastic containers may be used for holding
and culture chambers, and in the water-supply
system. These materials should be washed in
detergent, acid-rinsed, and soaked in flowing water
for a week or more before use. Cast-iron pipe
should not be used in water-supply systems
because colloidal iron will be added to the
overlying water and strainers will be needed to
remove rust particles. Copper, brass, lead,
galvanized metal, and natural rubber must not
contact overlying water or stock solutions before or
during a test. Items made of neoprene rubber and
other materials not mentioned above should not be
used unless it has been shown that their use will
not adversely affect survival, growth, or
reproduction of the test organisms.
6.3.2 New lots of plastic products should be tested
for toxicity by exposing organisms to them under
ordinary test conditions before general use.
6.3.3 General Equipment
6.3.3.1 Environmental chamber or equivalent
facility with photoperiod and temperature control
(20°C to 25°C).
6.3.3.2 Water purification system capable of
producing at least 1 mega-ohm water (USEPA,
1991 a).
6.3.3.3 Analytical balance capable of accurately
weighing to 0.01 mg.
6.3.3.4 Reference weights, Class S—for
documenting the performance of the analytical
balance(s). The balance(s) should be checked
with reference weights that are at the upper and
lower ends of the range of the weighings made
when the balance is used. A balance should be
checked at the beginning of each series of
weighings, periodically (such as every tenth
weight) during a long series of weighings, and after
taking the last weight of a series.
6.3.3.5 Volumetric flasks and graduated
cylinders—Class A, borosilicate glass or nontoxic
plastic labware, 10 to 1000 mL for making test
solutions.
6.3.3.6 Volumetric pipets—Class A, 1 to 100 ml.
6.3.3.7 Serological pipets—1 to 10 mL, graduated.
6.3.3.8 Pipet bulbs and fillers.
6.3.3.9 Droppers, and glass tubing with fire
polished edges, 4- to 6-mm ID—for transferring
test organisms.
6.3.3.10 Wash bottles—for rinsing small
glassware, instrument electrodes, and probes.
6.3.3.11 Electronic (digital) thermometers—for
measuring water temperature. Mercury-filled glass
thermometers should not be used.
6.3.3.12 National Bureau of Standards Certified
thermometer (see USEPA Method 170.1; USEPA,
1979b).
6.3.3.13 Dissolved oxygen (DO), pH, and salinity
meters for routine physical and chemical
measurements (portable field-grade instruments
are acceptable unless a test is conducted to
specifically measure the effects of one of these
measurements). A temperature-compensated
salinity refractometer is useful for measuring
salinity of water overlying field-collected sediment.
6.3.3.14 Ammonia-specific probe with a functional
range between 1 and >100 mg/L total ammonia.
6.3.3.15 Table 6.1 lists additional equipment and
supplies.
6.3.4 Test Chambers
6.3.4.1 Test chambers may be constructed in
several ways and of various materials, depending
on the experimental design and the contaminants
of interest. Clear silicone adhesives, suitable for
aquaria, sorb some organic compounds that might
be difficult to remove. Therefore, as little adhesive
as possible should be in contact with the test
material. Extra beads of adhesive should be on
the outside of the test chambers rather than on the
inside. To leach potentially toxic compounds from
the adhesive, all new test chambers constructed
using silicone adhesive should be held at least 48
h in overlying water before use in a test.
25
-------
Table 6.1 Equipment and Supplies for Culturing and Testing L. plumulosus
Biological Supplies
Brood stock of test organisms
TetraMin®
Live microalgae (e.g., Pseudoisochrysis paradoxa, Phaeodactylum tricomutum, Isochrysis galbana, Chaetoceros calcitrans,
Skeletonema sp., or ttallashsims spp. [optional food items for culturing L. plumulosus])
Containers and Glassware
Culture chambers (e.g.. 35-cm x 30-cm x 15-cm plastic wash bin)
Test chambers (1-L glass jar or beaker)
Glass bowls
Wide-bore pipets, droppers, or glass tubing (4- to 6-mm ID) for organism transfer
Glass disposable serological pipets or digital equivalent
Graduated cylinders (assorted sizes, 10 mL to 2 L)
Instruments and Equipment
Dissecting microscope
Stainless-steel (for culture or contaminated sediment) or Nytex (for culture sediment only) sieves (U.S. Standard No. 18,35, and
60 mesh or 1.0, 0.5 or 0.6, and 0.25 mm)
Photoperiod timer
Light meter
Environmental chamber, water bath, or equivalent with photoperiod and temperature control
Thermometer, electronic (digital)
Continuous recording thermometer
Dissolved oxygen meter
pH meter
Meter with Ion-specific ammonia electrode (or functional equivalent)
Salinity meter or temperature compensating salinity refractometer
Drying oven
Desiccator
Balance (0.01 mg sensitivity)
Refrigerator
Freezer
Light box
Hemacytometer (optional)
Mortar and pestle, blender, grain mill, coffee grinder
Pump for water exchanges
Miscellaneous
Ventilation system for test chambers
Ventilation system for counts of alcohol-preserved samples
Air supply and air stones/pipets (oil free and regulated)
Weighing pans
Fluorescent light bulbs
Delonized water
Air line tubing
Plastic dish pan
Sieve cups
Chemicals
Detergent (nonphosphate)
Acetone (reagent grade)
Hydrochloric or nitric acid (reagent grade)
Reagents for preparing synthetic seawater (reagent grade CaCI2 -2 H2O, KBr, KCI, MgCI2-6 H2O, Na2B4O7-10 H2O, NaCI, Na
HCO3, Na2SO4, SrCI2 -6 H2O [optional])
Alcohol (either ethyl or isopropyl)
Rosebengal
Reference toxicant (ammonia, copper sulfate, cadmium chloride, sodium dodecyl sulfate)
26
-------
6.3.5 Cleaning
6.3.5.1 All nondisposable sample containers, test
chambers, and other equipment that have come in
contact with sediment should be washed after use
in the manner described below to remove surface
contaminants.
1. Soak 15 min in tap water and scrub with
detergent or clean in an automatic
dishwasher.
2. Rinse twice with tap water.
3. Carefully rinse once with fresh, dilute (10%,
V:V) hydrochloric or nitric acid to remove
scale, metals, and bases. To prepare a
10% solution of acid, add 10 mL of
concentrated acid to 90 mL of deionized
water.
4. Rinse twice with deionized water.
5. Rinse once with full-strength, pesticide-
grade acetone to remove organic
compounds (uses a fume hood or canopy).
Hexane might also be used as a solvent for
removing nonionic organic compounds.
However, acetone is preferable if only one
organic solvent is used to clean equipment.
6. Rinse three times with deionized water.
6.3.5.2 All test chambers and equipment should
be thoroughly rinsed or soaked with the dilution
water immediately before use in a test. See
USEPA (2001) for information on equipment
decontamination procedures with regards to
collecting sediments in the field.
6.3.5.3 Many organic solvents (e.g., methylene
chloride) leave a film that is insoluble in water. A
dichromate-sulfuric acid cleaning solution can be
used in place of both the organic solvent and the
acid (see ASTM, 2000f), but the solution might
attack silicone adhesive and leave chromium
residues on glass. An alternative to use of
dichromate-sulfuric acid could be to heat
glassware for 8 h at 450°C.
27
-------
Section 7
Water, Reagents, and Standards
7.1 Water
7.1.1 Requirements
7.1.1.1 Seawater used to test and culture
organisms should be uniform in quality and from
the same source. Acceptable seawater must allow
satisfactory survival, growth, and reproduction of
the test organisms. If problems are observed in
the culturing or testing of organisms, the
characteristics of the water should be evaluated.
See USEPA (1991 a) and ASTM (2000b) for a
recommended list of chemical analyses of the
water supply.
7.1.2 Source
7.1.2.1 Culture and testing water can be natural or
synthetic seawater. The source of water will
depend to some extent on the objective of the test.
All natural waters should be obtained from an
uncontaminated source beyond the influence of
known discharges. Suitable water sources should
have intakes that are positioned to (1) minimize
fluctuations in quality and contamination,
(2) maximize the concentration of dissolved
oxygen, and (3) ensure low concentrations of
sulfide and iron. Natural seawater should be
collected at slack high tide, or within 1 h after high
tide if taken from an semi-enclosed or urbanized
area. It might be desirable or necessary to dilute
full strength seawater with an appropriate
freshwater source to achieve 5%o or 20%o.
7.1.2.2 Sources of freshwater (i.e., 0%o) for dilution
include distilled or deionized water, reverse
osmosis water, and uncontaminated well or spring
water (USEPA, 1991a). Municipal water supplies
can be variable and might contain unacceptably
high concentrations of materials such as copper,
lead, zinc, fluoride, chlorine, or chloramines.
Chlorinated water should not be used to dilute
seawater used for culturing or testing because
residual chlorine and chlorine-produced oxidants
are toxic to many aquatic organisms.
Dechlorinated water should only be used as a last
resort, because dechlorination is often incomplete
(ASTM, 2000d).
7.1.3 Water Treatment and Quality
7.1.3.1 Seawater and dilution water should be
filtered ($5 mm) shortly before use to remove
suspended particles and organisms. Water that
might be contaminated with pathogens should be
treated shortly before use by filtration (<;0.45 mm),
either alone or in combination with UV sterilization.
7.1.3.2 Water should be aerated using air stones,
surface aerators, or column aerators. Adequate
aeration will stabilize pH, bring concentrations of
DO and other gases into equilibrium with air, and
minimize oxygen demand and concentrations of
volatiles. The initial concentration of DO in test
water should be ;>6 mg/L to help ensure that DO
concentrations are acceptable in test chambers.
7.1.3.3 DO, salinity, and pH should be measured
on each batch of water before it is used in cultures
and tests. Batches of salinity-adjusted culture
water can be held for approximately 1 week; a
lower holding temperature (about 4°C) helps
maintain acceptable water quality. Other
investigators have reported success in holding
reconstituted seawater for toxicity testing for over
1 month (Ingersoll et al., 1992).
7.1.3.4 For site-specific investigations, it might be
desirable to have the water-quality characteristics
of the overlying water (i.e., salinity) as similar as
possible to the site water. Other applications may
require use of water from the site where sediment
is collected. In estuarine systems, however, the
pore water salinity of sediment might not be the
same as the overlying water at the time of
collection (Sanders et al., 1965).
28
-------
7.1.4 Reconstituted/Synthetic Seawater
7.1.4.1 Although reconstituted seawater is
acceptable, natural seawater is preferable,
especially for tests in which the bioavailability of
chemicals is affected by seawater chemistry.
Reconstituted seawater is prepared by adding
specified amounts of reagent-grade chemicals.to
high-purity distilled or deionized water (ASTM,
2000f; USEPA, 1991 c). Suitable salt reagents can
be reagent-grade chemicals, or commercial sea
salts, such as Crystal Sea Marinemix®, Instant
Ocean®, or HW Marinemix®. Preformulated brine
(e.g., 60%0 to 90%o), prepared with dry ocean salts,
or by heat-concentrating or freezing natural
seawater, can also be used.
7.1.4.2 A synthetic sea formulation called GP2
can be prepared with reagent-grade chemicals
and diluted with a suitable high-quality water to the
desired salinity (Section 7.1.2.2; USEPA, 1994c).
7.1.4.3 The suitability and consistency of a
particular salt formulation for use in holding and
testing should be verified by laboratory tests
because some formulations can produce
unwanted toxic effects or sequester contaminants
(Environment Canada, 1992). In controlled tests
with the salt formulations mentioned above, Emery
etal. (1997) found differences in survival, growth,
and reproduction, and that laboratories can have
acceptable performance (i.e., survival) with any of
the salts evaluated. Because of higher growth
rates observed in the Crystal Sea Marinemix®
seasalt, they recommended its use for culturing
and testing (Emery et al., 1997).
7.1.4.4 Deionized, distilled, or reverse-osmosis
water should be obtained from a system capable
of producing at least 1 mega-ohm water. If large
quantities of high quality water are needed, it might
be advisable to precondition water with a mixed-
bed water treatment system. Some investigators
have observed that aging of reconstituted water
prepared from deionized water for several days
before use in sediment tests may improve
performance of test organisms. Other investigators
have reported success in holding reconstituted
seawater for toxicity testing for over 1 month
(Ingersoll et al., 1992).-
7.1.4.5 Salinity, pH, and DO should be measured
on each batch of reconstituted water. The
reconstituted water should be aerated before use
to adjust pH and DO to the acceptable ranges
(e.g., Table 11.1). Reconstituted sea water should
be filtered (s5 mm) shortly before use to remove
suspended particles and should be used within 24
h of filtration. USEPA (1991 a) recommends
holding a batch of reconstituted water for no longer
than 2 weeks due to the potential for
bacteriological growth. Other investigators have
reported success in holding reconstituted seawater
for toxicity testing for over 1 month (Ingersoll etal.,
1992).
7.2 Reagents
7.2.1 Material safety data sheets should be
followed for reagents and other chemicals
purchased from supply houses. The test
material(s) should be at least reagent grade,
unless a test using a formulated commercial
product, technical-grade, or use-grade material is
specifically needed. Reagent containers should be
dated when received from the supplier, and the
shelf life of the reagent should not be exceeded.
Working solutions should be dated when prepared
and the recommended shelf life should not be
exceeded.
7.3 Standards
7.3.1 Appropriate standard methods for chemical
and physical analyses should be used when
possible. For those measurements for which
standards do not exist or are not sensitive enough,
methods should be obtained from other reliable
sources.
29
-------
Section 8
Sample Collection, Storage, Manipulation, and Characterization
8.1 Collection
8.1.1 Before the preparation or collection of
sediment, a procedure should be established for
the handling of sediment that might contain
unknown quantities of toxic contaminants
(Section 5).
8.1.2 Sediments are spatially and temporally
variable (Stemmer et al., 1990a). Replicate
samples should be collected to determine variance
in sediment characteristics. Sediment should be
collected with as little disruption as possible;
however, subsampling, compositing, or
homogenization of sediment samples might be
necessary for some experimental designs.
Sampling can cause loss of sediment integrity,
change in chemical speciation, or disruption of
chemical equilibrium (ASTM, 2000c). Benthic
grabs (i.e., Ponar, Smith-Maclntyre, Van Veen) or
core samplers should be used rather than a
dredge to minimize disturbance of the sediment
sample. Sediment should be collected to a depth
that will represent expected exposure
concentration. For example, samples collected for
evaluations of dredged material should include
sediment cores to the depth of removal. Surveys
of the toxicity of surficial sediment are often based
on samples of the upper 2 cm of sediment.
8.1.3 Exposure to direct sunlight during collection
should be minimized, especially if the sediment
contains photolytic compounds (e.g., PAHs).
Collect, manipulate, and store sediments using
tools made of chemically inert materials to
minimize contamination of the sample
(ASTM, 2000b). Sediment samples should be
cooled to 4°C as quickly as possible in the field
before shipment or return to the laboratory .(ASTM,
2000b). Coolers with gel packs, ice, or dry ice can
be used to cool samples in the field; however,
sediment should never be frozen. Continuous-
recording monitors can be used to measure
temperature during shipping (e.g., TempTale
Temperature Monitoring and Recording System,
Sensitech, Inc., Beverly, MA).
8.1.4 For additional information on sediment
qollection and shipment, refer to methods
published by USEPA (2001) and ASTM (2000c).
8.2 Storage
8.2.1 Because the contaminants of concern and
influencing sediment characteristics are not always
known, it is desirable to hold the sediments after
collection in the dark at 4°C. Traditional
convention has held that toxicity tests should be
initiated as soon as possible following collection
from the field, although actual recommended
storage times range from 2 weeks (ASTM, 2000c)
to less than 8 weeks (USEPA-USACE, 1998).
Discrepancies in recommended storage times
reflected a lack of data concerning the effects of
long-term storage on the physical, chemical, and
toxicological characteristics of the sediment.
However, numerous studies have recently been
conducted to address issues related to sediment
storage (Dillon et al., 1994; Becker and Ginn,
1995; Carr and Chapman, 1995; Moore et al.,
1996; Sarda and Burton, 1995; Sijm et al., 1997;
DeFoe and Ankley, 1998). The conclusions and
recommendations offered by these studies vary
substantially and appear to depend primarily upon
the type or class of contaminant(s) present.
Considered collectively, these studies suggest that
the recommended guidance that sediments be
tested sometime between the time of collection
and 8 weeks storage is appropriate. Additional
guidance is provided below.
8.2.2 Extended storage of sediments that contain
high concentrations of labile contaminants (e.g.,
ammonia, volatile organics) may lead to a loss of
30 .
-------
these contaminants and a corresponding reduction
in toxicity. Under these circumstances, the
sediments should be tested as soon as possible
after collection, but not later than within 2 weeks
(Sarda and Burton, 1995). Sediments that exhibit
low-level to moderate toxicity often exhibit
considerable temporal variability in toxicity,
although the direction of change is often
unpredictable (Carr and Chapman, 1995; Moore et
al., 1996; DeFoe and Ankley, 1998). For these
types of sediments, the recommended storage
time of <8 weeks may be most appropriate. In
some situations, a minimum storage period for
low-to-moderately contaminated sediments may
help reduce variability. For example, DeFoe and
Ankley (1998) observed high variability in survival
during early testing periods (e.g., <2 weeks) in
sediments with low toxicity. DeFoe and Ankley
(1998) hypothesized that this variability partially
reflected the presence of indigenous predators
that remained alive during this relatively short
storage period. Thus, if predatory species are
known to exist, and the sediment does not contain
labile contaminants, it may be desirable to
store the sediment for a short period before testing
(e.g., 2 weeks) to reduce potential for
interferences with indigenous organisms.
Sediments that contain comparatively stable
compounds (e.g., high-molecular-weight
compounds such as polychlorinated biphenyls
[RGBs]) or that exhibit a moderate-to-high level of
toxicity, typically do not vary appreciably in toxicity
in relation to storage duration (Moore et al., 1996;
DeFoe and Ankley, 1998). For these sediments,
long-term storage (e.g., >8 weeks) can be
undertaken.
8.2.3 Researchers may wish to conduct additional
characterizations of sediment to evaluate
possible effects of storage. Concentrations of
contaminants of concern could be measured
periodically in pore water during the storage period
and at the start of the sediment test (Kemble et al.,
1994). Ingersoll et al. (1993) recommend
conducting a toxicity test with pore water within
2 weeks from sediment collection and at the start
of the sediment test. Freezing might further
change sediment properties such as grain size or
contaminant partitioning and should be avoided
(ASTM, 2000c; Schuytema et al., 1989). Sediment
should be stored with no air over the sealed
samples (no head space) at 4°C before the start
of a test (Shuba et al., 1978). Sediment may be
stored in containers constructed of suitable
materials as outlined in Section 6.
8.3 Manipulation
8.3.1 Homogenization and Sieving
8.3.1.1 Samples tend to settle during shipment.
As a result, water above the sediment should not
be discarded but should be mixed back into the
sediment during homogenization. Sediment
samples should only be sieved to remove
indigenous organisms if there is a good reason to
believe indigenous organisms may influence the
response of the test organism. Sieving
procedures are outlined in Section 4.3.2.3.
However, large indigenous organisms and large
debris can be removed using forceps. Reynoldson
et al. (1994) observed reduced growth of
amphipods, midges, and mayflies in sediments
with elevated numbers of oligochaetes and
recommended sieving sediments suspected to
have high numbers of indigenous oligochaetes. If
sediments must be sieved, it may be desirable to
analyze samples before and after sieving (e.g.,
pore water metals, dissolved organic carbon
[DOC], AVS, TOC) to document the influence of
sieving on sediment chemistry.
8.3.1.2 If sediment is collected from multiple field
samples, the sediment can be pooled and mixed
by stirring or using a rolling mill, feed mixer, or
other suitable apparatus (see ASTM, 2000c).
Homogenization of sediment can be accomplished
by hand with a teflon paddle or using a variable-
speed hand-held drill outfitted with a stainless-
steel auger.
8.3.2 Sediment Spiking
8.3.2.1 Test sediment can be prepared by
manipulating the properties of a control sediment.
Mixing time (Stemmer et al., 1990a) and aging
(Word et al., 1987; Landrum, 1989; Landrum and
Faust, 1992) of spiked sediment can affect
bioavailability of contaminants in sediment. Many
studies with spiked sediment are often started only
a few days after the chemical has been added to
31
-------
the sediment. This short time period may; not be
long enough for sediments to equilibrate with the
spiked chemicals (Section 8.3.2.2.3), Consistent
spiking procedures should be followed in order to
make interlaboratory comparisons. See USEPA
(2001) and ASTM (2000c) for additional detail
regarding sediment spiking.
8.3.2.1.1 The cause of sediment toxicity and the
magnitude of interactive effects of contaminants
can be estimated by spiking a sediment with
chemicals or complex waste mixtures (Lamberson
and Swartz, 1992). Sediments spiked with a range
of concentrations can be used to generate either
point estimates (e.g., LC50) or a minimum
concentration at which effects are. observed
(LOEC). The influence of sediment physico-
chemical characteristics on chemical toxicity can
also be determined with sediment-spiking studies
(Adams etal., 1985).
8.3.2.2 The test material(s) should be at least
reagent grade, unless a test using a formulated
commercial product, technical-grade, or use-grade
material is specifically needed. Before a test is
started, the following should be known about the
test material: (1) the identity and concentration of
major ingredients and impurities; (2) water
solubility in test water; (3) log Kow,
bioconcentration factor (BCF) from other test
species, persistence, hydrolysis, and photolysis
rates of the test substances; (4) estimated toxicity
to the test organism and to humans; (5) if the test
concentration^) are to be measured, the precision
and bias of the analytical method at the planned
concentration(s) of the test material; and
(6) recommended handling and disposal
procedures. Addition of test material(s) to
sediment may be accomplished using.various
methods, such as a (1) rolling mill, (2) feed mixer,
or (3) hand mixing (ASTM, 2000c; USEPA, 2001).
Modifications of the mixing techniques might be
necessary to allow time for a test material to
equilibrate with the sediment. Mixing time of
spiked sediment should be limited from minutes to
a few hours, and temperature should be kept low
to minimize potential changes in the physico-
chemical and microbial characteristics of the
sediment (ASTM, 2000c). Duration of contact
between the chemical and sediment can affect
partitioning and bioavailability (Word et al.., 1987).
Care should be taken to ensure that the chemical
is thoroughly and evenly distributed in the
sediment. Analyses of sediment subsamples are
advisable to determine the degree of mixing
homogeneity (Ditsworth et al., 1990). Moreover,
results from sediment-spiking studies should be
compared to the response of test organisms to
chemical concentrations in natural sediments
(Lamberson and Swartz, 1992).
8.3.2.2.1 Organic chemicals have been added to
sediments using the following procedures:
(1) directly in a dry (crystalline) form; (2) coated on
the inside walls of the container (Ditsworth et al.,
1990); or (3) coated onto silica sand (e.g., 5% w/w
of sediment) which is added to the sediment (D.R.
Mount, USEPA, Duluth, MN, personal
communication). In Techniques 2 and 3, the
chemical is dissolved in solvent, placed in a glass
spiking container (with or without sand), then the
solvent is slowly evaporated. The advantage of
these three approaches is that no solvent is
introduced .to the sediment, only the chemical
being spiked. When testing spiked sediments,
procedural blanks (sediments that have been
handled in the same way, including solvent
addition and evaporation, but that contain no
added chemical) should be tested in addition to
regular negative controls.
8.3.2.2.2 Metals are generally added in an
aqueous solution (ASTM, 2000c; Carlson et al.,
1991; Di Toro et al., 1990). Ammonia has also
been successfully spiked using aqueous solutions
(Moore et al., 1997; Besser et al., 1998). Inclusion
of spiking blanks is recommended.
8.3.2.2.3 Sufficient time should be allowed after
spiking for the spiked chemical to equilibrate with
sediment components. For organic chemicals, it
is recommended that the sediment be aged at
least 1 month before starting a test; 2 months or
more may be necessary for chemicals with a high
log Kow (e.g., >6; D.R. Mount, USEPA, Duluth,
MN, personal communication). For metals, shorter
aging times (1 to 2 weeks) may be sufficient.
Periodic monitoring of chemical concentrations in
pore water during sediment aging is highly
recommended as a means to assess the
32
-------
equilibration of the spiked sediments. Monitoring
of pore water during spiked sediment testing is
also recommended.
8.3.2.3 Direct addition of a solvent (other than
water) to the sediment should be avoided if
possible. Addition of organic solvents may
dramatically influence the concentration of
dissolved organic carbon in pore water. If an
organic solvent is to be used, the solvent should
be at a concentration that does not affect the test
organism. Further, both solvent control and
negative control sediments must be included in the
test. The solvent control must contain the highest
concentration of solvent present and must be from
the same batch used to make the stock solution
(see ASTM, 20000- ' . •
8.3.2.3.1 If direct addition of organic solvent is to
be used, the same concentration of solvent should
be used in all treatments. If the concentration of
solvent is not the same in all treatments, a solvent
test should be conducted to determine whether
survival, growth, or reproduction of the test
organisms is related to the concentration of the
solvent.
8.3.2.4 If the test contains both a negative control
and a solvent, control, the survival, growth, or
reproduction of the organisms tested should be
compared. If a statistically significant difference is
detected between the two controls, only the
solvent control may be used for meeting the
acceptability of the test and as the basis for
calculating results. The negative control might
provide additional information on the general
health of the organisms tested. If no statistically
significant difference is detected, the data from
both controls should be used for meeting the
acceptability of the test and as the basis for
calculating the results (ASTM, 2000g). If
performance in the solvent control is markedly
different from that in the negative control, it is
possible that the data are compromised by
experimental artifacts and may not accurately
reflect the toxicity of the chemical in natural
sediments.
8.3.3 Test Concentration(s) for Laboratory
Spiked Sediments
8.3.3.1 If a test is intended to generate an LC50, a
toxicant concentration series (0.5 or higher) should
be selected that will provide partial mortalities at
two or more concentrations of the test chemical.
The LC50 of a particular compound may vary
depending on physical and chemical sediment
characteristics. It may be desirable to conduct a
range-finding test in which the organisms are
exposed to a control and three or more
concentrations of the test material that differ by a
factor of ten. Results from water-only tests could
be used to establish concentrations to be tested in
a whole-sediment test based on predicted pore
water concentrations (Di Toro et al., 1991).
8.3.3.2 Bulk-sediment chemical concentrations
might be normalized to factors other than dry
weight. For example, concentrations of nonpolar
organic compounds might be normalized to
sediment organic-carbon content, and
simultaneously extracted metals might be
normalized to acid volatile sulfides (Di Toro et al.,
1990; Di Toro etal., 1991).
8.3.3.3 In some situations it might be necessary to
simply determine whether a specific concentration
of test material is toxic to the test organism, or
whether adverse effects occur above or below a
specific concentration. When there is interest in a
particular concentration, it might only be necessary
to test that concentration and not to determine an
LC50.
8.4 Characterization
8.4.1 All sediment should be characterized and at
least the following determined: salinity, pH, and
ammonia of the pore water; TOC; .particle-size
distribution (percent sand, silt, clay); and percent
water content (ASTM, 2000b; Plumb; 1981). See .
Section 8.4.4.7 for methods to isolate pore water.
8.4.2 Other analyses on sediment might include
biological oxygen demand (BOD), chemical
oxygen demand (COD), cation exchange capacity,
redox potential (Eh), total inorganic carbon, total
volatile solids (TVS), AVS, metals, synthetic
organic compounds, oil and grease, petroleum
33
-------
hydrocarbons, as well as interstitial water analyses
for various physico-chemical parameters.
8.4.3 Macrobenthos can be evaluated by
subsampling the field-collected sediment. If direct
comparisons are to be made, subsamples for
toxicity testing should be collected from the same
sample to be used for analysis of physical and
chemical characteristics. Qualitative descriptions
of the sediment can include color, texture,
presence of hydrogen sulfide, and presence of
indigenous organisms. Monitoring the odor of
sediment samples should be avoided because of
potential hazardous volatile contaminants. It may
be desirable to describe color and texture
gradients that occur with sediment depth.
8.4.4 Analytical Methods
8.4.4.1 Chemical and physical data should be
obtained using appropriate standard methods
whenever possible. For those measurements for
which standard methods do not exist or are not
sensitive enough, methods should be obtained
from other reliable sources.
8.4.4.2 The precision, accuracy, and bias of each
analytical method used should be determined in
the appropriate matrix: sediment, water, or tissue.
Reagent blanks and analytical standards should
be analyzed and recoveries should be calculated.
8.4.4.3 Concentration of spiked test material(s) in
sediment, interstitial water, and overlying water
should be measured as often as practical during a
test. If possible, the concentration of the test
material in overlying water, interstitial water, and
sediment should be measured at the start and end
of a test. Measurement of test material(s)
degradation products might also be desirable.
8.4.4.4 Separate chambers should be set up at the
start of a test and destructively sampled during
and at the end of the test to monitor sediment
chemistry. Test organisms and food might be
added to these extra chambers.
8.4.4.5 Measurement of test material(s)
concentrations in water can be accomplished by
pipeting water samples from about 1 cm to 2 cm
above the sediment surface in the test chamber.
Overlying water samples should not contain any
surface debris, any material from the sides of the
test chamber, or any sediment.
8.4.4.6 Measurement of concentrations of test
material(s) in sediment at the end of a test can be
taken by siphoning most of the overlying water
without disturbing the surface of the sediment,
then removing appropriate aliquots of the sediment
for chemical analysis.
8.4.4.7 Interstitial Water
8.4.4.7.1 Interstitial water (pore water), defined as
the water occupying the spaces between sediment
or soil particles, is often isolated to provide either
a matrix for toxicity testing or to provide an
indication of the concentration or partitioning of
contaminants within the sediment matrix. Draft
USEPA ESGs are based on the presumption that
the concentration of chemicals in the interstitial
water are correlated directly to their bioavailability
and, therefore, their toxicity (Di Toro et al., 1991).
Of additional importance is contaminants in
interstitial waters can be transported into overlying
waters through diffusion, bioturbation, and
resuspension processes (Van Rees et al., 1991).
The usefulness of interstitial water sampling for
determining chemical contamination or toxicity will
depend on the study objectives and nature of the
sediments at the study site.
8.4.4.7.2 Isolation of sediment interstitial water can
be accomplished by a wide variety of methods,
which are based on either physical separation or
on diffusion/equilibration. The common physical-
isolation procedure can be categorized as
(1) centrifugation, (2) compression/squeezing, or
(3) suction/vacuum. Diffusion/equilibrium
procedures rely on the movement (diffusion) of
pore water constituents across semipermeable
membranes into a collecting chamber until an
equilibrium is established. A description of the
materials and procedures used in the isolation of
pore water is included in the reviews by Bufflap
and Allen (1995a), ASTM (2000c), and USEPA
(2001).
8.4.4.7.3 When relatively large volumes of water
are required (>20 mL) for toxicity testing or
chemical analyses, appropriate quantities of
sediment are generally collected with grabs or
34
-------
corers for subsequent isolation of the interstitial
water. Several isolation procedures, such as
centrifugation (Ankley and Scheubauer-Berigan,
1994), squeezing (Carr and Chapman, 1995) and
suction (Winger and Lasier, 1991; Winger et al.,
1998), have been used successfully to obtain
adequate volumes for testing purposes. Peepers
(dialysis) generally do not produce sufficient
volumes for most analyses; however, larger sized
peepers (500-mL volume) have been used for
collecting interstitial water in situ for chemical
analyses and organism exposures (Burton, 1992;
Sarda and Burton, 1995).
8.4.4.7,4 There is not one superior method for the
isolation of interstitial water used for toxicity testing
and associated chemical analyses. Factors to
consider in the selection of an isolation procedure
may include (1) volume of pore water needed,
(2) ease of isolation (materials, preparation time,
and time required for isolation), and (3) artifacts in
the pore water caused by the isolation procedure.
Each approach has unique strengths and
limitations (Bufflap and Allen, 1995a; 1995b;
Winger et al., 1998), which vary with sediment
characteristics, chemicals of concern, toxicity test
methods, and desired test resolution (i.e., the data
quality objectives). For suction or compression
separation, which use a filter or a similar surface,
there may be changes to the characteristics of the
interstitial water compared with separation using
centrifugation (Ankley et al., 1994; Horowitz et al.,
1996). For most toxicity test procedures, relatively
large volumes of interstitial water (e.g., liters) are
frequently needed for static or renewal exposures
with the associated water chemistry analyses.
Although centrifugation can be used to generate
large volumes of interstitial water, it is difficult to
use centrifugation to isolate water from coarser
sediment. If smaller volumes of interstitial water
are adequate and logistics allow, it may be
desirable to use peepers, which establish an
equilibrium with the pore water through a
permeable membrane. If logistics do not allow
placement of peeper samplers, an alternative
procedure could be to collect cores that can be
sampled using side port suctioning or
centrifugation (G.A. Burton, Wright State
University, personal communication). However, if
larger samples of interstitial water are needed, it
would be necessary to collect multiple cores as
quickly as possible using an inert environment and
to centrifuge samples at ambient temperatures.
See USEPA (2001) and ASTM (2000c) for
additional detail regarding isolation of interstitial
water.
35
-------
Section 9
Quality Assurance and Quality Control
9.1 Introduction
9.1.1 Developing and maintaining a laboratory
quality assurance (QA) program requires an
ongoing commitment by laboratory management
and also includes the following: (1) appointment of
a laboratory quality assurance officer with the
responsibility and authority to develop and
maintain a QA program; (2) preparation of a
Quality Assurance Project Plan with Data Quality
Objectives; (3) preparation of written descriptions
of laboratory Standard Operating .Procedures
(SOPs) for test organism culturing, testing,
instrument calibration, sample chain-of-custody,
laboratory sample tracking system, and other
procedures, as required; and (4) provision of
qualified technical staff and suitable space and
equipment to assure reliable data. Additional
guidance for QA can be obtained in USEPA
(1989b; 1999), and Moore et al., 1994.
9.1.2 QA practices within a testing laboratory
should address all activities that affect the quality
of the final data, such as (1) sediment sampling
and handling, (2) the source and condition of the
test organisms, (3) the condition and operation of
equipment, (4) test conditions, (5) instrument
calibration, (6) replication, (7) use of reference
toxicants^ (8) record keeping, and (9) data
evaluation.
9.1.3 Quality Control (QC) practices, on the other
hand, consist of the more focused, routine, day-to-
day activities conducted within the scope of the
overall QA program. For a more detailed
discussion of quality assurance and general
guidance on good laboratory practices related to
testing see FDA (1978), USEPA (1979a; 1980a;
1980b; 1991 a; 1994b; 1995; 2001), DeWoskin
(1984), and Taylor (1987).
9,2 Performance-Based Criteria
9.2.1 The USEPA EMMC recommended the use
of performance-based methods in developing
standards for chemical analytical methods
(Williams, 1993). Performance-based methods
were defined by EMMC as a monitoring approach
that uses methods that meet pre-established,
demonstrated performance standards. Minimum
required elements of performance, such as
precision, reproducibility, bias, sensitivity, and
detection limits should be specified, and the
method should be demonstrated to meet the
performance standards.
9.2.2 In developing guidance for culturing
L. plumulosus, it was determined that no single
method has to be used to culture organisms.
Success of a test relies on the health of the culture
from which organisms are taken for testing.
Having healthy organisms of known quality and
age (i.e., size) for testing is the key consideration
relative to culture methods. Therefore, a
performance-based criteria approach is the
preferred method by which individual laboratories
should evaluate culture health, rather than a
control-based criteria approach. Performance-
based criteria were chosen to allow each
laboratory to optimize culture methods that provide
organisms that produce reliable and comparable
test results. Performance criteria for culturing and
testing L. plumulosus are listed in Table 11.3. .
9.3 Facilities, Equipment, and Test
Chambers
9.3.1 Separate areas for test organism culturing
and testing must be provided to avoid loss of
cultures from cross-contamination. Ventilation
systems should be designed and operated to.
prevent recirculation or leakage of air from
chemical analysis laboratories or sample storage
36
-------
and preparation areas into test organism culturing
or sediment testing areas, and from sediment
testing laboratories and sample preparation areas
into culture rooms.
9.3.2 Equipment for temperature control should be
adequate to maintain recommended test-water
temperatures. Recommended materials should be
used in the fabricating of the test equipment that
comes in contact with the sediment or overlying
water.
9.3.3 Before a sediment test is conducted in a new
facility, a "noncontaminant" test should be
conducted in which all test chambers contain a
control sediment and overlying water. This
information is used to demonstrate that the facility,
control sediment, water, and handling procedures
provide acceptable responses of test organisms
(Section 9.14).
9.4 Test Organisms
9.4.1 The organisms should appear healthy,
behave normally, feed well, and have low mortality
in test controls (s20%). The species of test
organisms should be positively identified. Test
organisms should not show signs of disease or
apparent stress (e.g., discoloration, unusual
behavior).
9.5 Water
9.5.1 The quality of water used for organism
culturing and testing is extremely important.
Overlying water used in culturing, holding,
acclimation, and testing organisms should be
uniform in quality. Acceptable water should allow
satisfactory survival, growth or reproduction of the
test organisms. L plumulosus should not show
signs of disease or apparent stress (e.g.,
discoloration, unusual behavior). See Section 7
for additional details.
9.6 Sample Collection and Storage
9.6.1 Sample holding times and temperatures
should conform to conditions described in
Section 8.
9.7 Test Conditions
9.7.1 It is desirable to measure temperature
continuously in at least one chamber during each
test. Temperatures should be maintained within
the limits specified in Section 11. DO, temperature,
salinity, ammonia, and pH should be checked as
prescribed in Section 11.3.
9.8 Quality of Test Organisms
9.8.1 It may be desirable for laboratories to
periodically perform 96-h water-only reference-
toxicity test to assess the sensitivity of culture
organisms (Section 9.16). Data from these
reference-toxicity tests could be used to assess
genetic strain or life-stage sensitivity to select
chemicals. The previous requirement for
laboratories to conduct monthly reference -toxicity
tests (USEPA, 1974a; 1994d) has not been
included as a requirement for sediments due to the
inability of reference-toxicity tests to identify
stressed populations of test organisms (McNultyet
al., 1999; McGee et al., 1998). Physiological
measurements such as lipid content might also
provide useful information regarding the health of
the cultures.
9.8.2 Test animals should only be obtained from
cultures. It is likely to be impractical to obtain test-
sited neonates directly from a supplier because of
their sensitivity to physical disturbances and their
rapid growth. Instead, test laboratories will likely
want to establish their own cultures of
L. plumulosus from which to harvest neonates. It
is desirable to determine the sensitivity of
L plumulosus obtained from an outside source.
For cultured organisms, the supplier should
provide data with the shipment describing the
history of the sensitivity of organisms from the
same source culture. For field-collected
organisms, the supplier should provide data with
the shipment describing the collection location, the
time and date of collection, the water salinity and
temperature at the time of collection, and collection
site sediment for holding and acclimation
purposes. The supplier should also certify the
species identification of the test organisms and
provide the taxonomic references (e.g.,
Schoemaker, 1932; Bousfield, 1973)orname(s)of
the taxonomic expert(s) consulted.
37
-------
9.8.3 All organisms in a test must be from the
same source (Section 10.2.2) (Table 10.1).
Organisms may be obtained from laboratory
cultures or from commercial or government
sources. The test organisms used should be
identified using an appropriate taxonomic key, and
verification should be documented. The use of
field-collected amphipods to start cultures is
discussed in Section 10.4. Obtaining organisms
from wild populations is useful for enhancement of
genetic diversity of existing cultures or to establish
new cultures. McGee et al. (1998) found seasonal
variability in sensitivity to cadmium in field-
collected L plumulosus. Therefore field-collected
organisms should not be used for toxicity testing
unless organisms are cultured through several
generations in the laboratory. In addition, the
ability of the wild population of sexually
reproducing organisms to cross-breed with the
existing laboratory population should be
determined (Duan et al., 1997). Sensitivity of the
wild population to select contaminants (see
Section 9.16.4) should also be documented.
9.9 Quality of Food
9.9.1 Problems with the nutritional suitability of the
food will be reflected in the survival, growth, or
reproduction of L. plumulosus in cultures or in
sediment tests.
9.10 Test Acceptability
9.10.1 Test acceptability requirements related to
these endpoints are provided in Table 11.3. Test
acceptability requirements for the 28-d
L. plumulosus test are as follows: (1) survival at
28-d must equal or exceed 80% in the control
sediment and (2) measurable growth and
reproduction must be found in all replicates of the
negative control treatment. Additional
requirements for acceptability of the tests are
presented in Table 11.3. An individual test may be
conditionally acceptable if temperature, dissolved
oxygen, and other specified conditions fall outside
specifications, depending on the degree of the
departure and the objectives of the tests (see
Table 11.1). The acceptability of a test will depend
on the experience and professional judgment of
the laboratory analyst and reviewing staff of the
regulatory authority. Any deviation from test
specifications should be noted when reporting data
from a test.
9.11 Analytical Methods
9.11.1 All routine chemical and physical analyses
for culture and testing water, food, and sediment
should include established quality assurance
practices outlined in USEPA methods manuals
(1979a; 1979b; 1991c; 1994a; 1994b; 1994c;
1994d).
9.11.2 Reagent containers should be dated when
received from the supplier, and the shelf life of the
reagent should not be exceeded. Working
solutions should be dated when prepared and the
recommended shelf life should not be exceeded.
9.12 Calibration and Standardization
9.12.1 Instruments used for routine measurements
of chemical and physical characteristics such as
pH, DO, temperature, total ammonia, and salinity
should be calibrated before use according to the
instrument manufacturer's procedures as indicated
in the general section on quality assurance (see
USEPA Methods 150.1, 360.1, 170.1, and 120.1;
USEPA, 1979a). Calibration data should be
recorded in a permanent log.
9.12.2 The analytical balance(s) should be
checked with reference weights, which are at the
upper and lower ends of the range of weight
values used. A balance should be checked at the
beginning of each series of weighing, periodically
(such as every tenth weight) during a long series
of weighing, and after taking the last weight of
series.
9.13 Replication and Test Sensitivity
9.13.1 The sensitivity of sediment tests will depend
in part on the number of replicates/treatment, the
significance level selected, and the type of
statistical analysis. If the variability remains
constant, the sensitivity of a test will increase as
the number of replicates is increased. The
minimum recommended number of replicates for
the 28-d test with L plumulosus is five, which was
calculated by a cost-power analysis of test results
38
-------
(see Section 13.5.1.6; DeWitt et al., 1997b). The
minimum recommended number of replicates
varies with the objectives of the test and the
statistical method used for analysis of the data
(Section 12).
9.14 Demonstrating Acceptable
Performance
9.14.1 Intralaboratory precision, expressed as a
coefficient of variation (CV) of the range in
response for each type of test to be used in a
laboratory, can be determined by performing five
or more tests with different batches of test
organisms, using the same reference toxicant, at
the same concentrations, with the same test
conditions (e.g., the same test duration, type of
water, age of test organisms) and same data
analysis methods. This should be done to gain
experience for the toxicity tests and to serve as a
point of reference for future testing. A reference-
toxicity concentration series (^0.5) should be
selected that will provide partial mortalities at two
or more concentrations of the test chemical
(Section 8.3.3). Information from previous tests
can be used to improve the design of subsequent
tests to optimize the dilution series selected for
future testing.
9.14.2 Before conducting tests with contaminated
sediment, it is strongly recommended that the
laboratory conduct the tests with control
sediment(s) alone. Results of these preliminary
studies should be used to determine if use of the
control sediment and other test conditions (i.e.,
water quality) result in acceptable performance in
the test outlined in Table 11.1.
9.14.3 Laboratories should demonstrate that their
personnel are able to recover an average of at
least 90% of the organisms of a range of size
classes (including neonates) from whole sediment.
For example, test organisms could be added to
control sediment or test sediment and recovery
could be determined after 1 h (Tomasovic et al.,
1994).
9.15 Documenting Ongoing Laboratory
Performance
9.15.1 For a given reference-toxicity test,
successive tests should be performed with the
same reference toxicant, at the same
concentrations, in the same type of water,
generating LCSOs using the same data analysis
method (Section 12).
9.15.2 Outliers, which are data falling outside the
control limits, and trends of increasing or
decreasing sensitivity are readily identified. If the
reference-toxicity results from a given test falls
outside the "expected" range (e.g., +2 standard
deviations [SD]), the sensitivity of the organisms
and the credibility of the test results may be
suspect. In this case, the test procedure should
be examined for defects and should be repeated
with a different batch of test organisms.
9.15.3 A sediment test may be acceptable if
specified conditions of a reference-toxicity test fall
outside the expected ranges (Section 9.10.2).
Specifically, a sediment test should not be judged
unacceptable if the LC50 for a given reference-
toxicity test falls outside the expected range or if
mortality in the control of the reference-toxicity test
exceeds 10%. All the performance criteria outlined
in Table 11.3 must be considered when
determining the acceptability of a sediment test.
The acceptability of the sediment test would
depend on the experience and judgement of the
investigator and the regulatory authority.
9.15.4 Performance should improve with
experience, and the control limits should gradually
narrow as the statistics stabilize. However, control
limits of a mean +2 SD, by definition, will be
exceeded 5% of the time regardless of how well a
laboratory performs. For this reason, laboratories
that develop very narrow control limits can be
penalized if a test result that falls just outside the
control limits is rejected de facto. The width of the
control limits should be considered in decisions
regarding rejection of data (Section 13).
9.16 Reference Toxicants
9.16.1 Historically, reference-toxicity testing has
been thought to provide three types of information
relevant to the interpretation of toxicity test data:
(.1) an indication of the relative "health" of the test
organisms used in the test; (2) a demonstration
that the laboratory can perform the test procedure
in a reproducible manner; and (3) information to
39
-------
indicate whether the sensitivity of the particular
strain or population in use at the laboratory is
comparable to those in use in other facilities. With
regard to the first, recent work by McNulty et al.
(1999) and McGee et al. (1998) suggests that
reference-toxicity tests may not be effective in
identifying stressed populations of test organisms.
In addition, reference-toxicity tests recommended
for use with sediment toxicity tests are short-term,
water column tests, owing in part to the lack of a
standard sediment for reference-toxicity testing.
Because the test procedures for reference-toxicity
tests are not the same as for the sediment toxicity
test of interest, the applicability of reference-
toxicity tests to demonstrate ability to reproducibly
perform the sediment test procedures is greatly
reduced. Particularly for long-term sediment
toxicity tests, with L. plumulosus performance of
control organisms over time may be a better
indicator of success in handling and testing these
organisms (Section 11).
9.16.2 Although the requirement for monthly
testing has been removed in this manual, periodic
reference-toxicity testing should be conducted as
an indication of comparability of results among
laboratories (minimumly one test every six months
should be conducted to evaluate potential
differences in genetic strain of organisms). In
particular, reference-toxicity tests should be
performed when organisms are obtained from
outside sources, when there are changes in
culture practices, or when brood stock from an
outside source is incorporated into a culture.
9.16.3 In many instances, reference-toxicity tests
have been conducted every time the
L plumulosus 28-d test was run. This may
provide additional quality assurance data
regarding the toxicological sensitivity of the test
organism. However, the decision whether to
conduct reference-toxicity tests every time the
L. plumulosus 28-d test is run is dependent on the
goal of the study (Section 9.16.2).
9.16.4 Reference toxicants such as cadmium
(available as cadmium chloride [CdCy), and
ammonia, are suitable for use. Care must be taken
with cadmium due to its carcinogenic nature and
with ammonia because it is very labile. Use of
nonhazardous alternatives for reference toxicants
is recommended (Section 5.6.3). No one
reference toxicant can be used to measure the
sensitivity of test organisms with respect to
another toxicant with a different mode of action
(Lee, 1980). However, it may be unrealistic to test
more than one or two reference toxicants routinely.
9.16.5 Test conditions for conducting reference-
toxicity tests with L. plumulosus are outlined in
Table 9.1.
*v
9.16.6 Based on 96-h, water-only reference-
toxicity tests at 20%0 with neonate L. plumulosus,
one should expect a mean LC50 value for
cadmium of approximately 0.5 mg/L (range:
0.2 mg/L to 0.7 mg/L) and LC50 values for total
ammonia between 25 mg/L and 60 mg/L (DeWitt
et al., 1997a). At 5%o, one should expect a mean
LC50 value for cadmium of approximately
0.05 mg/L (range: 0.01 mg/L to 0.09 mg/L) and
LC50 values for total ammonia between 37 mg/L
and 53 mg/L (Emery et al., 1997; Moore et al.,
1997).
9.17 Record-Keeping
9.17.1 Proper record-keeping is essential to the
scientific defensibility of a testing program. A
complete file should be maintained for each
individual sediment test or group of tests on
closely related samples. This file should contain a
record of the sample chain-of-custody; a copy of
the sample log sheet; the original bench sheets for
the test organism responses during the sediment
test(s); chemical analysis data on the sample(s);
control data sheets for reference toxicants;
detailed records of the test organisms used in the
test(s), such as species, source, age, date of
receipt, and other pertinent information relating to
their history and health; information on the
calibration of equipment and instruments; water
quality monitoring records; test conditions used;
and results of reference-toxicity tests. Laboratory
data should be recorded immediately to prevent
the loss of information or inadvertent introduction
of errors into the record. Original data sheets
should be signed and dated by the laboratory
personnel performing the tests. For additional
detail, see Section 12.
40
-------
Table 9.1 Recommended Test Conditions for Conducting Reference-toxicity Tests
Parameter
1. Test Type:
2. Dilution series:
3. Toxicant:
4. Temperature:
5. Salinity:
6. Light quality:
7. Illuminance
8. Photoperiod:
9. Renewal of water:
10. Age and size of test organisms:
11. Test chamber:
12. Volume of water:
13. Number of organisms/chamber:
14. Number of replicate chambers/
treatment:
15. Aeration:
16. Dilution water:
17. Water quality monitoring frequency:
18. Test duration:
19. Endpoint:
20. Test acceptability:
Conditions
Static, water-only test
Control and at least 5 test concentrations (>0.5 dilution factor)
Cd, Ammonia
25°C+2°C
5%o or 20%o (±2%o), matched to salinity of 28-d sediment toxicity
test (Section 11.3.6.6)
Wide-spectrum fluorescent lights
500-1000 lux
16 h light: 8 h dark ,
None
size-selected: between Oi25 mm and 0.6 mm
250 mL to 1 -L glass beaker or jar
80% of chamber volume (minimum)
n = 20 if 1 per replicate; n = 10 (minimum) if >1 replicate
1 minimum; 2 recommended
Not recommended; but aerate as necessary to maintain >60%
DO saturation (>4.4 mg/L)
Culture water, surface water, site water, or reconstituted water
Salinity and pH, at beginning and end of test; temperature and
dissolved oxygen daily
96 h
Survival (LC50)
>90% control survival
41
-------
Section 10
Collection, Culture, and Maintaining of Test Organisms
10.1 Life History
10.1.1 Leptocheirus plumulosus
10.1.1.1 L. plumulosus is a burrow-building member of
the family Aoridae (Figure 10.1). It is an infaunal
amphipod found in subtidal portions of Atlantic
Coast brackish estuaries from Cape Cod,
Massachusetts, to northern Florida (Bousfield,
1973; DeWitt et al., 1992a). It is common in
protected embayments, but has been collected in
channels of estuarine rivers at water depths up to
13 m (Schoemaker, 1932; Holland et al., 1988;
Schlekat et al., 1992). In Chesapeake Bay,
densities of L plumulosus can reach 24,000/m2 to
29,000m2 (Holland et al., 1988).
10.1.1.2 L. plumulosus is a relatively large
amphipod (adult length up to 13 mm) with a
cylindrically shaped body that is brownish-grey
incolor. A distinguishing feature is a series of dark
bands or stripes that cross the dorsal surface of
the pareons and pleons. It feeds on particles that
are in suspension and on the sediment surface
(DeWitt et al., 1988). Two studies have shown
that L plumulosus population abundance in
Chesapeake Bay is negatively correlated with
sediment contamination (Holland et al., 1988;
McGee and Fisher, 1997). Thus, this amphipod
would appear to be a good candidate to be an
environmental indicator.
A.
Figure 10.1 Leptocheirus plumulosus morphology (A) and characteristics used to determine the
gender (B-C) of the amphipod. A: Adult male L. plumulosus. B: First gnathopod of the
male (GN1), showing notched palm under dactyl. C: First gnathopod of female, showing flat
palm under pactyl. Illustration of L. plumulosus, by E.L. Bousfield, reproduced with
permission of the Canadian Museum of Nature, Ottawa, Canada.
42
-------
10.1.1.3 L. plumulosus is found in both oligohaline
and mesohaline regions of east coast estuaries;
ambient water salinity at collection sites has
ranged from 0%o to 15%o (Holland et al., 1988;
DeWitt et al., 1992a; Schlekat et al., 1992; McGee
et al., 1994). Laboratory studies have
demonstrated that L. plumulosus 28-d test can be
conducted at salinity values ranging from 1%o to
35%0 (Section 11.4.4; Schlekat et al., 1992; SAIC,
1993b; DeWitt et al., 1992a, 1997a; Emery et al.,
1997).
10.1.1.4 This amphipod is most often found in
fine-grained sediment with a relatively high
proportion of particulate organic material, although
it has been collected in fine sand with low organic
content (Jordan and Sutton, 1984; Holland et al.,
1988; Marsh and Tenor, 1990; DeWitt et al.,
1992a; Schlekat et al., 1992; McGee et al., 1994).
Laboratory studies with L. plumulosus revealed no
effect of sediment grain size on survival in control
sediment containing 5% to 100% silt-clay content
(DeWitt et al., 1997a). However, Emery et al.
(1997) found significantly reduced survival in
sediments in which clay content exceeded 84%.
10.1.1.5 Populations of L. plumulosus can be
seasonally ephemeral with major population
growth in fall and spring and large population
declines in the summer (Holland et al., 1988;
Marsh and Tenore, 1990; McGee, 1998). This
pattern appears to be driven by changes in
temperature and food availability and subsequent
effects on life history traits (Marsh and Tenore,
1990; McGee, 1998). Short-term population
fluctuations are also a function of the amphipod's
relatively short generation time (DeWitt et al.,
1992a). At 28°C, the age of the first brood release
is approximately 24 d (DeWitt et al., 1992a).
10.1.1.6 L plumulosus has been collected for
cultures from several areas in the Maryland portion
of Chesapeake Bay, including the Magothy,
Chester, Corsica, and Wye Rivers. Organisms
have been collected for culturing year-round from
the Magothy River subestuary of Chesapeake Bay
(C. Schlekat, University of South Carolina, and B.
McGee, U.S. Fish and Wildlife Service, Annapolis,
MD, unpublished data, personal communication).
10.2 General Culturing Procedures
10.2.1 Acceptability of a culturing procedure is
based in part on performance of organisms in
culture and in the sediment test (Section 1.4 and
9.2). No single technique for culturing test
organisms is required. What may work well for
one laboratory may not work as well for another
laboratory. Although a variety of culturing
procedures are outlined in Section 10.3 for
L. plumulosus, organisms must meet the test
acceptability requirements listed in Table 11.3.
10.2.2 All organisms in a test must be from the
same source. Organisms may be obtained from
laboratory cultures or from commercial or
government sources; a partial list sources is
provided in Table 10.1. The test organism used
should be identified using an appropriate
taxonomic key, and verification should be
documented (Section 9.8.2).
Table 10.1 Sources of Starter Cultures of Test
Organisms
Aquatic Biosystems, Inc.
1300 Blue Spruce Road, Suite C
Fort Collins, Colorado 80524
Scott Kellman
phone: 800/331-5916; fax: 970/484-2514
email: SRK@riverside.com
Chesapeake Cultures, Inc.
P.O. Box 507
Hays, Virginia 23702
Elizabeth Wilkins, President
phone: 804/693-4046; fax: 804/694-4703
email: growfish@c-cultures.com
website: www.c-cultures.com
Aquatic Research Organisms
P.O. Box 1271
Hampton, New Hampshire 03842-1271
Stan Sinitski or Mark Rosenqvist
phone: 800/927-1650; fax: 603/926-5278
website: www.arocentral.com
43
-------
Obtaining organisms from wild populations should
be avoided unless organisms are cultured through
several generations in the laboratory before use in
testing (Section 10.4). In addition, the ability of the
wild population of sexually reproducing organisms
to crossbreed with the existing laboratory
population should be determined (Duan et al.,
1997).
10.2.3 Test organisms obtained from commercial
sources should be shipped in well-oxygenated
water without sediment in insulated containers to
maintain temperature during shipment.
Temperature, salinity and DO of the water in the
shipping containers should be measured at the
time of shipment and on arrival to determine if the
organisms might have been subjected to low DO,
salinity change, or temperature and salinity
fluctuations. The temperature and salinity of the
shipped water should be gradually adjusted to the
desired culture temperature and salinity at rates
not exceeding 3"C or 3%o per 24 h.
10.2.4 A group of organisms should not be used
for a test if they appear to be unhealthy,
discolored, or otherwise stressed (e.g., >20%
mortality for 48 h before the start of a test). If the
organisms fail to meet these criteria, the entire
batch should be discarded and a new batch should
be obtained. All organisms should be as uniform
as possible in age and life stage. Test organisms
should be handled as little as possible. When
handling is necessary, it should be done as gently,
carefully, and as quickly as possible.
10.3 Culturing Procedure for
Leptocheirus plumulosus
10.3.1 The culturing method below is based on
procedures described in DeWitt et al. (1997a). A
periodic-renewal culture system is used. It
consists of culture bins that contain aerated water
over a thin (about 1 cm) layer of clean, fine-
grained sediment in which the amphipods burrow.
Culturing areas must be separate from testing
areas to avoid exposing the cultures to
contaminants. Before L. plumulosus are received
at a testing facility, appropriate permits or
approvals for import of live organisms should be
obtained, if necessary. If culturing is to occur in an
area where L. plumulosus are not indigenous to
local waters, precautions should be taken to
prevent release of living organisms to the outside
environment (Section 10.6). Test animals should
be destroyed at the end of toxicity test.
10.3.2 Starting a Culture
10.3.2.1 Amphipods for starting a laboratory
culture of L. plumulosus should be obtained from
a source with an established culture in which the
species has been verified (see Table 10.1 for
commercial sources of L. plumulosus).
Alternatively, L. plumulosus can be obtained from
field populations (see Section 10.4). Upon receipt
of amphipods, the temperature and salinity of the
water in shipping container(s) should be gradually
adjusted to 20°C and desired culture salinity, at
rates not exceeding 3°C or 3%o per 24-h period.
Feeding and regular maintenance should begin
once the acclimation period is complete. Separate
organisms into three size classes by gentle
sieving: adults (retained on 1.0-mm mesh),
subadults (pass through 1.0-mm mesh and
retained on 0.6-fnm mesh), and neonates (pass
through 0.6-mm mesh and retained on 0.25-mm
mesh). Seed each culture bin with approximately
equal numbers of adults, subadults, and neonates
to achieve a population density between 0.25/cm2
to 0.35/cm2 (2500/m2 to 3500/m2). Select only
actively moving, healthy-looking organisms.
Cultures should not be stocked at densities greater
than 0.5/cm2 (5000/m2). See Section 10.3.8.4 for
guidance on maintaining culture densities. Field-
collected organisms should be added periodically
to the culture population to maintain genetic
diversity of the cultured amphipods (see
also 9.8.3).
10.3.3 Culture Bins
10.3.3.1 Culture bins should be easy to maintain.
Plastic wash tubs (approximately 35 cm x 30 cm x
15 cm) have been used successfully by several
laboratories (DeWitt et al., 1992a). They are
relatively light when filled with water and sediment,
broad enough to allow for easy viewing of
amphipod burrows, easily cleaned, inexpensive,
and readily available. A wide variety of containers
and materials may work just as well for culturing
this species. New plasticware should be soaked
in running water for several days prior to use in the
44
-------
cultures to leach out potentially toxic compounds.
Previously used culture bins usually can be
satisfactorily cleaned using hot water and a scrub
brush or pad, without the use of a chemical
cleanser. Culture bins should not be washed with
soap or detergent except in extreme conditions. If
such a cleaning is deemed necessary, culture bins
must be rinsed and soaked thoroughly after
cleaning to remove residual cleanser.
10.3.4 Culture Sediment
10.3.4.1 Cultures should be established with a
thin layer (1 cm to 1.5 cm) of sediment spread on
the bottom of a culture bin. Sediment used for
culture purposes should be the same as the
control sediment used in sediment toxicity tests.
Suitable sources for culture sediment include the
amphipod collection site or an area adjacent to salt
marsh vegetation. Culture sediment should be
uncontaminated, organic-rich, fine-grained marine
or estuarine sediment that is not anoxic. The
organic carbon content (% TOC) should range
between 1.5% and 4%. The sediment should be
press-sieved through a 0.25-mm screen before
use to facilitate the harvesting of neonates and to
remove indigenous macroinvertebrates. Culture
sediment can also be wet sieved. Wet-sieving
involves agitating or swirling the sieve containing
sediment in water so that particles smaller than the
selected mesh size are washed through the sieve
into a container (ASTM, 2000a). The sieve may
be placed on a mechanical shaker, or the
sediments on the screen can be stirred with a
nylon brush to facilitate the process. Alternatively,
the particles may be washed through the sieve
with a small volume of running water. Culture
sediment can also be frozen (>48 h) to provide
additional assurance that viable
macroinvertebrates are not present. Frozen
sediment should be homogenized after thawing
and before use. Culture sediment can be stored
frozen for approximately 1 year.
10.3.5 Culture Water
10.3.5.1 Culture water used for holding and
acclimating test organisms and for conducting
toxicity tests should be of uniform quality and from
the same source. See Section 7.1.2 for
acceptable sources of water. Cultures of
L plumulosus are maintained at a salinity of either
5%o or 20%0. Culture salinity will depend on the
anticipated pore water salinity of test sediment and
desired overlying water salinity to be used in the
test (Section 11.3.6.6). To obtain these salinity
values, natural or reconstituted seawater should
be diluted with nonchlorinated well water,
deionized water, distilled water, or reverse-
osmosis water. Seawater and dilution water
should be filtered (<;5 m). Water that might be
contaminated with pathogens should be treated
shortly before use by filtration (sO.45 m), either
alone or in combination with ultraviolet sterilization.
DO, salinity, and pH should be checked before the
water,is used in cultures. Batches of salinity-
adjusted culture water can be held for
approximately 1 week; a lower holding
temperature (<20°C) helps maintain acceptable
water quality. Water depth in culture bins should
be at least 10 cm. Aeration, provided through an
air stone or pipet, should be moderate and
constant, but not so vigorous as to resuspend
sediment. Overlying water should be replaced the
day after a new culture is established; thereafter,
it should be renewed two or three times per week
(Section 10.3.7.2).
10.3.6 Temperature and Photoperiod
10.3.6.1 Cultures should be maintained at20°C to
25°C. The reproductive rate of L. plumulosus
increases at temperatures greater than 20°C,
necessitating more frequent culture thinning.
Higher temperatures also can promote unwanted
growth of nuisance- organisms (such as
nematodes, small worms, copepods, etc.).
Temperatures below 20°C may not foster
sufficiently prolific reproductive rates. Fluorescent
lights should be on a 16 h light : 8 h dark
photoperiod at a light intensity of 500 to 1000 lux.
An efficient procedure is to maintain long-term
cultures at 20°C, and increase culture temperature
to about 25°C a few weeks in advance of testing.
10.3.7 Food and Feeding
10.3.7.1 This method recommends the simplest
effective diet for routine use for L. plumulosus
culture: finely milled TetraMin® provided two or
three times per week. TetraMin® is a dry fish food
(flake or powder) widely available in retail pet
45
-------
stores. The food is prepared by milling, grinding,
or chopping the flakes to a fine powder. A small
flour mill, blender, or coffee grinder is useful for
this. Ground powder is then sifted through a
0.25-mm mesh screen, retaining and using only
the material that passes through the sieve. Use of
a respirator or fume hood will minimize aspiration
of dust. When establishing a new culture bin, do
not add food for 3 to 4 days after amphipods are
placed in new sediment. This will encourage the
organisms to consume labile organic matter in
sediment and to turn over the sediment by
burrowing and feeding.
10.3.7.2 Culture bins should be provided with food
in conjunction with water renewal. Two or three
times a week, approximately 60% of culture water
should be removed from each culture bin (by
decanting, siphoning, or pumping) and replaced
with the same volume of renewal water. Each
culture bin is provided with approximately 0.4 g of
dry food sprinkled evenly over the water surface,
or as a slurry in culture water two or three times
per week (e.g., Monday-Wednesday-Friday or
Monday-Thursday). The amount of dry food
added will depend on the density of each culture
bin. Newly started culture bins should receive
slightly less food (e.g., 0.3 g) than bins containing
mature cultures. Excess food can decompose
encouraging microbial and fungal growth on the
sediment surface deteriorating water quality.
10.3.7.3 Some laboratories have experienced
success in culturing L. plumulosus when other
food is provided (i.e., live microalgae or a mixed
dried food; DeWitt et al., 1992a). Modifications to
the diet can be used by laboratories in order to
optimize culture practices as long as performance
criteria are satisfied (Table 11.3).
10.3.7.4 One feeding alternative is to supply
renewal water consisting of seawater, cultured
phytoplankton, and deionized water combined to
the proper salinity and adjusted to an algal density
of approximately 10s cells/mL (DeWitt et al.
1992a). Proportions will vary depending upon the
salinity of the seawater and the density of the
cultured phytoplankton. Live algae also can be
used periodically to supplement a routine supply of
dry food. The algae used can include a single or
multiple species (e.g., Pseudoisochrysisparadoxa,
Phaeodactylum tricornutum, Isochrysis galbana,
Chaetoceros calcitrans, Skeletonema sp.,
Dunalicella tertiolecta, and/or Thallasiosirus spp.).
Other algal species might be used if it can be
demonstrated that they foster amphipod growth
and reproductive rates equal to those of the
aforementioned food alternatives. A mixture of
algal species is recommended, if live algae is
included in the diet.
10.3.8 Culture Maintenance
10.3.8.1 Observations and Measurements.
Cultures should be observed daily to ensure that
temperature is acceptable and aeration is
adequate in all culture bins. Inspection for the
presence of oligochaetes, polychaetes, copepods,
infaunal sea anemones, or chironomids should be
conducted weekly. The presence of excessive
densities of these or other competing or predatory
organisms should prompt renewal of culture
sediment after separating L plumulosus from the
invasive organisms. During routine maintenance,
cultures should be inspected for the presence of
microbial and fungal build-up on the sediment
surface. This build-up appears as a white or gray
growth that may originate near uneaten food.
Presence of microbial build-up may indicate that
more food is being provided than is required by the
amphipods. No additional food should be provided
to culture bins with surficial microbial build-up until
the build-up is no longer present. Sieving of
sediment and renewal of culture bins can expedite
removal of decaying material.
10.3.8.2 Healthy cultures are characterized by an
abundance of burrow-openings on the sediment
surface and turbid water from amphipod activity.
Although amphipods may leave their burrows to
search for food or mates, they will ordinarily
remain in their burrows during the illuminated
portion of the photoperiod. Amphipod density
should therefore only be estimated by examining
the number of burrow openings. An abundance of
organisms on the sediment surface (e.g., >15 per
culture bin) could indicate inadequate sediment
quality, low DO concentrations, or overcrowding.
A culture bin with an abundance of amphipods or
unhealthy individuals on the sediment surface
46
-------
should be examined closely, and the dissolved
oxygen concentration should be measured in the
overlying water. If the DO concentration is below
60% saturation (<4.4 mg/L), the culture bin should
be sieved, and the population and culture
sediment examined. If the population is too dense
(i.e., >1.5/cm2), the culture should be thinned as
described in Section 10.2.8.4. If the sediment
becomes an unacceptable habitat because it is
anaerobic or black and sulphidic below the
sediment surface, or contains an excess of
competitive or predatory organisms, the healthy
surviving amphipods should be placed in a new
culture bin with newly prepared culture sediment.
10.3.8.3 Water temperature and DO should be
measured in culture bins on a regular basis,
approximately every week. Cultures should be
continuously aerated. Salinity should be
measured after water renewal. Ammonia and pH
in overlying water should be measured with each
new batch of sediment before organisms are
added.
10.3.8.4 Renewal of Cultures. L plumulosus can
be prolific, and care must be taken to ensure that
culture bins do not get overcrowded. Amphipods
in overcrowded culture bins can be stressed
because of food and space limitations, causing the
fecundity of females to drop below five
eggs/brood/female and the abundance of
neonates and subadults to decline dramatically.
Culture density should not exceed 1.5
amphipods/cm2 and should ideally be maintained
at approximately 0.5 amphipods/cm2. A typical
indication of overcrowding is a fairly uniform size
distribution of amphipods (mostly small adults) and
the presence of only two to four eggs in the brood
pouches of gravid females. If sediment is not
replaced occasionally, the cultures may become
infested with undesirable species, such as worms
or copepods. These "pests" may compete for
food, bind sediment as fecal pellets, or produce
mucus, thereby reducing culture productivity or
increasing the effort required to harvest
amphipods. Field-collected organisms should be
added to the culture population periodically
(approximately annually) in order to maintain
genetic diversity of the culture organisms.
10.3.8.5 To avoid overcrowding, cultures should
be thinned every 6 to 8 weeks by sieving through
a 0.25-mm mesh screen to remove sediment.
Sediment can be used for a total of 2 to 4 months
before it should be replaced. Discard old
sediment, prepare new culture bins and sediment,
and restock each bin as described in
Section 10.3.2.
10.4 Field Collection
10.4.1 Although established cultures of
L plumulosus are the recommended source of
organisms for new cultures, it is recognized that
field collection of amphipods might be necessary
to enhance genetic diversity of existing cultures or
to establish new cultures at a laboratory. The
taxonomy of the organisms must be confirmed
before they are introduced into existing laboratory
populations. New organisms must be carefully
inspected, and all other species of amphipods
must be removed. The ability of a wild population
of sexually reproducing organisms to crossbreed
with existing laboratory populations of
L. plumulosus must be confirmed through long-
term culture maintenance (Duan et al., 1997).
Collection areas should- be relatively free of
contamination. Field collection of L. plumulosus
neonates for immediate use in a chronic toxicity
test is not recommended.
10.4.2 Collection Methods
10.4.2.1 L. plumulosus is subtidal and can be
collected with a small dredge or grab (e.g., Ponar,
Smith-Mclntyre, or Van Veen). In very shallow
water, sediment containing L plumulosus can be
collected with a shovel or scoop, or using a suction
dredge (DeWitt et al., 1992a).
10.4.2.2 All apparatus used for collecting, sieving,
and transporting amphipods and control-site
sediment should be clean and made of nontoxic
material. They should be marked "live only," must
never be used for working with formalin or any
other toxic materials, and should be stored
separately to avoid cross-contamination. The
containers and other collection equipment should
be cleaned and rinsed with distilled water,
deionized water, dechlorinated laboratory water,
reconstituted seawater, or natural seawater from
47
-------
the collection site or an uncontaminated seawater
source before use.
10.4.2.3 To minimize stress, amphipods should be
handled carefully, gently, quickly, and only when
necessary. Sieving should be performed by slow
immersion in collection-site water. Once sieved,
the amphipods should remain submersed in
collection-site water at the ambient temperature at
all times. Amphipods that are dropped or injured
should be discarded. Once separated from the
sediment, amphipods should not be exposed to
direct sunlight.
10.4.2.4 L. plumulosus can be isolated easily from
collection-site sediment by gentle sieving. Ideally,
amphipods will be separated into adults, subadult,
and neonates as described in Section 10.3.2. To
reduce field processing time, 1.0-mm and 0.6-mm
mesh sieves can be used to isolate adults and
subadults with which to start a culture. Sediment
passing through the 0.6-mm sieve could be
temporarily used for holding until further
processing of the sediment is practical. The final
sieving of collection-site sediment through
0.25-mm mesh can be deferred until materials are
returned to the laboratory. Collection-site water
should be used to sieve sediment in the field.
10.4.2.5 No sediment should be placed in
transport containers, with collection-site water.
Detritus and predators recovered by sieving
should be removed, and the collected amphipods
should be gently washed into the transport
containers with collection-site water. An adequate
portion of collection-site sediment should be
returned with the amphipods to serve as both
laboratory holding sediment and control sediment
in toxicity tests.
10.4.2.6 Water salinity and temperature at the
surface and bottom of the collection site should be
measured and recorded.
10.4.2.7 During transport to the laboratory,
amphipods should be held at or slightly below the
collection-site temperature. Containers of
amphipods and sediment should be transported to
the laboratory in coolers; ice-packs might be
necessary to maintain temperature. The water in
the containers of amphipods should be aerated if
transport time exceeds 1 h.
,10.4.2.8 Holding and acclimation procedures are
the same as those described in Sections 10.3.2
through 10.3.7 for initiation of a culture.
10.4.3 Shipping Methods
10.4.3.1 It is critical that demonstrated shipping
methods are used to ensure that organisms arrive
in a healthy condition. Additionally, the amphipods
that are received by a laboratory should meet the
shipping acceptance criteria recommended in
Section 10.4.4.3.
10.4.3.2 L. plumulosus should be shipped in water
only. Care must be taken to select containers with
a firm seal that is not easily broken in shipment.
The containers are filled to the top with well-
aerated water. No more than 100 amphipods/L
should be added to each container. For shipping,
scalable plastic bags, cubitainers, and other
sealable plastic containers can be used. The
containers should be filled with well-aerated
collection-site water or culture water before they
are sealed. The double packing bags should be
placed in a container that has a protective layer of
material (i.e., Styrofoam or newspaper) sufficiently
thick to prevent excessive movement with an
underlayer of ice packs. The shipping container
should be marked to prevent it from being inverted.
10.4.3.3 Performance Criteria for Shipped
Amphipods. The process of ensuring the
availability of healthy amphipods for starting
cultures begins before the organisms arrive in the
laboratory from the supplier. • It is desirable to
assess the quality and acceptability of each batch
of shipped amphipods using the criteria that follow.
Biological criteria should include an exhibition of
active swimming, crawling, or burrowing behavior
upon placement in water, and an acceptable color.
L. plumulosus should be brownish or orangish-
gray. Mortality among the shipped organisms
should not exceed 5%. The shipping containers
should arrive intact, and the temperature of water
in shipping containers should be between 10°C
and 20°C. Information on culture conditions,
including at least temperature and salinity, should
be provided by the supplier. Finally, a quantity of
48
-------
collection-site sediment should be included as
substratum for amphipods for use during the
acclimation period, in culturing, and in toxicity tests
as control sediment. It might be desirable for the
testing facility to specify these criteria for the
supplier. If these criteria are not met, the
organisms might have experienced stress during
shipment, and culturing success might be
jeopardized.
10.5 Obtaining Amphipods for Test
10.5.1.1 The cultures usually can be harvested
approximately 4 to 5 weeks after initiation or up
until the cultures are thinned and renewed (6 to 8
weeks after initiation). Neonates used for testing
may be selected on the basis of size or age. For
size-selected neonates, the contents of culture
bins are gently sieved through 0.60-mm and 0.25-
mm screens. Juveniles passing through the 0.6-
mm mesh and retained on the 0.25-mm mesh are
used for testing, and individual neonates typically
have a dry weight of approximately 0.03 mg to
0.06 mg and body length of approximately 1.3 mm
to 1.7 mm. Culture bins of about 35 cm X 30 cm
typically produce at least 300 to 400 neonates with
a healthy culture. Selecting neonates for testing
.based on size is the preferred option for method
comparability. For age-selected neonates, gravid
females are isolated from cultures 5 d before test
initiation. Gravid females are placed in separate
culture bins with sediment and are fed. Two days
prior to test initiation, these females are then
transferred to bins containing only water (at 25°C
and 5%o or 20%0). On the day of test initiation, the
contents of these bins are gently passed through
a 1-mm screen on which adults are retained.
Neonates that pass through this screen are
transferred to a shallow glass container for sorting.
Special care must be taken to ensure that the
neonates are handled gently, selecting and
transferring them with wide-bore pipets only, and
maintaining the water temperature and salinity
within recommended test conditions.
10.5.1.2 Approximately one-third more amphipods
than are needed for the test should be sieved from
the sediment and transferred to a sorting tray. The
additional organisms allow for the selection of
healthy, active individuals. Organisms not used in
toxicity tests can be used to establish new
cultures.
10.5.2 Acceptability of Organisms
10.5.2.1 Amphipods placed in the holding bins
should be active and healthy. Sluggish or
apparently dead individuals should be discarded.
If greater than 5% of the amphipods in the holding
bins appear unhealthy or are dead, the entire
group should be discarded and not used in tests.
10.6 Minimization of Risk of Release of
Nonindigenous Organisms
10.6.1 If L. plumulosus is not endemic to the local
estuarine environment, containment and water
treatment procedures should be implemented to
minimize the chance of accidental release of
organisms or pathogens to local waters. The
same precautions might also be required if the
culture population of L. plumulosus is not derived
from local sources. Some local or state authorities
might require special permits and procedures to
allow receipt and culturing of nonindigenous
species. Containment and treatment policies and
procedures could include the procedures
described below. All test animals should be
destroyed at the end of toxicity tests.
10.6.2 Culturing and holding of the amphipods
should only occur in specially designated
laboratory areas, separate from those used to
hold, culture, or experiment with native species.
These areas should have no access to drains
leading directly to local surface waters. Handling
of nonindigenous species should be limited
to trained and authorized personnel. The
amphipods should be cultured in a static-renewal
manner to minimize the amount of water that must
be treated. Any seawater removed from culture
bins should be treated with chlorine bleach or
ozonation to kill any escaping organisms and
pathogens. All equipment and labware used to
culture or handle the amphipods should be
cleaned thoroughly. Any excess or dead
amphipods should be placed in bleach or treated
by ozonation or heat killed (boiling water) to
ensure they are killed prior to disposal as sanitary
waste.
49
-------
Section 11
Leptocheirus plumulosus 28-d Chronic Test for Sediment
11.1 Introduction
11.1.1 Leptocheirus plumulosus has been used
extensively to test the toxicity of estuarine and
marine sediment. The choice of this amphipod
species as a test organism is based on sensitivity
to sediment-associated contaminants, availability
and ease of collection and culturing, tolerance of
environmental conditions (e.g., temperature,
salinity, grain-size), ecological importance, ease of
handling in the laboratory, and ease of measuring
test endpoints. Additionally, this species is
intimately associated with sediment by nature of its
burrowing and feeding habits. L plumulosus is
tolerant of salinity values between >1%o to 35%o
and sediment from fine- to coarse-grained. Field
validation studies have shown that amphipods are
absent or have reduced abundances at sites
where toxicity has been demonstrated in
laboratory tests. Amphipod sediment toxicity tests
have been successfully performed for regulatory
and research purposes by numerous laboratories,
including state and federal government agencies,
private corporations, and academic institutions
(see Section 1 for additional details).
11.1.2 Guidance for L. plumulosus has been
developed previously (ASTM, 2000c; USEPA,
1994d). Most standard whole sediment toxicity
tests have been developed to produce a survival
endpoint with potential for a sublethal endpoint
(reburial) with some species. Methods that
measure sublethal effects have either not been
previously available or used routinely to evaluate
sediment toxicity (Craig, 1984; Dillon and Gibson,
1986; Ingersolland Nelson, 1990; Ingersoll, 1991;
Burton et al., 1992). Most assessments of
contaminated sediment rely on short-term lethality
testing methods (e.g., <;10 d; USEPA-USACE,
1991, 1998). Short-term lethality tests are useful
in identifying "hot spots" of contamination, but may
not be sensitive enough to evaluate moderately
contaminated areas. However, sediment quality
assessments using sublethal responses of benthic
organisms, such as growth and reproduction, have
been used to successfully evaluate moderately
contaminated areas (Scott, 1989; Niewlony et al.
1997;DeWittetal. 1997c).
11.1.3 The 28-d toxicity test with L plumulosus is
a test with a lethality endpoint and two sublethal
endpoints: growth and reproduction. These
sublethal endpoints have potential to provide a
toxic response to chemicals that might not cause
acute effects or significant mortality in a test.
Sublethal response in 28-d exposures is also
valuable for population modeling of contaminant
effects. This data can be used for population-level
risk assessments of benthic pollutant impacts.
11.1.4 Section 11.2 describes guidance for
conducting the 28-d test with L plumulosus that
can be used to evaluate the effects sediment
contaminants on survival, growth, and
reproduction. Refinement of these methods may
be described in future editions of this manual, after
additional laboratories have successfully used this
method (Section 13.5). These methods are based
on procedures described in DeWitt et al. (1997a;
1997b) and Emery et al. (1997).
11.1.5 Results of tests using procedures different
from the procedures described in Section 11.2
may not be comparable, and these different
procedures may alter contaminant bioavailability.
Comparisons of results obtained using modified
versions of these procedures might provide useful
information concerning new concepts and
procedures for conducting sediment tests with
estuarine and marine organisms. If tests are
conducted using procedures different from those
described in this manual, additional tests are
required to determine comparability of results
(Section 1.3).
50
-------
11.2 Procedure for Conducting a
Leptocheirus plumulosus 28-d
Test for Measuring Sublethal
Effects of Sediment-associated
Contaminants
11.2.1 Recommended conditions for conducting
a 28-d chronic sediment toxicity test with
L. plumulosus are summarized in Table 11.1. A
general activity schedule is outlined in Table 11.2.
Decisions concerning the various aspects of
experimental design, such as the number of
treatments and water quality characteristics,
should be based on the purpose of the test and the
methods of data analysis (Section 12).
11.2.2 The 28-d chronic sediment toxicity test with
L. plumulosus is conducted at 25°C and a salinity
of either 5%o or 20%o with a 16 h light : 8 h dark
photoperiod at an illuminance of about 500 to 1000
lux (Table 11.1). Test chambers are 1-L glass
chambers containing 175 mL of sediment and
about 725 mL of overlying seawater. Twenty
neonate amphipods are added to each test
chamber at the start of a test. Five replicate test
containers per treatment are recommended for
routine testing (Section 13.5.1.2). Exposure is
static-renewal with water exchanges and feeding
three times per week, on Monday, Wednesday,
and Friday. The test organisms are fed after water
renewals. Overlying water can be culture, surface,
site, or reconstituted water adjusted to the test
salinity. For site-specific evaluations, the
characteristics of the overlying water should be as
similar as possible to the site where sediment is
collected. Requirements for test acceptability are
summarized in Table 11.3.
11.3 General Procedures
11.3.1 Sediment into Test Chambers
11.3.1.1 The day before the addition of amphipods
(Day-1), each test sediment, including control and
reference sediment, should be homogenized
among replicate beakers. This can be achieved
by mixing, by stirring manually, or by using a
rolling mill, feed mixer, or other apparatus
(Section 8.3.1.2) or by serially spooning out small
aliquots of sediment to each test chamber. If a
quantitative confirmation of homogeneity is
required, replicate subsamples should be taken
from the sediment batch and analyzed for TOG,
chemical concentrations, and particle size.
Ammonia can be measured in the pore water.
11.3.1.2 A 175-mL aliquot of sediment is added to
each test chamber with five replicates per
sediment treatment. It is important that an
identical volume be added to each replicate test
chamber; the volume added should provide a
sediment depth of 2 cm in the test chamber. The
sediment added to the test chamber should be
settled by tapping the bottom or side of the test
chamber against the palm of the hand or another
soft object. Alternatively, sediment can be
smoothed with a nylon, fluorocarbon, glass, or
polyethylene spatula. Sediment known or
suspected to be contaminated should be added to
test chambers in a certified laboratory fume hood.
11.3.2 Addition of Overlying Water
11.3.2.1 The procedure for addition of overlying
water should not suspend significant portions of
the sediment in test chambers. A turbulence
reducer can be used to minimize disruption of
sediment as test water is added. The turbulence
reducer can be either a disk cut from polyethylene,
nylon, or Teflon® sheeting (4 to 6 mil) attached to
a nylon monofilament line (or nontoxic equivalent),
or a glass or plastic plate attached (open face up)
to a glass or plastic rod. The turbulence reducer
needs to fit inside the test chamber. It is
positioned just above the sediment surface and
raised as water is added. It is convenient to mark
each test chamber on the side at 900 mL and to fill
with water to reach the mark. A turbulence
reducer can be rinsed with clean water between
replicates of a treatment, but a separate
turbulence reducer should be used for each
treatment. The test chambers should be covered,
and placed in a temperature controlled water bath
(or acceptable equivalent) in randomly assigned
positions. Aeration is started when suspended
sediment has settled (often overnight). A test
.begins when the test organisms are added to the
test chambers (Day 0).
51
-------
Table 11.1 Test Conditions for Conducting a 28-d Sediment Toxicity Test with Leptocheirus plumulosus
Parameter Conditions
1. Test type:
2. Test sediment grain size:
3. Test sediment pore water salinity:
4. Overlying water salinity:
5. Test sediment pore water ammonia:
6. Test sediment pore water sulfides:
7. Temperature:
8. Light quality:
9. Illuminance:
10. Photoperiod:
11. Test chamber:
12. Sediment volume:
13. Sediment preparation:
14. Overlying water volume:
15. Renewal of overlying water:
16. Source:
17. Life stage and size:
18. Number test organisms/chamber:
19. Number of replicate chambers/
treatment:
20. Diet:
21. Feeding schedule:
22. Aeration and dissolved oxygen (DO):
23. Overlying water:
Whole sediment toxicity test, static-renewal
>5% silt and clay to <85% clay
1%oto35%o
Daily limits: 5%o (±3%o) if pore water is 1%o to 10%o, 20%o (+3%o) if
pore water is >10%o to 35%0; 28-d mean: 5%o (±2%o) or 20%o
(±2%o)
< 60mg/L (total mg/L, pH 7.7); < 0.8 mg/L (unionized mg/L, pH
7.7)
Not established.
Daily limits: 25'C (±3'C); 28-d mean: 25'C (±2'C)
Wide-spectrum fluorescent lights
500-1000 lux
16 h light: 8 h dark
1-L glass beaker or jar with 10-cm inner diameter
175 mL (about 2 cm depth)
Press-sieved through 0.25-mm (see Section 4.3.2.3)
Fill to 900 mL mark in test chamber (c.725 mL H2O)
3 times per week: siphon off and replace 400 mL
Laboratory cultures
Neonates: age-selected (<48 h old) or size-selected; retained
between 0.25-mm and 0.6-mm mesh screens
20
5 for toxicity test; >2 dummy chambers for pore water
ammonia (Day 0 and Day 28)
Days 0-13, 20 mg TetraMin® per test chamber; Days 14-28, 40 mg
TetraMin® per test chamber
3 times per week (M-W-F) after water renewal.
aerate constantly with tickle flow of bubbles
Daily limits: >3.6 mg/L (50% saturation)
28-d mean: >4.4 mg/L (60% saturation)
Clean seawater, natural or reconstituted water; same source as
used for culturing.
52
-------
Table 11.1 (continued)
Parameter
Conditions
24. Overlying water quality and monitoring
frequency:
25. pH:
26. Pore water quality:
27. Test duration:
28. Test organism observations:
29. Endpoints:
30. Test acceptability:
Daily temperature in water bath or test or dummy chamber, daily
min/max recommended; salinity, temperature, DO, and pH at test
initiation and termination, and in one replicate per sediment
treatment preceding water renewal during the test (three times per
week); aeration rate daily in all containers; total ammonia on Days
0 and 28 in one replicate per treatment.
7.0 to 9.0 pH units
Total ammonia, salinity, temperature, and pH of pore water from
surrogate containers on Days 0 and 28; recommended in bulk
sediment prior to testing.
28 d
Observe condition and activity in each test chamber preceding
water renewal (3 times per week).
Survival, growth rate, and reproduction.
Minimum mean control survival of 80%, growth and reproduction
measurable in all control replicates, and satisfaction of
performance-based criteria outlined in Table 11.3.
53
-------
Table 11.2 General Activity Schedule for Conducting a 28-d Sediment Toxicity Test with
Leptochelrus plumulosus
Day
Preparation
Pretest
Pretest
Day-1
Initiation
Day 0
Activity
Start or renew cultures approximately 6 to 8 weeks in advance of test initiation. Increase culture
water temperature to about 25°C approximately 2 weeks in advance of test initiation.
Determining pore water salinity of test sediment and acclimate L. plumulosus cultures to overlying
water salinity to be used in testing.
Layer sediment in test chambers, add overlying water. Measure pore water total ammonia in bulk
sediment and begin purging procedures, if appropriate (Section 11.4.5). Measure tare weight of
weigh boats for dry weights. Set up positive control reference-toxicity test chambers if
appropriate.
Measure pore water total ammonia, temperature, salinity, and pH from dummy jar. Measure
salinity, temperature, DO, and pH in all test chambers. If water quality parameters are within test
ranges, proceed with initiation; if not, correct problem and re-measure water quality. Obtain
neonate test organisms, initiate test, and initiate positive control reference toxicant test if
conducted. Only feed if a Monday, Wednesday, or Friday. Prepare 3 sets of 20 neonates for
initial weight of growth rate endpoint; rinse in deionized water; dry overnight at 70°C, and weigh
or measure length on Day 1 or later.
Positive Control Reference-toxicitv Test
Day 1 to 3
Day 4
Measure and record water quality parameters in one replicate test chamber from each positive
control treatment.
Measure water quality parameters and record observations of amphipod activity in all positive
control test chambers. Terminate the positive control references-toxicity control test if conducted.
Maintenance of 28-d Test
Daily
Check aeration in all test chambers and test temperature (water bath, environmental chamber,
or dummy chamber). If aeration is interrupted in a test chamber, measure and record DO prior
to resumption of aeration. Check photoperiod controllers.
3 Times per Week Measure water quality in one replicate test chamber per sediment treatment. Record
(M-W-F) observations of amphipod activity and condition of sediment and water in all test chambers.
Siphon off and replace 400 mL of water in all test chambers. Add food to all test chambers.
Termination of 28-d Test
Day 28 Measure salinity, temperature, DO, and pH in all test chambers. Measure tare weight of weight
boats for dry weight measurements. Terminate 28-d test: sieve adults and offspring from
sediment, count surviving adults, prepare adults for drying, and dry to constant weight at 70°C.
Count offspring, or preserve and stain offspring.
Day 29 or later Measure dry weight or length of adults. If offspring were preserved, count them.
54
-------
Table 11.3 Test Acceptability Requirements for a 28-d Sediment Toxicity Test with Leptocheirus plumulosus
A. It is recommended for conducting the 28-d test with L. plumulosus that the following performance criteria be met:
1. Neonate L. plumulosus, size-selected (retained between 0.25-mm and 0.6-mm screens) or age selected
(<24h old), are used to initiate the test(s).
2. Average survival of amphipods in the negative control sediment must be greater than or equal to 80% at
the end of the test, with no single replicate having 60% survival or less.
3. Measurable growth and reproduction should be observed in all replicates of the negative control treatment.
4. The time-weighted average of daily temperature readings must be within +2°C of the desired temperature.
The instantaneous temperature must always be within ±3°C of desired temperature.
5. The time-weighted average of daily salinity readings must be 5%o ±2%o or 20%o ± 2%o. The instantaneous
salinity readings must always be 5%o ±3%o or 20%o ± 3%o.
B. Performance-based criteria for culturing L. plumulosus include the following:
1. Laboratories should perform periodic 96-h water-only reference-toxicity tests (at a minimum, one test every
six months) to assess the sensitivity of culture organisms (Section 9.16).
2. Records should be kept on the frequency of restarting cultures.
3. Laboratories should record the pH and ammonia of the culture water at least quarterly. Dissolved oxygen
and salinity should be measured weekly. Temperature should be recorded daily.
4. Laboratories should characterize and monitor background contamination and nutrient quality of food if
problems are observed in culturing or testing organisms.
C. Additional requirements:
1. A negative-control sediment and appropriate solvent controls must be included in a test. The concentration
of solvent used must not adversely affect test organisms.
2. All organisms in a test must be from the same source.
3. All test chambers should be identical and should contain the same amount of sediment and overlying
water.
4. Natural physico-chemical characteristics of test sediment collected from the field should be within the
tolerance limits of the test organisms.
5. Storage of sediments collected from the field should follow guidance outlined in Section 8.2.
6. Salinity, pH, and DO, in the overlying water, ammonia in pore water and test sediment grain size should
be within test condition limits of the test species (Table 11.1), or else effects of the variables need to be
considered during interpretation of test results.
55
-------
11.3.3 Initial Measurements
11.3.3.1 On Day 0, water quality must be
measured in all test chambers prior to adding
amphipods to test chambers. If any water quality
parameter is outside acceptable limits
(Table 11.1), correct the problem in all replicate
containers of that treatment, re-measure water
quality parameters, and continue test initiation if
water quality values are acceptable. Aberrant pH
values might be caused by characteristics of
certain sediments and therefore may be
impractical to correct.
11.3.4 Acclimation
11.3.4.1 Test organisms should be cultured at a
temperature near 25°C. Amphipod cultures held
below 23°C need to be acclimated to test
temperature of 25°C (±3°C) before test initiation.
Ideally, test organisms should be cultured in the
same water that will be used in testing.
11.3.4.2 Occasionally there is a need to perform
evaluations at temperatures or salinity's different
than those recommended in Table 11.1. Under
these circumstances, it may be necessary to
acclimate organisms to the desired test
temperature or salinity to prevent thermal shock
that could result when organisms are moved
immediately from the culture temperature or
salinity to the test temperature or salinity (ASTM,
2000a). Reproduction and growth rates in cultures
may be greatly reduced at temperatures <20°C.
However, reproduction and growth is not effected
by salinity's ranging from 5%o and 20%o (DeWitt et
al., 1997a). Acclimation can be achieved by
exposing organisms to a gradual change in
temperature or salinity; however, the rate of
change should be relatively slow to prevent shock.
A change in temperature or salinity not exceeding
3°C or 3%o per 24-h period is strongly
recommended (see Sections 10.2.3 and 10.3.2).
Tests at temperatures other than 25°C need to be
preceded by studies to determine expected
performance under alternate conditions.
11.3.5 Addition of Amphipods
11.3.5.1 The test is initiated when amphipods are
added to the test chambers. See Section 10.5 for
procedures for obtaining neonates for testing.
Amphipods should be randomly selected and
placed in transfer containers (small dishes or eye
cups) containing a small amount of test water. The
number of amphipods in each dish should be
verified by recounting before organisms are added
to test chambers. To facilitate recounting,
amphipods may be distributed to test chambers in
batches of 5 or 10 instead of the full complement
of 20. Because neonates are very small, extreme
caution should be taken to ensure that each test
chamber receives all 20 amphipods at test
initiation. The distribution of amphipods to the test
chambers needs to be done in a randomized
fashion. Animals need to be added to test
chambers as soon as possible following their
collection to minimize handling stress and
exposure to temperature changes. Three
randomly selected sets of 20 neonates for initial
weight determination needs to be set aside during
initiation of the test.
11.3.5.2 To facilitate the initiation process, aeration
should be stopped in test chambers immediately
prior to adding the neonates. Sediment in test
chambers should not be disrupted during the
initiation procedure. Neonates from a transfer
container should be poured into a test chamber.
Any neonates remaining in transfer containers can
be washed immediately into the test chamber
using a gentle stream of water at appropriate
temperature and salinity. Neonates trapped at the
water's surface can be submerged by using a
blunt probe or by gently dribbling a few drops of
test or culture water onto the amphipod from
above. A disk of 6-mil polyethylene, nylon, or
Teflon® can be used on the water surface to
minimize disruption of the sediment surface, if
necessary. Rinse the disk after amphipods are
added to ensure that none have stuck to the disk.
The disk should be removed once the amphipods
have been introduced. A separate disk should be
used for each treatment to avoid cross
contamination. Aeration is continued after
amphipods are added to test chambers.
11.3.5.3 After the test organisms have been
added, the test chambers should be examined for
individuals that did not burrow into the sediment
and might have been stressed or injured during the
isolation, counting, or initiation processes. Injured
56
-------
or stressed test organisms will not burrow into
sediment and should be removed. Neonates that
have not burrowed within 1 h should be replaced
with test organisms from the same sieved
population, unless they are repeatedly burrowing
into the sediment and immediately emerging in an
apparent avoidance response. In that case, the
amphipods are not replaced. The number of
amphipods that are replaced in each test chamber
needs to be recorded.
11.3.6 Test Conditions
11.3.6.1 Test limits for the 28-d L. plumulosus
test are provided in Table 11.1. Test sediments
with characteristics that exceed these limits are
subject to noncontaminant effects that needs to be
considered during interpretation of test results.
11.3.6.2 Aeration. The overlying water in each
test chamber needs to be aerated continuously
after an initial settling period, except during
introduction of the test organisms. Filtered, dry,
clean air should be bubbled through a glass or
plastic pipet via plastic tubing (about 3
bubbles/sec). The tip of the pipet should be
suspended 2 cm to 3 cm above the surface of the
sediment so that it does not disturb the sediment
surface. The concentration of DO in the water
overlying the 'sediment in the test chambers is
maintained at or near saturation by gentle
aeration. Ideally, air is bubbled through the water
at a rate that maintains a high percentage of
saturation (e.g., about 90%) but does not disturb
the sediment surface. If air flow to one or more
test chambers is interrupted (i.e., for more than
1 h), DO should be measured in those test
chambers to determine whether DO
concentrations have fallen below 4.4 mg/L. The
28-d mean should be >4.4 mg/L DO, and daily DO
measurements should be >3.6 mg/L (50%
saturation). Results may be unacceptable for test
chambers in which aeration is interrupted or DO
concentrations fall to below 50% of saturation.
11.3.6.3 Lighting. Laboratory lighting should be
maintained on a 16 h light: 8 h dark photoperiod
cycle throughout the test at an intensity of 500 to
1000 lux.
11.3.6.4 Feeding and Water Renewal. A
TetraMin®-only diet is recommended for the 28-d
sediment toxicity test with L. plumulosus. With this
diet, 400 mL of overlying water is replaced three
times per week (Monday-Wednesday-Friday), after
which a TetraMin® slurry is delivered to each
chamber in 1-mL aliquots. Water removal and
replacement must be completed using procedures
that minimize disturbance to sediment in the test
chambers. Water can be removed by siphoning
through a tube with fine-meshed screening over
the intake to prevent uptake of amphipods. A
pump can also be used to remove water. Water
should not be poured from test chambers because
this practice can resuspend and disturb the
sediment. A separate turbulence-reducer should
be used for each treatment when water is replaced
to avoid cross contamination (see section 11.3.2).
TetraMin® is fed at a rate of 20 mg per test
chamber between Days 0-13 and 40 mg per test
chamber between Days 14-28. To prepare the
slurry, TetraMin® is finely ground with a food mill
(blender, mortar and pestle, or a similar device)
and sieved through a 0.25-mm screen. Test water
is added to the appropriate amount of TetraMin®,
and the slurry is mixed on a stir plate for 15 min.
Appendix A provides a sample calculation for
preparation of food rations. The slurry is prepared
fresh for each use and needs to be mixed
continuously during feeding to prevent the
TetraMin® from settling.
11.3.6.4.1 Laboratory experimentation has shown
that food ration can affect the response of test
animals to sediment-associated contaminants.
The food ration of TetraMin® recommended in this
protocol was evaluated with two other food rations
in an experiment in which test animals were
exposed to sediments spiked with PCB29 at
concentrations between 15 and 240 ppm
(T. Bridges, USAGE, personal communication).
The feeding rates evaluated at each PCB29
concentration included 30 mg/60 mg (Days 0-
13/Days 14-18), 20 mg/40 mg.and 10 mg/20 mg
per test chamber. Significant reductions in
survival and growth were evident only in the
highest PCB29 concentration for each of the food
rations. Decreased reproduction was also evident
at 240 ppm PCB29 at each food ration as well as
at 120 ppm for the 20 mg/40 mg and 10 mg/20 mg
rations (T. Bridges, USAGE, personal
communication). Given the generally lower
57
-------
reproductive rates observed at the lowest food
ration, the 20 mg/40 mg feeding rate is
recommended for use in this protocol.
11.3.6.5 Water Temperature. The test
temperature was selected to approximate
summertime temperature experienced by
L plumulosus in the wild (Holland et al., 1988;
McGee, 1998). The test temperature is 25°C with
a daily maximum range of ±3°C and a 28-d
weighted mean of 25°C ±2°C. Water used for
renewal of test chambers needs to be adjusted to
test temperature before use in renewals.
11.3.6.6 Salinity. The salinity of the water
overlying the test sediment should be 5%o ±3%o
(daily readings) when test sediment pore water
salinity is 1%o to 10%<>; an overlying water salinity
should be 20%o ±3%o when test sediment pore
water salinity is >10%o. Selection of which
overlaying water salinity should be based on the
pore water salinity of the samples to be tested. If
the suite of samples includes sediments with pore
water salinity values spanning the range of both
less-than and greater-than 10%o, use the
appropriate overlying water salinity for each
sample (i.e., 5%o or 20%o), and include control-
sediment treatments for both 5%o and 20%o
overlying water salinity values. The 28-d mean
salinity values should deviate no more than 2%0
from the recommended salinity (5%o or 20%o).
Pore water salinity of each test sediment should be
measured prior to the initiation of a test. Sediment
pore water can be measured in water overlying
sediment in sample containers before
homogenization of sediment. Alternatively, pore
water salinity can be obtained by centrifugation
(see Section 8.4.4.7).
11.3.7 Measurements and Observations
11.3.7.1 Temperature should be measured at
least daily in a dummy chamber or from the water
bath or environmental chamber. The temperature
of the water bath or a test chamber should be
continuously monitored with minimum and
maximum temperature recorded daily. A dummy
container identical to test containers is
recommended for continuous temperature
monitoring. The time-weighted average of daily
temperature readings must be 25°C ±2°C. The
instantaneous temperature must always be within
±3°C of the desired temperature.
11.3.7.2 Salinity, DO, temperature, and pH of the
overlying water should be measured three times
per week in at least one test chamber per
treatment before renewal of water.
11.3.7.3 Total ammonia should be measured in
overlying and pore water at test initiation (Day 0 or
Day -1 for pore water) and at test termination
(Day 28). Salinity, pH, and temperature should be
measured with each ammonia measurement.
Simultaneous measurements of ammonia, salinity,
pH, and temperature in sediment pore water
should be taken before test initiation. If test
sediments are sieved (Section 4.3.2.3), pore water
samples for ammonia should be collected before
and after sieving. Pore water can be obtained by
centrifugation or from overlying water in sample
containers (prior to pretest homogenization). If
ammonia levels exceed recommended limits
(Table 11.1), then ammonia reduction procedures
are advisable before test initiation. However, if
ammonia is the chemical of concern in the test
sediments, pore water ammonia concentrations
should not be deliberately manipulated.
11.3.7.4 Each test chamber should be examined
daily to ensure that airflow to the overlying water is
acceptable. Daily checks for amphipods trapped
at the water surface are recommended for the first
three days of a test. Amphipods caught in the air-
water interface should be gently pushed down into
the water using a blunt glass probe or drops of
dilution water. The number of amphipods
swimming in the water column and trapped in the
air-water interface should be noted and amphipods
submerged before each water renewal. The
number of apparently dead test organisms should
be noted, but organisms should not be removed or
otherwise disturbed during the test. Exuviae may
be mistaken for dead amphipods; therefore, care
should be taken in identifying animals as dead.
11.3.8 Ending a Test
11.3.8.1 The contents of each test chamber are
sieved to isolate the test organisms. The mesh
sizes for sieving the contents of the test chambers
is 0.5 to 0.6 mm to isolate adults and 0.25 mm to
isolate offspring. The 0.6-mm sieve should not be
58
-------
stacked atop the 0.25-mm sieve for this process.
Test water should be used for sieving. Material
retained on each sieve should be washed into a
sorting tray with clean test water. L plumulosus
are easily removed from the sediment by the
sieving process.
11.3.8.2 Material that has been washed from the
sieve into a sorting tray should be carefully
examined for the presence of amphipods. A small
portion of the material should be sorted through at
a time, and amphipods should be removed as they
are found. Amphipods and residual sediment
retained on the 0.25-mm sieve should be rinsed
briefly with freshwater to remove salts and washed
into a labeled sample jar (;>8 oz) using 70%
alcohol (either ethyl or isopropyl). Use of a wide
funnel supported by a ring stand facilitates this
process. Because offspring are very small, great
care is needed to ensure that all organisms are
transferred from the screen to the sample jar. Add
sufficient 70% alcohol to preserve the amphipods,
and add about 3 mL of rose bengal solution (about
1 g/L) to stain the organisms. Offspring may be
counted on test termination day, but waiting 2 to 3
d allows the amphipods to be more darkly stained.
11.3.8.3 Survival. Numbers of live and dead
adult amphipods should be determined and
recorded for each test chamber (see Figure 11.1).
Missing adult organisms are assumed to have
died, decomposed, and disintegrated during the
test; they should be included in the number dead
in calculations of the percentage survival for each
replicate treatment. Amphipods that are inactive
but not obviously dead are observed using a low-
power dissecting microscope or a hand-held
magnifying glass. Any organism that fails to
exhibit movement (i.e., neuromuscular twitch of
pleopods or antennae) upon gentle prodding with
a probe should be considered dead. An
independent count of survival in 10% of test
chambers should be completed by a second
observer. . Based one the experience of one
laboratory, the intralaboratory median CV for
survival (sample size of 88 treatments) can be
expected to be 11% (DeWitt et al. 1997b; see
Section 13.5.1). Based on one study involving 10
laboratories, the interlaboratory CV for survival
ranged from 4% to 19% (DeWitt el a. 1997b; see
Section 13.5.2). It should be expected that
intralaboratory CV for survival will decrease over
time as a laboratory gains experience using this
method. Similarly, the interlaboratory CV for
survival should decrease from reported values
here as more laboratories gain experience using
this method.
11.3.8.4 Growth Rate. Growth rate of amphipods
can be reported as daily change of average
individual length or weight. However, measuring
length is more laborious and therefore more
expensive than measuring weight to determine
growth rate, and does not result in an increase in
sensitivity in L plumulosus 28-d test (DeWitt et al.,
1997a). Dry weight of amphipods can be
determined as follows: (1) transferring the archived
amphipods from a replicate out of the preservative
into a crystallizing dish; (2) rinsing amphipods with
deionized water; (3) transferring these rinsed
amphipods to a preweighed aluminum pan; (4)
drying these samples to constant weight at 60°C;
and (5) weighing the pan and dried amphipods on
a balance to the nearest 0.01 mg. Average dry
weight of individual amphipods in each replicate is
calculated from these data. Due to the small size
of the amphipods, caution should be taken during
weighing 20 dried amphipods after 28-d sediment
exposure may weigh less than 25 mg). The
average per-capita dry weight of adult amphipods
for each replicate is the difference between the
tared weight of the boat and the total weight of the
boat plus dried amphipods, divided by the number
of amphipods in the weigh boat. The growth rate
endpoint (mg/d) is the difference between per
capita adult and neonate dry weights, divided by
28 d. In other words, for each replicate, calculate:
Growth Rate (mg/individual/day) = (mean adult dry
weight - mean neonate dry weight)/28. Weigh pans
need to be carefully handled using powder-less
gloves and the balance should be calibrated with
standard weights with each use. Use of small
aluminum pans will help reduce variability in
measurements of dry weight. Weigh boats can
also be constructed from sheets of aluminum foil.
Amphipod body length (±0.1 mm) can be
measured from the base of the first antennae to
the tip of the third uropod along the curve of the
dorsal surface. The use of a digitizing system and
microscope to measure length has been described
59
-------
in Kemble et al. (1994) for Hyalella azteca and
DeWitt et al. (1992a and 1997a) for Leptocheirus
plumulosus. Based on the experience of one
laboratory, the intralaboratory median CV for
growth (sample size of 87 treatments) can be
expected to be 3% (DeWitt et al. 1997b; see
Section 13.5.1). Based on one study involving 10
laboratories, the interlaboratory CV for growth
ranged from 14% to 38% (DeWitt el a. 1997b; see
Section 13.5.2). It should be expected that
intralaboratory CV for growth rate will decrease
over time as a laboratory gains experience using
this method. Similarly, the interlaboratory CV for
growth rate should decrease from reported values
here as more laboratories gain experience using
this method.
11.3.8.5 Reproduction. The offspring should be
counted within 2 weeks of terminating the test. It
may be possible to count the offspring the day the
experiment is broken down. If not, preserve
offspring in 70% alcohol (either ethyl or isopropyl).
Transfer preserved, stained offspring to a fine
screen (<0.25-mm mesh) and rinse with
freshwater to remove alcohol and excess stain.
Rinse the live or preserved neonates into a
shallow dish and count them under magnification,
such as a dissecting microscope. Record the
number of offspring. For QA, 10% of the samples
should be recounted by a second analyst. The
reproduction endpoint is calculated as the number
of offspring per living adult. Based on the
experience of one laboratory, the intralaboratory
median CV for reproduction (sample size of 88
treatments) can be expected to be 18% (DeWitt et
al. 1997b; see Section 13.5.1). Based on one
study involving 10 laboratories, the interlaboratory
CV for survival ranged from 35% to 102% (DeWitt
el a. 1997b; see Section 13.5.2). It should be
expected that intralaboratory CV for reproduction
will decrease over time as a laboratory gains
experience using this method. Similarly, the
interlaboratory CV for reproduction should
decrease from reported values here as more
laboratories gain experience using this method.
11.3.9 Control Performance Issues and
Revisions to the Protocol
11.3.9.1 The Leptocheirus plumulosus 28-d
sediment toxicity test, like all experimental
systems, is subject to occasional failures.
Because the L plumulosus 28-d sediment toxicity
test is more complex and of longer duration than
any of the marine amphipod short-term sediment
toxicity tests, there are more opportunities for
problems to occur in this long-term test than in the
short-term tests. Problems with the test are most
readily detected by failure to meet test
acceptability criteria in the control treatment
(Tables 11.1 and 11.3), such as mortality <20% or
failure of amphipods to grow or reproduce. Test
failures usually can be attributed to a failure to
maintain one or more test requirements described
in Tables 11.1 and 11.3; however, tests
sometimes fail inexplicably. Possible causes for
unaccountable test failures have included
overfeeding (e.g., leading to anoxia or increased
production of hydrogen sulfide), poor health of test
animals (i.e., culture failure), or accidental
introduction of toxic materials into test chambers.
Scientists from the USEPA and the USAGE
observe that the frequency of failure decreases as
the laboratory and staff using the test gain more
experience through conducting the test; however,
neither agency has explicit data on the frequency
of failure. Users of this test should be aware of
this possibility and prepare for the possibility to
rerun the test on occasion. Both agencies expect
that the protocol for the L. plumulosus 28-d
sediment toxicity test will be revised periodically,
especially as new experimental data reveal test
conditions that reduce the probability of possible
test failure.
11.4 Interpretation of Results
11.4.1 This section describes information that is
useful in helping to interpret the results of
sediment toxicity tests with L. plumulosus. Section
12 provides additional information on analyses and
reporting of toxicity test data.
11.4.2 Influence of Indigenous Organisms
11.4.2.1 Indigenous organisms may be present in
field-collected sediment. The presence of
organisms taxonomically similar to the test
organism can make interpretation of treatment
effects difficult. Predatory organisms can
adversely affect test organism survival. For
example, Redmond and Scott (1989) showed that
60
-------
the polychaete Nephtys incisa will consume
amphipods under test conditions. All control,
reference, and test sediment should be press-
sieved through 0.25-mm mesh to avoid these
complications. If test sediment is not sieved, the
number and species of indigenous organisms
should be determined to better interpret results.
11.4.3 Effects of Sediment Grain Size
11.4.3.1 L. plumulosus tolerates a wide range of
sediment types. There is generally little effect on
survival, growth rate, or reproduction when coarse-
grained (sand) or fine-grained (predominantly silt
and clay) sediment is used. In some studies,
L plumulosus has exhibited >90% survival in
clean sediment ranging from nearly 100% sand to
nearly 100% silt + clay (SAIC, 1993a; 1993b;
Schlekat et al., 1992; J. Kavanaugh, University of
West Florida, Gulf Breeze, FL, personal
communication). However, adverse effects can
occur in sediment with very high levels of clay or
sand. Laboratory studies have shown significant
reduction in survival when clay content exceeded
84%, and survival, growth and reproduction were
significantly reduced in 100% sand (Emery et al.,
1997). Results have been equivocal from
controlled tests with mixed grained sediments
(between 10% and 90% silt/clay). Emery et al.
(1997) found an increase in growth as sediment
coarseness increased up to 75% sand. DeWitt et
al. (1997a) reported enhanced growth in finer-
grained sediment as compared with more coarse-
grained material, but the difference in growth was
not considered to be biologically significant (DeWitt
et al., 1997a). Therefore, L. plumulosus should be
tested with sediment with silt/clay content between
5% and 85% (Table 11.1). If sediment
characteristics exceed these bounds, an
appropriate clean control/reference sedimerit
should be incorporated into the test to separate
effects of sediment-associated contaminants from
effects of particle size.
11.4.4 Effects of Pore Water Salinity
11.4.4.1 The range of salinity in which a given
species can survive when the overlying water
salinity is matched to that of the pore water salinity
is the salinity tolerance range. The potential for a
toxic response caused by salinity alone exists if a
species is exposed to conditions outside of its
range of tolerance. For estuarine sediment, it is
important to know the pore water salinity of each
sediment before testing is started and to use
overlying water of an appropriate salinity.
L plumulosus is not recommended for testing with
truly freshwater sediments (0%o pore water salinity)
or with sediments having pore water salinity >35%0
until further testing is completed to confirm
acceptable response in organisms (DeWitt et al.,
1997a). This methods manual recommends use
of standard salinity of overlying water for testing
(i.e., 5%o or 20%0; Table 11.1).
11.4.4.2 L. plumulosus, a euryhaline species, can
survive and thrive in a wide range of salinity
conditions. The salinity tolerance and application
range for this amphipod is 1%o to 35%o (DeWitt et
al., 1989; DeWitt et al., 1992a; SAIC, 1993b;
Schlekat et al., 1992; DeWitt et al., 1997a).
Although there is some evidence of salinity-related
stress for L. plumulosus at salinity extremes, the
breadth of salinity tolerance exhibited by this
species is most likely sufficient for application to
the majority of sediments that might be
encountered in an estuarine system (i.e., interstitial
salinity from 1%o to >30%o).
11.4.4.3 This method recommends testing with an
overlying water salinity of either 5%o or 20%<>; the
choice of overlying water salinity is dependant on
the pore water salinity of test sediment.
11.4.4.4 Although matching overlying and pore
water salinity values in test containers might be
appropriate for some study designs, this practice is
logistically complicated and normally impractical to
accomplish. Acclimation of amphipod cultures to
the appropriate salinity is required. Moreover, if
sediment samples to be tested have different pore
water salinity values, extreme care needs to be
exercised to ensure that renewals are completed
with water of the appropriate salinity.
11.4.5 Effects of Sediment-associated
Ammonia
11.4.5.1 Field-collected sediment may contain
concentrations of pore water ammonia that are
toxic to amphipods. The water-only NOEC for
L. plumulosus is 60 mg/L (USEPA, 1994d). If
ammonia concentrations are above this value at
61
-------
test initiation, mortality may be due in part to
effects of ammonia. Depending on test
application, it might be desirable to lower the
ammonia concentration by manipulating the test
system prior to introduction of test organisms if
measured ammonia in the pore water or overlying
water is greater than the NOEC. However, if
ammonia is the chemical of concern in the test
sediments, pore water ammonia concentrations
should not be deliberately manipulated. If
sediment toxicity tests are conducted to evaluate
the acceptability of dredged material for disposal,
the manipulations could be performed. Section
12.3.6 discusses methods for conducting TIEs to
determine whether ammonia is contributing to the
toxicity of sediment samples. Manipulations
involve flushing the test system by renewing a
specified amount of overlying water until ammonia
concentrations are reduced. The effects of dilution
of ammonia on pore water concentration is not
known. Due to this uncertainty, one option could
be to monitor pore water concentrations.
11.4.5.2 If ammonia is of concern to the regulatory
application associated with the sediment toxicity
test, overlying water should be sampled
approximately 1 cm above the sediment surface
prior to introduction of test organisms on Day 0.
Pore water ammonia should be measured when
sediment samples are prepared for testing. If both
the pore water and overlying water ammonia
concentrations are <60 mg/L, then the test may
proceed normally. If the ammonia concentration is
>60 mg/L in a given sample, then ammonia level
can be reduced by aerating the sample to
saturation and replacing 2 volumes of overlying
water per day. Purging pore water ammonia (up
to 60 mg/L) from test sediments prior to starting
the toxicity test, and employing the routine
replacement of overlying water in each test
chamber every other day (M-W-F) did result in a
consistently reduced pore water ammonia
concentration throughout the 28 days from
approximately 60 mg/L to approximately 1 mg/L
(DeWitt et al., 1997a). Similar results were
obtained by other researchers (Moore et al. 1997;
Moore et al. 1995). The analyst should measure
the pore water ammonia concentration each day
until it is <60 mg/L. The pore water ammonia
threshold for the chronic sediment toxicity test was
based on 28-d exposures of the amphipods to
sediments with experimentally-elevated pore water
ammonia (up to 60mg/L), employing the specified
purging technique prior to starting the toxicity test
exposure, and employing the routine replacement
of overlying water (M-W-F) (DeWitt et al., 1997a).
No lethal or sublethal toxicity was observed in this
experiment at any one of the tested pore water
ammonia concentrations, which is most likely
caused by loss of ammonia from the test system
due to diffusion of pore water ammonia from the
sediments to the overlying water and the
replacement of the overlying water three times per
week. Because dummy test containers are
required for pore water measurements, a minimum
of two dummy containers are required (one for
Day 0 and one for Day 28). Additional dummy
containers should be prepared if pore water
ammonia levels are high enough to require several
successive days for pore water ammonia
reduction. When ammonia concentrations are
reduced to <60 mg/L, testing should be initiated by
adding test organisms.
11.4.6 Future Research
11.4.6.1 Research to find methods that reduce the
variability of the growth rate and reproduction
endpoints could lead to improvements in the
statistical power of the L. plumulosus chronic
toxicity test. A second "round-robin" study, using
only laboratories with considerable experience
running this toxicity test, could provide improved
estimates of the interlaboratory accuracy and
precision of each endpoint. Additional research is
needed to evaluate the relative toxicological
sensitivity of the lethal and sublethal endpoints to
a wide variety of chemicals spiked in sediment and
to mixtures of chemicals in sediments from
pollution gradients in the field. Additional research
is needed to evaluate the ability of the test's lethal
and sublethal endpoints to estimate the responses
of populations and communities of benthic
invertebrates to contaminated sediments.
Research is also needed to link the toxicity test's
endpoints to a field-validated population model of
L. plumulosus that would then generate estimates
of population-level responses of the amphipod to
test sediments and thereby provide additional
ecologically relevant interpretive guidance for the
toxicity test.
62
-------
Section 12
Data Recording, Data Analysis and Calculations, and Reporting
12.1 Data Recording
12.1.1 Quality assurance project plans with data
quality objectives and SOPs should be developed
before starting a test. Procedures should be
developed by each laboratory to record, verify, and
archive data (USEPA, 1999).
12.1.2 A file should be maintained for each
sediment test or group of tests on closely related
samples (Section 9). This file should contain a
record of the sample chain-of-custody; a copy of
the sample log sheet; the original bench sheets for
the test organism responses during the sediment
test(s); chemical analysis data on the sample(s);
control data sheets for reference toxicants;
detailed records of the test organisms used in the
test(s), such as source, age, date of receipt, and
other pertinent information relating to their history
and health; information on the calibration of
equipment and instruments; test conditions used;
and results of reference toxicity tests. Original
data sheets should be signed and dated by the
laboratory personnel performing the tests. A
record of the electronic files of data should also be
included in the file.
12.1.3 Example data sheets are in Appendix A.
12.2 Data Analysis
12.2.1 Statistical methods are used to make
inferences about populations, based on samples
from those populations. In most sediment toxicity
tests, test organisms are exposed to contaminated
sediment to estimate the response of the
population of laboratory organisms. The organism
response to these sediments is usually compared
with the response to a control or reference
sediment. In any toxicity test, summary statistics,
such as means and standard errors for response
variables (e.g., survival), should be provided for
each treatment (e.g., pore water concentration,
sediment).
12.2.1.1 Types of Data
12.2.1.1.1 Two types of data and three endpoihts
(survival, growth rate, and reproduction) will be
obtained from the 28-d L plumulosus chronic test.
Survival is a dichotomous or categorical type of
data. Growth rate and reproduction are
representative of continuous data.
12.2.1.2 Sediment Testing Scenarios
12.2.1.2.1 Sediment tests are conducted to
determine whether contaminants in sediment are
harmful to benthic organisms. Sediment tests are
commonly used in studies designed to 1) evaluate
dredged material, 2) assess site contamination in
the environment (e.g., to rank areas for cleanup),
and 3) determine effects of specific contaminants,
or combinations of contaminants, through the use
of sediment spiking techniques. Each of these
broad study designs has specific statistical design
and analytical considerations, which are detailed
below.
12.2.1.2.2 Dredged Material Disposal Suitability
In these studies, each site is compared individually
to a reference sediment. The statistical
procedures appropriate for these studies are
generally pairwise comparisons. Additional
information on toxicity testing of dredged material
and analysis of data from dredged material
disposal suitability evaluations is available in
USEPA-USACE (1991; 1998).
12.2.1.2.3 Site Assessment of Field
Contamination. Surveys of sediment toxicity
often are included in more comprehensive
analyses of biological, chemical, geological, and
hydrographic data. Statistical correlation can be
improved and costs can be reduced if subsamples
63
-------
are taken simultaneously for sediment toxicity or
bioaccumulation tests, chemical analyses, and
benthic community structure determinations.
There are several statistical approaches to field
assessments, each with a specific purpose. If the
objective is to compare the response or residue
level at all sites individually with a control or
reference sediment, then the pairwise comparison
approach described below is appropriate. If the
objective is to compare all sites in the study area,
then a multiple comparison procedure that
employs an experiment-wise error rate is
appropriate. If the objective is to compare among
groups of sites, then orthogonal contrasts are a
useful data analysis technique.
12.2.1.2.4 Sediment-Spiking Experiments.
Sediment spiked with known concentrations of
contaminants can be used to establish cause-and-
effect relationships between chemicals and
biological responses. Results of toxicity tests with
test materials spiked into sediment at different
concentrations can be reported in terms of an
LC50, EC50, IC50, NOEC, or LOEC. The
statistical approach outlined above for spiked-
sediment toxicity tests also applies to the analysis
of data from sediment dilution experiments or
water-only reference-toxicity tests.
12.2.2 Experimental Design
12.2.2.1 The guidance outlined below on the
analysis of sediment toxicity test data is adapted
from a variety of sources including USEPA (1991 a;
1994a; 1994b; 1994c), and USEPA-USACE
(1998). The objectives of a sediment toxicity test
are to quantify contaminant effects on test
organisms exposed to natural or spiked sediment
or dredged materials and to determine whether
these effects are statistically different from those
occurring in a control or reference sediment. Each
experiment consists of at least two treatments: the
control and one or more test treatment(s). The
test treatment(s) consist(s) of the contaminated or
potentially contaminated sediment(s). A control
sediment is always required to ensure that no
contamination is introduced during the
experimental set-up and that test organisms are
healthy. A control sediment is used to judge the
acceptability of the test (Table 11.3). Designs
other than those for sediment-spiking experiments
also require a reference sediment that represents
an environmental condition or potential treatment
effect of interest. The reference sediment is
defined as a relatively uncontaminated sediment
and is used as the standard with which all test
sediments are compared. Testing a reference
sediment provides a site-specific basis for
evaluating toxicity of the test sediments.
Comparisons of test sediments to multiple
reference or control sediments representative of
the physical characteristics of the test sediment
(i.e., grain size, organic carbon) may be useful in
these evaluations (Section 2.1.2).
12.2.2.2 Experimental Unit
12.2.2.2.1 During toxicity testing, each test
chamber to which a single application of treatment
is applied is an experimental unit. The important
concept is that the treatment (sediment) is applied
to each experimental unit as a discrete unit.
Experimental units should be independent and
should not differ systematically.
12.2.2.3 Replication
12.2.2.3.1 Replication is the assignment of a
treatment to more than one experimental unit. The
variation among replicates is a measure of the
within-treatment variation and provides an es-
timate of within-treatment error for assessing the
significance of differences between treatments.
12.2.2.4 Minimum Detectable Difference (MOD)
12.2.2.4.1 As the minimum difference between
treatments which the test is required or designed
to detect decreases, the number of replicates
required to meet a given significance level and
power increases. Because no consensus
currently exists on what constitutes a biologically
acceptable minimum detectable difference (MOD),
the appropriate statistical minimum significant
difference should be a data quality objective
(DQO) established by the individual user (e.g.,
program considerations) based on their data
requirements, the logistics and economics of test
design, and the ultimate use of the sediment
toxicity test results.
64
-------
12.2.2.5 Minimum Number of Replicates
12.2.2.5.1 Five replicates are recommended for
28-d marine and estuarine sediment toxicity testing
for each control or experimental treatment.
However, it is always prudent to include as many
replicates in the test design as are economically
and logistically possible. Both the 10-d and 28-d
sediment toxicity testing methods recommend the
use of 20 organisms per replicate for marine
testing (USEPA, 1994a). An increase in the
number of organisms per replicate in all
treatments, including the control, is allowable only
if (1) test performance criteria for the
recommended number of replicates are achieved,
and (2) it can be demonstrated that no change
occurs in contaminant availability as a result of
increased organism loading. See Table 11.1 for a
description of the number of replicates and test
organisms/replicate recommended for long-term
testing of L. plumulosus.
12.2.6.6 Randomization
12.2.2.6.1 Randomization is the unbiased
assignment of treatments within a test system and
to the exposure chambers ensuring that no
treatment is favored and that observations are
independent. It is also important to (1) randomly
select the organisms (but not the number of
organisms) for assignment to the reference and
test treatments (e.g., a bias in the results might
occur if all the largest animals are placed in the
same treatment), (2) randomize the allocation of
sediment (e.g., not take all the sediment in the top
of a jar for the control and the bottom for spiking),
and (3) randomize the location of exposure units.
12.2.2.7 Pseudoreplication
12.2.2.7.1 The appropriate assignment of
treatments to the replicate exposure chambers is
critical to the avoidance of a common error in
design and analysis termed "pseudoreplication"
(Hurlbert, 1984). Pseudoreplication occurs when
inferential statistics are used to test for treatment
effects even though the treatments are not
replicated or the replicates are not statistically
independent (Hurlbert, 1984). The simplest form
of pseudoreplication is the treatment of
subsamples of the experimental unit as true
replicates. For example, two aquaria are
prepared, one with reference sediment, the other
with test sediment, and 10 organisms are placed
in each aquarium. Even if each organism is
analyzed individually, the 10 organisms only
replicate the biological response and do not
replicate the treatment (i.e., sediment type). In this
case, the experimental unit is the 10 organisms
and each organism is a subsample. A less
obvious form of pseudoreplication is the potential
systematic error due to the physical segregation of
exposure chambers by treatment. For example, if
all the reference exposure chambers are placed in
one area of a room and all the test exposure
chambers are in another, spatial effects (e.g.,
different lighting, temperature) could bias the
results for one set of treatments. Random
physical intermixing of the exposure chambers or
randomization of treatment location might be
necessary to avoid this type of pseudoreplication.
Pseudoreplication can be avoided or reduced by
properly identifying the experimental unit, providing
replicate experimental units for each treatment,
and applying the treatments to each experimental
unit in a manner that includes random physical
intermixing (interspersion) and independence.
However, avoiding pseudoreplication may be
difficult or impossible given resource constraints.
12.2.2.8 Optimum Design of Experiments
12.2.2.8.1 An optimum design is one which obtains
the most precise answer for the least effort. It
maximizes or minimizes one of many optimality
criteria, which are formal, mathematical
expressions of certain properties of the model that
are fit to the data. Optimum design of experiments
using specific approaches described in Atkinson
and Donev (1992) has not been formally applied to
sediment testing; however, it might be desirable to
use the approaches in experiments. The choice of
optimality criterion depends on the objective of the
test, and composite criteria can be used when a
test has more than one goal. A design is optimum
only for a specific model, so it is necessary to
know beforehand which models might be used
(Atkinson and Donev, 1992).
65
-------
12.2.2.9 Compositing Samples
12.2.2.9.1 Decisions regarding compositing of
samples depend on the objective of the test.
Compositing is used primarily in bioaccumulation
experiments when the biomass of an individual
organism is insufficient for chemical analysis.
Compositing consists of combining samples (e.g.,
organisms, sediment) and chemically analyzing
the mixture rather than the individual samples.
The chemical analysis of the mixture provides an
estimate of the average concentration of the
individual samples making up the composite.
Compositing also may be used when the cost of
analysis is high. Each organism or sediment
sample added to the composite should be of equal
size (i.e., wetweight) and the composite should be
completely homogenized before taking a sample
for chemical analysis. If compositing is performed
in this manner, the value obtained from the
analysis of the composite is the same as the
average obtained from analyzing each individual
sample (within any sampling and analytical errors).
If true replicate composites (not subsample
composites) are made, the variance of the
replicates will be less than the variance of the
individual samples, providing a more precise
estimate of the mean value. This increases the
power of a test between means of composites
over a test between means of individuals or
samples for a given number of samples analyzed.
If compositing reduces the actual number of
replicates, however, the power of the test will be
reduced. If composites are made of individuals or
samples varying in size, the value of the
composite and the mean of the individual
organisms or sediment samples are no longer
equivalent. The variance of the replicate
composites will increase, decreasing the power of
any test between means. In extreme cases, the
variance of the composites can exceed the
population variance (Tetra Tech, 1986).
Therefore, it is important to keep the individuals or
sediment samples comprising the composite
equivalent in size. If sample sizes vary, consult
the tables in Schaeffer and Janardan (1978) to
determine whether replicate composite variances
will be higher than individual sample variances,
which would make compositing inappropriate.
12.2.3 Hypothesis Testing and Power
12.2.3.1 The purpose of the 28-d L. plumulosus
chronic toxicity test is to determine whether the
biological response to a treatment sample differs
from the response to a control or reference
sample. Figure 12.1 presents the possible
outcomes and decisions that can be reached in a
statistical test of such a hypothesis. The
nullypothesis is that no difference exists among
the mean reference and treatment responses.
The alternative hypothesis of greatest interest in
sediment tests is that the treatments are toxic
relative to the reference sediment.
12.2.3.2 Statistical tests of hypotheses can be
designed to control for the chances of making
incorrect decisions. In Figure 12.1, alpha (a)
represents the probability of making a Type I
statistical error. A Type I statistical error in this
testing situation results from the false conclusion
that the treated sample is toxic or contains
chemical residues not found in the control or
reference sample. Beta (P) represents the
probability of making a Type II statistical error, or
the likelihood that one erroneously concludes there
are no differences among the mean responses in
the treatment and control or reference samples.
Traditionally, acceptable values have ranged from
0.1 to 0.01, with 0.05 or 5% used most commonly.
This choice should depend upon the
consequences of making a Type I error.
Historically, having chosen a, environmental
«
1.
DECISION
TR = Control
TR > Control
True State of Nature
TR = Control TR > Control
Correct
1-ot
Type 1 Error
a
Type II Error
P
Correct
1-P
(Power)
NOTE: Treatment response (TR); alpha (a) represents the prob-
ability of making a Type I statistical error (false positive); beta (ji)
represents the probability of making a Type II statistical error
(false negative).
Figure 12.1 Treatment response for a Type I
and Type II error
66
-------
researchers have ignored P and the associated
power of the test (1-0).
12.2.3.3 Fairweather (1991) presented a review of
the need for, and the practical implications of,
conducting power analysis in environmental
monitoring studies. This review also includes a
comprehensive bibliography of recent publications
on the need for, and use of, power analyses in
environmental study design and data analysis.
The consequences of a Type II statistical error in
environmental studies should never be ignored
and may, in fact, be the most important criteria to
consider in experimental designs and data
analyses that include statistical hypothesis testing.
According to Fairweather (1991), the commitment
of time, energy and people to a false positive (a
Type I error) will only continue until the mistake is
discovered. In contrast, the cost of a false
negative (a Type II error) will have both short- and
long-term costs (e.g., ensuring environmental
degradation and the cost of its rectification).
12.2.3.4 The critical components of the
experimental design associated with the test of
hypotheses outlined above are (1) the required
MOD between the treatment and reference
responses, (2) the variance among treatment and
reference replicate experimental units, (3) the
number of replicate units for the treatment and
reference samples, (4) the number of animals
exposed within a replicate exposure chamber, and
(5) the selected probabilities of Type I (a) and
Type II (3) errors.
12.2.3.5 Sample size or number of replicates might
be fixed because of cost or space considerations,
or might be varied to achieve a priori probabilities
of a and (3. The MOD should be established
ahead of time based upon biological and program
considerations. The investigator has little control
of the variance among replicate exposure
chambers. However, this variance component can
be minimized by selecting test organisms that are
as biologically similar as possible and maintaining
test conditions within prescribed QC limits.
12.2.3.6 The MOD can be expressed as the
absolute or relative (i.e., percentage) change from
the mean reference response. In this technical
manual, MOD is expressed as the absolute
change from the mean reference response
(Section 13). To test the equality of the reference
and a treatment response, a two-sample t-test with
its associated assumptions is an appropriate
parametric analysis. If the desired MOD, the
number of replicates per treatment, the number of
organisms per replicate, and an estimate of typical
among replicate variability (CV) are available, it is
possible to use a graphical approach as in
Figure 12.2 to determine how likely it is that a 20%
reduction will be detected in the treatment
response relative to the reference response. The
CV is defined as 100% x (standard deviation
divided by the mean). In a test design with eight
replicates per treatment and with an a level of
0.05, high power (i.e., >0.8) to detect a 20%
reduction from the reference mean occurs only if
the CV is 15% or less (Figure 12.2). The choice of
these variables also affects the power of the test.
If five replicates are used per treatment
(Figure 12.3), the CV needs to be 10% or lower to
detect a 20% reduction in response relative to the
reference mean with a power of 90%.
12.2.3.7 Relaxing the a level of a statistical test
increases the power of the test. Figure 12.4
duplicates Figure 12.3 except that a is 0.10
instead of 0.05. Selection of the appropriate a
level of a test is a function of the costs associated
with making Type I and II statistical errors.
Evaluation of Figure 12.3 illustrates that with a CV
of 15% and an a level of 0.05, there is
approximately 60% probability (power) of detecting
a 20% reduction in the mean treatment response
relative to the reference mean. However, if a is
set at 0.10 (Figure 12.4) and the CV remains at
15%, then there is approximately 80% probability
(power) of detecting a 20% reduction relative to
the reference mean. The latter example would be
preferable if an environmentally conservative
analysis and interpretation of the data is desirable.
12.2.3.8 Increasing the number of replicates per
treatment will increase the power to detect a 20%
reduction in treatment response relative to the
reference mean (Figure 12.5). Note, however, that
for less than eight replicates per treatment, it is
difficult to have high power (i.e., >0.80) unless the
CV is less than 15%. If space or cost limit the
number of replicates to fewer than eight per
67
-------
1.0
0.8 -
0.6-•
0.4 -
0.2
0.0
CV=5%
-t-
-t-
0% 10% 20% ' 30% 40% 50%
% Reduction of Control Mean
60% 70%
Figure 12.2 Power of the test vs. percentage reduction in treatment response relative to the control
mean at various CVs (eight replicates, a = 0.05 [one-tailed]
1.0
0.8 •
0.6
0.4
0.2
0.0
CV=5%
0% 10% 20% 30% 40% 50%
% Reduction of Control Mean
60% 70%
Figure 12.3 Power of the test vs. percentage reduction in treatment response relative to the
control mean at various CVs (five replicates, a = 0.05 [one-tailed]
68
-------
M « a m n 10
% Reduction of Control Mean
Figure 12.4 Power of the test vs. percentage reduction in treatment response relative to the control
mean at various CVs (5 replicates, a = 0.10 [one-tailed])
< a 10 12
No. of Replicates (n)
14 16
Figure 12.5 Effect of CV and number of replicates on the power to detect a 20% decrease in
treatment response relative to the control mean (a = 0.05 [one-tailed])
69
-------
treatment, then it may be necessary to find ways
to reduce the among-replicate variability.
Options that are available to increase the power
of the test include selecting more uniform
organisms to reduce biological variability and/or
increasing the a level of the test. For CVs in the
range of 30% to 40%, even eight replicates per
treatment is inadequate to detect small
reductions (<20%) in response relative to the
control mean.
12.2.3.9 The effect of the choice of a and 3 on
number of replicates for various CVs, assuming
the combined total probability of Type I and
Type II statistical errors is fixed at 0.25, is
illustrated in Figure 12.6. An a of 0.10,
therefore, establishes a 3 of 0.15. In
Figure 12.6, if a = 3 = 0.125, the number of
replicates required to detect a difference of 20%
relative to the reference is at a minimum. As
a or 3 decrease, the number of replicates
required to detect the same 20% difference
relative to the reference increases. However,
the curves are relatively flat over the range of
0.05 to 0.20, and their shape will change
dramatically if the combined total of a + 3 is
changed. Limiting the total of a + 3 to 0.10
greatly increases the number of replicates
necessary to detect a preselected percentage
reduction in mean treatment response relative to
the control mean.
12.2.4 Comparing Means
12.2.4.1 Figure 12.7 outlines a decision tree for
analysis of survival, growth rate, and
reproduction data subjected to hypothesis
testing. In the tests described herein, samples
or observations refer to replicates of treatments.
Sample size n is the number of replicates (i.e.,
exposure chambers) in an individual treatment,
not the number of organisms in an exposure
chamber. Overall sample size N is the
combined total number of replicates in all
treatments. The statistical methods discussed in
this section are described in general statistics
texts such as Steel and Torrie (1980), Sokal and
Rohlf (1981). Dixon and Massey (1983), Zar
(1984), and Snedecor and Cochran (1989). It is
recommended that users of this manual have at
least one of these texts and associated statistical
tables on hand. A nonparametric statistics text
such as Conover (1980) might also be helpful.
12.2.4.2 Mean
12.2.4.2.1 The sample mean is the average value,
or £x//n where
n = number of observations (replicates)
x, = ith observation
Ex, = every x summed = x, + x2 + x3 + . . . + xn
12.2.4.3 Standard Deviation
12.2.4.3.1 The sample standard deviation (S) is a
measure of the variation of the data around the
mean and is equivalent to - The sample
variance, s2, is given by the following "machine" or
"calculation" formula:
12.2.4.4 Standard Error of the Mean
12.2.4.4.1 The standard error of the mean (SE or
s/Jn) estimates variation among sample means
rather than among individual values. The SE is an
estimate of the standard deviation among means
that would be obtained from several samples of n
observations each. Most of the statistical tests in
this manual compare means with other means
(e.g., dredged sediment mean with reference
mean) or with a fixed standard (e.g., Food and
Drug Administration [FDA] action level; ASTM,
2000d). Therefore, the "natural" or "random"
variation of sample means (estimated by SE),
rather than the variation among individual obser-
vations (estimated by S), is required for the tests.
12.2.4.5 Tests of Assumptions
12.2.4.5.1 In general, parametric statistical
analyses such as t tests and analysis of variance
are appropriate only if (1 ) there are independent,
replicate experimental units for each treatment, (2)
the observations within each treatment follow a
normal distribution, and (3) variances for both
treatments are equal or similar. The first
70
-------
25
20- •
15- •
I
10 ••
5--
0 I I I I I I I I I 1 1 I I I 1 I l l I I I I I
SSBBSPSStSSB
o d ri d d d d d o e> o o
Alpha (Beta = 0.25-Alpha)
Figure 12.6 Effect of alpha and beta on the number of replicates at various CVs (assuming
combined a + (B = 0.25)
Data-Survival. Growth Rate, and Reproduction
I
Test for Normality
Normal
Shapiro-Wilk's Test (N<50)
Tests for Homogeneity of Variance
Bartlett's I | Hartley's]
Heterogeneous Variances
Non Normal—
•^Transformation?
T(, . No, n=
Homogeneous Variances
Yes, n>2 1
Ran kits
ransfor
>3 Replicates
Mest for
Unequal Variances
Equal Replication
No
Bonferroni's |
Yes
Comparison-Wise Alpha
Fisher's LSD, Duncan's
Experiment-Wise Alpha
Dunnett's
Yes
Equal Replication
Yes
No
Steel's Many-One
Rank Test
Wicoxon
with Bonferroni's
Endpoint
Figure 12.7 Decision tree for analysis of survival, growth rate, and reproduction data subjected to
hypothesis testing
71
-------
assumption is an essential component of
experimental design. The second and third
assumptions can be tested using the data
obtained from the experiment. Therefore, before
conducting statistical analyses, tests for normality
and equality of variances should be performed.
12,2.4.5.2 Outliers. Extreme values and
systematic departures from a normal distribution
(e.g., a log-normal distribution) are the most
common causes of departures from normality or
equality of variances. 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, by plotting, or by analysis
of residuals. An explanation should be sought for
any questionable data points. Sometimes an
investigator knows from past experience that
occasional wild observations occur, though the
process is otherwise stable. Except in such cases,
statisticians warn against automatic rejection rules
based on tests of significance, particularly if there
appear to be several outliers. The apparent
outliers may reflect distributions of the
observations that are skewed or have long tails
and are better handled by methods being
developed for nonnormal distributions. (Snedecpr
and Cochran, 1989). 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. An appropriate
transformation, such as the arcsine square root
transformation, will normalize many distributions
(USERA, 1985). Problems with outliers can
usually be solved only by using nonparametric
tests, but careful laboratory practices can reduce
the frequency of outliers.
12.2.4.5.3 Tests for Normality. The most
commonly used test for normality for small sample
sizes (n<50) is the Shapiro-Wilk's Test. This test
determines whether residuals are normally
distributed. Residuals are the differences between
individual observations and the treatment mean.
Residuals, rather than raw observations, are
tested because subtracting the treatment mean
removes any differences among treatments. This
scales the observations so that the mean of
residuals for each treatment and over all
treatments is zero. The Shapiro-Wilk's Test pro-
vides a test statistic W, which is compared to
values of W expected from a normal distribution.
W will generally vary between 0.3 and 1.0, with
lower values indicating greater departure from
number of replicates (n) and design. A balanced
design means that all treatments have an equal
number (n) of replicate exposure chambers. A
design is considered normality. Because
normality is desired, one looks for a high value of
W with an associated probability greater than the
pre-specified alpha level.
12.2.4.5.3.1 Table 12.1 provides alpha levels to
determine whether departures from normality are
significant. Normality should be rejected when the
probability associated with W (or other normality
test statistic) is less than a for the appropriate total
unbalanced when the treatment with the largest
number of replicates (nmax) has at least twice as
many replicates as the treatment with the fewest
replicates (nmin). Note that higher a levels are
used when the number of replicates is small or
when the design is unbalanced, because these are
the cases in which departures from normality have
the greatest effects on t tests and other parametric
comparisons. If data fail the test for normality,
even after transformation, nonparametric tests
should be used for additional analyses (Section
12.2.7.17 and Figure 12.7).
12.2.4.5.3.2 Tables of quantiles of W can be found
in Shapiro and Wilk (1965), Gill (1978), Conover
(1980), and other statistical texts. These
references also provide methods of calculating W,
although the calculations can be tedious. For that
reason, commonly available computer programs or
statistical packages are preferred for the
calculation of W.
12.2.4.5.4 Tests for Homogeneity of Variances.
There are a number of tests for equality of
variances. Some of these tests are sensitive to
departures from normality, which is why a test for
normality should be performed first. Bartlett's test
or other tests such as Levene's test or Cochran's
test (Winer, 1971; Snedecor and Cochran, 1989)
all have similar power for small, equal sample
sizes (n=5) (Conover etal., 1981). The data must
be normally distributed for Bartlett's test. Many
72
-------
Table 12.1 Suggested alpha Levels to Use for Tests of Assumptions
Test
Normality
Number of
Observations1
N = 2 to 9
N = 10 to 19
N = 20 or more
a
Balanced
0.10
0.05
0.01
When Design Is
Unbalanced2
0.25
0.10
0.05
Equality of
Variances
N = 2 to 9
N = 10 or more
0.10
0.05
0.25
0.10
1 N = total number of observations (replicates) in all treatments combined; n = number of observations
(replicates) in an individual treatment.
software packages for t tests and ANOVA provide
at least one of the tests.
12.2.4.5.4.1 If no tests for equality of variances are
included in the available statistical software,
Hartley's Fmax can easily be calculated:
Fmax = (larger of s?, s|) / (smaller of s?, sf)
When Fmax is large, the hypothesis of equal
variances is more likely to be rejected. F^ is a
two-tailed test, because it does not matter which
variance is expected to be larger. Some statistical
texts provide critical values of Fmax (Winer, 1971;
Gill, 1978; Rohlf and Sokal, 1981).
12.2.4.5.4.2 Levels of a for tests of equality of
variances are provided in Table 12.1. These
levels depend upon the number of replicates in a
treatment (n) and allotment of replicates among
treatments. Relatively high a values (i.e., ;>0.10)
are recommended, because power of the above
tests for equality of variances is rather low (about
0.3) when n is small. Equality of variances is
rejected if the probability associated with the test
statistic is less than the appropriate a.
12.2.4.6 Transformations of the Data
12.2.4.6.1 When the assumptions of normality or
homogeneity of variance are not met,
transformations of the data may remedy the
problem, so that the data can be analyzed by
parametric procedures, rather than by a
nonparametric technique. The first step in these
analyses is to transform the responses, expressed
as the proportion surviving, by the arcsine-square
root transformation. The arcsine-square root
transformation is commonly used on
proportionality data to stabilize the variance and
satisfy the normality requirement. If the data do
not meet the assumption of normality and there
are four or more replicates per group, then the
nonparametric test, Wilcoxon Rank Sum Test, can
be used to analyze the data. If the data meet the
assumption of normality, Bartlett's Test or Hartley's
F test for equality of variances is used to test the
homogeneity of variance assumption. Failure of
the homogeneity of variance assumption leads to
the use of a modified t test and the degrees of
freedom for the test are adjusted.
12.2.4.6.2 The arc sine-square root transformation
consists of determining the angle (in radians)
represented by a sine value. In this
73
-------
transformation, the proportion surviving 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. When
the proportion surviving is 0 or 1, a special
modification of the transformation should be used
(Bartlett, 1937). An example of the arcsine-square
root transformation and modification are provided
below.
1. Calculate the response proportion (RP) for
each replicate within a group, where
RP = (number of surviving organisms)/(number
exposed)
2. Transform each RP to arcsine, as follows:
a. For RPs greater than zero or less than one:
Angle (in radians) = arc sine ^(RP)
b. Modification of the arc sine when RP = 0.
Angle (in radians) = arc sine J—
V 4n
where n = number animals/treatment
replicate.
c. Modification of the arc sine when RP = 1.0.
Angle = 1.5708 radians - (radians for RP=0)
12.2.4.7 Two Sample Comparisons (n=2)
12.2.4.7.1 The true population mean (u) and
standard deviation (a) are known only after
sampling the entire population. In most cases,
samples are taken randomly from the population,
and the S calculated from those samples is only an
estimate of o. Student's f-values account for this
uncertainty. The degrees of freedom for the test,
which are defined as the sample size minus one
(n-1), should be used to obtain the correct f-value.
Student f-values decrease with increasing sample
size because larger samples provide a more
precise estimate of |J and a.
12.2.4.7.2 When using a t table, it is crucial to
determine whether the table is based on one-tailed
probabilities or two-tailed probabilities. In
formulating a statistical hypothesis, the alternative
hypothesis can be one-sided (one-tailed test) or
two-sided (two-tailed test). The null hypothesis
(H0) is always that the two values being analyzed
are equal. A one-sided alternative hypothesis (Ha)
is that there is a specified relationship between the
two values (e.g., one value is greater than the
other) versus a two-sided alternative hypothesis
(Ha), which is that the two values are simply
different (i.e., either larger or smaller). A one-
tailed test is used when there is an a priori reason
to test for a specific relationship between two
means, such as the alternative hypothesis that the
treatment mortality or tissue residue is greater
than the reference mortality or tissue residue. In
contrast, the two-tailed test is used when the
direction of the difference is not important or
cannot be assumed before testing.
12.2.4.7.3 Because control organism mortality and
sediment contaminant concentrations are
presumed lower than reference or treatment
sediment values, conducting one-tailed tests is
recommended in most cases. For the same
number of replicates, one-tailed tests are more
likely to detect statistically significant differences
between treatments (e.g., have a greater power)
than two-tailed tests. This is a critical
consideration when dealing with a small number of
replicates (such as five/treatment). The other
alternative for increasing statistical power is to
increase the number of replicates, which increases
the cost of the test.
12.2.4.7.4 There are cases when a one-tailed test
is inappropriate. When no a priori assumption can
be made as to how the values vary in relationship
to one another, a two-tailed test should be used.
An example of an alternative two-sided hypothesis
is that the reference sediment TOG content is
different (greater or lesser) from the reference
sediment TOG.
12.2.4.7.5 The t value for a one-tailed probability
can be found in a two-tailed table by looking up t
under the column for twice the desired one-tailed
probability. For example, the one-tailed f value for
a = 0.05 and df = 20 is 1.725, and is found in a
two-tailed table using the column for a = 0.10.
12.2.4.7.6 The usual statistical test for comparing
two independent samples is the two-sample t test
(Snedecorand Cochran, 1989). The f statistic for
74
-------
testing the equality of means x, and x2 from two
independent samples with n., and n2 replicates and
unequal variances is where s^ and s| are the
t = (x, - x2) / -JS? / n, + sf / riz
sample variances of the two groups. Although the
equation assumes that the variances of the two
groups are unequal, it is equally useful for
situations in which the variances of the two groups
are equal. This statistic is compared with the
student's t distribution with degrees of freedom
given by Satterthwaite's (1946) approximation:
df =
This formula can result in fractional degrees of
freedom, in which case one should round the
degrees of freedom down to the nearest integer in
order to use a t table. Using this'approach, the
degrees of freedom for this test will be less than
the degrees of freedom for a t test assuming equal
variances. If there are unequal numbers of
replicates in the treatments, the t test with
Bonferroni's adjustment can be used for data
analysis (USEPA, 1994b; 1994c). When variances
are equal, an F test for equality is unnecessary.
12.2.4.8 Nonparametric Tests
12.2.4.8.1 Tests such as the t test, which analyze
the original or transformed data, and which rely on
the properties of the normal distribution, are
referred to as parametric tests. Nonparametric
tests, which do not require normally distributed
data, analyze the ranks of data and generally
compare medians ratherthan means. The median
of a sample is the middle or fiftieth percentile
observation when the data are ranked from
smallest to largest. In many cases, nonparametric
tests can be performed simply by converting the
data to ranks or normalized ranks (rankits) and
conducting the usual parametric test procedures
on the ranks or rankits. Rankits are simply the z-
scores expected for the rank in a normal
distribution. Thus, using rankits imposes a normal
distribution over all the data, although not
necessarily within each treatment. Rankits can be
obtained by ranking the data, then converting the
ranks to rankits using the following formula:
where z is the normal deviate and N is the total
rankit =
number of observations. Alternatively, rankits may
be obtained from standard statistical tables such
as Sokal and Rohlf (1 981 ).
12.2.4.8.2 Nonparametric tests are useful because
of their generality, but have less statistical power
than corresponding parametric tests when the
parametric test assumptions are met. If
parametric tests are not appropriate for
comparisons because the normality assumption is
not met, data should be converted to normalized
ranks (rankits).
12.2.4.8.3 If normalized ranks are calculated, the
ranks should beTconverted to rankits using the
formula above. In comparisons involving only two
treatments (n=2), there is no need to test
assumptions on the rankits or ranks; simply
proceed with a one-tailed t test for unequal
variances using the rankits or ranks.
12.2.4.9 Analysis of Variance (n=2)
12.2.4.9.1 Some experiments are set up to
compare more than one treatment with a control,
whereas others may also be interested in
comparing the treatments with one another. The
basic design of these experiments is the same as
for experiments evaluating pairwise comparisons.
After the applicable comparisons are determined,
the data must be tested for normality to determine
whether parametric statistics are appropriate and
whether the variances of the treatments are equal.
If normality of the data and equal variances are
established, then ANOVA may be performed to
address the hypothesis that all the treatments
including the control are equal. If normality or
equality of variance are not established, then
transformations of the data might be appropriate,
or nonparametric statistics can be used to test for
equal means. Tests for normality of the data
should be performed on the treatment residuals.
A residual is defined as the observed value minus
the treatment mean, that is, rik = oik -(kth treatment
mean). Pooling residuals provides an adequate
sample size to test the data for normality.
75
-------
12.2.4.9.2 The variances of the treatments should
also be tested for equality. Currently there is no
easy way to test for equality of the treatment
means using analysis of variance if the variances
are not equal. In a toxicity test with several
treatments, one treatment might have 100%
mortality in all of its replicates, or the control
treatment may have 100% survival in all of its
replicates. These responses result in 0 variance
fora treatment that results in a rejection of equality
of variance in these cases. No transformation will
change this outcome. In this case, the replicate
responses for the treatment with 0 variance should
be removed before testing for equality of
variances. Only those treatments that do not have
0 replicate variance should be used in the ANOVA
to get an estimate of the within treatment variance.
After a variance estimate is obtained, the means of
the treatments with 0 variance can be tested
against the other treatment means using the
appropriate mean comparison. Equality of
variances among the treatments can be evaluated
with the Hartley F^ test or Bartlett's test. The
option of using nonparametric statistics on the
entire set of data is also an alternative.
12.2.4.9.3 If the data are not normally distributed
or the variances among treatments are not
homogeneous, even after data transformation,
nonparametric analyses are appropriate. If there
are four or more replicates per treatment and the
number of replicates per treatment is equal, the
data can be analyzed with Steel's Many-One Rank
test. Unequal replication among treatments
requires data analysis with the Wilcoxon Rank
Sum test with Bonferroni's adjustment. 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 test and can be
applied to data when the normality assumption has
not been met. Steel's test requires equal
variances across treatments and the control, but is
thought to be fairly insensitive to deviations from
this condition (USEPA, 1991 a). Wilcoxon's Rank
Sum test is a nonparameteric test to be used as
an alternative to the Steel's test when the number
of replicates are not the same within each
treatment. A Bonferroni's adjustment of the
pairwise error rate for comparison of each
treatment versus the control is used to set an
upper bound of alpha on the overall error rate.
This is in contrast to the Steel's test with a fixed
overall error rate for alpha. Thus, Steel's tests is
a more powerful test (USEPA, 1991 a).
12.2.4.9.4 Different mean comparison tests are
used depending on whether an a percent
comparison-wise error rate or an a percent
experiment-wise error rate is desired. The choice
of a comparison-wise or experiment-wise error
rate depends on whether a decision is based on a
pairwise comparison (comparison-wise) or from a
set of comparisons (experiment-wise). For
example, a comparison-wise error rate would be
used for deciding which stations along a gradient
were acceptable or not acceptable relative to a
control or reference sediment. Each individual
comparison is performed independently at a
smaller a than that used in an experiment-wise
comparison, such that the probability of making a
Type I error in the entire series of comparisons is
not greater than the chosen experiment-wise a
level of the test. This results in a more
conservative test when comparing any particular
sample to the control or reference. However, if
several samples were taken from the same area
and the decision to accept or reject the area were
based on all comparisons with a reference, then
an experiment-wise error rate should be used.
When an experiment-wise error rate is used, the
power to detect real differences between any two
means decreases as a function of the number of
treatment means being compared with the
reference treatment.
12.2.4.9.5 The recommended procedure for
pairwise comparisons that have a comparison-
wise a error rate and equal replication is to do an
ANOVA followed by a one-sided Fisher's Least
Significant Difference (LSD) test (Steel and Torrie,
1980). A Duncan's mean comparison test should
give results similar to the LSD. If the treatments
do not contain equal numbers of replicates, the
appropriate analysis is the t test with Bonferroni's
adjustment. For comparisons that maintain an
experiment-wise a error rate, Dunnett's test is
recommended for comparisons with the control.
76
-------
12.2.4.9.6 Dunnett's test has an overall error rate
of a, which accounts for the multiple comparisons
with the control. Dunnett's procedure uses a
pooled estimate of the variance, which is equal to
the error value calculated in an ANOVA.
12.2.4.9.7 To perform the individual comparisons,
calculate the t statistic for each treatment and
control combination, as follows:
where Y/ = mean for each treatment
Y,- = mean for the control
Sw = square root of the within mean
square
n, = number of replicates in the control
n, = number of replicates for treatment"!"
To quantify the sensitivity of the Dunnett's test, the
minimum significant difference (MSD = MOD) may
MSD = d S^ fl/n,} + (1/n)
be calculated with the following formula:
where d = critical value for the Dunnett's
Procedure
Sw = square root of the within mean
square
n = number of replicates per treatment,
assuming an equal number of
"replicates at all treatment
concentrations
n, = number of replicates in the control
12.2.5 Methods for Calculating LCSOs,
ECSOs, and ICps
12.2.5.1 Figure 12.8 outlines a decision tree for
analysis of point estimate data. USEPA manuals
(USEPA, 1991 a; 1994b; 1994c) discuss in detail
the mechanics of calculating LC50 (or EC50) or
values using the most current methods. The most
commonly used methods are the Graphical, Probit,
trimmed Spearman-Karber, and the Linear
Interpolation Methods. Methods for evaluating
point estimate data using logistic regression are
outlined in Snedecor and Chochran (1989). In
general, results from these methods should yield
similar estimates. Each method is outlined below,
and recommendations are presented for the use of
each method.
12.2.5.2 Data for at least five test concentrations
and the control should be available to calculate an
LC50, although each method can be used with
fewer concentrations. Survival in the lowest
concentration must be at least 50%, and an LC50
should not be calculated unless at least 50% of the
organisms die in at least one of the serial dilutions.
When less than 50% mortality occurs in the
highest test concentration, the LC50 is expressed
as greater than the highest test concentration.
12.2.5.3 Due to the intensive nature of the
calculations for the estimated LC50 and
associated 95% confidence interval using most of
the following methods, it is recommended that the
data be analyzed with the aid of computer
software. Computer programs to estimate the
LC50 or ICp values and associated 95%
confidence intervals using the methods discussed
below (except for the Graphical Method) were
developed by USEPA and can be obtained by
sending a diskette with a written request to
USEPA, National Exposure Research Laboratory,
26 W. Martin Luther King Drive, Cincinnati, OH
45268 or calling 513/569-7076.
12.2.5.4 The Graphical Method
12.2.5.4.1 This procedure estimates an LC50 (or
EC50) by linearly interpolating between points of a
plot of observed percentage mortality versus the
base 10 logarithm (Iog10) of treatment
concentration. The only requirement for its use is
that treatment mortalities bracket 50%.
12.2.5.4.2 For an analysis using the Graphical
Method, the data should first be smoothed and
adjusted for mortality in the control replicates. The
procedure for smoothing and adjusting the data is
detailed in the following steps: Let p0, p1t ..., pk
denote the observed proportion mortalities for the
control and the k treatments. The first step is to
smooth the PI if they do not satisfy p0-pr...-pk. The
smoothing process replaces any adjacent p^ that
77
-------
Data Survival Point Estimates
i
Two or More Partial Mortalities
Yes
r
Significant Chi-Square Test
Yes
No
4
One Partial Mortality
No
Yes
Linear Interpolation
Trimmed Spearman-Karber
LC50 and 95% Confidence Intervals
Figure 12.8 Decision tree for analysis of point estimate data
1
aph
T
do not conform to p0-pr...-pk with their average.
For example, if p, is less than pM then:
where pf = the smoothed observed proportion
mortality for concentration i.
Adjust the smoothed observed proportion mortality
in each treatment for mortality in the control group
using Abbott's formula (Finney, 1971). The
adjustment takes the form:
where
the smoothed observed
proportion mortality for the
control
pf = the smoothed observed
proportion mortality for
concentration i.
12.2.5.5 The Probit Method
12.2.5.5.1 This method is a parametric statistical
procedure for estimating the LC50 (or EC50) and
the associated 95% confidence interval (Finney,
1971). The analysis consists of transforming the
observed proportion mortalities with a Probit
transformation, and transforming the treatment
concentrations to Iog10. Given the assumption of
normality for the Iog10 of the tolerances, the
relationship between the transformed variables
mentioned above is about linear. This relationship
allows estimation of linear regression parameters,
using an iterative approach. A Probit is the same
as a z-score: for example, the Probit
corresponding to 70% mortality is z70 or = 0.52.
78
-------
The LC50 is calculated from the regression and is
the concentration associated with 50% mortality or
z=0. To obtain a reasonably precise estimate of
the LC50 with the Probit Method, the observed
proportion mortalities must bracket 0.5 and the
Iog10 of the tolerance should be normally
distributed. To calculate the LC50 estimate and
associated 95% confidence interval, two or more
of the observed proportion mortalities must be
between zero and one. The original percentage of
mortalities should be corrected for control mortality
using Abbott's formula (Section 12.2.5.4.2; Finney,
1971) before the Probit transformation is applied to
the data.
12.2.5.5.2 A goodness-of-fit procedure with the
Chi-square statistic is used to determine whether
data fit the Probit model. If many data sets are to
be compared with one another, the Probit Method
is not recommended because it may not be
appropriate for many of the data sets. This
method is also only appropriate for percent
mortality data sets and should not be used for
estimating endpoints that are a function of the
control response, such as inhibition of growth or
reproduction. Most computer programs that
generate Probit estimates also generate
confidence interval estimates for the LC50. These
confidence interval estimates on the LC50 might
not be correct if replicate mortalities are pooled to
obtain a mean treatment response (USEPA-
USACE, 1998). This can be avoided by entering
the Probit-transformed replicate responses and
doing a least-squares regression on the
transformed data.
12.2.5.6 The Trimmed Spearman-Karber
Method
12.2.5.6.1 The trimmed Spearman-Karber Method
is a modification of the Spearman-Karber,
nonparametric statistical procedure for estimating
the LC50 and the associated 95% confidence
interval (Hamilton et a!., 1977). This procedure
estimates the trimmed mean of the distribution of
the Iog10 of the tolerance. If the log tolerance
distribution is symmetric, this estimate of the
trimmed mean is equivalent to an estimate of the
median of the log tolerance distribution. Use of the
trimmed Spearman-Karber Method is only
appropriate for lethality data sets and when the
requirements for the Probit Method are not met
(USEPA, 1994b; 1994c).
12.2.5.6.2 To calculate the LC50 estimate with the
trimmed Spearman-Karber Method, the smoothed,
adjusted, observed proportion mortalities must
bracket 0.5. To calculate a confidence interval for
the LC50 estimate, one or more of the smoothed,
adjusted, observed proportion mortalities must be
between zero and one.
12.2.5.6.3 Smooth the observed proportion
mortalities as described for the Probit Method.
Adjust the smoothed observed proportion mortality
in each concentration for mortality in the control
group using Abbott's formula (see Probit Method;
Section 12.2.5.5). Calculate the amount of trim to
use in the estimation of the LC50 as follows:
Trim = max (pf, 1-
where pf = the smoothed, adjusted proportion
mortality for the lowest treatment
concentration, exclusive of the
control
pj^ = the smoothed, adjusted proportion
mortality for the highest treatment
concentration
k = the number of treatment
concentrations, exclusive of the
control.
12.2.5.7 The Linear Interpolation Method
12.2.5.7.1 The Linear Interpolation Method
calculates a toxicant concentration that causes a
given percent reduction (e.g., 25%, 50%, etc.) in
the endpoint of interest and is reported as an ICp
value (where p = the percent effect). The
procedure was designed for general applicability in
the analysis of data from chronic toxicity tests and
for the generation of an endpoint from a
continuous model that allows a traditional
quantitative assessment of the precision of the
endpoint, such as confidence limits for the
endpoint of a single test or a mean and coefficient
of variation for the endpoints of multiple tests.
79
-------
12.2.5.7.2 As described in USEPA (1994b;
1994c), the Linear Interpolation Method of
calculating an ICp assumes that the responses (1)
are monotonically nonincreasing, where the mean
response for each higher concentration is less
than or equal to the mean response for the
previous concentration, (2) follow a piecewise
linear response function, and (3) are from a
random, independent, and representative sample
of test data. If the data are not monotonically
nonincreasing, they are adjusted by smoothing
(averaging). In cases where the responses at the
low toxicant concentrations are much higher than
those in the controls, the smoothing process may
result in a large upward adjustment in the control
mean. In the Linear Interpolation Method, the
smoothed response means are used to obtain the
ICp estimate reported for the test. No assumption
is made about the distribution of the data except
that the data within a group being resampled are
independent and identically distributed.
12.2.8.7.3 The Linear Interpolation Method
assumes a linear response from one concentration
to the next. Thus, the 1C is estimated by linear
interpolation between two concentrations whose .
responses bracket the response of interest, the (p)
percent reduction from the control.
12.2.5.7.4 If the assumption of monotonicity of test
results is met, the observed response means (Y,)
should stay the same or decrease as the toxicant
concentration increases. If the means do not
decrease monotonically, the responses are
"smoothed" by averaging (pooling) adjacent
means. Observed means at each concentration
are considered in order of increasing
concentration, starting with the control mean (Y,).
If the mean observed response at the lowest
toxicant concentration (v|) is equal to or smaller
than the control mean (Y,), it is used as the
response. If it is larger than the control mean, it is
averaged with the control, and this average is
used for both the control response (M,) and the
lowest toxicant concentration response (M2). This
mean is then compared with the mean observed
response for the next higher toxicant concentration
vj) • Again, if the mean observed response for the
next higher toxicant concentration is smaller than
the mean of the control and the lowest toxicant
concentration, it is used as the response. If it is
higher than the mean of the first two, .it is averaged
with the first two, and the resulting mean is used
as the response for the control and two lowest
concentrations of toxicant. This process is
continued for data from the remaining toxicant
concentrations. Unusual patterns in the deviations
from monotonicity might require an additional step
of smoothing. Where Y, decrease monotonically,
the (Y|) become M, without smoothing.
12.2.5.7.5 To obtain the ICp estimate, determine
the concentrations Cj and CJ+1 that bracket the
response M^ (1 - p/1 00), where M1 is the smoothed
control mean response and p is the percent
reduction in response relative to the control
response. These calculations can easily be done
by hand or with a computer program as described
below. The linear' interpolation estimate is
calculated as follows:
where Cj = tested concentration whose
observed mean response is
greater than M,(1 - p/1 00)
CJ + 1 = tested concentration whose
observed mean response is less
thanM,(1 -p/100)
M., = smoothed mean response for the
control
Mj = smoothed mean response for
concentration J
MJ + 1 = smoothed mean response for
concentration J + 1
p = percent reduction in response
relative to the control response
ICp = estimated concentration at which
there is a percent reduction from
the smoothed mean control
response
12.2.5.7.6 Standard statistical methods for
calculating confidence intervals are not applicable
80
-------
for the ICp. The bootstrap method, as proposed
by Efron (1982), is used to obtain the 95%
confidence interval for the true mean. In the
boostrap method, the test data Yj, are randomly
resampled with replacement to produce a new set
of data Yji* that is statistically equivalent to the
original data, but which produces a new and
slightly different estimate of the ICp (ICp*). This
process is repeated at least 80 times (Marcus and
Holtzman, 1988), resulting in multiple "data" sets,
each with an associated ICp* estimate. The
distribution of the ICp* estimates derived from the
sets of resampled data approximates the sampling
distribution of the ICp estimate. The standard
error of the ICp is estimated by the standard
deviation of the individual ICp* .estimates.
Empirical confidence intervals are derived from the
quantiles of the ICp* empirical distribution. For
example, if the test data are resampled a minimum
of 80 times, the empirical 2.5% and the 97.5%
confidence limits are about the second smallest
and second largest ICp* estimates (Marcus and
Holtzman, 1988). The width of the confidence
intervals calculated by the bootstrap method is
related to the variability of the data. When
confidence intervals are wide, the reliability of the
1C estimate is in question. However, narrow
intervals do. not necessarily indicate that the
estimate is highly reliable, because of undetected
violations of assumptions and the fact that the
confidence limits based on the empirical quantiles
of a bootstrap distribution of 80 samples may be
unstable.
12.3 Data Interpretation
12.3.1 Sediments spiked with known
concentrations of contaminants can be used to
establish cause -and -effect relationships between
chemicals and biological responses. Results of
toxicity tests with test materials spiked into
sediments at different concentrations may be
reported in terms of an LC50, an EC50, an IC50,
or as an NOEC or LOEC (Section 3). Consistent
spiking procedures should be followed in order to
make interlaboratory comparisons (Section 8.3).
The data interpretation of USEPA program specific
regulatory decisions will be developed by the
respective USEPA program office.
12.3.2 Evaluating effect concentrations for
chemicals in sediment requires knowledge of
factors controlling the bioavailability. Similar
concentrations of a chemical in units of mass of
chemical per mass of sediment dry weight often
exhibit a range in toxicity in different sediment
(Di Toro et al., 1991; USEPA, 1992c). Effect
concentrations of chemicals in sediment have
been correlated to interstitial water concentrations,
and effect concentrations in interstitial water are
often similar to effect concentrations in water-only
exposures. The bioavailability of nonionic organic
compounds are often inversely correlated with the
organic carbon concentration of the sediment.
Whatever the route of exposure, the correlations of
effect concentrations to interstitial water
concentrations indicate that predicted or measured
concentrations in interstitial water can be useful for
quantifying the exposure concentration to an
organism. Therefore, information on partitioning of
chemicals between solid and liquid phases of
sediment can be useful for establishing effect
concentrations.
12.3.3 Toxic units can be used to help interpret the
response of organisms to multiple chemicals in
sediment. A toxic unit is the concentration of a
chemical divided by an effect concentration. For
example, a toxic unit of exposure can be
calculated by dividing the measured concentration
of a chemical in pore water by the water-only LC50
for the same chemical (Ankley et al., 1991).
Toxicity expressed as toxic units can be summed,
and this may provide information on the toxicity of
chemical mixtures (Ankley et al., 1991).
12.3.4 Field surveys can be designed to provide
either a qualitative reconnaissance of the
distribution of sediment contamination or a
quantitative statistical comparison of contamination
among sites (Burton and Ingersoll, 1994). Surveys
of sediment toxicity are usually part of more
comprehensive analyses of biological, chemical,
geological, and hydrographic data. Statistical
correlation can be improved and costs reduced if
subsamples are taken simultaneously for sediment
toxicity or bioaccumulation tests, chemical
analyses, and benthic community structure.
81
-------
12.3.5 Descriptive methods, such as toxicity tests
with field-collected sediment, should not be used
alone to evaluate sediment contamination. An
integration of several methods using the weight of
evidence is needed to assess the effects of
contaminants associated with sediment (Long and
Morgan, 1990; Ingersoll et al., 1996; 1997;
Macdonald et al., 1996). Hazard evaluations
integrating data from laboratory exposures,
chemical analyses, and benthic community
assessments provide strong complementary
evidence of the degree of pollution-induced
degradation in aquatic communities (Burton, 1991;
Canfield et al., 1994; 1996; 1998; Chapman et al.,
1992; 1997).
12.3.6 TIE procedures can be used to provide
insights as to specific contaminants responsible for
toxicity in sediment (USEPA, 1991 b; Ankley and
Thomas, 1992). For example, the toxicity of
contaminants such as metals, ammonia, hydrogen
sulfide, and non ionic organic compounds can be
identified using TIE procedures.
12.4 Reporting
12.4.1 The record of the results of an acceptable
sediment test should include the following
information either directly or by referencing
available documents:
12.4.1.1 Name of test and investigator(s), name
and location of laboratory, and dates of start and
end of test.
12.4.1.2 Source of control, reference, or test
sediment, and method for collection, handling,
shipping, storage, and disposal of sediment.
12.4.1.3 Source of test material, lot number if
applicable, composition (identities and
concentrations of major ingredients and impurities
if known), known chemical and physical properties,
and the identity and concentration(s) of any
solvent used.
12.4.1.4 Source and characteristics of overlying
water, description of any pretreatment, and results
of any demonstration of the ability of an organism
to survive or grow in the water.
12.4.1.5 Source, history, and age of tes(t
organisms; source, history, and age of brood
stock, culture procedures; and source and date of
collection of test organisms, scientific name, name
of person who identified the organisms and the
taxonomic key used, age or life stage, means and
ranges of weight or length, observed diseases or
unusual appearance, treatments used, and holding
procedures.
12.4.1.6 Source and composition of food;
concentrations of test material and other
contaminants; procedure used to prepare food;
and feeding methods, frequency and ration.
12.4.1.7 Description of the experimental design
and test chambers, the depth and volume of
sediment and overlying water in the chambers,
lighting, number of test chambers and number of
test organisms/treatment, date and time test starts
and ends, temperature measurements, DO (pg/L)
and any aeration used before starting a test and
during the conduct of a test.
12.4.1.8 Methods used for physical and chemical
characterization of sediment.
12.4.1.9 Definition(s) of the effects used to
calculate LC50 or EC50s, biological endpoints for
tests, and a summary of general observations of
other effects.
12.4.1.10 A table of the biological data for each
test chamber for each treatment, including the
control(s), in sufficient detail to allow independent
statistical analysis.
12.4.1.11 Methods used for statistical analyses of
data.
12.4.1.12 Summary of general observations on
other effects or symptoms.
12.4.1.13 Anything unusual about the test, any
deviation from these procedures, and any other
relevant information.
12.4.2 Published reports should contain enough
information to clearly identify the methodology
used and the quality of the results.
82
-------
Section 13
Precision and Accuracy
13.1 Determining Precision and
Accuracy
13.1.1 Precision is a term that describes the
degree to which data generated from replicate
measurements differ and reflects the closeness of
agreement between replicates. Accuracy is the
difference between the value of the measured data
and the true value and is the closeness of
agreement between an observed value and an
accepted reference value. Quantitative
determination of precision and accuracy in
sediment testing of aquatic organisms is difficult or
may be impossible in some cases, as compared to
analytical (chemical) determinations. This is due
in part to the many unknown variables that affect
organism response. Determining the accuracy of
a sediment test using field samples is not possible
since the true values are not known. Because
there is no acceptable reference material suitable
for determining the accuracy of sediment tests,
accuracy of the test methods has not been
determined (Section 13.2).
13.1.2 Sediment tests exhibit variability due to
several factors (Section 9). Test variability can be
described in terms of two types of precision, either
single laboratory precision (intralaboratory or
repeatability; Section 13.5.1) or multilaboratory
(interlaboratory or reproducibility; Section 13.5.2)
precision. Intralaboratory precision reflects the
ability of trained laboratory personnel to obtain
consistent results repeatedly when performing the
same test on the same organism using the same
toxicant. Interlaboratory precision (also referred to
as round-robin or ring tests) is a measure of
reproducibility of a method when tests are
conducted by a number of laboratories using that
method and the same organism and samples.
Generally, intralaboratory results are less variable
than interlaboratory results (USEPA,
1991; 1991d; 1994b; 1994c; Hall et al., 1989;
Grothe and Kimerle, 1985).
13.1.3 A measure of precision can be calculated
using the mean and relative standard deviation
(percent coefficient of variation, or CV% =
standard deviation/mean x 100)of the calculated
endpoints from the replicated endpoints of a test.
However, precision reported as the CV should not
be the only approach used for evaluating precision
of tests and should not be used for the NOEC
levels derived from statistical analyses of
hypothesis testing. The CVs can be very high
when testing extremely toxic samples. For
example, if there are multiple replicates with no
survival and one with low survival, the CV might
exceed 100%, yet the range of response is
actually quite consistent. Therefore, additional
estimates of precision should be used, such as
range of responses and minimum detectable
difference (MOD) compared with control survival or
growth rate. Several factors' can affect the
precision of the test, including test organism age,
condition and sensitivity; handling and feeding of
the test organisms; overlying water quality; and the
experience of the investigators in conducting tests.
For these reasons, it is recommended that trained
laboratory personnel conduct the tests in
accordance with the procedures outlined in
Section 9. Quality assurance practices should
include the following: (1) single laboratory
precision determinations that are used to evaluate
the ability of the laboratory personnel to obtain
precise results using reference toxicants for test
organisms and (2) preparation of control charts
(Section 13.4) for each reference toxicant and test
organism. The single laboratory precision
determinations should be made before conducting
a sediment test and should be periodically
performed as long as whole-sediment tests are
being conducted at the laboratory.
83
-------
13.1.4 Intralaboratory precision data are routinely
calculated for test organisms using water-only
96-h exposures to a reference toxicant, such as
cadmium chloride (CdCI2). Intralaboratory
precision data should be tracked using a control
chart. Each laboratory's reference-toxicity data will
reflect conditions unique to that facility, including
dilution water, culturing, and other .variables
(Section 9). However, each laboratory's
reference-toxicity CVs should reflect good
repeatability.
13.1.5 One interlaboratory precision (round-robin)
test has been completed on the 28-d chronic test
with Leptochelrus plumulosus (DeWitt et al.,
1997b). Ten laboratories participated in the round-
robin study, which used a dilution series of highly
contaminated Black Rock Harbor sediment from a
Superfund site in Connecticut mixed with
uncontaminated, diluent sediment from Sequim
Bay, Washington. The results of this round-robin
study are described in Section 13.5.
13.2 Accuracy
13.2.1 The relative accuracy of toxicity tests
cannot be determined because there is no
acceptable reference material. The relative
accuracy of the reference-toxicity tests can only be
evaluated by comparing test responses to control
charts.
13.3 Replication and Test Sensitivity
13.3.1 The sensitivity of sediment tests will depend
in part on the number of replicates per
concentration, the selected probability levels (a
and P) selected, and the type of statistical
analysis. For a specific level of variability, the
sensitivity of the test will increase as the number of
replicates is increased. The minimum
recommended number of replicates varies with the
objectives of the test and the statistical method
used for analysis of the data (Section 12).
13.4 Demonstrating Acceptable
Laboratory Performance
13.4.1 Intralaboratory precision, expressed as a
CV, can be determined by performing five or more
tests with different batches of test organisms,
using the same reference toxicant, at the same
concentrations, with the same test conditions (e.g.,
the same test duration, type of water, age.of test
organisms, feeding), and same data analysis
methods. A reference-toxicity concentration series
(dilution factor of 0.5 or higher) should be selected
that will provide partial mortalities at two or more
concentrations of the test chemical (Section
9.14,Table 9.1). See Section 9.16 for additional
detail on reference-toxicity testing.
13.4.2 Test animals should only be obtained from
culture. It is likely to be impractical to obtain test-
sized neonates directly from a supplier because of
their sensitivity to physical disturbances and their
rapid growth. Instead, test laboratories will likely
want to establish their own cultures of
L plumulosus from which to harvest neonates.
13.4.3 Before conducting tests with potentially
contaminated sediment, it is strongly
recommended that the laboratory conduct the
tests with control sediment(s) alone. Results of
these preliminary studies should be used to
determine if use of the control sediment and other
test conditions (i.e., water quality) result in
acceptable performance in the tests as outlined in
Tables 11.1 and 11.3.
13.4.4 A control chart should be prepared for each
combination of reference toxicant and test
organism. Each control chart should include the
most current data. Endpoints from five tests are
adequate for establishing the control charts. In
this technique, a running plot is maintained for the
values (Xj) from successive tests with a given
reference toxicant (Figure 13.1), and the endpoints
(LC50, NOEC, ICp) are examined to determine
whether they are within prescribed limits. Control
charts as described in USEPA (1991 a) and
USEPA (1993b) are used to evaluate the
cumulative trend of results from a series of
samples. The mean and upper and lower control
limits (±2 SD) are recalculated with each
successive test result. After 2 years of data
collection, or a minimum of 20 data points, the
control (cusum) chart should be maintained using
only the 20 most recent data points.
84
-------
UPPER CONTROL LIMIT
CENTRAL TENDENCY
LOWER CONTROL LIMIT
, , , , I , , , , I , , , i I , i , i I ^
0 5 10 15 20
Q
o
-------
estimate of the toxicant concentration that is lethal
to 50% of the test organisms in the time period
prescribed by the test. The LC50 is determined by
an appropriate procedure, such as the trimmed
Spearman-Karber Method, Probit Method,
Graphical Method, or the Linear Interpolation
Method (Section 12).
13.4.9 The point estimation analysis methods
recommended in this manual have been chosen
primarily because they are well-tested, well-
documented, and are applicable to most types of
test data. Many other methods were considered in
the selection process, and it is recognized that the
methods selected are not the only possible
methods of analysis for toxicity data.
13.5 Precision of the 28-d Chronic
Sediment Toxicity Test Method
13.5.1 Intralaboratory Performance
13.5.1.1 Studies described in DeWitt et al. (1997b)
provide additional data to characterize
intralaboratory precision with the 28-day long-term
toxicity test with L plumulosus. This data set
provides an estimate of intralaboratory precision
from a single laboratory from a total of 88
treatments (Table 13.1). To be consistent with
standard statistical procedures, these data were
transformed to reduce the heterogeneity of within
class variance. Percent survival was transformed
to the arcsine-square root of the value; growth rate
was transformed to the natural logarithm of the
value; and reproduction (offspring per survivor)
was transformed to the arcsine -square root of the
value. A CV was calculated on the transformed
data for each treatment within an experiment. The
observed distribution obtained from the resulting
sample of CVs from all experiments was then
characterized. This distribution of CVs then
provides an appropriate range on which to base
sample size calculations for future experiments.
The median CVs were 11% for survival, 3% for
growth rate, and 18% for reproduction (Table
13.1). The range between the first and third
quartiles provides a useful nonparametric interval
bounding the distribution. This range was 8% to
14% for survival, 2% to 6% for growth rate, and
13% to 36% for reproduction (Table 13.1).
13.5.1.2 These Values are similar to CVs for
intralaboratory precision calculated for survival
from 10-d tests with control sediment using
Hyalella azteca and Chironomus tentans (7.2%
and 5.7%, respectively; USEPA, 2000).
13.5.2 Interlaboratory Precision
13.5.2.11nterlaboratory precision for L. plumulosus
in the 28-d whole sediment toxicity test using the
methods described in this manual (Table 11.1)
was evaluated by round-robin testing (DeWitt et
al., 1997b). Ten laboratories, including federal and
state government laboratories, contract
laboratories, and academic laboratories with
demonstrated experience in chronic toxicity testing
using L plumulosus, participated in round-robin
toxicity testing (DeWitt et al., 1997b).
Table 13.1 Intralaboratory Precision Distribution of the Coefficient of Variation for Each Test
Endpoint (DeWitt et al. 1997a)
Sample 1st 3rd
End Point Size Mean Median Minimum Maximum Quartile Quartile
% Survival (Arcsine 88
transformed)
Growth rate (log 87
transformed)
Reproduction (square 88
root transformed)
14% 11%
4% 3%
31% 18%
0%
0%
0%
173%
16%
8%
14%
2% 6%
141% 13% 36%
86
-------
The experimental design required each laboratory
to conduct the 28-d chronic test using a dilution
series of Black Rock Harbor sediment (BRH; a
Superfund site in Connecticut) mixed with clean,
diluent sediment from Sequim Bay, Washington.
Each sediment treatment was prepared in a single
batch that was subsampled and shipped to testing
laboratories. A total of four concentrations of BRH
sediment and one negative control sediment were
tested. Across all treatments, total organic carbon
averaged 2.6% dry weight, total solids averaged
33%, and grain size averaged 15% sand, 42% silt,
and 43% clay. In general, cadmium, chromium,
copper, lead, nickel, and zinc, as well as total
PAHs, increased along the dilution series gradient.
Table 13.2 summarizes the concentration ranges
for the inorganic contaminants.
13.5.2.2 Approximately 4 months before the start
of the round-robin study, laboratories not currently
maintaining cultures of L plumulosus were
supplied with amphipods, sediment, food, and
culturing methods by the Battelle Marine Sciences
Laboratory (MSL). Each laboratory maintained
cultures following the culturing method detailed in
DeWitt et al. (1997a). Each laboratory used its
own source of clean seawater.
Table 13.2 Ranges of the BRH Sediment
Dilution Series Chemical
Concentrations (mg/kg dry wt;
from DeWitt et al., 1997b)
Cadmium
Chromium
Copper
Lead
Nickel
Zinc
Total PAHs
Low (BRH
treatment)
4.09 (0.0%)
104 (0.0%)
104(0.0%)
31.1 (0.0%)
91.2(0.0%)
189(0.0%)
9.85(1.4%)
High (BRH
treatment)
13.5(15.1%)
767(15.1%)
1503(15.1%)
209(15.1%)
150(15.1%)
736(15.1%)
17.5(15.1%)
13.5.2.3 Of the ten laboratories participating in the
round-robin, only five laboratories had ;>80%
survival in the negative control sediment, and
thereby met this performance criterion for test
acceptability (Top of Table 13.3). Analysis of the
data resulting from the round-robin included only
these five laboratories. Mean survival in the
negative control sediment was 93.6%, the CV was
4.2%, and the range was from 89% to 98%
(Table 13.3). The CVs across laboratories from
the five treatments ranged from 3.1% to 12.8%,
with a mean of 8.4%, and increased with dose.
None of the laboratories produced less than 70%
survival, even in the highest concentration of BRH
sediment. Further, none of the laboratories
produced a monotonic dose response for survival.
This suggests that the test did not contain a wide
enough series of dilutions to adequately measure
the response of survival. .For those laboratories
that showed a statistically significant decrease in
survival in the highest concentration of BRH (n=4),
an average of 16% change in survival was
produced between the control and the highest
concentration of BRH sediment.
13.5.2.4 For the five laboratories that met the
performance criterion, interlaboratory precision for
this study was characterized by the maximum and
minimum CV for each endpoint. The minimum
interlaboratory CV averaged about 4% for survival,
14% for growth rate, and 35% for reproduction
(Table 13.4). Maximum interlaboratory CV
averaged 19% for survival, 38% for growth rate,
and 102% for reproduction. The interlaboratory
MOD for survival ranged from 8% to 31%, and the
intralaboratory MOD for survival ranged from 10%
to 26%. The interlaboratory MOD for growth rate
ranged from 0.011 to 0.017 mg/ind/d, and the
intralaboratory MOD for growth rate ranged from
0.009 to 0.024 mg/ind/d. The interlaboratory MOD
for reproduction ranged from 0.33 to 2.86 offspring
per survivor, and the intralaboratory MOD for
reproduction ranged from 0.92 to 2.73 offspring
per survivor. These MDD's should be interpreted
cautiously, because they are derived from one
study consisting of a small number of
comparisons. Although the technical staff for
laboratories participating in the round-robin had
extensive sediment toxicity testing experience,
many had limited testing experience specifically
with L. plumulosus. Therefore, these values for
87
-------
Table 13.3 Results of Round-robin Intel-laboratory Precision of Endpoint Sensitivity for L. plumulosus in a
28-d Long-term Toxicity Test using Black Rock Harbor Sediments (DeWitt et al., 1997b)
A) Results for Laboratories that met Control Performance Criteria
Concentration of Black Rock Harbor Sediment
Lab 0.0% 1.4% 4.6% 8.3% 15.1%
4
6
7
8
9
Mean
%CV
MOD %
6
7
8
9
Mean
%CV
MOD
Mean Percent Survival (%CV)
89(11.5) 92(3.0) 82(17.6) 76(16.4)
96(6.8) 93(2.9) 97(4.6) 95(7.4)
90(6.8) 88(9.5) 84(12.9) 92(6.2)
95(6.4) 92(6.2) 72(42.4) 74(42.0)
98(2.8) 96(2.3) 84(15.4) 91(10.6)
93.6 92.2 83.8 85.6
4.2 3.1 10.6 11.5
10 7 26 24
Mean Growth Rate mg/d(%CV)
0.059
0.084
0.045
0.089
0.063
35.8
0.014
(9.8)
(4.4)
(18.3)
(8.7)
0.054
0.075
0.031
0.078
0.057
35.7
0.014
(6.0)
(4.9)
(12.7)
(13.4)
0.046
0.063
0.036
0.065
0.049
29.8
0.017
(19.0)
(8.5)
(25.1)
(12.7)
0.039
0.053
0.024
(11.7)
(7.2)
(27.5)
0.060(12.0)
0.039
45.1
0.012
73(13.4)
96 (5.7)
82(11.9)
70(18.2)
86 (14.5)
81.4
12.8
16
0.020
0.035
0.014
0.045
0.025
59.4
0.011
(24.1)
(28.0)
(14.1)
(11.6)
MOD %
16
8
13
31
14
MOD
mg/ind/d
0.009
0.009
0.010
0.012
Mean Offspring per Survivor (%CV)
4
6
7
8
9
Mean
%CV
MOD
0.27
4.37
5.22
1.66
7.09
3.72
73.8
2.86
(141)
(41.0)
(55.7)
(65.8)
(30.8)
2.26 (72.3)
2.96 (53.8)
3.99 (40.5)
1.10(54.2)
5.43(21.9)
3.15
52.5
2.10
0.65
2.58
3.61
1.52
3.48
2.37
53.8
1.53
(149)
(27.5)
(42.5)
(29.8)
(29.8)
0.35
1.70
2.21
0.25
1.65
1.23
71.2
1.42
(56.5)
(43.4)
(75.4)
(91.5)
(60.7)
0.33
0.18
0.48
0.10
0.19
0.25
59.5
0.33
MOD # offspring
(81.2)
(76.6)
(65.6)
(108 )
(99.0)
1
1
2
0
1
.33
.77
.73
.92
.96
B) Results for Laboratories that did not meet the Control Performance Criteria
Concentration of Black Rock Harbor Sediment
ah 00% 14% 4.6% 8,3% 15.1%
1
2
3
5
10
1
2
3
5
10
1
2
3
5
10
53(31.7)
0(— )
72 (34.6)
60 (56.5)
69 (29.6)
0.024(81.7)
0(— )
0.050 (50.2)
0.058 (16.0)
0.006 (54.5)
0.7 (45.2)
0(-)
4.8 (42.5)
3.1 (80.8)
0 1 (131}
Mean Percent Survival (%CV)
74(13.0) 65(38.5) 58(18.9) 39(64.4)
10 (— ) 27137.1) 15 (— ) 0(— )
85(17.1) 74(15.4) 61(21.2) 55(24.9)
88(18.7) 66(29.5) 84(24.7) 76(11.8)
59(49.9) 58(44.2) 37(70.0) 25(58.3)
Mean Growth Rate mg/ind/d (%CV)
0.032(37.7) 0.012(74.9) 0.012(67.9) 0.008(71.2)
0.027 (—) 0.028(49.0) 0.01 7 (—) 0 (— )
0.067(21.0) 0.055(33.3) 0.034(52.4) 0.025(32.0)
0.062(31.7) 0.037(67.6) 0.036(43.0) 0.024(12.0)
0.014(139) 0.007(47.1) 0.003(54.0) 0.003(80.2)
Mean Offspring per Survivor (%CV)
1.7(57.0) 0.4(206) 0.1(163) 0 (— )
1.3(— ) 1.2(18.0) 0.6 (—) 0(— )
3.7(51.5) 3.4(34.9) 0.4(92.4) 0(138)
2.3(25.5) 1.1(136) 0.8(113) 0.6(117)
1.4(111) 0.5(98.3) 0.8(157) 0.3(144)
88
-------
Table 13.4 Summary of Intel-laboratory Precision at Five Laboratories for the 28-Day Leptocheirus
plumulosus Chronic Test Using Five Dilutions of Black Rock Harbor Sediment (DeWitt et al.
1997b)
Lab-4 Lab-6 Lab-7 Lab-8 Lab-9
Survival
Min CV (%)
Max CV (%)
Growth rate
Min. CV (%)
Max CV (%)
Offspring per Survivor
Min CV (%)
Max CV {%)
3
18
36
96
56
149
3
7
6
24
27
77
6
13
4
28
40
75
6
42
13
27
30
108
2
15
9
13
22
99
interlaboratory precision may be higher than would
be expected from laboratories with routine
experience testing with this species.
13.5.2.5 A cost-power analysis was conducted on
round-robin data to determine the number of
replicates required per treatment for the 28-d
whole-sediment standard testing using
L plumulosus (DeWittetal., 1997b). This analysis
involved evaluating both the improvement in
statistical power of the test to detect a difference
between treatment means and the additional
expense of adding more replicates. For this
analysis, the cost of a replicate was assumed to
be proportionate to the time required to conduct all
of the tasks associated with one treatment. If cost
was not a concern, 14 replicates would be optimal
and would provide 80% power for detecting a
30% difference in reproduction at a CV of
approximately 36%. This number of replicates is
impractical because of costs and logistics. The
cost-power analysis for the L. plumulosus chronic
test indicated that six replicates per treatment
gives the greatest statistical power at the most
efficient cost. However, this conclusion was based
on the assumption that every 1% increase in
improved detection equals a 1 % increase in cost.
The decision to specify 5 replicates per treatment
in this manual was based primarily on an effort to
keep the cost of performing this test to a minimum.
Based on the median CVs for growth rate,
survival, and reproduction calculated from a large
data set (3%, 11%, and 18%, respectively; see
Section 13.5.1.4), five replicates will provide high
power (sO.80) to detect a 20% decrease in
survival and growth rate endpoints relative to the
control (Figure 12.5). For the reproduction
endpoint, the power to detect a 20% decrease will
be closer to 0.40 using five replicates and 0.50
using six replicates. With power fixed at 80% and
at a CV of 20%, the median CV demonstrated for
reproduction with five replicates would be suitable
to detect approximately 18% reduction in
reproduction and with six replicates approximately
16% reduction. Thus, there is relatively little
gained by increasing the number of replicates from
five to six. Nevertheless, if reproduction is the
assessment endpoint of most concern, then
incorporation of more than five replicates should
be considered. Because space and cost
considerations make use of five replicates
desirable, this method would benefit from
additional research to find ways to reduce the
89
-------
among-replicate variability for the reproduction
endpoint.
13.5.2.6 The mean growth rates across the
laboratories for each dose decreased with
increasing concentration of BRH sediment (Table
13.4). Thus, the growth rate was a more sensitive
measure to the concentration of BRH survival.
The CVs across the laboratories from the five
treatments ranged from 29.8% to 59.4% , with a
mean of 41.2%, and were on average five times
greater for growth rate than for survival (Table
13.4). Of the five laboratories that met the
performance criterion for control survival, three
laboratories produced a monotonic dose response
to growth rate. The percentage of change in the
growth rate between control and the highest
concentration of BRH sediment was on average
58% for these three laboratories.
13.5.2.7 The mean reproduction across
laboratories for each dose decreased with
increasing concentration of BRH sediment. Thus,
the measure of reproduction was a more sensitive
to the concentration of BRH than was survival;
however, the CVs across laboratories are on
average eight times greater for reproduction than
for survival. The CVs for the five treatments
ranged from 52.5% to 73.8%, with a mean of
62.2%. Of the five laboratories that met the
performance criteria for control survival, three
laboratories produced a monotonic dose response
in reproduction. The percentage of change in
reproduction (offspring/survivor) between the
control and the highest concentration of BRH
sediment was on average 95% for these three
laboratories.
13.5.2.8 USEPA (2000) included a review of a
series of round-robin studies from which
interlaboratory precision was analyzed. CVs for
survival in 10-d whole-sediment tests with H.
azteca ranged from 6% to 114% in three test
sediments. Similar tests with C. tentans produced
CVs of 8% to 181 % in three test sediments. In 28-
d whole-sediment tests with H. azteca, CVs from
five test sediments ranged from 7% to 28% for
survival, from 52% to 78% for growth (dry weight),
and from 66% to 193% for reproduction.
13.5.2.9 The Leptocheirus round-robin study
exhibited similar or better intra- and interlaboratory
precision than many chemical analyses and
toxicity test methods (USEPA, 1991 a; 1991d;
1998). The cause(s) of the high failure rate among
laboratories participating in the round-robin study
is not known. Several of the laboratories had not
conducted this toxicity test previously, and
inexperience with the procedures may have
contributed to some of the test failures. Some of
the laboratories suggested that uneaten food might
have accumulated during early days of the
experiment, which might have led to lethal low-
dissolved oxygen stress to the young amphipods
(DeWitt et al 1997b). Because of this potential
problem, additional experiments were conducted
(Section 11.3.6.4.1) to find the minimum food
ration that would minimize the build-up of excess
food, minimize mortality, produce significant
growth rate and reproduction endpoints of the 28-d
L plumulosus sediment toxicity test. The diet
recommended in this manual (Section 11.3.6.4) is
based on the results of that experiment.
90
-------
Section 14
References
Adams, W.U., R.A. Kimerle, and R.G. Mosher.
1985. An approach for assessing the
environmental safety of chemicals sorbed to
sediments. In Aquatic Toxicology and Hazard
Assessment: 7th Volume, ASTM STP 854, R.D.
Cardwell, R. Purdy, and R.C. Banner, eds.,
pp. 429-453.
American Chemical Society. Waste Management
Manual for Laboratory Personnel, Department of
Government Relations and Science Policy,
Washington, D.C.
Ankley, G.T. and M. K. Schubauer-Berigan. 1994.
Comparison of techniques for the isolation of pore
water for sediment toxicity testing. Arch. Environ.
Contam. Toxicol. 27:507-512.
Ankley, G. and N. Thomas. 1992. Interstitial water
toxicity identification evaluation approach. In
Sediment Classification Methods Compendium,
pp. 5-1 to 5-14. EPA-823/R-92-006, Washington,
D.C.
Ankley, G.T., D.A. Benoit, J.C. Balough, T.B.
Reynoldson, K.E. Day, and R.A. Hoke. 1993.
Development and evaluation of test methods for
benthic invertebrates and sediments: Effects of
flow rate and feeding on water quality and
exposure conditions. Arch. Environ. Contam.
Toxicol. 25:12-19.
Ankley, G.T., D.A. Benoit, J.C. Balough, T.B.
Reynoldson, K.E. Day, and R.A. Hoke. 1994a.
Evaluation of potential confounding factors in
sediment toxicity tests with three freshwater
benthic invertebrates. Environ. Toxicol. Chem. 13:
627-635.
Ankley, G.T., S.A. Collyard, P.O. Monson, and
P.A. Kosian. 1994b. Influence of ultraviolet light
on the toxicity of sediments contaminated with
polycyclic aromatic hydrocarbons. Environ.
Toxicol. Chem. 11:1791-1796.
Ankley, G.T., P.M. Cook, A.R. Carlson, D.J. Call,
J.A. Swenson, H.F. Corcoran, and R.A. Hoke.
1992a. Bioaccumulation of PCBs from sediments
by oligochaetes and fishes: Comparison of
laboratory and field studies. Can. J. Fish. Aquat.
Sci. 49:2080-2085.
Ankley, G.T., M.K. Schubauer-Berigan, and J.R.
Dierkes. 1991. Predicting the toxicity of bulk
sediments to aquatic orgnisms using aqueous test
fractions: Pore water versus elutriate. Environ.
Toxicol. Chem. 10: 1359-1366.
Ankley, G.T., M.K. Schubauer-Berigan, and R.A.
Hoke. 1992b. Use of toxicity identification
evaluation techniques to identify dredged material
disposal options: A proposed approach. Environ.
Management. 16:1-6.
ASTM. 2000a. Standard guide for conducting 10-
day static sediment toxicity tests with marine and
estuarine amphipods. E1367-99 Annual Book of
ASTM Standards Vol. 13.04, West Conshohocken,
PA.
ASTM. 2000b. Standard test methods for
measuring the toxicity of sediment-associated
contaminants with freshwater invertebrates.
E1706-95b. In Annual Book of ASTM Standards,
Vol. 11.05, West Conshohocken, PA.
ASTM. 2000c. Standard guide for collection,
storage, characterization, and manipulation of
sediments for toxicological testing. E1391-94. In
Annual Book of ASTM Standards, Vol. 11.05, West
Conshohocken, PA.
ASTM. 2000d. Standard guide for the
determination of bioaccumulation of sediment-
associated contaminants by benthic invertebrates.
E1688-00. In Annual Book of ASTM Standards,
Vol. 11.05, West Conshohocken, PA.
91
-------
ASTM. 2000e. Standard guide for designing
biological tests with sediments. E1525-94a. In
Annual Book of ASTM Standards, Vol. 11.05, West
Conshohocken, PA.
ASTM. 2000f. Standard guide for conducting acute
toxicity tests with fishes, macroinvertebrates, and
amphibians. E729-96 In Annual Book of
Standards, Vol. 11.05, West Conshohocken, PA.
ASTM.2000g. Standard guide for conducting early
life-stage toxicity test with fishes. E1241-98 In
Annual Book of Standards, Vol. 11.05, West
Conshohocken, PA.
ASTM. 2000h. Standard terminology relating to
biological effects and environmental fate. E 943-
OOa. In Annual Book of ASTM Standards,
Vol. 11.05, West Conshohocken, PA.
ASTM. 2000I. Standard practice for conducting
bioconcentration tests with fishes and saltwater
bivalve mollusks. E1022-94, In Annual Book of
Standards Vol. 11.05, West Conshohocken, PA.
Atkinson, A.C. and Donev, A.N. 1992. Optimum
Experimental Designs. Oxford: Clarendon Press.
328 p.
Bagarinao, T. 1992. Suifide as an environemental
factor and toxicant: tolerance and adaptations in
aquatic organisms. Aquatic Tox. 24:21-62.
Bartlett, M.S. 1937. Some examples of statistical
methods of research in agriculture and applied
biology. J. Royal Statist. Soc. Suppl. 4:137-183.
Becker, D.S., and T.C. Ginn. 1995. Effects of
storage time on toxicity of sediments from Puget
Sound, Washington. Environ. Toxicol. Chem.
14:829-836.
Besser, J.M., C.G. Ingersoll, E. Leonard, and D.R.
Mount. 1998. Effect of zeolite on toxicity of
ammonia and metals to freshwater invertebrates
and implications for sediment toxicity identification
evaluation (TIE) procedures. Environ. Toxicol.
Chem.: In review.
Boese B.L., Lee II, H., Specht, D.T., Randall, R.C.,
and Winsor, M.H. 1990. Comparison of aqueous
and solid-phase uptake for hexachlorobenzene in
the tellinid clam Macoma nasuta A mass balance
approach. Environ. Toxicol. Chem. 9:221-231.
Bousfield, E.L. 1973. Shallow Water
Gammeridean Amphipoda of New England.
Ithaca, NY: Cornell University Press, 312 pp.
Bridges, T.S., J.D. Farra, and B.M. Duke. 1997.
The influence of food ration on sediment toxicity in
Neanthes arenaceodentata (Annelida:Polychaeta).
Environ. Toxicol. Chem. 16:1659-1665.
Bufflap, ST., and Allen, H.E. 1995a. Sediment
pore water collection methods for tace metal
analysis: A review. Wat Res. 29:165-177.
Bufflap, ST. and Allen, H.E. 1995b. Comparison
of pore water sampling techniques for trace
metals. Wat Res. 29:2015-2054.
Bureau of National Affairs, Inc. 1986. U.S.
Environmental Protection Agency General
Regulation for Hazardous Waste Management.
Washington, D.C.
Burton, G.A. 1991. Assessment of freshwater
sediment toxicity. Environ. Toxicol. Chem. 10:
1585-1627.
Burton, G.A., Jr. 1992. Sediment collection and
processing factors affecting realism. In Sediment
Toxicity Assessment, ed. G.A. Burton, Jr., 37-66.
Lewis Publ. Boca Raton, FL.
Burton, G.A., Jr. and C.G. Ingersoll. 1994.
Evaluating the toxicity of sediments. In The
Assessment of Contaminated Great Lakes
sediment. US Environmental Protection Agency
Report, Region V, Chicago, IL. EPA-905-B94-002.
Burton, G.A., M.K. Nelson, and C.G. Ingersoll.
1992. Freshwater benthic toxicity tests. In
Sediment Toxicity Assessment, G.A. Burton, ed..
Lewis Publishers, Chelsea, Ml. pp. 213-240.
Canfield, T.J., E.L. Brunson, F.J. Dwyer, C.G.
Ingersoll, and N.E. Kemble. 1998. Assessing
sediments from the upper Mississippi river
navigational pools using benthic invertebrate
community evaluation and the sediment quality
triad. Arch. Environ. Contam. Toxicol. 35:202-212
92
-------
Canfield, T.J., F.J. Dwyer, J.F. Fairchild, P.S.
Haverland, C.G Ingersoll, N.E. Kemble, D.R.
Mount, T.W. LaPoint, G.A. Burton, and M.C. Swift.
1996. Assessing contamination in Great Lakes
sediments using benthic invertebrate communities
and the sediment quality triad approach. J. Great
Lakes Res. 22:565-583.
Canfield, T. J., N.E. Kemble, W.G. Brumbaugh, F.J.
Dwyer, C.G. Ingersoll, and J.F. Fairchild. 1994.
Use of benthic invertebrate community structure
and the sediment quality triad to evaluate metal-
contaminated sediment in the Upper Clark Fork
River, Montana. Environ. Toxicol. Chem. 13:1999-
2012.
Carlson, A.R., G.L. Phipps, V.R. Mattson, P.A.
Kosian, and A.M. Cotter. 1991. The role of acid-
volatile sulfide in determining cadmium
bioavailability and toxicity in freshwater sediments.
Environ. Toxicol. Chem. 10:1309-1319.
Carr, R.S. and D.C. Chapman. 1995. Comparison
of methods for conducting marine and estuarine
sediment porewater toxicity tests, extraction,
storage, and handling techniques. Arch. Environ.
Contam. Toxicol. 28: 69-77.
Chapman, P.M.. 1989. Current approaches to
developing sediment quality criteria. Environ.
Toxicol. Chem. 8:589-599.
Chapman, P.M., B. Anderson, S. Carr, V. Engle,
R. Green, J. Hameedi, M. Harmon, P. Haverland,
J. Hyland, C. Ingersoll, E. Long, J. Rodgers, M.
Salazar, P.K. Sibley, P.J. Smith, R.C. Swartz, B.
Thompson, and H. Windom. 1997. General
guidelines for using the sediment quality triad.
Mar. Pollut. Bull. 34:368-372.
Chapman, P.M., E.A. Power, and G. A. Burton, Jr.
1992. Integrative assessments in aquatic
ecosystems. In Sediment Toxicity Assessment.
G.A. Burton, Jr., ed. Lewis Publishers, Ann Arbor,
Ml. pp. 313-340.
Ciarelli, S., W.A..P.M.A. Vonck, N.M. van Straalen,
and J. Stronkhorst. 1998. Ecotoxicology
assessment of contaminated dredged material with
the marine amhpipod Corophium volutator. Arch
Environ. Contam. Toxicol. 34:350-356.
CWA (Clean Water Act). Citation Pending
Conover, W.J. 1980. Practical Nonparametric
Statistics. Second edition. John Wiley and Sons,
New York, NY, 493 p.
Conover, W.J., M.E. Johnson, and M.M. Johnson.
1981. A comparative study of tests for
homogeneity of variances, with applications to the
outer continental shelf bidding data.
Technometrics 23: 351-361.
Craig, G.R. 1984. Bioassessment of sediments:
Review of previous methods and
recommendations for future test protocols. lEC
Beak consultants, Ltd. Mississauga, Ontario.
Davenport, R. and A. Spacie. 1991. Acute
phototoxicity of harbor and tributary sediments
from lower Lake Michigan. J. Great Lakes Res.
17:51-56.
Day, K.E., R.S. Kirby, and T.B. Reynoldson. 1992.
The effects of sediment manipulations on chronic
sediment bioassays with three species of benthic
invertebrates. Abstract. Presented at the 13th
Annual Meeting, Society of Environmental
Toxicology and Chemistry, Cincinnati, OH,
Nov. 8-12.
DeFoe, D.L. and G.T. Ankley. 1998. Influence of
storage time on toxicity of freshwater sediments to
benthic macroinvertebrates. Environ. Pollut.
99:123-131.
DeWitt, T.H., G.R. Ditsworth, and R.C. Swartz.
1988. Effects of natural sediment features on
survival of the Phoxocephalid amphipod,
Rhepoxynius abronius. Marine Envir. Res., 25:99-
124.
DeWitt, T.H., M.S. Redmond, J.E. Sewall, and
R.C. Swartz. 1992a. Development of a chronic
sediment toxicity test for marine benthic
amphipods. U.S. Environmental Protection
Agency. CBP/TRS/89/93.
DeWitt, T.H., B.L. McGee, R. Alden, H. Phelps,
and A. Pinkney. 1992b. A multi-species, multi-lab
comparison of the toxicity of a Chesapeake Bay
sediment. Presented at the Thirteenth Annual
Meeting of the Society for Environmental
Toxicology and Chemistry, Cincinnati, OH,
November 8-12.
93
-------
DeWitt, T.H., L.A. Niewolny, S.L. Nieukirk, B.D.
Gruendell, W.W. Gardiner, and A.B. Borde.
1997a. Support for development of a standard
chronic sediment toxicity protocol with the
estuarine amphipod, Leptocheirus plumulosus.
Prepared for the U.S. Environmental Protection
Agency, Office of Water. PNWD-2369. Battelle
Pacific Northwest Division, Richland, WA.
DeWitt, T.H., M.R. Pinza, L.A. Niewolny, V.I.
Cullinan, and B.D. Gruendell. 19975.
Development and evaluation of standard
marine/estuarine chronic sediment toxicity test
method using Leptocheirus plumulosus. Prepared
for the U.S. Environmental Protection Agency,
Office of Science and Technology. PNNL-11768.
Pacific Northwest National Laboratory, Richland,
WA.
DeWitt, T.H, M.R. Pinza, and L.A. Niewonly.
1997c. Phase I (Dilution Series Study) and Phase
II (Field-Collected Sediment Study) of the PTI
Chronic Leptocheirus plumulosus Project.
Prepared for PTI Environmental, Bellevue,
Washington. PNWD-2384. Battelle Pacific
Northwest Division, Richland. WA..
DeWitt, T.H., R.C. Swartz, and J.O. Lamberson.
1989. Measuring the acute toxicity of estuarine
sediments. Environ. Toxicol. Chem. 8:1035-1048.
DeWoskin, R.S. 1984. Good laboratory practice
regulations: A comparison. Research Triangle
Institute, Research Triangle Park, NC. 63 p.
Dickson, K.L., A.W. Maki, and W.A. Brungs (eds).
1987. Fate and effects of sediment-bound
chemicals in aquatic systems: proceedings of the
sixth Pellston Environmental Workshop, 1984.
Pergamon Press, New York, NY.
Dillon, T.M. and A.B. Gibson. 1986.
Bioassessment methodologies for the regulatory
testing of freshwater dredged material.
Miscellaneous Paper EL-86-6, U.S. Army Engineer
Waterways Experiment Station, Vicksburg, MS.
Dillon, T.M., D.W. Moore, and A.S. Jaruis. 1994.
The effects of storage temperature and time on
sediment toxicity. Arch. Environ. Contam. Toxicol.
27:51-53.
Di Toro, D.M., J.D. Mahony, D.J. Hansen, K.J.
Scott, M.B. Hicks, S.M. Mays, and M.S. Redmond.
1990. Toxicity of cadmium in sediments: The role
of acid volatile sulfide. Environ. Toxicol. Chem.
9:1489-1504.
Di Toro, D.M., C.S Zarba, D.J. Hansen, W.J.
Berry, R.C. Swartz, C.E. Cowan, S.P. Pavlou, H.E.
Allen, N. A. Thomas, and P.R. Paquin. 1991.
Technical basis for establishing sediment quality
criteria for nonionic organic chemicals using
equilibrium partitioning. Environ. Toxicol. Chem.
10:1541-1583.
Ditsworth, G.R., D.W. Schults, and J.K.P. Jones.
1990. Preparation of benthic substrates for
sediment toxicity testing. Environ. Toxicol. Chem.
9:1523-1529.
Dixon, W.J. and F.J. Massey, Jr. 1983.
Introduction to Statistical Analysis. Fourth edition.
McGraw-Hill Book Company, New York, NY. p.
678.
Duan, Y., S.I. Guttman, and J.T.Oris. 1997.
Genetic differentiation among laboratory
populations of Hyalella azteca : Implications for
toxicology. Environ. Toxicol. Chem. 16:691-695.
Efron, B. 1982. The Jackknife, the Bootstrap, and
other resampling plans. CBMS 38, Soc. Industr.
Appl. Math., Philadelphia, PA.
Emery, V.L, Jr., and D.W. Moore. 1996.
Preliminary protocol for conducting 28-day chronic
sublethal bioassays using the estuarine amphipod
Leptocheirus plumulosus (Shoemaker). Technical
Note EEDP-01 -36. U.S. Army Corps of Engineers,
Vicksbury, MS.
Emery, V.L., Jr., D.W. Moore, B.R. Gray, B.M.
Duke, A.B. Gibson, R.B. Wright, and J.D. Farrar.
1997. Development of a chronic sublethal
sediment biodassay using the estuarine amphipod
Leptocheirus plumulosus (Shoemaker). Environ.
Toxic. Chem. 16:1912-1920.
Environment Canada. 1992. Biological test
method: Acute test for sediment toxicity using
marine or estuarine amphipods. Environment
Canada, Conservation and Protection, Ottawa,
Ontario. 83 p., EPS 1/RM/26.
94
-------
Enviroment Canada. 1994. Guidance Document
on Measurement of Toxicity Test Precision Using
Control Sediments Spiked with a Reference
Toxicant. Report EPS 1/RM/30.
Fairweather, P.O. 1991. Statistical power and
design requirements for environmental monitoring.
>Ausf. J. Mar. Freshwater Res. 42:555-567.
Finney, D.J. 1971. Probit Analysis. Third edition,
Cambridge, University Press, London. 333 p.,
1971.
FDA. 1978. Good laboratory practices for non-
clinical laboratory studies. Part 58. Fed. Reg.
43(247):60013-60020 (December 22, 1978).
Gill, J.L. 1978. Design and Analysis of
Experiments in the Animal and Medical Sciences.
Vol. 3. Appendices. The Iowa State University
Press, Ames, I A. 173 p.
Gray, B.R., V.L. Emery, D.L Brandon, R.B. Wright,
B.M. Duke, J.D. Farrar, and D.W. Moore. 1998.
Selection and optimal measures of growth and
reproduction for the sublethal Leptocheirus
plumulosus sediment bioassay. Environ. Toxicol.
and Chemistry. 17:2288-2297.
Green, A.S. and G.T. Chandler. 1996. Life-table
evaluation of sediment-associated chlorpyrifos
chronic toxicity to the benthic copepod',
Amphiascus tenuiremis. Arch. Environ. Contarri.
Toxicol. 31:77-83.
Green, R.H. 1979. Sampling Design and
Statistical Methods for Environmental Biologists.
Wiley Interscience. New York, NY. 257 p.
Grothe, D.R. and Kimerle, R.A. 1985. Inter-and
Intra-laboratory variability in Daphnia magna
effluent toxicity test results. Environ. Toxicol.
Chem. 4:189-192.
Hall, W.S., Patroczka, J.B., Mirenda, R.J.,Porter,
B.A., and Miller, E. 1989. Acute toxicity of
industrial surfactants to Mysidopisis bahia. Arch.
Environ. Contam. Toxicol. 18:765-772.
Hamilton, M.A., R.C. Russo, and R.V. Thurston.
1977. Trimmed Spearman-Karber method for
estimating median lethal concentrations in toxicity
bioassays. Environ. Sci. Technol. 11:714-719.
Correction 12:417(1978).
Harkey, G.A., S. Kane Driscoll, and P. Landrum.
1997. Effect of feeding in 30-day bioaccumulation
assays using Hyalella azteca in fluoranthene-
dosed sediment. Environ. Toxicol. Chem. 16:762-
769.
Harkey, G.A., P.F. Landrum, and S.J. Klaine.
1994. Preliminary studies on the effect of feeding
during whole sediment bioassays using
Chironomus riparius larvae. Chemosphere,
28:597-606.
Hines, A.H., P.J. Haddon, JJ. Miklas, L.A.
Wiechert, and A.M. Haddon. 1986. Estuarine
invertebrates and fish: Sampling design and
constraints for long-term measurement of
population dynamics. In New Approaches to
Monitoring Aquatic Ecosystems. STP 940, T.P.
Boyle, ed., American Society for Testing and
Materials.
Holland, A.F., AT. Shaughnessy, L.C. Scott, V.A.
Dickens, J.A. Ranasinghe, and J.K. Summers.
1988. Progress report: Long-term benthic
monitoring and assessment program for the
Maryland portion of Chesapeake Bay (July 1986-
October1987). PPRP-LTB/EST-88-1. VERSAR,
Columbia, MD.
Horwitz, A.J., Elrick, K.A., and Colberg, M.R. 1992.
The effect of membrance filtration artifacts on
dissolved trace element concentrations. Wat Res.
26:753-763.
Hurlbert, S.H. 1984. Pseudoreplication and the
design of ecological field experiments. Ecol.
Mono. 54: 187-211.
Ingersoll, C.G. 1991. Sediment toxicity and
bioaccumulation testing. Standardization News
19:28-33.
Ingersoll, C.G. 1995. Sediment toxicity tests. In,
Fundamentals of Aquatic Toxicology , Second
edition, G. Rand, ed. Taylor and Francis,
Washington, D.C. pp. 231-255.
Ingersoll, C.G., F.J. Dwyer, S.A, Burch, M.K.
Nelson, D.R. Buckler, and J.B. Hunt. 1992. The
use of freshwater and saltwater animals to
distinguish between toxic effects of salinity and
contaminants in irrigation drainwater. Environ
Toxicol Chem. 11:503-511.
95
-------
Ingersoll, C.G. and M.K. Nelson. 1990. Testing
sediment toxicity with Hyalella azteca (Amphipoda)
and Chironomus riparius (Diptera). In Aquatic
Toxicology and Risk Assessment, 13th volume.
ASTM STP 1096 W.G. Landis and W.H. van der
Schalie, eds.). Philadelphia, PA. pp. 93-109,1990.
Ingersoll, C.G., T. Dillon, R.G. Biddinger (eds.).
1997. Methodological uncertainty in sediment
ecological risk assessment. In Ecological Risk
Assessments of Contaminated Sediment. SETAC
Press, Pensacola, FL. 389 p.
Ingersoll, C.G., E.L. Brunson, F.J. Dwyer, O.K.
Hardesty, and N.E. Kemble. 1998. Use of
sublethal endpoints in sediment toxicity tests with
the amphipod Hyalella azteca. Environ. Toxicol.
Chem. 17:1508-1523.
Ingersoll, C.G., D.R. Buckler, E.A. Crecelius, and
T.W. LaPoint. 1993. U.S. Fish and Wildlife
Service and Battelle final report for the USEPA
GLNPO assessment and remediation of
contaminated sediment (ARCS) project: Biological
assessment of contaminated Great Lakes
sediment. EPA-905-R93-006, Chicago, IL.
Ingersoll C.G., F.J. Dwyer, and T.W. May. 1990.
Toxicity of inorganic and organic selenium to
Daphnia magna (Cladocera) and Chironomus
riparius (Diptera). Environ. Toxicol. Chem. 9:1171-
1181.
Ingersoll, C.G., P.S. Haverland, E.L. Brunson, T.J.
Canfield, F.J. Dwyer, C.E. Henke, and N.E.
Kemble. 1996. Calculation and evaluation of
sediment effect concentrations for the amphipod
Hyalella azteca and the midge Chironomus
riparius. J. Great Lakes Res. 22:602-623.
Jordan, R.A. and C.E. Sutton. 1984. Oligohaline
benthic invertebrate communities at two
Chesapeake Bay power plants. Estuaries 7:192-
212.
Karickhoff, S.W. and K.R. Morris. 1985. Sorption
dynamics of hydrophobic pollutants in sediment
suspensions. Environ. Toxicol. Chem. 4:469-479.
Kemble, N.E., W.G. Brumbaugh, E.L. Brunson,
F.J. Dwyer, C.G. Ingersoll, D.P. Monda, and D.F.
Woodward. 1994. Toxicity of metal-contaminated
sediments from the upper Clark Fork River, MT, to
aquatic invertebrates in laboratory exposures.
Environ. Toxicol. Chem. 13:1985-1997.
Kemp, P.F. and R.C. Swartz. 1988. Acute toxicity
of interstitial and particle-bound cadmium to a
marine infaunal amphipod. Mar. Environ. Res.
26:135-153.
Knezovich, J.P., F.L. Harrison, and R.G. Wilhelm.
1987. The bioavailability of sediment-sorbed
organic chemicals: A review. Water Air Soil Pollut.
32:233-245.
Knezovich, J.P., D.J. Steichen, J.A. Jelinski, and
S.L.Anderson. 1996. Sulfide tolerance of four
marine species used to evaluate sediment and
pore-water toxicity. Bull. Environ. Contam.
Toxicol. 57: 450-457.
Kohn, N.P., J.Q. Word, O.K. Niyogi, L.T. Ross, T.
Dillon, and D.W. Moore. 1994. Acute toxicity of
ammonia to four species of marine amphipod.
Mar. Environ. Res. 38: 1-15.
Lamberson, J.O. and R.C. Swartz. 1988. Use of
bioassays in determining the toxicity of sediment
to benthic organisms. In Toxic Contaminants and
Ecosystem Health: A Great Lakes Focus., M.S.
Evans, ed. John Wiley and Sons,New York, NY.
pp. 257-279.
Lamberson, J.O. and R.C. Swartz. 1992. Spiked-
sediment toxicity test approach. In Sediment
Classification Methods Compendium. EPA-823-R-
92-006, Washington, D.C. pp. 4-1 to 4-10.
Landrum, P.F. 1989. Bioavailability and
toxicokinetics of polycyclic aromatic hydrocarbons
sorbed to sediments for the amphipod,
Pontoporeia hoyi. Environ. Sci. Tech. 23:588-
595.
Landrum, P.F. and W.R. Faust. 1992. Variation in
the bioavailability of polycyclic aromatic
hydrocarbons sorbed to sediments for the
amphipod Pontoporeia hoyi. Environ. Toxicol.
Chem. 11:1197-1208.
96
-------
Lee, D.R. 1980. Reference toxicants in quality
control of aquatic bioassays. In, Aquatic
invertebrate bioassays. ASTM STP 715, A.L.
Buikema and J. Cairns, Jr., eds, Philadelphia, PA.
pp. 188-199.
Lee, II, H., B.L. Boese, and P.P. Landrum. 1994.
Guide for determination of the bioaccumulation of
sediment-associated contaminants by benthic
invertebrates (8/94 draft in ASTM review).
Lee, II, H., B.L. Boese, J. Pelletier, M. Winsor, D.T.
Specht, and R.C. Randall. 1989. Guidance
manual: Bedded sediment bioaccumulation tests.
USEPA Report No. 600/X-89/302, ERL-
Narragansett. Rl. 232 p.
Levin, L., H. Caswell, T. Bridges, C. Dibacco, D.
Cabrera, and G. Plaia. 1996. Demographic
responses of estuarine polychaetes to
pollutantsrlife table response experiments.
Ecological Applications. 6:1295-1313.
Long, E.R. and L.G. Morgan. 1990. The potential
for biological effects of sediment-sorbed
contaminants in the NOAA National Status and
Trends program. NOAA Technical Memorandum
NOS DMA 52, Seattle, WA. 175 pp + appendices.
Long, E.R., M.F. Buchman, S.M. Bay, R.J.
Breteler, R.S. Carr, P.M. Chapman, J.E. Hose,
A.L. Lissner, J. Scott, and D.A. Wolfe. 1990.
Comparative evaluation of five sediment toxicity
tests with sediments from San Francisco Bay and
Tomales Bay, CA. Environ Toxicol. Chem.
9:1193-1214.
MacDonald, D.D., R.S. Carr, F.D. Calder, E.R.
Long, and C.G. Ingersoll. 1996. Development and
evaluation of sediment quality guidelines for
Florida coastal waters. Ecotoxicology 5:253-278.
Marcus, A.M. and A.P. Holtzman. 1988. A robust
statistical method for estimating effects
concentrations in short-term fathead minnow
toxicity tests. Manuscript submitted to the Criteria
and Standards Division, U. S. Environmental
Protection Agency, by Battelle Washington
Environmental Program Office, Washington, D.C.,
June 1988, under EPA Contract No. 69-03-3534.
39 p.
Marsh, A.G. and K.R. Tenore. 1990. The role of
nutrition in regulating the population dynamics of
opportunistic, surface deposit feeders in a
mesohaline community. Limnol. Oceanogr.
35(3):710-724.
McGee, B.L., C.E. Schlekat, D.M Boward, and T.L.
Wade. 1995. Sediment contamination and
biological effects in a Chesapeake Bay marina.
Ecotoxicology. 4:39-59.
McGee, B.L., C.E. Schlekat, and E. Reinharz.
1993. Assessing sublethal levels of sediment
contamination with the estuarine amphipod,
Leptocheirus plumulosus. Environ Toxicol. Chem.
12:577-588.
McGee, B.L. 1998. Population dynamics of the
estuarine amphipod Leptocheirus plumulosus:
Implications for sediment toxicity tests. Ph.D.
dissertation, University of Maryland, College Park,
MD.
McGee, B.L, D.A Wright and D.J. Fisher. 1998.
Biotic factors modifying the acute toxicity of
aqueous cadmium to the estuarine amphipod
Leptocheirus plumulosus. Archives of
Environmental Contamination and Toxicology
35:34-40.
McGee, B.L. and D.J. Fisher. 1999. Field
validation of the chronic sediment bioassay with
the estuarine amphipod Leptocheirus plumulosus.
Final Report. Prepared for the U.S. Environmental
Protection Agency, Office of Science and
Technology by the University of Maryland, Wye
Research and Education Center, Queenstown
MD.
McGee, B.L, D.J. Fisher, L.T. Yonkos, G.P.
Ziegler, and S. Turley. 1999. Assessment of
sediment contamination, acute toxicity and
population viability of the estuarine amphipod
Leptocheirus plumulosus in Baltimore Harbor.
Environ. Toxicol. Chem. 18:2151-2160.
McNulty, E.W., F.J. Dwyer, M.R. Ellersieck, I.E.
Greer, C.G. Ingersoll, and C.F. Rabeni. 1999.
Evaluation of the ability of reference-toxicity tests
to identify stress in laboratory populations of the
amphipod Hyatella azteca. Environ. Toxicol.
Chem. 18:544-548.
97
-------
Meador, J.P. and C.A. Rice. 2001. Impaired
growth in the polychaete Armandia brevis exposed
to tributyltin in sediment. Marine Environmental
Research. 51:113-129.
Moore, D.M., T. M. Dillon, J.Q. Word, and J.A.
Ward. 1994. Quality assurance/quality control
(QA/QC) guidance for laboratory dredged material
bioassays; Results of a QA/QC workshop held
May 26-27, 1993, in Seattle, Washington.
Miscellaneous Paper D-94-3, U.S. Army Engineer
Waterways Experiment Station, Vicksbury, MS.
Moore, D.M. and V.L. Emery, Jr. 1996.
Preliminary protocol for conducting 28-day chronic
sublethal sediment bioassays using the estuarine
amphipod Leptocheirus plumulosus (Shoemaker).
Environmental Effects of Dredging Technical
Notes. EEDP-01-36. March 1996. U.S. Army
Engineer Watersays Experiment Station,
Vicksburg, MS.
Moore, D.W., T.S. Bridges, B.R. Gray, and B.M.
Duke. 1997. Risk of ammonia toxicity during
sediment bioassays with the estuarine amphipod
Leptocheirus plumulosus. Environ. Toxicol. Chem.
16:1020-1027.
Moore, D.W., T.M. Dillon, and E.W. Gamble. 1996.
Long-term storage of sediments: Implications for
sediment toxicity testing. Environ. Pollut. 89:147-
155.
National Research Council (NRC). 1989.
Contaminated Marine Sediments-Assessment
and Remediation, NRC, National Academy Press,
Washington, D.C.
Niewolny, L.A., T.H. DeWitt, and B.D. Gruendell.
1997. ALCOA Point Comfort Sediment Quality
Triad Project: Leptocheirus plumulosus Chronic
Survival Sediment Toxicity Test. Prepared for
Parametrix, Inc. Kirkland, Washington. PNWD-
2381. Battelle Pacific Northwest Division,
Richland. WA..
Oris, J.T. and J.P. Giesy. 1985. The
photoenhanced toxicity of anthracene to juvenile
sunfish (Lepomisspp.).Aquat. Toxicol. 6:133-146.
Pfitezenmeyer, H. 1975. Benthos. \r\ABiological
Study of Baltimore Harbor, T. Koo, ed.,
Contribution No. 621. Center for Environmental
and Estuarine Studies, University of Maryland,
Solomons, MD.
Plumb, R.H. 1981. Procedures for handling and
chemical analyses of sediment and water
samples. U.S. Environmental Protection
Agency/Corps of Engineers Technical Committee
on Criteria for Dredged and Fill Material. Contract
EPA-4805572010, EPA/CE-81-1. 478p.
Redmond, M.S. and K.J. Scott. 1989. Amphipod
predation by the infaunal polychaete, Nephtys
incisa. Estuaries. 12(3):205-207.
Reinharz, E. 1981. Animal sediment
relationships: A case study of the Patapsco River.
Open File No.6. Maryland Geological Survey,
Baltimore, MD.
Reish, D.J. 1988. The use of toxicity testing in
marine environmental research, Chapter 10. In
Marine Organisms as Indicators. D.F. Soule and
G.S. Kleppel eds. Springer-Verlag, New York, NY.
pp. 213-245.
Rohlf, F.J. and R.R. Sokal. 1981. Statistical
Tables. W.H. Freeman and Company, New York,
NY.
SAIC. 1993a. Draft data report on comparative
sensitivity of four amphipod species to sediment-
associated contaminants. Submitted to USEPA.
EPA contract No. 68-CO-0044. Prepared by
Science Applications International Corporation,
Narragansett, Rl. 15 pp.
SAIC. 1993b. Draft data report on salinity
tolerance and application for four amphipod
species. Submitted to USEPA. EPA contract No.
68-CO-0044. Prepared by Science Applications
International Corporation, Narragansett, Rl. 22 pp.
SAIC. 1993c. Draft data report on comparative
sensitivity of four amphipod species to aqueous
ammonia. Submitted to USEPA. EPA contract
No. 68-CO-0044. Prepared by Science
Applications International Corporation,
Narragansett, Rl. 30 pp.
98
-------
Sanders, H.L., P.C. Mangelsdorf, Jr., and G.R.
Hampson. 1965. Salinity and faunal distribution in
the Pocasset River, Massachusetts. Limnology
and Oceanography Supplement to Vol 10:R216-
R229.
Sarda, N., and G.A. Burton 1995. Ammonia
variation in sediments: Spatial, temporal, and
method-related effects. Environ. Toxicol. Chem.
14:1499-1506.
Sasson-Brickson, G. and G.A. Burton. 1991. In-
Situ and laboratory sediment toxicity testing with
Ceriodaphnia dulsia. Environ. Toxicol. Chem.
10:201-208.
Satterthwaite, F.W. 1946. An approximate
distribution of estimates of variance components.
Biom.Bull. 2:110-114.
Schaeffer, D.J. and Janardan, K.G. 1978.
Theoretical comparision of grab and composite
sampling programs. Biom. J. 20:215-227.
Schlekat, C.E., B.L. McGee, and E. Reinharz.
1992. Testing sediment toxicity in Chesapeake
Bay with the amphipod Leptocheirus plumulosus:
An evaluation. Environ Toxicol. Chem. 11:225-
236.
Schlekat C.E, K.J. Scott, R.C. Swartz, B. Albrecht,
L. Antrim, K. Doe, S. Douglas, J.A. Ferretti, D.J.
Hansen, D.W. Moore, C. Mueller, A. Tang. 1995.
Interlaboratory comparison of a 10-day sediment
toxicity test method using Ampelisca abdita,
Eohaustorius estuarius and Leptocheirus-
plumulosus. Environmental Toxicology and
Chemistry 14:2163-2174.
Schlekat, C.E. 1998. Assimilation of metals by
sediment-ingesting invertebrates: Effects of natural
sediment qualities and intrinsic metal
characteristics, Ph.D. Thesis. University of South
Carolina, Columbia, SC.
Schults, D.W., S.P. Ferraro, L.M. Smith, F.A.
Roberts, and C.K. Poindexter. 1992. A
comparison of methods for collecting interstitial
water for trace organic compounds and metal
analyses. Wat. Res. 26:989-995.
Schuytema, G.S., A.V. Nebeker, W.L. Griffis, and
C.E. Miller. 1989. Effects of freezing on
sediments contaminated with DDT and endrin.
Environ. Toxicol. Chem. 8:883-891.
Scott, J., C. Mueller, W. Starkel, and R. Mayer.
1996. Chronic response of Leptocheirus
plumulosus in 28-day exposures to contaminated
estuarine sediments. Poster presented at the
Society of Environmental Toxicology and
Chemistry 17lh Annual Meeting, November 17-21,
1996, Washington, D.C.
Scott, K.J. 1989. Effects of contaminated
sediments on marine benthic biota and
communities. In Contaminated Marine Sediments
Assessment and Remediation. National Research
Council. National Academy Press, Washington,
D.C. pp. 132-154.
Scott, K.J. and M.S. Redmond. 1989. The effects
of a contaminated dredged material on laboratory
populations of the tubicolous amphipod, Ampelisca
abdita. In Aquatic Toxicology and Hazard
Assessment: 12th Volume, ASTMSTP1027, U.M.
Cowgill and L.R. Williams, eds. American Society
for Testing and Materials, Philadelphia, PA.
pp. 289-303.
Scott. K.J. 1998. Citation pending.
Shapiro, S.S. and M.B.Wilk. 1965. An analysis of
variance test for normality (complete samples).
Biometrika 52:591-611.
Shoemaker, C.R. 1932. A new amphipod of the
genus Leptocheirus from Chesapeake Bay.
Wash. Acad. Sci. 22: 548-551.
Shuba, P.J., H.E. Tatem, and J.H. Carroll. 1978.
Biological assessment methods to predict the
impact of open-water disposal of dredged material.
U.S. Army Corps of Engineers Technical Report
D-78-5Q, Washington, D.C.
Sims, J.G. and D.W. Moore. 1995. Risk of pore
water hydrogen sulfide toxicity in dredged material
bioassays. Miscellaneous paper, D-95-4, U.S.
Army Engineer Waterways Experiment Station,
Vicksburg, MS.
99
-------
Sijm, R.T.H., M. Haller, and S.M. Schrap. 1997.
Influence of storage on sediment characteristics
and of drying sediment on sorption coefficients of
organic contaminants. Bull. Environ. Contam.
Tox/co/. 58:961-968.
Snrchek, J.C. and Zeeman, M. 1993. Assessing
risks to ecological systems from chemicals.
Chapter 3 In: Handbook for Environmental Risk
Assessment and Management, ed. P.Colow, pp.
24-90.
Snedecor, G.W. and G.C. Cochran. 1989.
Statistical Methods. Eighth edition. The Iowa State
University Press, Ames, IA. 507 p.
Sokal, R.R. and F.J. Rohlf. 1981. Biometry,
second edition. W.H. Freeman and Company, New
York, NY.
Southerland, E., M. Kravitz, and T. Wall. 1992.
Management framework for contaminated
sediments (The USEPA Sediment Management
Strategy). In Sediment ToxicityAssessment, G.A.
Burton, ed. Lewis Publishers, Ann Arbor, Ml.
pp. 341-370.
Steel, R.G.D. and J.H. Torrie. 1980. Principles
and Procedures of Statistics. McGraw Hill Book
Co, New York, NY.
Stemmer, B.L., G.A. Burton Jr., and G. Sasson-
Brickson. 1990a. Effect of sediment; spatial
variance and collection method on cladoceran
toxicity and indigenous microbial activity
determinations. Environ. Tox/co/. Chem. 9:1035-
1044.
Stemmer, B. L., G.A. Burton, Jr., and S. Leibfritz-
Frederick. 1990b. Effectofsedimenttestvariables
on selenium toxicity to Daphnia magna. Environ.
Toxicol. Chem. 9:1035-1044.
Stephan, C.E., D.I. Mount, D.J. Hansen, J.H.
Gentile, G.A. Chapman, and W.A. Brungs. 1985.
Guidelines for deriving numerical national water
quality criteria for the protection of aquatic
organisms and their uses. PB85-227049, National
Technical Information Service, Springfield, VA.
Swartz, R.C. 1989. Marine sediment toxicity tests.
In Contaminanted Marine Sediments-Assessment
and Remediation. National Research Council,
National Academy Press, Washington, DC. pp.
115-129.
Swartz, R.C., W.A. DeBen, J.K.P. Jones, J.O.
Lamberson, and F.A. Code. 1985. Phoxocephalid
amphipod bioassay for marine sediment toxicity.
In Aquatic Toxicology and Hazard Assessment:
7th Volume, ASTM STP 854, R.D. Cardwell, R.
Purdy, and R.C. Bahner, eds., American Society
for Testing and Materials,Philadelphia, PA.
pp. 284-307.
Swartz, R.C., P.F. Kemp, D.W. Schultz, and J.O.
Lamberson. 1988. Effects of mixtures of
sediment contaminants on the marine infaunal
amphipod, Rhepoxynius abronius. Environ.
Toxicol. Chem. 7:1013-1020.
Taylor, J.K. 1987. Quality Assurance of Chemical
Measurements. Lewis Publ., Inc., Chelsea, Ml.
Tetra Tech, Inc. 1986. ODES statistical power
analysis. Draft report, prepared for the Office of
Marine and Estuarine Protection, USEPA Contract
No. 68-01-6938, TC-3953-03. Bellevue, WA.
Tomasovic, M. et al. 1994. Recovery of Hyalella
azteca from sediment. Environ. Toxicol. Chem.
13:1311-1314.
USEPA. 1979a. Handbook for analytical quality
assurance in water and wastewater laboratories.
EPA-600/4-79-019, Cincinnati, OH.
USEPA. 1979b. Methods for chemical analysis of
water and wastes. EPA-600/4-79-020, Cincinnati,
OH.
USEPA. 1980a. Proposed good laboratory
practice guidelines for toxicity testing. Paragraph
163.60-6. Fed. Reg. 45:26377-26382 (April 18,
1980).
USEPA. 1980b. Physical, chemical, persistence,
and ecological effects testing; Good laboratory
practice standards (proposed rule). 40 CFR 772,
Fed. Reg. 45:77353-77365 (November 21,1980).
USEPA. 1985. Methods for measuring the acute
toxicity of effluents to freshwater and marine
organisms. Third edition. EPA-600/4-85/013,
Cincinnati, OH.
100
-------
USEPA. 1986. Occupational health and safety
manual. Office of Administration, EPA,
Washington, D.C.
USEPA. 1989a. Evaluation of the apparent
effects threshold (AET) approach for assessing
sediment quality. Report of the sediment criteria
subcommittee. SAB-EETFC-89-027, Washington,
D.C.
USEPA. 1989b. Toxic Substance Control Act
(TSCA); Good laboratory practice standards, Final
Rule, Federal Register 54: 34034-34050. August
17, 1989.
USEPA. 1990a. Evaluation of the equilibrium
partitioning (EqP) approach for assessing
sediment quality. Report of the sediment criteria
subcommittee of the ecological processes and
effects committee. EPA-SAB-EPEC-90-006,
Washington, D.C.
USEPA. 1990b. Evaluation of the sediment
classification methods compendium. Report of the
sediment criteria subcommittee of the ecological
processes and effects committee. EPA-SAB-
EPEC-90-018, Washington, D.C.
USEPA. 1990c. Managing contaminated
sediments. EPA decision-making process.
EPA/506/6-90/002.
SEPA. 1991 a. Methods for measuring the acute
toxicity of effluents to freshwater and marine
organisms. Fourth edition. EPA 600/4-90/027F,
Cincinnati, OH.
USEPA. 1991b. Sediment toxicity identification
evaluation: Phase I (Characterization), Phase II
(Identification) and Phase III (Confirmation).
Modifications of effluent procedures. EPA 600/6-
91/007, Duluth, MN.
USEPA. 1991c. Technical support document for
water quality-based toxic control. EPA-505/2-
90/001, Washington, D.C.
USEPA. 1992a. Proceedings from workshop on
tiered testing issues for freshwater and marine
sediments. Washington, D.C., September: 16-18,
1992.
USEPA. 1992b. An SAB report: Review of
sediment criteria development methodology for
non-ionic organic contaminants. Report of the
sediment criteria subcommittee of the ecological
processes and effects committee. EPA-SAB-
EPEC-93-002, Washington, D.C.
USEPA. 1992c. Sediment classification methods
compendium. EPA-823-R-92-006, Washington,
D.C.
USEPA. 1994a. Procedures for conducting
sediment toxicity and bioaccumulation tests using
freshwater invertebrates. First Edition. EPA-
600/R-94/024, Duluth, MN.
USEPA. 1994b. Short-term methods for
estimating the chronic toxicity of effluents and
surface waters to freshwater organisms. Third
edition. EPA-600/4-91/002, Cincinnati, OH.
USEPA. 1994c. Short-term methods for
estimating the chronic toxicity of effluents and
surface waters to marine and estuarine organisms.
Second edition. EPA-600/4-91/003, Cincinnati,
OH.
USEPA. 1994d. Methods for assessing the
toxicity of sediment-associated contaminants with
estuarine and marine amphipods. EPA/600/R-
94/025, Office of Research and Development,
Washington, D.C.
USEPA. 1995. QA/QC guidance for sampling and
analysis of sediments, water, and tissues for
dredged material evaluations—chemical
evaluations. EPA823-B-95-001. Office of Water,
Washington, D.C.
USEPA. 1996. Marine toxicity identification
evaluation (TIE). Phase I Guidance document.
EPA/600/R-96/054. Burgess, R.M., K.T. Ho, G.E.
Morrison, G. Chapman, D.L. Denton, eds. U.S.
Environmental Protection Agency, National Health
and Environmental Effects Research Laboratory,
Narragansett, Rl.
USEPA. 1997a. The Incidence and Severity of
Sediment Contamination in Surface Waters of the
United States, Volume 1: National Sediment
Quality Survey. EPA 823-R-97-006. Office of
Science and Technology, Washington, D.C.
101
-------
USEPA. 1997b. The Incidence and Severity of
Sediment Contamination in Surface Waters of the
United States, Volume 2: Data summaries for
areas of probable concern. EPA 823-R-97-007.
Office of Science and Technology, Washington,
D.C.
USEPA. 1997c. The Incidence and Severity of
Sediment Contamination in Surface Waters of the
United States, Volume 3: National sediment
contaminant point source inventory. EPA 823-R-
97-008. Office of Science and Technology,
Washington, D.C.
USEPA. 1998. Contaminated sediment
management strategy. EPA823-R-98-001. Office
of Water, Washington, D.C.
USEPA. 1999. EPA requirements for quality
assurance project plans for environmental data
operations. Draft interim final. EPA QA/R-5.
Office of Research and Development, Washington,
D.C.
USEPA. 2000. Methods for measuring the toxicity
and bioaccumulation of sediment-associated
contaminants with freshwater invertebrates,
second edition EPA600/R-99/064, Duluth, MN.
USEPA. 2001. Methods for collection, storage,
and manipulation of sediments for Chemical and
Toxicological Analyses (January, 2001 draft).
USEPA-USACE. 1977. Ecological evaluation of
proposed discharge of dredged material into ocean
waters. Technical commitee on criteria for
dredged and fill material. Environmental Effects
Laboratory. U.S. Army Engineer Waterways
Experiment station. Vicksburg, MS.
USEPA-USACE. 1991. Evaluation of dredged
material proposed for ocean disposal (Testing
manual). EPA-68-C8-0105. USEPA, Office of
Marine and Estuarine Protection, Washington,
D.C.
USEPA-USACE. 1998. Evaluation of dredged
material proposed for discharge in waters of the
U.S. testing manual. EPA-823-B-98-004
Washington, D.C.
Van Rees, K.C.J., E.A. Sudlicky, P. Suresh, C.
Rao, and K.R. Reddy. 1991. Evaluation of
laboratory techniques for measuring diffusion
coefficients in sediments. Environ. Sci. Technol.
25:1605-1611.
Walters, D.B. and C.W. Jameson. 1984. Health
and Safety for Toxicity Testing. Butterworth
Publications, Woburn, MA.
Wiederholm, T., A.M. Wiederholm, and M. Goran.
1987. Fresh-water oligochaetes. Water Air Soil
Pollut. 36:131-154.
Winger, P.V. and Lasier, P.J. 1993. Vacuum-
extracted pore water toxicity testing. In
Environmental Toxicology and Risk Assessment,
2nd volume, eds. J.W. Gorsuch, F.J. Dwyer, C.G.
Ingersoll, T.W. LaPoint, 640-662. ASTM STP
1173, Philadelphia, PA.
Winger, P.V., Lasier, P.J., and Jackson, B.P. 1998.
The influence of extration procedure on ion
concentration in sediment pore water. Arch
Environ. Contam. Toxicol. 35:8-13.
Williams, D.A. 1971. A test for differences
between treatment means when several dose
levels are compared with a zero dose control.
Biometrics. 27: 103-117
Williams, L.G., P.M. Chapman, and T.C. Ginn.
1986. A comparative evaluation of marine
sediment toxicity using bacterial luminescence,
oyster embryo, and amphipod bioassays. Marine
Environmental Research, 19: 225-249.
Williams, L.R. 1993. Methods for the EPA's
regulatory program. Environmental Testing and
Analysis. 2:36.
Winer, B.J. 1971. Statistical Principles in
Experimental Design. Second edition. McGraw-Hill
Book Company, New York, NY. 907 p.
Word, J.Q., J.A. Ward, L.M. Franklin, V.I. Cullinan,
and S.L. Kiesser. 1987. Evaluation of the
equilibrium partitioning theory for estimating the
toxicity of the nonpolar organic compound DDT to
the sediment dwelling amphipod, Rhepoxynius
abronius. Battelle Marine Science Research
Laboratory Report, Task 1, WA56, Sequim, WA.
60pp.
102
-------
Word, J., W. Gardiner, and M. Pinza. 1997. Zar, J.H. 1984. Biostatistical Analysis. Second
Confounding factors related to marine sediment edition. Prentice-Hall, Inc., Englewood Cliffs, NJ.
ecotoxicological testing. PNWD-SA-4435. 1997 717 p.
Short Course, Society of Environmental
Toxicology and Chemistry.
103
-------
-------
Appendix A
Example Data Sheets For Use with the
28-d Chronic L plumulosus Test
-------
-------
Water Quality Measurements
Project Name:
Duration:
Project No:
Test Day:
Test type:
Date:
Species:
Page
of
Position
Treatment
Rep.
Temperature
(24-26°C)*
pH
(7.0 - 9.0)*
Dissolved
Oxygen
(>4.0 mg/l)
Salinity
(18-22 pot)
Recorder:
Test acceptability limits; take corrective action if values are outside limits.
Reveiewed by:
Date:
-------
Daily Observations
Project Name: Project No:
Duration: Test Day:
Test type: Date:
Species: Page of
Position Number on Number
Number Sediment Floating Comments
Recorder:
Reviewed by Date
-------
Overlying Water Renewal
Project Name: Project No:
Duration: Test Day:
Test type: Date:
Species: Page of
RENEWAL:
Date
Monday, Wednesday, and Friday
With the designated small peristaltic pump and correct hose for the
type of container, remove 400 mL overlying water from each jar and
then replace it with 400 mL of 20%o seawater at test temperature
Water Animals
M,W, or F Test Day Renewed Time Initials Fed Time Initials
Reviewed by: ' Date:
-------
Animal Feeding
Project Name:
Duration:
Test type:
Species:
Project No:
Test Day:
Date:
Page
of
FEEDING: Monday, Wednesday, and Friday
Test Days 0-13 = 20 mg ground Tetramin/test chamber
Test Days 14-28 = 40 mg ground Tetramin/test chamber
Example calculation If you have 60 test chambers- prepare
1200 mg of tetramin In 60 mLs of 20 ppt seawater for day
0-13 and 2400 mg of tetramin in 60 mLs of 20 ppt seawater
for days 14-28
Date Test Day Time Amt. of Food per Container Initials
Reviewed by: Date:_
-------
Termination Form
Project Name: Project No:
Duration: ' Test Day:
Test type: Date:
Species: Page of
Position Replicate Treatment # Live # Dead # Missing Comments
Recorder:
Reviewed by Date
-------
Neonate Counting Form
Project Name: Project No:
Duration: Test Day:
Test type: Pate:
Species: Page of
Position
Number Treatment Replicate Count 1 Initials Count 2 Initials
Reviewed by Date
-------
-------
-------
-------
O CD
•<
o'
S-'
CD
CO
CD
^^ m c |
co o <. ~
?£!3 S.
e
p
p
IV3
O
O3
O
I %
CD g.
S[ CO
-a
3
S"
o"
CD
CD
O
------- |