Microcosm sediment toxicity
Environ. Toxicol. Chem. 20, 2001
1949
experiment. In the Landis et al. [22,23] SAM experiments, the
structural variables were used exclusively, although Landis
Does a failure to detect statistically significant differences
in the parameters indicate that the ecological system has re-
turned to its prestressed state? The answer has been demon-
strated to be conclusively no by Kersting and Van Wijngaarden
[12] using a very different multispecies system. The ditch
microcosms were dosed twice with the herbicide Linuron
(Crescent Chemical, Hauppauge, NY, USA). The functional
parameters of the dosed systems were not different statistically
from the nondosed treatment. However, on redosing, the pre-
viously dosed systems were much more resistant to the her-
bicide. The change in the sensitivity of the system indicates
an important functional change in the system, although the
measured variables did not reveal the alteration.
There are important implications of this and related ex-
perimental multispecies tests for field research. These micro-
cosm experiments, because of their relatively low variability
and number of true replicates, can find treatment related im-
pacts at very low concentrations of materials. The statistical
power found in the microcosm experiments should be much
higher than that of a typical field study that, by its very nature,
lacks replicability and has no control over many important
variables. The lack of statistical power of field studies makes
an illusion of recovery or no significant difference using con-
ventional statistical techniques a likely outcome. Approaches
used to date in most field research and in experimental mi-
crocosms, including ours, do not incorporate changes in system
dynamics as an endpoint. Dealing with dynamic systems is
difficult statistically with typical ecological data sets, although
methods are under development [42].
[
42] used both structural and functional variables. In MFC,
we separated the functional and structural variables. Both sets
of variables detected treatment-related effects. The structural
variables were not as effective as the functional variables in
detecting treatment groups.
Persistence of effects
Persistence of detectable effects is a characteristic of all the
SAM and MFC experiments. Analytical chemistry of the water
column in the SAM experiments demonstrated the rapid deg-
radation of the water-soluble fraction. The MFC was dosed
with the neat material, and some bound to the sediment to be
released on mixing. Perhaps some of the observed effects in
the MFC could be due to materials not detectable in the water
column but bound to the components of the artificial sediment
or the accumulated detritus.
The variables that allow the detection of treatment effects
change over time in both types of experiments. Table 2 clearly
shows these changes as demonstrated by NCAA for the MFC.
Examination of the density plots also bears this fact. Para-
mecium bursaria was particularly sensitive to treatment group
during the middle of the experiment. Invertebrates and pH were
good indicators of treatment effects at the end. The change in
the indicators of effects is noted in the SAM experiments as
well and likely is a universal property of these and other eco-
logical systems.
The persistence of these treatment effects, the changes in
the indicators of these effects, and the repeatability of these
properties in SAM experiments led to formulation of the com-
munity conditioning hypothesis [29,43]. This hypothesis states
that ecological communities tend to preserve information
about every event in their history, long after the disturbance
occurred. The information can be carried in the structure of
the community, in the population dynamics and structure of
the constituent populations, and in the structure of their ge-
nomes. Corollaries to this are that ecological systems are com-
plex, nonequilibrium systems and that the best measures of
effects change over time [42].
As discussed previously, the MFC experiments demonstrat-
ed these properties. In the SAM experiments, the water-soluble
fraction of the jet fuel degraded during the course of the ex-
periment. In the MFC experiment, the toxicant could appar-
ently be stored in the sediment for subsequent degradation and
release into the water column. It is possible that the stressor
remained in the system, at low and transient concentrations in
the water column, and this residual could be directly causing
the effects at the end of the experiment.
Acknowledgement—We would like to thank the Institute of Environ-
mental Toxicology and Chemistry support staff, notably D. Hennessy,
A. Markiewicz, Mike Roze, L. Holmquist, and M.K. Moores. This
research was supported by U.S. Air Force Office of Scientific Research
Grant AFOSR-91-0291 DEF.
REFERENCES
1
. Lyman WJ. 1984. Establishing sediment criteria for chemicals—
Industrial perspectives. In Dickson KL, Maki AW, Brung WA,
eds, Fate and Effects of Sediment-Bound Chemicals in Aquatic
Systems. Pergamon, New York, NY, USA, pp 378–387.
2. Southerland E, Kravitz M, Wall T. 1992. Management framework
for contaminated sediments (U.S. EPA Sediment Management
Strategy). In Burton GA Jr, ed, Sediment Toxicity Assessment.
CRC, Boca Raton, FL, USA, pp 341–370.
3
. U.S. Environmental Protection Agency. 1987. An overview of
sediment quality in the United States. EPA 905/9-88-002. Office
of Water Regulations and Standards, Washington, DC.
. Burton GA Jr. 1991. Assessing the toxicity of freshwater sedi-
ments. Environ Toxicol Chem 10:1585–1627.
. Chapman PM. 1989. Current approaches to developing sediment
quality criteria. Environ Toxicol Chem 8:589–599.
4
5
6. Cairns J Jr. 1986. The myth of the most sensitive species.
In this and other microcosm experiments, we routinely fail
to find recovery as defined as concurrence with the nondosed
treatment. Admittedly, the lowest concentration in this and
other experiments tracks the nondosed treatments in most of
the parameters. However, examination of the graphs for pa-
rameters such as pH and total invertebrates demonstrates a
treatment-related effect late in the experiment. Total inverte-
brates exhibit two groups at the end of the MFC, the first group
BioScience 36:670–672.
. Kimball KD, Levin SA. 1985. Limitations of bioassays: The need
for ecosystem-level testing. BioScience 35:165–171.
7
8
. Giddings JM. 1981. Laboratory tests for chemical effects on
aquatic population interactions and ecosystem properties. In Ham-
mons AS, ed, Methods for Ecological Toxicology: A Critical
Review of Laboratory Multispecies Tests. Ann Arbor Science,
Ann Arbor, MI, USA, pp 23–91.
. Leffler JW. 1984. The use of self-selected, generic aquatic mi-
crocosms for pollution effects assessment. In Harris HH, ed, Con-
cepts in Marine Pollution Measurements. Maryland SeaGrant
College, University of Maryland, College Park, MD, USA, pp
9
being comprised of nondosed and 2-
ond group comprised of 10- and 25-
l treatments and the sec-
l treatments. Perhaps the
1
39–158.
variety of parameters that are examined, the statistical power
of six replicates per treatment, and the utilization of multi-
variate techniques allow sufficient power to detect subtle
changes in the treatment groups.
1
0. Taub FB, Read PL. 1982. Standardized aquatic microcosm pro-
tocol, Vol 2. 223-80-2352. Final Report. Food and Drug Admin-
istration, Washington, DC, USA.
11. Touart LW. 1988. Hazard Evaluation Division Technical Guidance