00526
Test Methods to Estimate the Acute and' Chronic Toxxcity and Bioaccumulation
•*" •
of Sediment-Associated Contaminants using the Aquatic Oligochaete,
Lumbriculus varieKat.ua
by
Gary L. Phipps and Gerald T. Ankley
U.S. EnviroiTJEsiital Protection Agency
Environmental Research Laboratory-Duluth
6201 Congdc?n Boulevard
Duluth, MTi 55804
-------
TABLE OF CONTENTS
Overview 1
Introduction 2
Background 2
Benthic Test Species 2
Oligochaetes in Aquatic Toxicology 4
Rationale for Species Selection 5
Methods and Materials 6
General 6
Life History and Life Cycle 6
Collection 6
Culturing ..... 7
Handling 8
Testing 8
Test System 8
Toxicity Tests 1CL
Bioaccumulation Tests 11
Feeding 12
Control Sediments 14
Sample Sizes and Statistical Techniques 15
Quality Assurance 17
Results and Discussion 17
Case Study One - Foundry Cove 17
Case Study Two - Keweenaw Waterway/Torch Lake 18
Summary and Conclusions 19
Acknowledgements 20
References 21
Figures
Tables
-------
OVERVIEW
There recently has been increased concern about the impacts of
contaminated sediments on aquatic ecosystems. An important shortcoming of
ongoing and planned sediment assessments, particularly from the standpoint of
the U.S. Environmental Protection Agency (EPA), is the lack of standard
techniques such as toxicity tests for evaluating potential impacts of
sediment-associated contaminants. In this report we describe methods
utilizing the aquatic oligochaete, Lumbriculus variegatus. to assess the acute
and chronic toxicity, and the presence of bioaccumulatable compounds in
contaminated sediments.
L. variegatus was chosen as a test species because: (1) it is
ecologically relevant (i.e., it has a wide distribution and is an important
component of aquatic food chains), (2) it is suitable for long-term testing
and evaluation of chronic toxicity endpoints (e.g., growth, reproduction), (3)
it is exposed via all important routes of concern, including ingestion of
contaminated particles, and (4) it has sufficient biomass to assess *
bioaccumulation of contaminants. Also, this species is easily cultured under
a variety of conditions (described herein) and is relatively easy to handle.
In addition to describing culture conditions, testing protocols (e.g.,
test lengths, sample sizes, feeding, etc.) and presenting the results of
actual tests with the worms, we give specifications for an automated renewal
exposure system which is suitable for testing contaminated sediments with a
variety of species.
-------
INTRODUCTION
Background - Recent surveys have amply demonstrated the extent of sediment
contamination in the U.S. (U.S. EPA 1987). Sediment contamination problems at
freshwater sites are prevalent throughout the country and include 41 major
areas of concern in the Great Lakes alone (International Joint Commission
1985). To address issues associated with contaminated sediments, extensive
remedial action planning, and in many cases Superfund activities, are underway
at a number of sites. Major efforts such as the Assessment and Remediation of
Contaminated Sediments program, through the EPA Great Lakes National Program
Office, have been specifically mandated by law. Although there is a
tremendous amount of activity focused upon contaminated sediments,
standardized tools such as toxicity tests, for addressing existing or
potential impacts of sediment-associated contaminants on aquatic ecosystems
are not available. This shortcoming is of particular concern to the EPA, not
only from an assessment standpoint, but because the lack of standardized test
methods seriously impedes cohesive regulatory activities. The objective of
this report is to present test methods, using a freshwater benthic ^
invertebrate, for assessing the toxicity and bioaccumulation of sediment-
associated contaminants.
Benthic Test Species - An important component of the assessment of
contaminated sediments is the use of toxicity tests to evaluate impacts on
benthic and upper-water column species. An "ideal" suite of toxicity tests
would have a number of attributes including: (1) ecological relevance, (i.e.,
the organisms are potentially important species in the system(s) of concern),
(2) ability to assess chronic endpoints such as reproductive effects, (3)
protective of the most sensitive species in the system, (4) incorporate all
possible routes of exposure (i.e., exposure to interstitial water, ingestion
of contaminated particles), (5)'-ability^to measure bioaccumulation to help
: "•' '•••'. -'•> f. -.;»...
assess possible impacts on higher .organisms.exposed' through food chain
transfer (biomagnification), and (6) utilize species amenable to culturing and
-------
handling. Many different organisms and test endpoints have been proposed for
assessing contaminated sediments (Giesy and Hoke 1990). However, the majority
of these tests do not conform to all of the above criteria. For example,
several studies have used standard upper-water column test organisms, such as
cladocerans and fish, to assess the toxicity of contaminated sediments;
however, these organisms generally are not relevant if species of concern are
benthic, particularly in terms of adequately addressing all possible routes of
exposure.
Three freshwater benthic organisms that have received significant
attention with respect to testing contaminated sediments, and for which some
test methods exist, are the amphipod Hyalella azteca, and the chironomids
Chironomus tentans and C. riparius (Nelson et al. 1990). These species are
ecologically-relevant, easily cultured and, because they are benthic and/or
epibenthic organisms, are exposed to contaminants through various routes.
Moreover, with these three species it is possible to examine a variety of
endpoints, including those traditionally used as measures of chronic toxicity
such as growth (C. tentans, C. riparius) and reproduction (H. azteca).
However it is difficult to use these three species to evaluate the impact
and/or presence of bioaccumulatable contaminants. Tissue masses obtained from
H. azteca generally are too small to collect enough sample to analyze, and the
life cycles of both chironomid species may be too rapid to adequately assess
the presence of bioaccumulatable contaminants (particularly if tests are
conducted with second instar larvae). Therefore, to compliment these three
test species it would be useful to use an organism that could be used in long-
term tests, and have sufficient biomass for analytical work. An excellent
first choice would be the mayfly (Hexagenia limbata), because it is considered
to be relatively sensitive and has a large tissue mass. However, H. limbata
currently cannot be cultured, so it is not a desirable species for routine
testing. A logical alternative for long-term testing and associated
analytical work involving bioaccumulatable compounds would be an oligochaete.
Oligochaetes are ubiquitous in a variety of freshwater benthic habitats, they
-------
have adequate body mass to perform analytical work with a logistically
reasonable number of organisms, and certain species are easily cultured.
Also, because of their prevalence in aquatic communities, oligochaetes are
particularly well suited for use in bioaccumulation studies and in assessing
potential for food chain transfer. Oligochaetes traditionally have been
considered to be relatively tolerant of certain classes of contaminants.
However, when assessing bioaccumulatable compounds, the tolerance of these
species perhaps is a positive attribute (Schuytema et al. 1990).
Oligochaetes in Aquatic Toxicology - Various investigators have evaluated the
use of distributions of oligochaete species as indicators of pollution in situ
(Brinkhurst 1980; Spencer 1980; Lauritsen et al. 1985; Robbins et al. 1989).
Results of these studies suggested that it was possible to identify
oligochaete assemblages representative of the degree of organic contamination
present at various study sites.
A number of researchers have used various oligochaete species in
laboratory toxicity tests with pure compounds and/or contaminated sediments.
Chapman et al. (1982a; 1982b) evaluated the toxicity of single chemicals, and
combinations of chemicals, to up to 12 different freshwater and marine
oligochaete species. They also investigated the influence of environmental
variables, e.g., pH, salinity, temperature and anoxia, on the toxicity of the
test chemicals to the oligochaetes. Wiederholm et al. (1987) described
testing protocols for five species of freshwater oligochaetes, and noted that
in sediments polluted by heavy metals, growth and reproduction were more
sensitive endpoints than survival. Keilty et al. (1988a; 1988b) reported the
toxicity of endrin to two species of freshwater oligochaetes, and concluded
that a behavioral response (burrowing) of the organisms appeared to be a
promising sublethal endpoint. Nebeker et al. (1989) investigated the toxicity
of hexachlorobenzene to the freshwater oligochaete species, Lumbriculus
variegatus, and found that tissue concentrations of the chlorinated compound
as high as 24 Mg/g had no adverse effects on survival, growth, or reproduction
-------
•of the worms. This agreed with observations made using "standard" test
species, e.g., the fathead minnow, PLmephales promelas.
Oligochaete species have also been used to evaluate the bioaccumulation
of nonpolar organic compounds in a laboratory setting. Oliver (1984)
investigated the uptake of 24 chlorinated chemicals by a "natural" oligochaete
assemblage exposed to Lake Ontario sediments for up to 110 d. He found that
the uptake of chemicals from test sediments by the oligochaetes was
approximately proportional to sediment concentrations of the chemicals. He
also noted that the Kow (octanol water partition coefficient) of the test
chemicals was integral in determining uptake kinetics. Connell et al. (1988)
reanalyzed the data of Oliver (1984), and concluded that this phenomenon could
be explained by the theory of equilibrium partitioning, assuming that the
primary exposure phase was sediment interstitial water. Schuytema et al.
(1990) exposed L. varLegatus to hexachlorobenzene, and based on
bioaccumulation, concluded that this oligochaete was a promising species for
monitoring bioaccumulatable nonpolar organics in environmental samples.
Rationale for Species Selection - We chose the oligochaete, L. variegatus. as
a test species for a number of reasons. L. variegatus is ecologically-
relevant, it is prevalent throughout the U.S. and Europe, it is present in all
five of the Great Lakes, and occurs in a great variety of sediment types
(Chekanovskaya 1962; Cook 1969; Spencer 1980). It is possible to use the
species to assess chronic toxicity endpoints such as growth and reproduction
(Nebeker et al. 1989). L. varLegatus, a truly benthic organism, is exposed to
contaminants via all routes of concern, including ingestion of contaminated
particles. Since a logistically reasonable number of organisms can provide
adequate tissue mass for residue analyses, (i.e., about 80-100 per g wet
weight) this species is ideal for long-term bioaccumulation studies. L.
varLegatus is extremely easy to culture and maintain in the laboratory year
round, and is relatively easy to handle (e.g., removal from test sediments).
-------
Finally, there is some information in the open literature describing the use
of L. variegatus in laboratory toxicity tests (Nebeker et al. 1989; Schuytema
et al. 1990).
METHODS AND MATERIALS
General
Life History and Life Cycle - L. variegatus normally dwells in silty and sandy
sediments at depths of 2 to 60 meters in reservoirs, rivers, lakes, ponds, and
marshes (Chekanovskaya 1962). The anterior half of an individual is usually
buried in the sediment while the posterior half undulates in the overlying
water to achieve respiratory exchange. While feeding on organic material in
the sediment the worms will tunnel to a depth corresponding to the aerobic
zone of the sediment, and do not penetrate the anaerobic zone appreciably.
L. variegatus has a sexual reproductive potential. However, little is
known about it, and Chekanovskaya (1962) states that individuals with sexual
organs are extremely rare. The worm's most common mode of reproduction is
architomy, where new individuals are budded off the anterior end of the
parent, and subsequently are replaced with eight new segments. Body length
varies from 40-90 mm and diameter varies from 1.0 to 1.5 mm. Newly hatched
worms have never been observed in our cultures, which consist solely of adults
of various sizes. In our laboratory, the population doubling rate is
approximately 10 to 14-d at 20°C. The population growth rate seems to be
related to how quickly the neonates grow to a size where they will again
divide.
Collection - L. variegatus may be found in many field locations. While it is
possible to start a culture with these organisms, the easiest way is to obtain
them from another laboratory. This eliminates the need to find an adequate
number of worms and positively identify the brood stock, which can be
difficult because of the morphic structures necessary for identification
(Pennak 1978). Brood stock can be successfully shipped by placing 100-200
-------
organisms in 200 mL of clean water and protecting them from being crushed or
exposed to extreme temperatures. Since these organisms are quite tolerant to
changes in temperature, dissolved oxygen (DO) and pH, elaborate acclimation
schemes generally are not necessary before they are introduced to new culture
systems. Several laboratories which currently maintain L. variegatus
cultures, as well as the appropriate contacts, are listed in Table 1.
Culturing - L. variegatus have been successfully cultured year round in our
laboratory using a variety of substrates and foods. The substrates used
include: shredded brown paper toweling (our recommended choice because of
ready availability, relatively well defined characteristics, and ease of
harvesting the worms), presoaked, dried maple and/or poplar leaves, or
organically rich, clean sediments. The substrates are placed on the bottom of
an aquarium at 3 to 6 cm depth. Filtered Lake Superior water flows through 57
L aquaria (100 mL/min) supplying bacteria, fungi, algae and other organisms,
all of which contribute to the culture/substrate food complex. This substrate
is supplemented twice weekly with about 3 mL of settled, hatched brine shrimp,
nauplii, and 1 to 2 g of U.S. Fish and Wildlife Service-certified salmon
starter every 1 to 2 wk depending upon biomass in the aquarium. We have had
excellent success in culturing the worms with about 50 snails, Helisoma sp..
which process the substrates reducing them to particle sizes that the worms
can easily ingest. The worms can be cultured without snails, but they seem to
reproduce more efficiently in their presence. Culture tanks do not require
frequent cleaning when the snails are present.
Cultures must not be allowed to become anaerobic; this possibility is
minimized by culturing the worms in a flow-through system. However, with
proper aeration and food limitations it is possible to culture L. variegatus
in a static system. A rule-of-thumb is, if the culture shows signs of
degraded water quality (e.g., smells offensive, has low DO, etc.) it has been
over-fed. This can be corrected by increasing the aeration rate and/or water
exchange rate. The optimal water flow in any particular system will be
-------
related Co biomass and biological oxygen demand, and therefore needs to be
monitored and adjusted as appropriate. We recommend that DO in the cultures
be maintained at > 60% saturation.
L. variegatus production increases with increasing temperature, but the
organisms should not be cultured above 25°C. Ideally they should either be
cultured within 2°C of the projected test temperature, or be temperature-
acclimated for at least 24-h before testing.
Light is supplied by a single cool white 20 watt fluorescent bulb for 3
aquaria, situated about 60 cm above the tank bottoms. The photoperiod is 16 h
light: 8 h dark throughout the year.
Handling - If paper toweling is used as the substrate, the worms and snails
will process it to a wet powder in 1 to 2 mo. The worms inhabiting this wet
powder are easily collected for testing. They can be removed from the culture
aquaria with a dipnet and placed in a flat, shallow-sided pan, rinsed, counted
and distributed. A pipette made from 4 mm I.D. glass, 20 cm long and fire
polished on both ends, fitted with a pipette bulb, is used for handling. The -
worms will clump and can be separated by a stream of water from the pipette.
It is not advisable to separate the worms mechanically since they can be
easily injured. Damaged or inactive worms should be discarded.
Testing
Test System - Toxicity and bioaccumulation tests are conducted in a system of
12 exposure tanks, 30 cm x 16 cm x 13 cm high (Fig. 1). This system is an
adaptation of the serial diluter described by Benoit et al. (1982). The tanks
have self-starting standpipe siphons 9.5 cm high on one end (Figure 1), which
automatically vary the height of the water in the tank by about 2.5 cm. The
tanks each hold eight 300 mL high form glass, berzelius exposure beakers.
Each beaker has two 2 cm holes drilled at a 180° angle from one another other
and 8 cm from the bottom of the beaker (Fig. 1). The holes are covered with
60 mesh stainless steel screen held in place with silicone adhesive. When a
8
-------
test is initiated, 100 mL of sediment is placed in the bottom of the beaker,
the beaker is then placed into the test tanks, leaving approximately 100-150
mL of overlying water in the beaker. Each different sediment in a test has
its own test tank to prevent cross contamination with other sediments. As the
water rises and falls in the tank (effected by the self starting siphons), the
water in each test beaker is exchanged equally with tank water. With 12 tanks
in a diluter system it is possible to test 12 different sediments with
multiple replicates and one species, or several species simultaneously. For
example, we have simultaneously tested L. variegatus. H. azteca. and C.
tentans larvae in this system.
A timer-controlled solenoid valve on the water supply line of the
diluter enables us to control the water replacement times in the test tanks.
This allows adjustment of water flow through the test tanks such that adequate
DO, pH, ammonia and test temperatures can be maintained without excessively
stripping the test sediments of the inherent toxic materials. This system
could be best described as an automatic renewal system, with a range of
renewal from static to 10 tank replacements per 24 h.
This basic test system has also been used with L. variegatus for
bioaccumulation studies with nonpolar organic chemicals (e.g., pesticides,
PCBs). In these studies, a larger volume of sediment is required to expose
enough organisms for analytical work. To accomplish this, the exposure
beakers are not used and varying amounts of sediment are added directly to the
test tank.
Before testing it is important to remove any toxic residues remaining in
the system. A good routine to clean residues from the test system is: (1)
soap and water wash, (2) dry, (3) rinse with acetone, (4) dry, (5) acid soak
for 30 minutes in 10% nitric acid, (6) rinse in distilled water and (7) dry.
The system then should be flushed with clean water for at least 4 d before
initiation of further tests. If it is known that there are no organic
contaminants the acetone rinse may be omitted.
-------
Although the exposure system described above is convenient for
evaluating toxicity and/or bioaccumulation of sediment-associated
contaminants, a variety of other types of systems can be used for testing L.
variegatus. For example, it is possible to test the organisms in beakers
containing sediment, with overlying water that is renewed often enough to
maintain adequate water quality. Alternatively, we have used L. variegatus in
water-only exposures with various chemicals and aqueous sediment test
fractions (e.g., sediment interstitial water), also in a renewal situation.
We feel that L. variegatus can be tested in a variety of systems and
situations, providing that appropriate controls are maintained.
Toxicity Tests - Whole sediments are homogenized so that all test beakers
receive uniform 100 ml samples. The beakers are then placed into the
appropriate test tanks where the water is flowed slowly into the beakers. The
system is operated for 24 h to equilibrate. A renewal rate of two to six
water turnovers daily is maintained. Organisms (usually 10) are randomly
placed into the beakers and the time noted as the start of the test. We have -
run toxicity tests ranging from 4 d to 28" d; however, we generally do not
recommend the use of tests of less than 10 d in duration. At the end of the
test, worms are removed by sieving with a standard test sieve, (No. 35, 500
/zm) counted, placed in dried, pre-weighed aluminum weigh boats, a drop of
ethanol added (to immobilize the worms), and the weigh boats placed into a
drying oven at 100°C for 4 h to 24 h, after which the worms are weighed.
Individual dry weight ranges from 0.5 to 2.2 mg per organism. The toxicity
test endpoints include survival, reproduction, and growth, expressed either as
total biomass or biomass per individual. Note that in tests greater than 10-d
in length, L. variegatus usually will reproduce, and given its mode of
reproduction (i.e. architomy), it is impossible to differentiate between young
and adult organisms. This necessitates the treatment of survival and
reproduction as a single endpoint.
10
-------
In some instances test sediments will contain oligochaetes which might
be confused with L. variegatus: however, in our experience we have not noted
that other oligochaete species present in test samples resembled L. variegatus
to the extent that problems were encountered (e.g., it is quite easy to
differentiate between L. variegatus and the ubiquitous tubificid species). If
problems are encountered where oligochaetes present in test sediments resemble
L. variegatus. more extensive morphological analysis may be required to
interpret test data. There is the possibility that L. variegatus may actually
be present in the test sample. In that case, the sediment probably is not
toxic to the species. Some have removed "extra" species in test sediments by
sieving, drying or freezing samples prior to testing; however, we strongly
advise against this because of the potential for dramatic changes in the
chemical and toxicological nature of the sediments (e.g., oxidation of sulfide
with the concomitant release of metals; Ankley et al. 1990; DiToro et al.
1990). To evaluate whether oligochaetes are present in test sediments, we
recommend routinely running an extra beaker with no organisms added, which
then is seived and checked for native oligochaetes when the exposure is
completed. This type of a control will thus alert the investigator to the
possibility of ambiguous test results.
Bioaccumulation Tests - Bioaccumulation tests generally are set up with larger
volumes of sediment because a larger biomass of worms is needed (particularly
in the case of nonpolar organics). Sediments are placed in the test tank at
1.5 to 2.0 L per tank. The sediments are allowed to settle for 24 h before
the test begins. The worms are weighed and placed in the sediment at the mass
required for analytical measurements. We use 1.0 g wet weight of worms (about
80-100 organisms) per tank. At this stocking rate, if it is assumed for
example, that 1 g of worms is exposed to 2 L of sediment, with an organic
carbon content of 5% (dry weight, assuming 50% water), this would yield a
sediment carbon/organism carbon of about 50 to 100, which should be high
enough to ensure that the worms do not exhaust the bioaccumulatable nonpolar
11
-------
organic compounds in Che sediments. As for toxicity tests, the system is
renewed with two to six water volume turnovers daily throughout the test. The
test is conducted for the desired length of time (we have run bioaccumulation
tests for up to 60-d) , the worms are removed from the sediment with the
standard testing sieve (No. 35, 500 pm) and put in clean water, where they are
held for 24 h to purge gut contents, and then rinsed again to remove remaining
debris. The worms are placed into acid- and solvent-cleaned glassware prior
to chemical analysis.
Feeding - We have maintained L. variegatus in Lake Superior water without
feeding for 35 d, with no loss of animals. Because feeding alters the organic
carbon content and possible bioavailability of contaminants in sediments,
feeding the worms during bioaccumulation tests is not recommended. However,
it is necessary to feed the worms a minimal amount of food when testing for
toxicity, because if one of the test sediments is low in organic carbon, it
may appear to be toxic, when, in fact, adverse impacts (i.e., low reproduction
and/or weight gain) may be caused by a lack of nutrients.
A series of tests was conducted to help define the influence of organic
carbon on L. variegatus growth and reproduction. A clean sediment with low
organic carbon (1 to 2% organic carbon) was compared to a clean sediment with
high organic carbon (8 to 10% organic carbon). In initial experiments,
sediments were tested for 12 days with variable worm loading rates (2, 5, 10
or 20 organisms per 100 mL sediment), using four replicate beakers per
treatment. Reproduction and biomass were consistently lowest at all stocking
rates in the low organic carbon content sediment. For example, reproduction
in the high organic carbon sediment was 62, 25, 63 and 72 greater than in the
low carbon sediment for the 2, 5, 10 and 20 worm stocking rates, respectively.
A second test was conducted, using various feeding rates, where L.
variegatus reproduction and growth in the sediment with low organic carbon, as
well as in sand with no organic carbon, were evaluated. Initial worm loading
was 10 animals per 100 mL of sediment. Feeding rates were 0, 0.11, 0.33, 1.0
12
-------
and 3.0 g of U.S. Fish and Wildlife Service-certified salmon starter stirred
into each 100 mL of sediment or sand at the beginning of the test. Water was
renewed at two volumes per day. Water quality was significantly impacted by
all treatments receiving more than 0.11 g of salmon starter, and most worms
died. However, animal numbers and biomass at the 0.11 g feeding level in both
the low organic carbon sediment and the sand were comparable to values
obtained using the high organic carbon sediment described above.
A third test was conducted in which triplicate treatments were fed 0,
20, 40 or 80 mg of salmon starter settled to the sediment surface every third
day, rather than the larger, initial amount stirred into the sediment, as in
the second test. The same low organic carbon sediment and sand as in the
second test were used. Table 2 gives the mean numbers of worms, total
biomass, and individual weights of worms for each treatment. Numbers of
worms, total biomass and individual weights were higher in the 20 mg/feeding
than in the treatments that were not fed. The number of worms, total biomass
and mean weight/worm were reasonably similar in the 20 mg/feeding for both the
low organic carbon sediment and sand. However, the biomass was slightly
greater in the sediment than in the sand.
When the feeding rate was increased from 20 mg/feeding to 40 mg/feeding
the mean number of worms decreased from 25.7 to 18.3 in the low carbon
sediment but increased from 23.3 to 25.0 in the sand. Although total biomass
in the low organic carbon sediment and sand was similar, mean weight per worm
was elevated in the low organic carbon sediment. The 80 mg/feeding treatment
showed an increase in individual worm weights but a decrease in worm numbers
and biomass, therefore, this feeding rate was definitely detrimental to the
worms. The choice between the 20 mg and 40 mg feeding rate is more difficult
to make. However, the 20 mg/feeding treatment provided fairly good agreement
between the three measured parameters in the two sediment types, while the 40
mg/treatment differed in worm numbers and mean weight/worm between the two
sediments, possibly indicating over-feeding.
13
-------
Based on results from this test, we recommend the addition of 20 mg of
trout starter per 100 mL sediment on every third day of a test. The food
should be added to the water overlying the sediment, stirred and allowed to
settle to the sediment surface. Our experience in feeding L. variegatus
during testing is somewhat limited; for example, there may be foods better
than salmon starter to use when testing and we are investigating this.
Because of varying biological oxygen demands for different sediments,
and the potential for exacerbation of biological oxygen demand by feeding, DO
must be closely monitored to assure that 60% of saturation is maintained. If
DO falls below 60% of saturation, the water flow rate must be increased or
static tests must be aerated; it also may be necessary to decrease the feeding
rate.
Control Sediments - The issue of control samples in toxicological research
with sediments is extremely complex; to date, there have been no entirely
satisfactory methods developed to address the issue of a suitable control for
sediment toxicity tests. A "true" control would exactly mimic the test sample-
in all respects except for contaminants; this is a nearly impossible
challenge. The basic problem in identifying suitable controls is related to
the fact that physical/chemical conditions, such as particle size or organic
carbon content, in the test sediments can markedly influence responses of test
organisms. Therefore, less than optimal physical/chemical characteristics in
the test sediments could result in adverse effects in the toxicity test, which
may be independent of the presence of toxic contaminants. Probably the best
treatment of the issue of appropriate control sediments was a statistical
technique presented by DeWitt et al. (1988); however, this approach requires
far more data than currently are available for L. variegatus.
The inability to define "true" controls for the L. variegatus test may
not be a serious problem. Based upon work by other researchers (e.g., DeWitt
et al. 1988), it appears that two of the most important physical/chemical
characteristics which influence the health of benthic species in clean
14
-------
sediments are (or are related to) particle size and organic carbon. In the
tests we have performed, it appears that L. variegatus is relatively tolerant
of a wide range of particle sizes. For example, we have tested the worm
successfully in sand. Of course, some common sense must be used, e.g., it
would be inappropriate to test the worm in a substrate consisting solely of
rocks. With respect to organic carbon content, as described above, L.
variegatus growth and reproduction can be quite dependent upon this variable.
However, the incorporation of routine feeding should help negate the effects
of organic carbon on the results of the worm test.
Although it is not now possible to define controls for sediment toxicity
tests in a traditional sense, the researcher must have something to compare
test results to in order to determine whether adverse effects have occurred.
Two options are available. The first approach is to generate baseline data
for statistical comparisons through a biological control, i.e., a clean
sediment in which the organism is known to live, grow and reproduce in an
acceptable manner. This type of control serves as a measure of test organism
health and therefore, serves as a basis with which to compare results
generated using test sediments. Typically, a biological control will be a
sediment routinely used by and readily available to the testing laboratory.
For example, our biological control for many studies is a sediment from West
Bearskin Lake in northern Minnesota. A second option for generating data for
statistical comparisons features the use of an ostensibly clean reference site
(or sites) from the study location (or a nearby system). Results of toxicity
tests with the test sediment(s) can then be compared statistically to results
generated using the reference samples. In all instances, a biological control
should be run simultaneously with the test sediment. A reference site also
may or may not be run, depending upon study objectives.
Sample Sizes and Statistical Techniques - The goal of any toxicity
test, must, in large part, dictate the number of replicates utilized
for testing. For example, in some instances it may be necessary to
15
-------
be able to identify statistically a 20% change from control values, while in
other cases the ability to discern a 100% difference is adequate. Rather than
attempting to make recommendations concerning sample sizes to be used, we
evaluated several of our data sets to determine approximate numbers of
replicates required to detect specific between sample differences in the
biological endpoints. The technique used for this exercise is described by
Steel and Torrie (1980; pg. 117); among the assumptions made for these
calculations was that our data were normally-distributed. The assumption of
data normality was evaluated and found to be reasonable using the NSCORES
function available through the MinitabR statistics package (Ryan et al. 1980).
Sample sizes were calculated assuming a two-tailed alternative with the
probability of Type I error («) of 0.05 and Type II error (0) of 0.20.
Variance estimates for the calculations for the various biological endpoints
in the L. variegatus test were derived from a series of 10-d exposures
conducted with multiple-replicates of sediments from sites exhibiting a wide
range of toxicities. Except in instances where all the worms were killed
(i.e., no variance), variance across the sites was reasonably similar. Table'
3 presents results of the sample size calculations for the 10-d exposures. In
our experience thus far, results of longer-term tests with the worm (e.g.,
28 d) are less variable than those from short-term tests (e.g., 10 d), so
these sample size estimates should be useful for tests of 10 d or longer.
When using these sample size estimates, notice that they were derived based on
variance observed under our test conditions, and may not apply to all tests in
which L. variegatus is used.
Results of toxicity tests with L. variegatus can be analyzed
statistically using a variety of standard techniques. We evaluate
nontransformed toxicity data using analysis of variance followed by a multiple
comparison technique, such as Dunnett's test. Recently, Hoke et al. (1990)
described the use of analysis of variance followed by linear orthogonal
contrasts to evaluate the results of sediment toxicity tests; this approach
may be more appropriate for assessing "groups" of samples than for the more
16
-------
commonly utilized multiple comparison techniques. In instances where only one
sample is being compared to a control or reference, standard t-tests also
would be suitable for data analysis. The reader is urged to consult a
standard biometrics text (e.g., Steel and Torrie 1980) for further guidance
concerning possible statistical tests.
Quality Assurance
Quality assurance procedures associated with good laboratory practices
were followed throughout this research. Because quality assurance is an
essential component of developing a new methodology, descriptions of these
procedures are included throughout the text.
RESULTS AND DISCUSSION
Presented below are two case studies in which L. variegatus was used to:
(1) assess the bioaccumulation of metals (cadmium, nickel) from sediments from
the Foundry Cove Superfund site in New York, and (2) assess the toxicity of
sediment samples collected from the copper-contaminated Keweenaw
Waterway/Torch Lake system in Michigan. For both studies, unless otherwise
noted, basic test conditions were those described in Materials and Methods.
Case Study One - Foundry Cove - Recent studies at Duluth and the EPA
laboratory at Narragansett have revealed that it is possible to predict the
bioavailability and toxicity of cationic metals in sediments through
determination of acid volatile sulfide (AVS) content (Ankley et al. 1990;
Carlson et al. 1990; DiToro et al. 1990). Briefly, sulfides in sediments,
normally precipitated as iron monosulfides, will preferentially bind to a
number of metals of environmental concern including cadmium, nickel, lead,
copper, and zinc, thereby reducing their bioavailability. These reactions
proceed on a unimolar basis, i.e., one mole of metal will react with one mole
of sulfide. Thus, as long as the molar ratio of metal/AVS is less than one,
17
-------
metals in sediments are not bioavailable; however, when the ratio exceeds one,
metals theoretically become bioavailable.
Initial studies which focused upon validating the metal/AVS hypothesis
used toxicity as an endpoint, i.e., determination of mortality of relatively
sensitive species, such as amphipods, in sediments with varying metal/AVS
ratios (DiToro et al. 1990). Bioaccumulation is another possible endpoint to
assess bioavailability of metals in sediments. Therefore, we designed a
series of studies with L. variegatus to determine whether metal/AVS
relationships in test sediments also could be used to predicted
bioaccumulation of metals by benthic species.
Test sediments used for these studies were from Foundry Cove, New York,
a Superfund site contaminated with both cadmium and nickel by a battery plant.
Detailed experimental protocols are given by Ankley et al. (1990). Briefly,
duplicate beakers with 20 L. variegatus were exposed to Foundry Cove sediments
from 17 sites for 10-d, after which the worms were depurated for 24 h in clean
Lake Superior water and then analyzed for metals. Figure 2 summarizes data
from these experiments. In the top panel (Fig. 2a), concentrations of metals ~
in L. variegatus (expressed as the molar sum of cadmium plus nickel) are
plotted against total metal concentrations in the test sediments; it is
apparent that the total metal concentration in the sediments was a poor
predictor of metal concentrations in the worms. In the lower panel (Fig. 2b),
concentrations of metals in L. variegatus are expressed relative to molar
metal/AVS ratios in the Foundry Cove sediments; overall, when metal/AVS ratios
were less than one, metal concentrations in the worms were uniformly low;
however, when the metal/AVS ratio exceeded unity, elevated tissue
concentrations of metals were observed. Thus, the metal bioaccumulation data
derived from the L. variegatus exposures provided additional support for the
critical role of AVS in determining metal bioavailability in sediments.
Case Study Two - Keweenaw Waterway/Torch Lake - As part of another ongoing
study at Duluth, a series of sediment samples were collected from the Keweenaw
18
-------
Waterway/Torch Lake system in the Keweenaw Peninsula of Michigan. Torch Lake
also is a Superfund site which has been contaminated with significant amounts
of copper associated with mine tailings. Sediments from the system reportedly
have low benthic diversity and are toxic to some upper-water column test
species (e.g., Daphnia magna) (Kraft and Sypniewski 1981; Malueg et al. 1984).
One of our objectives, therefore, was to determine whether sediments from the
system also were toxic to benthic species such as L. variegatus.
Exposures were conducted with three sediment samples from the system,
one from Torch Lake, and two from the Keweenaw Waterway; the biological
control for the test was sediment from West Bearskin Lake. Exposures were
conducted using 10 organisms per beaker with eight replicates per site (we
were interested in detecting relatively small differences in toxicity).
Sediments from all three test sites significantly decreased
survival/reproduction of the organisms in 10-d exposures (Table 3). (Note the
use of survival/reproduction as a single variable). Total biomass also was
significantly reduced in the Keweenaw/Torch Lake sediments relative to the
West Bearskin control; however, mean weight per worm was not significantly •*
decreased. Although in this particular test, survival/reproduction (and
subsequently total biomass) were significantly affected, while growth (weight
per organism) was not, with other samples we have seen the opposite effect,
i.e., growth, but not survival/reproduction was reduced. Overall, our test
results from the Keweenaw/Torch Lake sediments study indicate that sediment
toxicity may be an important factor contributing to low benthic diversity in
the system.
SUMMARY AND CONCLUSIONS
Methods are presented which detail the use of the aquatic oligochaete,
L. variegatus. in toxicity and bioaccumulation tests with contaminated
sediments. Although attention must be given to various issues in the broad
area of sediment testing (e.g., identification of suitable controls) as well
as specific factors in the L. variegatus test (e.g., feeding), we feel that,
19
-------
based on our experience with this species over the past 2 years, it is an
excellent organism for sediment assessment work. It offers great flexibility
with respect to toxicity endpoints as well as test lengths, and is suitable
for assessing bioaccumulative chemicals. Moreover, L. variegatus is readily
cultured under a variety of conditions and is amenable to handling.
Although test methods presented in this report for L. variegatus are as
far advanced as for any freshwater benthic species (e.g., H. azteca. C.
tentans. C. riparius: Nelson et al. 1990) it should not be construed that this
test has been field-validated. In order to provide regulators with a legally-
defensible set of tests, it is imperative that methods be validated by
comparing results obtained in the tests to benthic communities in situ. This
is the only way to ensure that the test methods are indeed protective of
benthic species. ERL-Duluth currently is initiating field validation studies
for L. variegatus. H. azteca. and C. tentans toxicity tests.
ACKNOWLEDGEMENTS
We would like to acknowledge the following persons who have contributed -
to this work - V. Mattson, D. Benoit, S. Collyard, T. Linnum, P. Kosian, E.
Leonard and A. Cotter. D. Benoit and R. Spehar provided valuable comments on
the report. R. LePage prepared the figures and S. Wojcik and D. Engberg typed
the report. This report was supported, in part, by the U.S. EPA Office of
Water (Criteria and Standards Division).
20
-------
REFERENCES
Ankley, G.T., G.L. Phipps, P.A. Kosian, A.M. Cotter, E.N. Leonard, D.A.
Benoit, V.R. Mattson and J.D. Mahony. 1990. Acid volatile sulfide as a
factor mediating the bioavailability of cadmium and nickel-contaminated
sediments from Foundry Cove, New York. In Preparation.
Benoit, D.A., V.R. Mattson and D.L. Olson. 1982. A continuous-flow
mini-diluter system for toxicity testing. Water Res. 16:457-464.
Brinkhurst, R.O. 1980. Pollution biology-the North American experience. In
"Proceedings of the First International Symposium on Aquatic Oligochaete
Biology," R.O. Brinkhurst and G.C. Cook, Eds. pp 471-475. Plenum Press,
New York, NY.
Carlson, A.R., G.L. Phipps, V.R. Mattson, P.A. Kosian and A.M. Cotter. The
role of acid volatile sulfide in determining cadmium bioavailability
in freshwater sediments. U.S. Environmental Protection Agency,
Environmental Research Laboratory-Duluth, Report No. 7971A.
Chapman, P.M., M.A. Farrell and R.O. Brinkhurst. 1982a. Relative tolerances
of selected aquatic oligochaetes to individual pollutants and
environmental factors. Aquat. Toxicol. 2:47-67.
Chapman, P.M., M.A. Farrell and R.O. Brinkhurst. 1982b. Relative tolerances
of selected aquatic oligochaetes to combinations of pollutants and
environmental factors. Aquat. Toxicol. 2:69-78.
Chekanovskaya, O.V. 1962. Aquatic oligochaeta of the U.S.S.R. Akademiya
Nauk SSSR. Moscow, U.S.S.R.
Cook, D.G. 1969. Observations on the life history and ecology of some
Lumbriculidae (Annelida, Oligochaeta). Hydrobiologia 34:561-574.
Connell, D.W., M. Bowman and D.W. Hawker. 1988. Bioconcentration of
chlorinated hydrocarbons from sediment by oligochaetes. Ecotoxicol.
Environ. Safety 16:293-302.
21
-------
DiToro, D.M., J.D. Mahony, D.S. Hansen, K.J. Scott, M.B. Hicks, S.M. Mayr an'd
M.S. Redmond. 1990. Toxicity of cadmium in sediments: the role of
acid volatile sulfide. Environ. Toxicol. Chem. In Press.
DeWitt, T.H., G.R. Ditsworth and R.C. Swartz. 1988. Effects of natural
sediment features on survival of the phoxocephalid amphipod, Rhepoxvnius
abronius. Mar. Environ. Res. 25:99-124.
Giesy, J.P. and R.A. Hoke. 1989. Freshwater sediment toxicity bioassessment:
Rationale for species selection and test design. J. Great Lakes Res.
15:539-567.
Hoke, R.A., J.P. Giesy and J.R. Adams. 1990. Use of linear orthogonal
contrasts in analysis of environmental data. Environ. Toxicol. Chem.
9:815-819.
International Joint Commission. 1985. Report on Great Lakes water quality.
Report of the Water Quality Board, International Joint Commission,
Kingston, Ontario.
Keilty, T.J., D.S. White and P.F. Landrum. 1988a. Short-term lethality and
sediment avoidance assays with endrin-contaminated sediment and two
oligochaetes from Lake Michigan. Arch. Environ. Contam. Toxicol. 17:95-
101.
Keilty, T.J., D.S. White and P.F. Landrum. 1988b. Sublethal responses to
endrin in sediment by Linmodrilus hoffmeisteri (Tubificidae), and in
mixed-culture with Stylodrilus heringianus (Lumbriculidae). Aquat.
Toxicol. 13:227-250.
Kraft, K.J. and R.H. Sypniewski. 1981. Effect of sediment copper on the
distribution of benthic macroinvertebrates in the Keweenaw Waterway. J.
Great Lakes Res. 7:258-263.
Lauritsen, D.D. S.C. Mozley and D.S. White. 1985. Distribution of
oligochaetes in Lake Michigan and comments on their use as indices of
pollution. J. Great Lakes Res. 11:67-76.
22
-------
Malueg, K.W., G.S. Schuytema, D.F. Krawczyk and J.H. Gakstatter. 1984.
Laboratory sediment toxicity tests, sediment chemistry and distribution
of benthic macroinvertebrates in sediments from the Keeweenaw Waterway,
Michigan. Environ. Toxicol. Chem. 3:233-242.
Nebeker, A.V., W.L. Griffis, C.M. Wise, E. Hopkins and J.A. Barbitta. 1989.
Survival, reproduction and bioconcentration in invertebrates and fish
exposed to hexachlorobenzene. Environ. Toxicol. Chem. 8:601-611.
Nelson, M.K., C.G. Ingersoll and F.J. Dwyer. 1990. Proposed guide for
conducting solid-phase sediment toxicity tests with freshwater
organisms. ASTM Document. Draft, Number 6.
Oliver, E.G. 1984. Uptake of chlorinated organics from anthropogenically
contaminated sediments by oligochaete worms. Can. J. Fish Aquat. Sci.
41:878-883.
Pennak, R.W. 1978. Freshwater invertebrates of the United States. John
Wiley and Sons, New York, NY.
Robbins, J.A., T.J. Keilty, D.S. White and D.N. Edgington. 1989.
Relationships among tubificid abundances, sediment composition and
accumulation rates in Lake Erie. Can. J. Fish. Aquat. Sci. 46:223-231.
Ryan, B.F., B.L. Jorner and T.A. Ryan. 1985. Minitab handbook, Second
Edition. Duxbury Press, Boston, MA.
Schuytema, G.S., D.F. Krawezyk, W.L. Griffis, A.V. Nebeker and M.L. Robideaux.
1990. Hexachlorobenzene uptake by fathead minnows and
macroinvertebrates in recirculating sediment/water systems. Arch.
Environ. Contain. Toxicol. 19:1-9.
Spencer, D.R. 1980. The aquatic oligochaetes of the St. Lawrence Great Lakes
Region. In "Proceedings of the First International Symposium on Aquatic
Oligochaete Biology," R.O. Brinkhurst and D.G. Cook, Eds. pp 115-164.
Plenum Press, New York, NY.
Steel, R.G.D. and J.H. Torrie. 1980. Principles and procedures of
statistics: a biometrical approach, Second Edition. McGraw-Hill Book
Company, New York, NY.
23
-------
U.S. Environmental Protection Agency. 1987. An overview of sediment qualify
in the United States. U.S. Environmental Protection Agency, EPA-905/9-
88-002. Washington, D.C.
Wiederholm, T., A.M. Wiederholm and G. Milbrink. 1987. Bulk sediment
bioassays with five species of fresh-water oligochaetes. Water Air Soil
Pollut. 36:131-154.
24
-------
Table 1. Some laboratories currently maintaining cultures of Lumbriculus
variegatus.
Laboratory
Contact Person
Phone
Environmental Research Laboratory-Duluth
6201 Congdon Boulevard
Duluth, MN 55804
Gary Phipps
218-720-5571
Environmental Research Laboratory-Corvallis Alan Nebeker
200 S.W. 35th Street
Corvallis, OR 97333
503-420-4875
Center for Lake Superior Environmental
Studies - University of Wisconsin-Superior
Superior, WI
Larry Brooke 715-394-8318
Environmental Sciences Section
Health and Environment Laboratories
Eastman Kodak Company
Building 306
Rochester, NY 14650
William Ewell 716-588-4528
-------
Table 2. Reproduction, biomass, and individual organism weights in a low organic carbon sediment and
sand with 4 different feeding rates. All values are expressed as the mean +
standard deviation for 3 replicates. Each replicate started with 10 animals.
Feeding
Rate
Number of Worms
Total Biomass (mg)
Biomass/Worm (mg)
Sediment
Sand
Sediment
Sand
Sediment
Sand
0 mg 16.7 ± 0.58 23.0 ± 2.6 13.94 + 0.50 13.41 ± 1.04 0.84 ± 0.05 0.59 + 0.05
20 mg 25.7 ± 2.5 23.3 ± 2.5 24.74 + 0.91 19.41 ± 1.91 0.97 ± 0.06 0.83 + 0.04
40 mg 18.3 ± 4.9 25.0 ± 5.0 24.90 + 2.48 21.92 ± 0.73 1.40 + 0.02 0.90 + 0.16
80 mg 10.7 + 2.1 15.3 ± 2.5 15.35 ± 1.05 15.24 + 1.53 1.46 + 0.02 1.00 + 0.06
-------
Table 3. Approximate sample sizes (n) for detecting among sample differences
in biological endpoints (8) for Lumbriculus variegatus toxicity
l_t •.
tests. Sample size
worms/beaker) .
Endpoint
Survival/Reproduction
(number of organisms)
Biomass
(mg/beaker)
Biomass
(mg/worm)
is expressed as
6
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
0.1
0.2
0.3
0.4
0.5
0.6
number of beakers (with 10
n
198
50
22
13
8
6
4
3
3
2
126
63
14
8
5
4
3
2
71
18
8
5
3
2
-------
Table 4. Toxicity of test sediments from Torch Lake (TL) and the Keweenaw -
Waterway (KW1, KW2) relative to the West Bearskin Lake (WBL)
control. Data are expressed as the mean + standard deviation for
eight replicate determinations.
Site Survival/Reproduction Total Biomass (mg) Biomass/Worm (rag)
WBL 14.40 ± 1.92 14.67+1.41 1.03+0.07
TL 7.13 ± 0.99* 5.02+1.92* 0.72 ± 0.30 .
KW1 6.90 ± 4.20* 5.08+3.51* 0.83 ± 0.20
KW2 9.13 ± 0.64* 8.65 ± 1.12* 0.95+0.10
* Differed significantly (p < 0.05) from the WBL control.
-------
Figure 1. A testing system for contaminated sediments.
Water Source
Timer Controlled
Valve
Test Tanks
Test Beaker
300ml. High Form
Test Tank
30x16x13cm. high
Self Starting
Siphon
U.S. Environmental Protection Agency
Great Lakes National Program Office
GLNPO Library
-------
Figure 2. Bioaccumulation of metals (cadmium, nickel) by Lumbriculus
varieqatus relative to sediment metal concentrations or sediment metal
concentrations normalized to acid volatile sulfide (AVS).
in
in
2"
ffl 4
"3
3
0.01
• •
0.1 1 10
log umol metal/g sediment
100
1.000
§
M
f.
0.01
0.1
1 10
log metai/AVS
•• 100
1.000
------- |