Ecological Risk A^yssm^n^
Final Report
Endpoints for
Ecological
Toxicity
ssb
TU
Submitted To:
Exposure Assessment Group
Office of Research and Development
U.S. Environmental Protection Agency
Submitted By:
Technical Resources, Inc.
3202 Monroe St., Rockville, MD 20852
Work performed under contract #68-02-4199
-------
Ecological Risk Assessment
ENDPOINTS FOR ECOLOGICAL TOXICITY
BY
Abe Mittelman, Project Manager
Joanne Settel
Kathleen Plourd
Rolland S. Fulton III
Gene Sun
Shakuntala Chaube
Patrick Sheehan*
TECHNICAL RESOURCES, INC.
3202 MONROE STREET, SUITE 200
ROCKVILLE, MARYLAND 20852
Contract No. 68-024199
September 30, 1988
Project Officer
John Segna
Exposure Assessment Group
Office of Research and Development
U.S. Environmental Protection Agency
Washington, D.C. 20460
*Aqua Terra Technologies
-------
TABLE OF CONTENTS
PAGE
Introduction iv
Chapter 1 Organism-Level Ecotoxicological Endpoints 1
1.0 Introduction 1
1.1 Acute Mortality 2
1.2 Biochemical Alterations 7
1.3 Osmoregulatory Activity 9
1.4 Respiratory Activity 12
1.5 Behavioral Alterations 12
1.5.1 Avoidance reactions 13
1.5.2 Predator-prey interactions 14
1.5.3 Reproductive behavior 15
1.5.4 Locomotor activity 16
1.5.5 Feeding activity 17
1.6 Reproductive Toxicity 21
1.7 Musculoskeletal Effects 24
1.8 Growth and Development 24
1.8.1 Developmental toxicity 27
1.8.2 Growth 31
1.9 Genotoxicity 32
1.10 Carcinogenicity 34
1.11 Modifying Factors 35
1.11.1 Microbial toxicity 37
1.11.2 Multifactorial interactions 41
1.12 Conclusion 43
Chapter 2 Population-Level Endpoints 43
2.0 Introduction 43
2.1 Population Dynamics 44
2.1.1 Birth rate 44
2.1.2 Death rate 45
2.1.3 Population growth models 45
2.1.4 Other population endpoints 51
-------
2.2 Life-history Strategy 52
2.2.1 Parental care and fecundity 54
2.2.2 Parental care and growth/development 54
2.2.3 Growth and age at maturity 55
2.2.4 Reproduction and life expectancy 55
2.2.5 Longevity and body size 56
2.2.6 Compensatory mechanisms of life-history
strategy 56
2.2.7 Life histories and response to contaminants 57
2.3 Conclusion 58
2.4 Methods for Selecting Appropriate Populations
to Monitor 59
2.4.1 Total population 60
2.4.2 Species dominance 62
2.4.3 Indicator species 63
2.4.4 Keystone species 65
2.4.5 Representative and important-species
approach 69
2.5 Conclusion 70
2.6 Species Interactions 71
2.6.1 Predator-prey interactions 72
2.6.2 Interspecies competition 74
2.7 Conclusion 76
Chapter 3 Ecosystem-Level Endpoints 77
3.0 Introduction 77
3.1 Ecosystem Function 78
3.2 Energy Flow 79
3.2.1 Primary productivity 79
3.2.2 Respiration 83
3.2.3 Photosynthesis/respiration ratio 84
3.2.4 Methods of measurement of primary
production 85
3.3 Biogeochemical Cycling 94
3.3.1 Nutrient analysis 95
3.3.2 Chlorophyll content 97
ii
-------
97
100
101
103
108
109
111
112
114
116
116
118
121
125
127
128
129
130
135
136
138
139
139
140
141
142
142
143
145
146
148
149
3.3.3 Leaching
3.3.4 Determination of a nutrient budget
3.3.5 Nutrient cycling: the nitrogen cycle
3.3.6 Methods for analysis
3.3.7 Decomposition
3.3.8 Methods of measurement
3.4 Conclusion
3.5 Ecosystem or Community Structure
3.6 Abundance and Biomass
3.7 Species Lists
3.8 Biological Indices (Pollution Indices)
3.9 Species Richness
3.10 Diversity
3.11 Comparative Indices (Similarity Indices)
3.12 Multivariate Analysis
3.13 Trophic Organization
3.14 Spatial Structure
3.15 Guilds
3.16 Conclusion
3.17 Stability
3.17.1 Resilience
3.17.2 Amplitude
3.17.3 Elasticity
3.17.4 Inertia
3.18 Conclusion
Summary
Introduction
Organism-Level Endpoints
Population-Level Endpoints
Ecosystem-Function Endpoints
Ecosystem-Structure Endpoints
Conclusions
iii
-------
INTRODUCTION
The performance of an ecological risk assessment requires a delineation of
endpoints suitable for measurement. Ecotoxicological endpoints are here defined
as physical or biological parameters, characteristic of the ecosystem, which are
measurably and predictably affected adversely by contaminants. Such endpoints
potentially include individual organism-, population-, and ecosystem-level
parameters. While organism-level effects have often been used in isolation to
assess environmental impact, these are not generally sufficient to describe
ecosystem response. The complex nature of interactions between the abiotic and
biotic components of an ecosystem necessitate a more integrative approach to
ecological risk assessment. Both population- and ecosystem-level effects need to
be considered, along with organism-level effects, in order to adequately describe
ecosystem response.
This document surveys the range of endpoints that have been or potentially
may be used in ecological risk assessment at the levels of the individual organism,
the population, and the ecosystem. For each endpoint discussed, strengths and
weaknesses as measures of ecological damage are presented, and examples of its
use in measuring pollutant effects are presented. The report focuses on the use
of endpoints in the detection of existing effects, particularly at the population
and ecosystem levels of organization, but many of the endpoints described are
also appropriate for pre-release prediction of contaminant effects through their
measurement in laboratory single-species bioassays, microcosm, mesocosm, or
experimental ecosystem studies. Test methods currently used in hazard assessment
are discussed in more detail in TRI (1988c).
iv
-------
A variety of endpoints have been used to assess the ecological hazards
associated with anthropogenic stress on ecosystems. The use of specific response
measures in ecological risk assessment depends, in part, upon the type of
ecosystem impacted, the type of stress, the proposed hypothesis, and the required
level of detection. This is to say that, although a general set of possible
response endpoints can provide data on effects at various levels of biological
organization, the selection of appropriate endpoints for each study must be
situation specific.
To evaluate the applicability and utility of response endpoints for ecological
risk assessments, the following criteria should be applied:
o Is there an appropriate theoretical base for the response endpoint?
o Do these endpoints provide ecologically meaningful information?
o Is the endpoint easy to measure?
o Is the endpoint sensitive to stress, and is the response concentration-
dependent?
o Can the endpoint be applied to analysis of a variety of situations and
different ecosystems?
o Is there a history of successful use of the response endpoint in impact
assessments?
o Does it have socio-economic-legal significance?
This document does not provide specific guidelines for selection of ecological
endpoints. However, further guidance is presented in another document,
"Ecological Endpoint Selection Criteria" (TRI, 1988a).
v
-------
Response endpoints at the individual organism level can be useful predictors
of stress effects at the population level, if the connection between individual
performance and population fitness is clear. Chemical injury to individuals
resulting in premature death and reduced reproductive success and recruitment
may be reflected in lowered abundance and altered distribution of exposed
populations. The connection between individual and population responses will
depend on the processes regulating population dynamics.
In turn, the relevance of population response data for predicting community-
and ecosystem-level impacts is related to the importance of the population in
structuring the community and in controlling functional processes. These
statements point to the importance of gathering ecologically meaningful
information and, if possible, response endpoint data at the appropriate biological
level to maximize confidence in extrapolations.
vi
-------
CHAPTER 1
ORGANISM-LEVEL ECOTOXICOLOGICAL ENDPOINTS
1.0 INTRODUCTION
Physiological parameters are important biological endpoints in ecological risk
assessment. The biological success of a population is based, in part, on the
physiological status of its individual members. Physiological endpoints include
mortality, biochemical alterations, osmoregulatory effects, respiratory effects,
behavioral effects, reproductive effects, musculoskeletal effects, effects on growth
and development, genotoxic effects, and carcinogenic effects. A major advantage
of using physiological endpoints is that they often respond rapidly to pollutant
stress. This early detection may allow corrective action to be applied before
irreparable damage has occurred. However, it is often not possible to relate
physiological endpoints to the population- or ecosystem-level effects of pollutants,
because these levels are strongly influenced by interactions among species and
between species and the physical environment (Levin et al., 1984).
1.1 ACUTE MORTALITY
Acute mortality, or toxicity, is the level of toxic chemical that produces
lethality in a specific proportion of test organisms in a given short period of time
(usually 48 or 96 hours). The most frequently used measure is the median lethal
dose or concentration that kills 50 percent of the organisms being tested (LDjq
or LC50). These tests are simple to perform, relatively inexpensive, and practical
to conduct, and an enormous background literature exists on them. Such tests
1
-------
have a bottom line appeal in that they can be presented as a single figure.
There are, however, many drawbacks associated with them. Test outcomes depend
on the physical and chemical environment (e.g., water hardness, water movement,
temperature, pH, photoperiod), the organisms used (species, degree of acclimation,
age, sex, genetic stock), synergistic and antagonistic interactions between
toxicants, and variations in procedural methods. These drawbacks are shared by
other physiological toxicological endpoints.
Acute lethality tests are often considered insufficiently sensitive to provide
information on sublethal or chronic effects; these occur at lower concentrations
and may be of considerable ecological significance. Sublethal impairment may
affect an organism's ability to cope with other naturally occurring stresses (such
as environmental variability or interactions with other organisms), and thus its
ability to survive in the natural environment. Acute toxicity testing represents
the quickest and least sophisticated way of obtaining a preliminary evaluation of a
hazardous substance (Monk, 1983).
1.2 BIOCHEMICAL ALTERATIONS
Biochemical responses to pollutants have received considerable attention
because they provide information on the mechanisms of action of toxic pollutants.
Biochemical responses to many different chemicals have been studied in fish. In a
study by Verma et al. (1981), Saccobranchus fossils were exposed to sublethal
concentrations of copper sulfate, the detergent swascofix E45, and chlordane and
metasystox for 40 days. These substances caused significant inhibition of the
enzyme acetylcholinesterase (AChE). Associated behavioral responses were violent
2
-------
movements followed by loss of equilibrium, possibly due to accumulation of
acetylcholine (ACh) at nerve endings, thus disrupting transmission of nerve
impulses. Enzymes involved in energy metabolism were also affected by the
toxicants. There was general inhibition of succinic dehydrogenase (SDH) and
pyruvic dehydrogenase (PDH), and stimulation of lactic dehydrogenase (LDH)
activity, indicating depression of aerobic metabolism and development of anaerobic
conditions at the tissue level.
Birds have also been used to monitor biochemical changes resulting from
pollutant exposures. In one study, the pesticide fenitrothion suppressed brain
cholinesterase (ChE) activity in several species of song birds (Busby et al., 1981).
In another study, Japanese quail exposed to the pollutant parathion exhibited ChE
inhibition. It was found that in the quail, inhibition in excess of 20 percent
indicated stress while inhibition of more than 50 percent usually caused death
(Ludke et al., 1975).
ChE inhibition is useful as an ecological endpoint because of its specificity
and clear association with pollutant-induced effects. While it can result in
disruption of neural action, it is not significantly affected by neural stress, and it
is associated with mortality. Mortality from ChE inhibition is preceded by various
symptoms, including anorexia, lethargy, antagonistic behavior, muscular
incoordination, and convulsions. Effects of sublethal inhibition of ChE can
include weight loss, impaired growth and reproduction, and increased susceptibility
to predation. A drawback associated with the use of ChE inhibition as an
ecotoxicological endpoint is that it is difficult to correlate with chemical
exposures in the field. This problem is, however, common to most
3
-------
ecotoxicological endpoints and arises from difficulties in determining field
exposures.
Brain neurotoxic esterase (NTE) activity has been used as a biochemical
measurement of delayed neurotoxicity in birds. In chickens, delayed neurotoxicity
is generally manifested as ataxia within 2 weeks of dosage, with eventual paralysis
of the legs. An excellent correlation has been established for organophosphate
pesticides, which cause delayed neurotoxicity and inhibition of NTE activity both
in vivo and in vitro (Lotti and Johnson, 1978).
Evidence for the effects of pesticides on NTE is provided by a study in
which the organophosphate pesticide EPN (phenylphosphonothioic acid-0-ethyl-0-4-
nitrophenyl ester) was fed to mallards (Anas platyrhynchos) for 90 days. The
treated birds experienced inhibition of brain NTE activity by 16, 69, 73, and 74
percent at concentrations of 10, 30, 90, and 270 ppm, respectively. Brain NTE
inhibition of 65 percent or greater was associated with severe ataxia or paralysis.
Treatment-related demyelination and degeneration of axons of the spinal cord
were also observed (Hoffman et al., 1984).
Other enzymes have also been used as endpoints to monitor pollutant stress.
One of these, delta-aminolevulinic acid dehydratase (ALA-D), is an essential
enzyme in the biosynthetic pathway of heme synthesis. Its inhibition may result
in significant reduction in heme synthesis and in neurological consequences.
However, the association of the ALA-D index with overall fitness of exposed
populations is not known. Depression of ALA-D activity has been observed in
4
-------
response to Pb exposure in fish (Jackim, 1973) and birds (Hutton, 1980; Hoffman
et al., 1981).
The mixed function oxygenase (MFO) enzyme system, catalyzed by the heme
protein cytochrome P-450, may also be a useful endpoint. It is responsible for
biotransformation of xenobiotics in vertebrates and invertebrates and has been
used to indicate pollution by aromatic hydrocarbons or other pollutants. For
example, elevation of MFO activity was detected in livers of brook trout
(Salvelinus fontinalis) that had been fed a single dose of PCB (Aroclor 1254)
(Addison et al., 1981). A number of studies have observed induction of hepatic
cytochrome P-450 activity by petroleum hydrocarbons (Stegeman, 1980). Though
changes in MFO activity appear to provide a useful endpoint for monitoring
effects of pollutant stress, results must be interpreted cautiously because many
factors can influence MFO activity, including diet, temperature, season, age, sex,
and species. The extent of MFO induction by different chemicals varies widely;
indeed some cause inhibition of MFO activity. While MFO induction may be used
as an indicator of certain pollutants, the relationship between it and
environmental damage remains unclear (Stegeman, 1980).
The physiological basis of stress response is endocrinological. Under stress,
the adenohypophysis is stimulated to produce more ACTH, which in turn results in
elevated levels of corticosteroids secreted by the adrenal cortex. Thus,
corticosteroid hormone levels can provide monitorable endpoints for pollutant-
induced stress. For example, corticosteroid changes were observed in Sockeye
salmon (Oncorhynchus nerka) exposed to low levels of Cu in freshwater aquaria.
Treated fish showed a rapid, dose-related increase in Cortisol, cortisone, and total
5
-------
corticosteroid levels (corticosteroid response). Fish exposed to 10"^ molar of Cu
died between 8 and 24 hours (Donaldson and Dye, 1975). Several other studies
have noted changes in corticosteroid levels in response to various pollutants,
including heavy metals, pesticides, coal dust, pulp mill effluent, and landfill
leachate (Donaldson, 1981). It has been suggested that disruption of the
endocrine balance (increased corticosteroids) is the cause of depressed growth in
oil-dosed birds (Sheehan 1984a). However, corticosteroid changes are not specific
responses to pollutants; levels can change in response to many environmental
stresses, including handling, anesthesia, temperature or salinity change, hypoxia,
confinement, crowding, or disease. Another difficulty is that organisms frequently
acclimate to sublethal stresses, resulting in a transitory corticosteroid response
(Donaldson, 1981; Schreck, 1981). The relationship between corticosteroid stress
responses and longer-term effects on growth and reproduction is unclear.
Lysosomal membrane stability is another biochemical endpoint for pollutant
effects. Membrane stability is important in maintaining cellular integrity and in
preventing autolytic cell damage by free hydrolases. In one study, pollutant
effects on lysosomes were monitored in the hydroid Campanularia flexuosa. An
increase in the levels of free glucosaminidase activity in endodermal cells was
observed after exposure to threshold concentrations of Cu, Cd, and Hg. This
change in glucosaminidase activity was hypothesized to result from decreased
stability of the lysosomal membrane (Moore and Stebbing, 1976). Such lysosomal
membrane stability has been shown to be correlated with growth indices both in
the laboratory and in the field (Sheehan, 1984a).
6
-------
Biochemical responses to toxic pollutants can be quite complex. A wide
spectrum of biochemical changes were observed in the eggs of brook trout (S.
fontinalis) exposed to PCB (Aroclor 1254) before and after hatching. These
changes decreased levels of hydroxyproline, vitamin C, collagen, and phosphorous
but increased Ca concentrations in the spine. The decreased growth rate, lowered
levels of hydroxyproline, and vitamin concentrations in sac fry exposed to PCB,
and later increased mortality in these fry, suggest that PCB exposure induced
competition between developmental and detoxification processes for the use of
vitamin C (Mauck et al., 1978).
In summary, biochemical responses are quite sensitive to short-term and sub-
lethal pollutant stresses and are often easily associated with the mechanisms of
toxicity. However, extrapolation to longer-term organismal responses, and to
population- or ecosystem-level responses, is not generally given with present
knowledge.
1.3 OSMOREGULATORY ACTIVITY
Osmoregulatory activity may be a good ecological endpoint because it can
provide an important measure of physiological stress in aquatic organisms. Its
function in aquatic species is prevention of loss of salts from the organism and
maintenance of the salt water balance over a range of salinities. Environmental
pollutants, by inhibiting or interfering with these adaptive mechanisms, might be
expected to reduce the ability of aquatic organisms to tolerate stressful salinity.
7
-------
Exposure of juvenile coho salmon (Oncorhynchus kisutch) to Cu in fresh
water reduced their tolerance to increased salinity. Effects of Cu on gill ATPase
(inhibition) resulting in impaired survival in sea water occurred within 24 to 72
hours. This decreased gill-ATPase activity was probably one of the factors
leading to loss in osmoregulatory ability and death (Lorz and McPherson, 1976).
Impairment of osmoregulatory capabilities was evident in a study of the
effects of sublethal concentrations of inorganic Hg and PCBs on chloride ion
levels in the blood of estuarine shrimp (Palaemonetes pugio). The study showed
significant effects on the ability of shrimp to adjust to rapid fluctuations in
environmental salinity. Thus a rapid change in salinity during exposure to
sublethal concentrations of the PCB could be lethal to the shrimp (Anderson et
al., 1974).
Studies of osmotic regulation in a variety of aquatic organisms have shown
that this is a sensitive measure of the effects of a number of different metals.
Jones (1975), for example, showed that Cd significantly lowered the blood osmotic
concentration of the isopod Idotea neglecta in seawater, while Cd, Zn, and Hg
significantly altered the osmoregulatory ability of another isopod, Jaera albifrons,
in water of dilute salinity. Increases in mortality are associated with
osmoregulatory changes at low salinities. Reduced osmotic regulation has also
been observed in a number of other estuarine species exposed to chlorinated
hydrocarbons (PCBs, DDT).
Osmoregulatory changes have also been observed in killifish (Fundulus
heteroclitus) and eels exposed to sublethal concentrations of DDT and Aroclor
8
-------
1221. Blood of exposed fish exhibited elevated Na+ and K+ levels and higher
osmolarity compared to controls (Anderson et al., 1974).
In another study, the ability of sodium-depleted killifish (F. heteroclitus) to
take up sodium was completely inhibited after a 24-hour exposure to sublethal
levels of inorganic Hg (Renfro et al., 1974).
The ability of organisms to acclimate to osmotic stress makes the
interpretation of the ecological significance of short-term osmoregulatory studies
difficult. The relationship between osmoregulatory changes and effects on
mortality is rarely well established. Although impairment of osmoregulatory
capabilities by pollutants has been repeatedly demonstrated in laboratory
experiments, it is questionable whether this is a suitable endpoint for
environmental monitoring studies. Osmotically impaired organisms have never been
found in contaminated estuarine environments, perhaps due to the rapid death of
such organisms following exposure to an osmotic stress (Bayne et al., 1980).
1.4 RESPIRATORY ACTIVITY
Several investigators have used the respiratory activities of fish and aquatic
macroinvertebrates as indicators of aquatic organism response to environmental
stress. Physiological parameters that have been used to monitor respiration
include ventilatory frequency and the cough response. Ventilatory frequency is
defined as the buccal and opercular opening and closing frequency. The cough
response has been frequently used as a short-term predictive parameter. It is
defined as a regularly recurring break in the ventilation rhythm. These
9
-------
parameters may be affected by variables other than pollutants, such as dissolved
oxygen, temperature, and predator influence. Thus, the interpretation of the
ecological significance of these respiratory responses can be problematic.
Respiratory responses provide rapid and inexpensive measures of response to
pollutants. However, it is unclear whether they accurately predict long-term,
chronic effects of pollutants. Among the most useful long-term chronic tests are
those employing full-life-cycle and critical-life-stage exposures for deriving a
maximum acceptable toxicant concentration (MATC). While the MATC facilitates
projection of "safe" concentrations of chemicals in natural waters, determination
of a MATC is expensive and time-consuming compared to many physiological
endpoints, although not in comparison to population- or ecosystem-level endpoints.
In one study, ventricular frequency was used to determine the diurnal
respiratory response of bluegills (Lepomis macrochirus) to surfactants. The results
were compared to previously existing full-life-cycle chronic toxicity data on other
species of fish. A good correlation was found between the chronic MATC for
fathead minnows and concentrations of surfactants that produced statistically
significant changes in the diurnal ventilatory frequencies in exposed bluegills
(Maki, 1979). These results support the notion that monitoring of ventilation
frequency may have predictive utility as a tool for early estimation of long-term
chronic effects.
In another study, elevated ventilatory rates and reduced food conversion
efficiency were observed in pinfish (Lagodon rhomboides) exposed to oil and
bleached kraft mill effluent (Stoner and Livingston, 1978). These changes indicate
10
-------
that the pollutant caused an elevated metabolic demand -- a situation that could
weaken the organism and leave it susceptible to additional stress.
Anderson et al. (1974) concluded that changes in oxygen consumption
provided a useful measure of pollutant effects in aquatic organisms. They noted
that polychlorinated biphenyls significantly reduced oxygen consumption in killifish
(Fundulus similus). In addition, the respiratory rate following exposure to PCBs
was only about 20 percent of the rate measured in the same fish prior to
exposure. In contrast, grass shrimp (P. pugio) exposed to PCBs showed an
increase in oxygen uptake. These authors also reported an age- and species-
specific response of crustaceans to petroleum hydrocarbons. Mussels (Mytilus
edulis) taken from areas polluted with heavy metals and organic compounds
displayed increased oxygen consumption rates.
The cough response has been used in several respiratory studies to measure
pollutant-induced stress. In one such study conducted by Bull and Mclnerney
(1974), exposure of juvenile coho salmon (O. kisutch) to sublethal concentrations
of the organophosphate insecticide fenitrothion caused an increase in the cough
response. Similarly, an increase in cough frequency in yearling brook trout was
observed within 2 to 24 hours after exposure of the fish to Cu. The mean cough
frequencies tended to increase with Cu concentration (Drummond et al., 1973).
The Maki (1979) study is suggestive of the utility of short-term respiratory
responses as predictors of long-term pollutant effects. However, further workv
needs to be done before respiratory responses can be used as indices of the
effects of pollutants on survival, growth, and reproduction.
11
-------
1.5 BEHAVIORAL ALTERATIONS
Behavioral changes provide potentially useful endpoints for measuring
pollutant effects, because they are elicited at very low pollutant concentrations
and often affect population survival and success. Changes in behavior affect
reproduction, migration, nesting, shelter construction, avoidance activities, and
vulnerability of prey. Difficulties with behavioral measures arise, however,
because they are not easily quantifiable. Practical assessment of pollution-related
behavioral changes depends on qualitative definition of normal behavior patterns.
1.5.1 Avoidance Reactions
Perception and avoidance of contamination is paramount for the survival of
species exposed to polluted ecosystems. In several studies, the Atlantic salmon
(Salmo solar) has been observed to avoid water contaminated with Cu or Zn. In 4
successive years of pollution, downstream returns of salmon were 22, 14, 10, and
15 percent of the upstream migrations, in contrast to the 1 to 3 percent
downstream returns during the 6 years before pollution. The estimated threshold
for the avoidance reaction was about 0.35 to 0.43 toxic units (1.0 toxic unit being
equal to the LD50). A unit of 0.8 blocked all upstream movement (Sprague et
al., 1965, cited in Sheehan, 1984a; Saunders and Sprague, 1967).
Differential avoidance of insecticides was observed in two mosquito-fish
(Gambusia affinis) populations differing greatly in insecticide tolerance (resistant
and susceptible populations). The study showed that both populations avoided
12
-------
endrin, toxaphene, and parathion, but only the susceptible population avoided DDT.
In addition, the concentrations at which the avoidance response occurred differed
markedly in the two populations. With the exception of DDT, the resistant
population showed a much greater avoidance of damaging concentrations of the
other three pesticides than the susceptible population (Kynard, 1974).
In another avoidance study, rainbow trout fry (Salmo gairdneri) were found
to avoid a selected group of toxicants such as copper sulfate and dalapon but not
others such as diquat (Folmar, 1976). Avoidance reactions have also been
observed in a number of different invertebrate species. In one study, the
distribution of burrowed marine bivalves (Macoma balthica) was assessed in a tank
containing sediment contaminated with various concentrations of metals, including
Cu, Pb, Zn, Cr, Hg, and Cd. At the highest pollutant concentrations, a
statistically significant avoidance response was observed (McGreer, 1979).
The midge larva (Chironomus tentans) was also found to display a linear
avoidance reaction to metal-contaminated (Cd, Zn) sediment of Lake Palestine in
northern Indiana (Wentsel et al., 1977). Other factors such as texture, organic
matter content of the sediment, and death of chironomids did not influence the
avoidance behavior.
1.5.2 Predator-Prey Interactions
Sublethal concentrations of pesticides have been reported to interfere with
prey escape and other antipredator behaviors. In a multiprey system, predators
may be expected to consume a higher than normal proportion of the more
13
-------
affected or impaired species. Thus, changes in predator-prey interactions can
provide an important indication of pollutant effects on a population. Examples of
effects of pollutants on predator-prey interactions are discussed in Chapter 2.
1.5.3 Reproductive Behavior
A number of different types of reproductive behavior have been used as
ecotoxicological endpoints. Behavioral changes are often mediated through
pollutant-induced changes in the endocrine system. These hormonal changes may
ultimately induce behavioral anomalies, such as abnormal courtship and nest
construction, decreased incubation attentiveness, nest desertion, impaired nest
defense, and decreased parental care.
Changes in parental behavior resulting from pollutant stress have been
observed in birds. Herring gulls (Larus argentatus) of Lake Ontario exposed to a
variety of pollutants, including DDE, PCBs, mirex, photomirex, and other
pesticides, were found to be inattentive and showed decreased nest defense
compared to birds from an uncontaminated site. Nests were left unattended for
long periods, thus exposing the eggs to predators and to less than optimal
temperature, which resulted in a high incidence of embryonic mortality (Fox et
al., 1978). Similarly, laughing gulls (Larus atricilla) exposed to a single sublethal
dose of the organophosphate insecticide parathion exhibited reduced parental care
activity and increased rate of nest desertion. Hatchability and defense behavior,
however, were not affected (King et al., 1984).
14
-------
Exposure of mallard hens (A. plalyrhynchos) to mercury has resulted in
abnormal maternal behavior. The exposed hens laidx their eggs outside their
nesting boxes and did not defend them against predators (Heinz, 1979). Likewise,
hens fed the pesticide Abate prior to initiation of laying until ducklings were 21
days of age, showed abnormal incubation behavior (Franson et al., 1983).
Courtship behavior has also been shown to be affected by toxicants. For
example, in one study, the mean number of seconds of total courtship activity
time displayed by male Ringed Turtle Doves (Streptopedia risoria) was reduced by
exposure to DDE in the diet (Haegele and Hudson, 1977). The degree of change
was found to be related to the level of DDE exposure. No change in courtship
behavior was observed at 10 ppm, while birds on the 50 ppm DDE contaminated
diet had reduction in activity time.
1.5.4 Locomotor Activity
Locomotor activity is one of a number of behavioral endpoints that has been
used to monitor pollutant stress on a population. Exposure to even low
concentrations of environmental pollutants are known to elicit adverse effects on
locomotor activity. Activities associated with changed behavior, however, vary
widely between subjects and over time and are not easily quantifiable. Therefore,
to properly measure a change in behavioral response, normal behavior patterns
must first be defined so that a quantal or gradual change can be demonstrated.
A number of studies document pollutant-induced changes in the locomotor
activity of fish. Davey et al. (1972) observed locomotor changes in goldfish
15
-------
(Carassius aurathus) exposed to DDT. In unexposed fish, a significant correlation
was established between the magnitudes of consecutive turns in opposite
directions, which initially decreased as the time between the turns increased and
then ceased abruptly. As a retention process is involved in this process, it is
assumed that this behavior is controlled by the central nervous system. It was
observed that DDT-exposed fish displayed a significant reduction in the
correlations between turns, implying an impairment of the retention process.
Returning the fish to clean water for up to 130 to 139 days did not restore
normal behavior.
Similarly, juvenile coho salmon (O. kisutch) exposed to the organophosphate
insecticide fenitrothion displayed changes in behavior patterns, indicating
physiological impairment (Bull and Mclnerney, 1974). All locomotor and some
comfort behaviors ceased, and many fish, unable to hold position, were swept
downstream.
In another study, effects on the locomotor activity of yearling brook trout
(S. fontinalis) exposed to Cu were observed. While locomotor activity varied
widely, all groups were from 4 to 6 times more active than the controls.
Increased activity lasted for 6 to 8 hours at all concentrations of Cu and then
returned to pretreatment values (Drummond et al., 1973).
1.5.5 Feeding Activity
Measures of changes in feeding response are potentially useful endpoints for
ecological risk assessment because they provide an early indication of
16
-------
physiological stresses that could result in growth retardation. Unfortunately, the
feeding response is difficult to measure, and the baseline data for this activity is
sparse.
One study in which feeding response proved to be an effective measure of
pollutant effects was conducted by Drummond et al. (1973). The investigators
monitored the feeding habit of the brook trout (S. fontinalis) following exposure
to Cu. Feeding was markedly depressed or ceased after 2 hours at Cu
concentrations of 9 ug/1 or higher and did not return to normal within 24 hours.
Feeding was also found to be more sensitive than other behavioral
parameters as a measure of the effects of sublethal concentrations of the
organophosphate insecticide fenitrothion on juvenile coho salmon (O. kisutch).
The chemical was found to depress the feeding response of the fish at
concentrations of 0.1 ppm (Bull and Mclnerney, 1974).
1.6 REPRODUCTIVE TOXICITY
Reproductive success, crucial for species perpetuation, primarily depends on
the normal functioning of the neuroendocrine system. This, in turn, is closely
associated with normal maturation of the reproductive organs and their
functioning during various phases of the reproductive cycle (e.g., gamete
formation and its maturation, fertilization, courting and mating behavior, and
rearing of the young (see the section on Behavior Alterations). Environmental
pollutants that adversely influence these reproductive parameters also influence
population dynamics. Birge et al. (1980, cited in Sheehan, 1984a) suggested that
17
-------
10 percent or greater increase in mortality in the developmental stages
(embryonic, larval) would significantly alter the population dynamics in natural
populations.
Reduced hatch success, one of the most commonly used indices of
reproductive success, was observed in a study of pollutant effects on barnacles
(Wu and Levings 1980). The investigators noted reduction in egg production and
survival in barnacles (Balanus glandula) exposed to bleached kraft pulp mill
effluent near a pulp mill outfall.
Similarly, a significant reduction in brood size and number of survivors was
observed among cladocerans (Moina macrocopa) exposed to sublethal
concentrations of Cd, with 50 percent reproductive impairment occurring at a
concentration of 0.78 ug/1 of Cd (Hatakeyama and Yasuno, 1981). Fish eggs
exposed to cyanide, to organochlorines such as Lindane, or to sublethal
concentrations of heavy metals such as Hg, Fe, and Cr showed decreased
fertilization rates (Billard, 1978, cited in Sheehan, 1984a).
Schofield (1976) reviewed studies of declining freshwater fish populations in
acidified waters in Scandinavia and eastern North America. He reported rapid
extinction rates of fish populations inhabiting acidified lakes (pH less than 4.5),
often resulting from chronic reproductive failure.
Reduced hatch success in birds also provides a sensitive endpoint for toxic
effects. Blus et al. (1972) found that dieldrin impaired reproduction in brown
pelicans (Pelecanus occodentalis). Hatching success was also reduced in nests of
18
-------
the red-breasted mergansers (Mergus serraior) contaminated with toxaphene (Heinz
et al., 1983). In a study of three generations of mallard ducks (A. platyrhynchos)
fed 0.5 ppm of methylmercury, exposed birds laid fewer hatchable eggs than
controls (Heinz, 1979). Similarly, screech owls exposed to endrin in their feed
experienced reductions in the number of eggs laid, number of eggs hatched, and
number of chicks fledged (Fleming et al., 1982).
Reduced hatch success in birds has been correlated with eggshell thickness.
Eggshell thinning thus provides a potentially useful endpoint for measuring
reproductive toxicity. Cooke (1973) reviewed experimental avian data and
concluded that environmental pollution can result in thin, easily cracked eggshells.
In addition, there are interspecific differences in response among birds, with
falcon eggs being more susceptible to thinning than eggs of gallinaceous birds.
Studies of the mechanisms involved in reduction of shell thickness by
organochlorines suggest general impairment of calcium metabolism, a reduction in
available carbonate in the shell gland lumen, a deleterious effect on the thyroid
and adrenal glands, and an alteration in organic matter being incorporated into
the developing shell. The extent to which each of these mechanisms is likely to
contribute to shell thinning will depend on the species, condition of the bird, and
environmental conditions.
These effects are illustrated in a study of red-breasted mergansers (M.
senator) nesting on islands in northwestern Lake Michigan. The birds
experienced a small degree of eggshell thinning, which was attributed to DDE
(Heinz et al., 1983; Peakall, 1983). In another study conducted on mallard ducks
19
-------
{A. platyrhynchos), shells of eggs laid by Hg-fed hens were significantly thinner
and of poorer quality, than the shells of those laid by controls (Heinz, 1979).
Egg shell thinning has been used as an indicator of population trends. In
the brown pelican, thinning of 15 to 20 percent has been associated with
declining populations on several widely separated islands (Blus et al., 1972).
Pollutant effects on the reproductive responses of animals are often mediated
through changes in the endocrine system. Rattner et al. (1984) recently reviewed
endocrine-induced effects on the reproductive response of birds to environmental
pollutants. They noted endocrine-related delays in breeding and reproduction in
several species exposed to DDT and its metabolites. For example, in one study,
ring doves (S. risoria) chronically fed DDE showed delayed egg laying and mating
behavior with subtle changes in the levels of the luteinizing hormone (LH) in
females.
Endocrine changes were also observed in several studies of Japanese quail.
One study showed that ingestion of kepone affected reproduction in Japanese quail
(Coturnix c. japonica) directly by interfering with normal egg production of the
oviducts and indirectly by altering hormone secretion of the hypothalamo-
hypophyseal-gonadal system, thus impairing ovarian follicular development.
Ingestion of PCBs affected egg laying capacity in female quails and reduced
testicular seminiferous elements in males (Rattner et al., 1984).
Reproductive effects were also correlated with endocrine changes in field
studies conducted on herring gulls (L. argentatus). Herring gulls exposed to DDE
20
-------
and PCBs exhibited abnormal incubation behavior and hyperactivity. Histological
examination and blood hormone levels indicated these effects were due to thyroid
dysfunction
There is very little information on the effect of environmental pollutants on
amphibian and reptile reproduction. Martin (1983) reviewed the experimental data
and reported that quinacrine exposure produced testicular lesions in lizards, Cd
suppressed spermatogenesis in sexually active frogs, and Pb and Cu disturbed the
germination of frog spawn.
A major problem in using reproductive success as an ecotoxicological
endpoint is that it can be influenced by many factors, including nutritional state,
weather, disease, and predation. Thus, the cause for reproductive failure may be
difficult to identify in field populations. In particular, there are few studies of
effects of toxic pollutants on reproductive success in wild mammalian populations.
1.7 MUSCULOSKELETAL EFFECTS
Among the numerous pathological signs that are associated with pollution
stress, skeletal abnormalities are most amenable to quantification. Some of the
most readily detectable and widely used abnormalities include vertebral and spinal
deformities in fish and shell erosion in crustaceans. The vertebral and spinal
cord deformities may be manifested as dorsoventral flexures (lordosis), lateral
flexures (scoliosis), or backward spinal curvature (kyphosis). These deformities
may have a variety of behavioral effects, including impaired swimming
performance, lowered feeding rates, impaired ability to avoid predators, decreased
21
-------
ability for territorial defense, reduced ability to compete for a sexual partner, and
general physiological weakness (Bengtsson, 1979).
Pollutant-induced stresses have been responsible for skeletal abnormalities in
fish. For example, skeletal abnormalities have been observed in minnows exposed
to Zn and Cd. Zn and Cd may affect the neuromuscular functioning of the fish,
resulting in overloading of the vertebrae and skeletal fracture. Interference with
calcium metabolism by heavy metal or chlorinated hydrocarbon contamination may
also contribute to fish skeletal deformities (Bengtsson et al., 1985). Holcombe et
al. (1976) observed a 20 percent increase in skeletal deformities in the second
generation of brook trout (S. fontinalis) exposed to Pb over three generations.
Spinal skeletal defects such as lordosis (16 percent), scoliosis (32 percent),
kyphosis (25 percent), and extreme rigidity and coiling of the vertebral column
(17 percent) have been described in the embryos and larvae of a variety of fish
exposed to pesticides, such as atrazine, chlorobenzene, and trisodium
nitrilotriacetic acid (Birge et al., 1979).
Increased incidences of skeletal abnormalities have also been described in
amphibian tadpoles (frogs, toads) exposed to DDT, oxamyl and dieldrin (Cooke,
1972, 1981; Martin, 1983) and in sea urchin (Paracentrotus livodus) embryos
exposed to Cu, Zn, Se, and Cd during the pre- and post-hatching stages of
development (Pagano et al., 1986).
Other kinds of physical abnormalities may serve as endpoints for monitoring
pollutant effects on a population. In one study, a high incidence of fin and gill
22
-------
raker abnormalities appeared in sea-spawning whitefish (C. lavaretus) exposed to
heavy metal pollution (Bengtsson et al., 1985).
In another study, dover sole (Microstomas pacificus), collected from the
vicinity of a major municipal waste-water discharge site in southern California,
exhibited fin erosion and tumor-like growth on the skin (Mearns and Sherwood,
1974). Histological examination showed the fin disease to be external (Wellings et
al., 1976) and apparently caused by contact with contaminated sediments around
the waste-water outfall.
One problem with the use of physical deformities as an ecotoxicological
endpoint is that they are not specific responses to pollutants. There may be
many causes of such abnormalities, including genetic factors, disease, and many
environmental variables, as well as interactions among these factors. Another
problem is that some of the commonest abnormalities (fin rot, shell erosion) are
difficult to quantify, while more quantifiable deformities rarely occur. The
potential effects of physical abnormalities on fitness of the affected organisms
remain mostly hypothetical. The very fact that deformed animals are captured
shows that the deformities have not prevented growth and survival in the affected
environment. Little is known how deformities affect such important population
parameters as growth rates, lifespan, reproduction, competitive ability, feeding
rates, and predator avoidance in the natural environment.
23
-------
1.8 GROWTH AND DEVELOPMENT
1.8.1 Developmental Toxicity
The developmental period of an organism's life cycle is the period extending
from fertilization of the egg through maturity. Pollutants can interfere with
development by disrupting early cell division, growth, and migration (embryonic
growth) or by interfering later with critical growth phases during organogenesis
(e.g., larval, fetal) causing embryotoxicity or lethality, altered growth (e.g.,
retardation), functional deficiencies (e.g., metabolic, behavioral), structural
abnormalities (e.g., terata), and diminished maturation of offspring. A number of
studies have examined pollutant effects on development and survival of the
organism through the embryonic and larval stages, because of the high sensitivity
of these developmental stages to pollutants, and because of the short duration of
early development.
A variety of studies demonstrate developmental effects resulting from
pollutant exposures. In one such study, fertilization success was significantly
reduced in sea urchins (e.g., Echinus esculentus) following exposure of sperm to
Cd or Zn (Pagano et al., 1986).
Amphibian embryonic and larval stages are often very sensitive to pollutants.
For example, low pH caused embryonic death in salamanders, while inorganic Cu
greatly reduced growth and prevented metamorphosis in frog tadpoles (Martin,
1983).
24
-------
In another study conducted with birds, injecting DDT and its metabolites
into gull eggs (Larus californicus) induced abnormal development of ovarian tissue
and oviduct in male gull embryos. This feminization process probably resulted
from estrogenic action (Rattner et al., 1984).
In addition to skeletal abnormalities discussed earlier, pollutant-induced
stresses have caused soft-tissue abnormalities. Soft tissue anomalies such as optic
malformations were observed in Atlantic silverside (Menidia menidia) embryos
exposed to insecticides, rainbow trout (S. gairdneri) embryos treated with
benzo(a)pyrene (B(a)P), and embryos of killifish (Fundulus heteroclitas) exposed to
magnesium chloride or methylmercury. Cardiac anomalies have been reported for
several species of fish exposed to pesticides (carbaryl, parathion, malathion,
mercury compounds) (Weis and Weis, 1987).
Cell and tissue pathology have often been used to demonstrate incidence of
internal injuries or anomalies caused by pollutants. Hose et al. (1984) exposed
rainbow trout (5. gairdneri) alevins (from fertilization through hatching) to B(a)P
and observed nuclear pycnosis and karyorrhexis in the brain, retina, and muscles;
microphthalmia associated with patent optic fissure; abnormal mitosis in hepatic,
neural, and muscular tissues; and skeletal malformations in the skull and vertebral
column. Possible ecological effects of such morphological abnormalities would be
decreased feeding and growth, and inability to escape predation. The teratogenic
responses described in these studies suggest that monitoring of abnormal egg,
embryos, and larvae may provide a good estimate of the severity of environmental
stresses.
25
-------
Behavioral abnormalities may also be elicited in offspring of parents exposed
to environmental contaminants. Thus, exposure of three generations of mallard
ducks (A. platyrhynchos) to 0.5 ppm of Hg resulted in a significant inhibition of
the ducklings' response to simulated maternal calls. The ducklings were hyper-
responsive to frightening stimuli in avoidance tests, but their locomotor activity
was normal (Heinz, 1979). A similar behavioral effect was observed in ducklings
from mallard hens chronically fed 3 ppm of DDE (Heinz, 1976).
Studies have demonstrated that the embryo-larval stages of aquatic organisms
are particularly susceptible to pollutant stress. Accumulation of organochlorine
pesticides (e.g., mirex, DDT) in several species of fish eggs, for example, resulted
in high embryonic and larval mortality with the embryos being more susceptible
than the larvae (Birge et al., 1979; Livingston, 1977). The rainbow trout was the
most sensitive species tested. Reduced survival rates were also observed in
embryos, larvae, and juveniles of fathead minnows (Pimephales promelas) exposed
to 4-methyl-2-pentanone, 1,4-dimethoxybenzene, benzophenone, and 3,4-
dichlorotoluene (Call et al., 1985).
In a recent review of early-life-stage toxicity tests on fish, McKim (1985)
found that, in 83 percent of 72 comparisons, estimates of MATC obtained from
early-life-stage tests were identical to MATCs established by longer, more
involved, and more costly complete- or partial-life-cycle toxicity tests. Early-
life-stage tests are thus considered to be particularly useful estimators of long-
term chronic toxicity. Limitations of utilizing early-life-stage tests are that the
duration of exposure may not be sufficient to observe cumulatively toxic effects,
26
-------
and they provide no mechanism for identifying alternative modes of action
resulting from toxicity to other life stages (Macek et al., 1978).
1.8.2 Growth
Growth provides an integrated index for measuring the physiological status
of an organism that has not as yet attained its maximum biomass. It is a widely
used endpoint for monitoring the effects of pollutants on both aquatic and
terrestrial organisms.
The use of changes in growth rate as a measure of pollutant stress in
aquatic organisms is illustrated in a study conducted by McKim and Benoit (1971).
These investigators monitored the growth rate of juvenile brook trout and found
it to be extremely sensitive to pollutant stress. Juvenile fish exposed to Cu for
14 months showed a significant reduction in growth rate. This reduction was
inversely correlated with Cu concentrations. By comparison, adult fish were
unaffected. The study indicated that long-term growth rates are an effective
measure of chronic stress.
Other studies have reported growth rate retardation in microinvertebrates at
high but sublethal pollution levels. Borgmann et al. (1980) studied the effects of
Cd, Cu, Hg, Pb, and As on the growth and survival of copepods. Experiments
were conducted seasonally, using naturally occurring water and food rather than
defined media, in order to determine the extent to which varying environmental
conditions affect toxicity. Growth rates of the copepod population were affected
27
-------
at sublethal metal concentrations. A seasonal cycle in toxicity was observed with
all metals except As.
Numerous studies indicate that growth rate provides a sensitive endpoint for
monitoring pollutant effects on plants. Pollutants may influence forest growth
and development via multiple pathways and mechanisms and over varying time
scales. This influence may be exerted via physiological disruption of the plant-
water balance (osmoregulatory activity), change in forest nutrition, effects on
growth and reproduction, or altered resistance to secondary stresses (McLaughlin,
1985). For example, ozone exposure disrupts the stomata-controlled leaf-water
balance (Heath, 1975). The physiologic disruption of the plant-water balance may
result in diminished capacity of the plant to take up water from the soil or
control its water loss to the atmosphere through its foliage.
In one plant study conducted by Miller et al. (1977), the vegetative growth
of corn shoots was suppressed by both Pb and Cd. Further, the concentration of
Cd in corn shoots was increased by the addition of Pb to the soil, while the
presence of Cd in the soil reduced the uptake of Pb.
Changes in rates of seedling germination and growth have provided sensitive
measures of pollutant effects. For example, reductions in growth rate, stem
elongation, leaf area, plant and root weight, fruit and seed set, and floral
productivity occurred in several plants grown in air with ozone levels ranging
from 8 to 10 ppm (Feder, 1973). Reductions in growth and germination also
occurred in plants exposed to heavy metal contaminated soil (Walley et al., 1974).
This last study is one of a number showing heritable variation in tolerance to
28
-------
heavy metals, thus indicating the potential for evolution of tolerance to toxic
pollutants.
Pollutant effects on seedlings were also observed by Constantinidou and
Kozlowski (1979) in a study of 4-month-old Ulmus americana seedlings exposed to
sulfur-dioxide (SO2) or ozone or both. Injury to the leaves was observed within
48 hours after all treatments. Expansion of new leaves was inhibited by the
mixture and SO2; expansion of young leaves was inhibited by all three treatments.
Leaf emergence was significantly reduced by SO2 and the mixture at the end of
the first week; stem dry weight and root dry weight after 5 weeks were reduced
by ozone, and/or SO2. Quiescent seedlings responded to all fumigation treatments
with severe defoliation. These experiments emphasize that the inhibitory effects
of pollutants on plant growth vary markedly for different organs and tissues and
at different stages of plant growth.
A variety of methods are available for evaluating and predicting forest
growth response to air pollution. One such method, developed in a 1982 project
entitled FORAST (Forest Response to Anthropogenic Stress), involves surveying
and characterizing forest damage based on long-term growth changes and effects
on wood density (McLaughlin and Braker, 1985). These dendroecological
techniques document growth responses of individual trees over their entire life-
spans. Techniques include measuring the width of annual rings and cross-dating
to rings measured in other trees of the same species. Confounding elements may
be eliminated by statistical analysis (e.g., effects of tree age, competitive status,
climate). Additionally, wood density can be measured annually and correlation of
growth decline with causal factors (e.g., emissions, meteorological data)
29
-------
can be combined with elemental analysis of tree rings to correlate pollutant
accumulation with measured growth changes.
Elemental analysis of tree rings provides another possible mechanism for
determining pollutant effects on forest trees. For example, increases in Fe, Ti,
Cd, Cu, and Mn in trees correspond to increased regional combustion of fossil
fuels (McLaughlin and Braker, 1985).
The sophisticated technology that is now available (e.g., PIXE-photon -
induced X-ray emission spectrophotometer, ICP, inductively coupled plasma
emission spectroscopy, laser spectrophotometer, and portable computer-based
systems) makes it possible to collect and analyze physiological data in the field,
in situ, on foliage-stressed and nonstressed trees.
Using forest growth as an ecotoxicological endpoint presents a number of
problems. Multiple surveys are necessary to determine changes in rates of growth
and wood density. Many potential confounding factors, such as soil
characteristics, geologic factors, elevation, aspect, wind, temperature, soil moisture
capacity, disease, grazing, and competition, cannot be controlled, so their effects
must be estimated statistically. There might be indirect as well as direct effects
of pollutants, such as pollutant-induced changes in nutrient availability and
altered resistance to secondary stresses.
30
-------
1.9 GENOTOXICITY
Mutation is a permanent change in genotype other than one brought about
by genetic recombination. Genetic agents can produce changes in cellular
deoxyribonucleic acid—a group of changes that are implicated as the initial events
in carcinogenesis. Analytic methods have been developed that make it possible to
measure genotoxic endpoints in the laboratory. The two most frequently used
methods for detecting genotoxic effects are cytogenetic analysis and sister
chromatid exchange (SCE). The two most popular in vivo test models are the
central mudminnow (Umbra limi) and the eastern mudminnow (Umbra pygmaea).
Both species are available in large numbers and are easily captured; they possess
a low number (22) of large, easily observed chromosomes; they survive well in a
laboratory atmosphere; and they possess the microsomal system necessary for
activating promutagens to genotoxic intermediates (Bishop and Valentine, 1982).
In one study, exposure (intraperitoneal injection) of the central mudminnow
{U. limi) to two direct-acting mutagens—methyl-methanesulfonate (MMS) and N-
methyl-N'-nitro-N-nitrosoguanidine (MMNG)--and to two indirect-acting mutagens
--cyclophosphamide (CP) and dimethylnitrosamine (DMN)--resulted in a significant
dose-dependent increase in the frequency of SCE (Bishop and Valentine, 1982). In
another study, the eastern mudminnow (U. pygmaea) exposed to polluted Rhine
water containing organic acids, esters, aldehydes, and phenolic compounds showed
a time-related increase in SCEs in gills and testicular tissues (Alink et al., 1980).
The SCE method of analysis was also successfully used to examine the
genotoxicity of mitomycin C (MMC) in the marine polycheate worm (Neanthes
31
-------
arenaceodentata). A dose of 5 x 10"^ mol/1 of MMC increased the rate of SCE
from a baseline frequency of 0.14/chromosome to 0.5/chromosome (Pesch and
Pesch, 1980). Other studies showed that the larvae of the mussel, M. edulis were
1.5 times more sensitive to the genotoxic effects of MMC than the polycheate
worm (Harrison and Jones, 1982).
One problem in using SCE as an ecotoxicological endpoint is that information
on SCE is available for only a few species. In addition, the mechanism for SCE
induction is not well understood, and the biological significance of SCE is
unknown. Some researchers, in fact, question whether SCE is correlated with a
mutational event (Bishop and Valentine, 1982).
1.10 CARCINOGENICITY
The results of some studies have demonstrated that selected fish and
invertebrates could be used as indicators of carcinogenic pollutants in the
environment through monitoring carcinogenic effects, such as the development of
benign or malignant neoplastic growth of cells in tissues.
Chemical induction of cancer has been demonstrated in numerous
environmental epidemiologic studies (Couch and Harshbarger, 1985). Approximately
300 species of fish and 15 species of bivalve molluscs have varied spontaneous and
experimentally induced tumors. The etiology of bivalve neoplasms is problematic
despite some studies that implicate one or a combination of viral, chemical, and
genetic factors. On the other hand, there is clear evidence of induction in fish.
Many of the fish neoplasms are idiopathic. Others, such as the lymphomas of
32
-------
northern pike, appear to be caused by a retrovirus. Chemical and genetic
impairment have been implicated in neoplasms originating from numerous cell
types in certain platyfish/swordfish hybrids.
Liver and skin cancers in various bottom fish have been associated with
point source concentrations of environmental chemicals. These cancers have been
produced in the laboratory by extracts of the bottom sediments of polluted
waterways. Studies have shown that fish liver metabolizes carcinogenic pollutants
into reactive intermediates and that fish experimentally exposed to known
carcinogens have frequently developed liver neoplasms (Couch and Harshbarger,
1985).
Although the most frequent tumor type reported in fish is the hepato-
carcinoma, other types such as fibrosarcomas of the skin, tumors of the gill, and
other organs have also been reported (Edwards and Overstreet, 1976; McCain et
al., 1977; Meyers and Hendricks, 1982; Sindermann, 1980).
The development of neoplasms has been strongly correlated with pollutant
exposures in other vertebrate classes. For example, neotenic tiger salamanders
(Ambystoma tigrinum) inhabiting a small isolated lagoon (Lubbock County, Texas),
which was heavily polluted with secondarily treated domestic sewage, developed
neoplastic and related skin lesions. The neoplasms were of epidermal, fibrotic,
and melanocytic origin. No neoplasms were discovered among the larvae sampled
from 16 proximal nonsewage lagoons (Rose and Harshbarger, 1977).
33
-------
It is important to develop methods for introducing and monitoring sensitive
fish/invertebrate species at or near suspected pollutant sources in an aquatic
ecosystem. The current histopathologic techniques for positive identification of
fish cancer should be backed up by identification of biochemical endpoints. These
endpoints would include specific enzyme inductions in selected tissues, early
detection of carcinogenic changes, and histopathological bioassays of fish liver
and bile extracts and invertebrate digestive glands.
Fish bioassays, especially those using embryos, have been advocated to pre-
screen chemicals and mixtures for carcinogenicity with considerable advantage in
cost and time over similar tests for carcinogenicity in rodents. Sensitivity is
excellent, and interpretation is rarely complicated by spontaneous tumors (Couch
and Harshbarger, 1985). However, synergistic effects and viral interactions
complicate analysis, especially in the natural habitat. Also, there have been few
studies of pollutant-induced carcinogenicity in avian and mammalian populations.
1.11 MODIFYING FACTORS
In an ecosystem, each organism lives within a number of physical, biological,
and chemical constraints because of exposure to many factors that may act
independently or together. Under these constrained circumstances of existence,
any additional environmental stress (such as toxic pollutants) may exacerbate or
ameliorate pre-existing conditions. These multifactorial interactions and stresses
commonly occur among inhabitants of all ecosystems.
34
-------
When there are multiple stresses, their effects may be additive or
interactive. If the effects are additive, the combined effect of two or more
stresses would be equal to the sum of the independent effects. If the effects are
interactive, they may be classed as either synergistic or antagonistic. In a
synergistic interaction, the combined effect of the multiple stresses is greater
than the sum of their independent effects. In an antagonistic interaction, the
combined effect of the multiple stresses is less than the sum of their independent
effects.
1.11.1 Microbial Toxicity
Multifactorial interactions, including pollutant toxicity and the stress
response of exposed microbial communities, have been extensively reviewed by
Babich and Stotzky (1980). These authors have shown that abiotic,
physicochemical factors influence pollutant toxicity to microbes and viruses either
by potentiation or attenuation and have emphasized the characteristics of the
specific environment into which the pollutants are deposited.
Both cationic and/or anionic elements of the environment can influence the
toxicity of heavy metals to microbes and viruses. For example, Mg reduced the
toxicity of Co, Cd, and Ni to bacteria (Escherichia coli) and filamentous green
algae (Hormidium rivulare); in a similar reaction, Ca reduced the toxicity of Zn or
Cd to H. rivulare and of Cd to Aspergillus niger and E. coli. In a synergistic
reaction, the combined toxicity of Cd and Zn to Klebsiella aerogenes was greater
than the sum of their individual toxicities.
35
-------
The inorganic anionic composition of the environment is known to influence
the chemical form of metals. For example, adding phosphate reduced the toxicity
of Pb to H. rivulare and of Cu to Aerobacter aerogenous. The reduction probably
resulted from the formation of sparingly soluble salts of both elements, i.e.,
Pb3(P04)2 and Cu(P04)2.
The pH of the environment into which the pollutant is deposited is important
because it may influence a chemical's form, solubility, and toxicity. For example,
as the pH is increased, Cd+2 is sequentially hydroxylated to other species, i.e.,
Cd(OH)^, Cd(OH)2, Cd(OH>3, and Cd(OH)42~. Each species is differentially toxic
to the microbiota and possesses differential affinities to viral particles, microbial
cells, and other particulates such as clay minerals, hydrous metal oxides, and
humic acids.
The oxidation-reduction potential (Ejj) of an environment may influence
toxicities to microbes because it may influence the form of some inorganic anions
(e.g., S2" vs. SC>42~), and thus influence the solubility, mobility, and toxicity of
heavy metals deposited into that environment. For example, Cu+ was more toxic
to E. colt than Cu2+, while Fe2+ showed a greater potential for mutagenicity in E.
coli than Fe^+.
The effect of temperature on the physiological and biochemical state of the
microbiota is best illustrated by the inhibition of the growth of the alga
Cyclotella meneghiniana by Cr, where increased inhibition accompanied an increase
in temperature from 5 to 25°C.
36
-------
In bacteria (Bacterium B III39), increases in hydrostatic pressure in the
environment resulted in increased toxicity of Ni to microbes. In addition, the
presence of organic particulates reduced the toxicity of Cd to a variety of
organisms (Bacillus megalerium, Aspergillus fischeri, Penicillium). These effects
resulted from the ability of particulates to remove heavy metal cations from
solution.
Microbial response to heavy metal pollutants in the environment may also be
influenced by the presence of other chemicals. This was demonstrated in
experiments in which bacterial (E. coli, Pseudomonas sp.) levels of respiration and
nitrogen fixation were altered in the presence of Cd to a variable extent
depending upon the species of Cd and the presence of the citrate ion (Lighthart,
1980).
1.11.2 Multifactorial Interactions
A strong positive correlation has been found between levels of selenium (Se)
and Hg in marine mammals and fish. The antagonistic interrelationship between
Se and Hg has been established in many studies (Beijer and Jernelov, 1978).
Synergism and antagonism have also been noted in several plant species
exposed to ozone and SO2. Tingey et al. (1973) observed significantly more foliar
injuries from mixed ozone/S02 treatment than from additive effects of single gas
treatment in tobacco and alfalfa. In other species such as cabbage, broccoli, and
tomato the foliar injuries from mixed gas exposure were additive or less than
additive.
37
-------
Plant injury from ozone, other oxidants, and air pollutants (SO2) may be
differentially affected by using agricultural chemicals, especially fungicides and
nematocides. Tobacco plants were protected against ozone-induced injury by
benomyl (methyl -1 -butyl-carbamyl-2-benzimidazolecarbamate), dodine
(dodecylguanidine acetate), and maneb (manganous ethylenebisdithiocarbamate).
Benomyl was most effective, reducing leaf injury by 59 percent; maneb decreased
the number of brown spot lesions per leaf by 66 percent. Smaller decreases were
observed with benomyl and dodine (Reinert and Spurr, 1972).
Ozone also enhanced Cd-induced injury in cress shoots (Lepidium sativum) by
stimulating early development of chlorosis and necrosis. The combined elements
induced these injuries at lower concentrations than the separate elements
(Czuba and Ormrod, 1984).
Assessment of nutritional stresses on populations of cladocerans (Daphnia
magna) maintained on a vitamin-enriched algal diet showed that these animals
were less sensitive to chronic Cu stresses than those fed a trout-granular diet
(Winner et al., 1977). This conclusion was based on the mean brood size,
survival, and the instantaneous rate of population growth.
Mortality of blue crabs (Callinectes sapidus) in a DDT-contaminated marsh
was increased when environmental temperatures were low (Koenig et al., 1976).
Examination of the hepatopancreas and swimmeret muscles from dead or dying
crabs revealed high concentrations of residual DDT (39.0 and 1.43 ppm,
respectively). These results suggested reduced temperatures and body burdens of
38
-------
DDT interact to produce acutely toxic effects. Several mechanisms of action were
proposed.
Survival was reduced and development was delayed in larval blue crabs (C.
sapidus) following exposure to 50 ppb of Cd at 25°C. Twelve combinations of Cd
and salinity at 25°C were examined in this study (Rosenberg and Costlow, 1976).
In similar tests with the mud-crab (Rhithropanpeus harrisii) larvae, where 63
combinations of Cd, salinity, constant temperatures, and cycling temperatures were
used, combination exposures of 50 or 150 ppb Cd and 10, 20, or 30°/oo salinity
significantly reduced survival (from hatched to first adult stage) of the larvae at
the 20 and 35°C temperature extremes. Results from cycling temperature
experiments (20 to 25° and 25 to 30°C) showed a stimulatory effect on survival of
larvae compared to constant temperature, both in presence and absence of chronic
Cd exposure.
Acidification may influence metal-organism interactions in at least two ways:
a decrease in pH increases solubility of heavy metals, making them more available
to be taken up by organisms; and a pH decrease may affect biological sensitivity
to metals at the level of the cell surface (Campbell and Stokes, 1985).
Elevated levels of aluminum exacerbate toxic effects of reduced pH on
crustacean zooplankton (Havas and Hutchinson, 1982). Examples of antagonism
also occur; aluminum can increase survival of some cladocerans and fish at some
pH levels. Proposed mechanisms for this are that trace metals may compete with
H+ ion for exchange sites on the gill surface, or they may reduce membrane
permeability (Havas, 1985).
39
-------
Heavy metals can reduce tolerance to other environmental stresses, such as
drought stress. For example, cadmium (20 ug/g soil) significantly reduced the dry
weight yield of two native plant species (Andropogon scoparius, Monarda fistulosa)
under drought stress. The effects of the stress and Cd treatment appeared to be
additive (Miles and Parker, 1980).
There are many examples of interactions between pollutants and biotic
environmental stresses. Studies on the effect of air pollutants on the plant-insect
relationship have demonstrated increased susceptibility of greenhouse-grown
soybeans to the Mexican bean beetle (Phaseolus vulgaris) following fumigation
with SO2 (Hughes et al., 1982). Adult females showed feeding preference for
fumigated leaves, produced eggs with higher viability, and showed increases in
both number of eggs per brood and in number of broods produced. A possible
mechanism of action involves modification of amino acid or sugar levels in the
soybean plants. Other examples of interactions between pollutants and predator-
prey relationships are covered in Chapter 2.
Toxic pollutants can also increase susceptibility to disease. In the marine
fish populations, the incidence of various nonspecific diseases (e.g., ulcers,
lymphocytosis, and fin rot in fish and shell erosion in crustaceans) are good
qualitative indicators of environmental stress. In one study, Hetrick et al. (1979)
reported increased susceptibility of rainbow trout to infectious hematopoietic
necrosis virus following exposure to less than 0.01 ppb of Cu. Mortality was
twice as high in Cu-exposed fish as in the controls. Suppression of the immune
response is probably the mechanism involved in causing increased susceptibility.
40
-------
In another study, eels exposed to water contaminated with 30 to 60 ppb of Cu
died with vibriosis (Vibrio anguillarum) infection, whereas eels kept in non-
contaminated water remained healthy (Rodsaether et al., 1977).
These interactions illustrate the many types of interactions that can take
place between environmental pollutants and abiotic and biotic components of the
ecosystem. This raises serious questions about the utility of single-species or
single-factor laboratory toxicity testing in developing ecotoxicological endpoints
for ecological risk assessment.
1.12 CONCLUSION
Ten major categories of physiological ecotoxicologic responses (endpoints)
that may be used to evaluate adverse effects of an environmental pollutant on
populations have been identified. They include acute mortality, biochemical
alterations, osmoregulatory effects, respiratory effects, behavioral effects,
reproductive effects, musculoskeletal abnormalities, and effects on growth and
development, genotoxicity, and carcinogenicity.
The ecotoxicological endpoints most directly related to individual success are
acute mortality, growth and development, and reproduction. A frequently
recommended approach is to begin with acute lethality tests to establish a crude
estimate of toxicity, followed by testing for chronic sublethal effects on growth
and reproduction (Macek et al., 1978; Monk, 1983). The primary advantage of
using growth and reproduction as endpoints is that they integrate all other
physiological processes necessary for individual success. The major disadvantage
41
-------
of these endpoints is the time and expense required to conduct full-life-cycle
chronic toxicity tests.
Other physiological and biochemical responses provide more rapid and
inexpensive measures of pollutant effects. These are potentially useful for early
detection of pollutant effects, thus allowing corrective action before irreparable
damage occurs. In addition, these endpoints are often useful in determining the
mechanisms of ecotoxicological effects. However, the relationship of these short-
term physiological responses to individual growth and reproductive success is
rarely well established. Among the few exceptions to this include the
demonstration that, in fish, early-life-stage toxic effects are highly correlated
with MATCs derived from life-cycle chronic toxicity tests. Other short-term
physiological responses cannot be recommended as ecotoxicological endpoints until
relationships between them and growth and reproductive success are determined.
Within an ecosystem, organisms are continually exposed to many
environmental stresses that may act independently or together. Any additional,
adverse environmental condition could further exacerbate pre-existing stresses. It
has been demonstrated that toxic pollutants may interact with other environmental
stresses. These interactions may be either synergistic or antagonistic and may
involve both abiotic and biotic factors in the environment. The numerous
examples of interactions among environmental factors make it clear that
laboratory single-species, single-factor tests will not be adequate to estimate all
the ecological effects of contaminants. This points to the necessity of further
relating ecotoxicologic effects on individual organisms to population- and
ecosystem-level effects of pollutants.
42
-------
CHAPTER 2
POPULATION-LEVEL ENDPOINTS
2.0. INTRODUCTION
Population-level endpoints form an important component of an ecological risk
assessment. These endpoints focus on the relationship between populations and
the physical, chemical, and biological factors in their environment. The first step
in a population-level assessment requires selection of the most appropriate
endpoints. Pollutants may affect both birth and death rates of populations.
These effects may cause changes in other measurable endpoints, such as
abundance, age structure, distribution, genetic makeup, and life history patterns.
Other potential endpoints are measures of such species interactions as predator-
prey relationships and interspecies competition.
Selection of the most appropriate population(s) to monitor is the next step
in a population-level assessment. Population-level endpoints provide essential
information about the structural dynamics of an ecosystem. In addition, they can
often provide useful indicators of whole ecosystem changes in response to
pollutant stress.
2.1 POPULATION DYNAMICS
The concept of population dynamics was first used by Elton in 1933 to
describe the area of study "concerned with rates of increase, fluctuations in
numbers, and the relation of problems of numbers to the environmental factors
43
-------
which influence the population." This subject deals with the influence of
environmental factors on the rates of birth, death, immigration, and emigration
(Odum, 1971). Population size is determined by the sum of gains from birth and
immigration minus the losses due to death and emigration (Moriarty, 1983).
2.1.1 Birth Rate
The birth rate of a population has long been used as a means of determining
its health. The population birth rate is determined by combining age-specific
fecundity rates with the age structure of the population (Warren, 1971). For
some species, there is a clear correlation of intraspecific competition (density-
dependent factor) with age-specific birth rates and age at maturity (Frank et al.,
1957, cited in Warren, 1971). Fisheries investigations generally do not measure
the birth rate because newly hatched/born individuals are often too small or too
dispersed to catch. Instead, fisheries studies measure recruitment, the young
large enough to be caught with fishing gear (Royce, 1972). Rate of birth (or
recruitment) is a good measure of contaminant stress, as a change can affect not
only the success of individual organisms, it can also be detrimental to population
abundance, age structure, and gene pool. Birge et al. (1980, cited in Sheehan,
1984a) suggested that a 10-percent-or-greater increase in mortality in the
developmental stages (embryonic, larval) would significantly alter population
dynamics in natural populations.
44
-------
2.1.2 Death Rate
The death rate is calculated as the number of individuals dying per unit of
population over time. As in the case of birth rates, the death rate is generally
described by age class. Death rates are inversely related to survivorship. A
survivorship curve is plotted as the number of individuals surviving to successive
age classes (Andrewartha and Birch, 1954; Warren, 1971). The death rate is not
easily measured in either terrestrial or aquatic animals. Dead and dying animals
are rarely seen in nature. As a result of this, death is generally estimated
indirectly from other data, such as historical information on the number of
survivors that are members of the same age class throughout the life cycle
(Royce, 1972).
2.1.3 Population Growth Models
Evaluation of populations has rested on the development of mathematical
descriptions of "populations in terms of abundance, mortality, and reproduction,
defining relationships between life history and population growth, explaining
fluctuations in abundance, and identifying regulatory mechanisms" (Barnthouse et
al., 1986). Many population models have been designed to manage exploited
populations (fisheries in particular) or to control nuisance species such as
Hydrilla.
The simplest population growth models describe exponential growth. However,
no population can grow in this fashion forever; in reality population sizes are
limited. There is little consensus in the scientific community as to the
45
-------
mechanisms regulating population size. The argument centers around the role of
density-dependent factors versus density-independent factors. Density-dependent
factors, e.g., intra- and interspecific competition, tend to stabilize the population
growth; density-independent factors, e.g., weather, act upon a population without
regard to its equilibrium size and are considered nonstabilizing influences.
Murray (1979) states that the prevailing view is that the size of populations is
generally regulated by density-dependent factors; however, in some populations
"density-dependent factors are inadequate, nonfunctional, or nonexistent; thus,
periods of exponential growth are followed by population crashes." However,
Andrewartha and Birch (1954) dispute a distinction between density-dependent and
density-independent factors. They argue that all factors can act in a density-
dependent fashion, but most populations are primarily influenced by nonstabilizing,
density-independent processes. A third view is that population growth is affected
by both density-dependent and density-independent factors (Ricklefs, 1973; Warren,
1971). Most population modeling deals only with density-dependent limitations on
growth.
Population growth rates are generally described by the logistic equations as
modified for competition by Lotka and Volterra. Most basic population models are
based on the logistic equation (Murray, 1979). The basic logistic equation is
represented by:
dN = rN (1-N) where N=number of individuals
dt K K=carrying capacity
t=time
r=per capita rate of increase
which may be restated verbally as:
46
-------
the population's = the population's
actual growth rate potential growth x
rate
the proportion by which
the population is below the
environment's carrying capacity
(Moriarty, 1983)
Carrying capacity (K) is defined as the number of individuals that the
resources of a habitat can support. It can be thought of as the total resources
available divided by the minimum maintenance requirement of each individual
(Ricklefs, 1973). Thus, carrying capacity is determined by characteristics of both
the environment and of the population (McNaughton and Wolf, 1979). The per
capita rate of increase (r) = b - d, where b = instantaneous birth rate and d =
instantaneous death rate.
The logistic equation describes a population growing in a sigmoid fashion,
rising to an equilibrium population size at the carrying capacity. This model
assumes a stable age structure in the population with overlapping generations
(Moriarty, 1983; Murray, 1979). Population growth rates are assumed to be
strictly density-dependent in the logistic model, with birth and death rates
linearly related to population size. Using this model, population growth rates
decline as population size increases, reaching zero as the carrying capacity is
attained. The model can be modified to include coefficients for competition,
predation, and other two-species interactions.
Population growth models used in fisheries often take a somewhat different
form. In the logistic equation, population growth rate is represented by the
number of individuals over a unit of time. Fishery models utilize changes in
population biomass over a unit of time. There are a variety of fishing yield
models; however, they all are based on the general relationship between the unit
47
-------
stock and the effects of additions and losses. In general, recruitment and growth
add to the unit stock while natural mortality and fishing mortality decrease it.
This is in complete agreement with the earlier descriptions of changes in
population numbers as described by Moriarty (1983) and Barnthouse et al. (1986).
However, the actual mathematical representation of this relationship is different
for fish populations. For example, this logistic model is often used in fishery
studies (Royce, 1972):
dP = kiP(L-P) where
dt
P=total biomass of fish in stock
t=time
L=limiting biomass to the stock
ki=growth rate constant
Another approach to predicting population growth is the Leslie matrix model.
This model predicts future population size and age structure of a population,
given both the structure at the present time and a matrix whose elements
represent age-specific fecundity and mortality (Usher, 1972). It can include
density-dependent factors as well as density-independent factors (Murray, 1979)
and can be modified to include some biological interactions, such as competition.
A common application of this model is in determining survivorship and recruitment
rates to estimate the maximum allowable harvest of fisheries, forests, or wildlife
(Boyce, 1977). In the basic Leslie matrix model, birth and death rates remain
constant at the initially measured values. As a result, the model describes
exponential population growth. However, the model can be modified by
incorporating logistic-type functions to make fecundity and mortality density-
dependent (Usher, 1972).
48
-------
There are other mathematical models that represent both terrestrial and
aquatic species. Different models are used depending on the type of population
being studied and the amount of data available on parameters, such as age
structure, size composition, and mortality rates. A brief classification of other
analytical methods of evaluating population interactions is given in Table 2.1.
These are covered in greater detail in a separate document, "Ecological Model
Selection Criteria" (TRI, 1988b).
The logistic equation is an adequate description of population growth of
organisms with simple life cycles in constant laboratory environments, but
populations with more complex life cycles and those in natural environments
seldom follow logistic growth. In particular, the stable carrying capacity
predicted by the logistic equation is not achieved in natural populations; in
reality, population numbers fluctuate (Krebs, 1985). As a result, efforts to model
toxicant effects on population growth have had limited success. Barnthouse et al.
(1986) concluded that population theory cannot now provide models that accurately
predict long-term consequences of toxic pollutant release.
The greatest success in using population growth models in studying pollutant
effects has been in laboratory studies of microbial growth. For example,
Christensen and Nyholm (1984) fit logistic, Weibull, and probit growth models to
growth rates of the alga Selenastrum capricornutum exposed to potassium
dichromate and copper, and of the alga Scenedesmus suspicatus affected by 3,5-
dichlorophenol and potassium dichromate. The concentrations of toxicant which
gave 10 percent and 90 percent growth rate reductions (EC10 and EC90) were
49
-------
Table 2.1 Model Classification System
Model Fully Described Quantified
type defined experimen-
tally
Partially
simpli-
fied and
quanti-
fied
Economic Objectives
criteria
quanti-
fied
STATISTICAL X
Information
from data
DIORISTIC
Distinguish
component
planning
research and
teaching
holistic
description
COMPONENT
Complete
quanti-
fication of
system.
Transfer
functions
experimen-
tally
determined
STRATEGIC
X
X
Analysis
or simu-
lation of
complex
systems
MANAGEMENT
Southwood (1978)
Decisions on
management
of complex
systems
50
-------
lowest for the Weibull model. Thus, the Weibull model was most sensitive in
detecting the EC 10, which is often considered a threshold level in environmental
risk assessment.
2.1.4 Other Population Endpoints
Although population growth models, at their current state of development, do
not appear to be useful for determining effects of pollutants on natural
populations, certain components of the models and other population characteristics
can be useful endpoints. As discussed earlier, birth and death rates are potential
measures of pollutant effects. Other population characteristics that are functions
of age-specific birth and mortality rates, such as population size, growth rates,
distribution, age structure, or genetic composition, are also potential
ecotoxicological endpoints.
Alterations in species abundance are easily observed and are clearly related
to the health of the population. A decline in population number is often an early
indicator of population stress. Distribution alterations may also be very apparent.
For example, populations may be most susceptible to pollutant stress near the
limits of their distribution and in environments in which they are only marginally
suited. Therefore, changes in spatial distribution may be an indicator of pollutant
stress (Sheehan, personal communication).
Effects of pollutants on per capita growth rates (r) in laboratory populations
have been determined from age-specific birth and mortality rates. For example,
Gentile et al., (1982) studied effects of mercury and nickel on population growth
51
-------
rates of the marine mysid, Mysidopsis bahia. Metal concentrations at which per
capita growth rates were reduced to zero were closely correlated with other
estimates of chronic toxicity.
The age structure of a population, which is determined by mortality and
natality rates in various age groups, can also be used as a population-level
endpoint. For example, acidification of lake waters often causes failures of
reproduction in fish populations, resulting in a shift in age and size structure of
the population to one consisting of older and larger fish. Despite the absence of
recruitment, large populations of long-lived fish may persist for several years
(Schofield, 1976; Schindler et al., 1985). So monitoring of age structure can
provide an early indicator of pollutant effects on populations.
Finally, pollutants may have effects on the genetic composition of
populations. Elimination of all but resistant genotypes may reduce genetic
variability within a population, which may impair the ability of the population to
respond to other naturally occurring stresses. As a possible example of this,
heavy-metal-tolerant genotypes of plants are frequently competitively inferior to
nontolerant genotypes on uncontaminated soils (Antonovics et al., 1971).
2.2 LIFE-HISTORY STRATEGY
Detailed knowledge of the life-history strategy of a species is essential to
the evaluation of the effects of chronic pollution on that population. Life-history
traits, including fecundity, growth and development, age at maturity, parental
care, and longevity of an organism, have some genetic basis and hence are subject
52
-------
to natural selection. Natural selection adjusts these traits to maximize the fitness
of each organism (Ricklefs, 1973).
The two most important factors in determination of a life-history strategy
are age-specific birth and survivorship (or death) rates within a population. As
described earlier, these factors are the driving variables in many population
models.
K- versus r-selection is a life-history theory describing evolutionary
responses to environmental variability and patterns of mortality. The terms r-
selection and K-selection come from the exponential and logistic representations
of population growth; r-selected species are characterized by high population
growth rates. They are dominant in early-succession ecosystems and in systems
following disturbance (environments in which density-independent mortality is
high) because they have the ability to disperse rapidly to unoccupied habitats and
increase population size rapidly. This is typical of invertebrates and many fishes.
K-selected species, on the other hand, are characterized by low mortality and
high competitive ability. They are favored in late-successional-stage ecosystems
where mortality is density dependent, abundance is near carrying capacity, and
abundance is largely determined by biological interactions (competition, predation).
Large mammals and trees are typical K-selected species.
Knowledge of a population's life-history strategy is extremely important
when selecting appropriate indicators of stress in ecological risk assessment. All
of the life-history parameters are interrelated; a change in any one of these
53
-------
factors may cause a complementary change in another. Relationships among life-
history characteristics are discussed below.
2.2.1 Parental Care and Fecundity
Intense parental care increases the survival rate of offspring (Ricklefs, 1973).
The more protection and feeding that the young receive, the greater their
survivorship. Any disruption of the ability of a parent to care for its young may
have an adverse effect on the survival of the young. Examples of disruptions of
parental behavior by pollutants are discussed in Chapter 1.
The intensity of parental care is inversely related to fecundity (Ricklefs,
1973). Lack (1954) proposed that the ability of parents to feed the young limits
brood size and that the average brood size maximizes the number of surviving
offspring. One experimental study has shown that offspring growth decreases as
brood size is increased. However, contrary to Lack's hypothesis, offspring
survivorship did not decrease as brood size was increased. The number of
surviving offspring per brood was maximized for larger broods than the average
naturally occurring brood size (Nur, 1984).
2.2.2 Parental Care and Growth/Development
Species that provide a high degree of parental care generally have broods
that are altricial (initially underdeveloped) (Ricklefs, 1973). Conversely, species
that provide little parental care generally have broods that are precocial (capable
of providing for themselves). A high degree of parental care permits the young
54
-------
to expend most of their energy on rapid growth and development. Any pollutant
that reduces the ability to care for altricial young can reduce their survival.
2.2.3 Growth and Age at Maturity
Many species, including certain invertebrates, fish, reptiles, amphibians, and
plants, continue to grow throughout life. In such species the fecundity rate is
directly proportional to the size of the organism (Ricklefs, 1973). Reproduction
requires energy and nutrients that otherwise could be allocated to growth.
Fecundity early in life reduces organism growth, and this can reduce future
fecundity. Survivorship patterns may determine the life-history strategy favored
by natural selection (McNaughton and Wolf, 1979). In organisms with higher
survivorship, natural selection may favor expenditure of energy on growth to
increase total fecundity. In organisms with low survivorship, natural selection
will favor early reproduction and high expenditure of energy on reproduction.
These are the reproductive strategies expected in K-selected and r-selected
species, respectively.
2.2.4 Reproduction and Life Expectancy
Larger litter size appears to reduce the life expectancy of females (and
nurturing males). Murray (1979) cites eight different studies of both vertebrates
and invertebrates in which virgin females were found to live longer than
reproducing females. He also cites several studies in which population mortality
increases during the reproductive season.
55
-------
2.2.5 Longevity and Body Size
Body size is strongly correlated with animal longevity. Larger mammals tend
toward longer life and smaller individual broods; they also exhibit extended
interbirth periods (Eisenberg, 1980). Diet also has an effect on longevity for
large animals; large herbivores tend to live longer than large carnivores. As body
size increases, mortality rate decreases due to a lack of predators large enough to
capture such prey, an increased ability to find food, and the increased buffering
against microclimatic changes that results from heavier body weight (Ricklefs,
1973).
2.2.6 Compensatory Mechanisms of Life-History Strategy
Stress to a population does not always cause a change in population numbers.
Populations have means of compensating for change just as ecosystems do. At
the ecosystem level, compensation may be in the form of replacement species; at
the population level, compensation occurs through changes in life-history strategy.
Such compensatory mechanisms may be very important in the success of a
population subjected to pollutants or other stresses. They should be considered
when assembling an ecological risk assessment.
Jensen and Marshall (1983) noted population-level compensation mechanisms
in a population of cladocerans exposed to metal contaminants. Population birth
rates of Daphnia galeata mendolae exposed to cadmium increased to compensate
for increased death rates. This relationship has also been observed in other
species, and life-history theory predicts that females are under selective pressure
56
-------
to produce more offspring at reduced population levels (Fowler, 1981). Decreased
availability of food or parental care for the young can cause phenotypic changes,
including asynchronous hatching (in birds), cannibalism of siblings, and selective
starvation of the weakest in the litter (McNaughton and Wolf, 1979).
Knowledge of a population's life-history strategy is vital to risk assessment
of the ecosystem (Sheehan, 1984a). As life-history traits are interrelated, any
pollutant stress affecting one trait may affect other aspects of life-history. For
example, a chemical that reduces reproduction may result in a longer life
expectancy for particular individuals; however, in the long run, it can lead to
extinction. Resources that would normally be devoted to producing and caring for
young can be used by the organism to enhance growth and survival. Thus,
compensatory mechanisms may ameliorate contaminant effects. However, if a
pollutant stress exceeds the compensatory capacities of the organisms, the
population will decline or disappear. In addition, energy expended on
compensatory responses may reduce a population's ability to respond to other
naturally occurring environmental stresses.
2.2.7 Life Histories and Response to Contaminants
There seem to be no universally applicable generalizations concerning
sensitivities of different taxa to pollutants (Sheehan, 1984a). However, life-
history strategies of organisms may be good predictors of their response to
pollutants. Pollution effects can be loosely separated into two categories:
intermittent acute toxic effects and chronic effects (alternatively termed
disturbance and stress, Gray, 1979). During recovery from intermittent toxic
57
-------
events, the community would be expected to be dominated by r-selected species.
This is because their entire life-history strategy is adapted to colonization and
growth in disturbed habitats: high dispersal, rapid growth and reproduction, and
short generation times. As recovery proceeds, these species are gradually out-
competed and replaced by K-selected species. An example is the polychaete
Capilella capitata, which develops large, transient populations following pollution
or other disturbances (Gray, 1979).
Organisms with stress-tolerant life histories would be expected to be least
affected by chronic pollution. Stress-tolerant organisms share some of the life-
history characteristics of K-selected species: low dispersal, low reproduction, and
long life spans. However, unlike K-selected species, stress-tolerant organisms
allocate more resources to maintenance of productive tissue in very stressful
environments. These characteristics would predispose organisms with stress-
tolerant life-histories to success during chronic pollution concentrations that may
be sublethal to them but lethal to r- or K-selected organisms (Grime, 1977; Gray,
1979).
2.3 CONCLUSION
Moriarty (1983) aptly states that "no population remains constant in number
forever." Population sizes increase and decrease in response to many factors,
such as weather, availability of food, and competition. Pollution is only one
factor that can influence population size. Even when an effect is directly
traceable to a particular pollutant, populations may have mechanisms to
compensate for the stress. Stress-tolerant life histories may be pre-adapted to
58
-------
chronic pollutant stress, while r-selected organisms appear to be pre-adapted for
recolonization during recovery from intermittent acute toxic disturbances.
However, life histories of organisms are adapted to a suite of physical, chemical,
and biological characteristics of the environment. Compensatory responses to a
pollutant may reduce a population's ability to cope with other environmental
regulating factors.
While population-level responses have been used for decades to determine the
effects of pollution, it is important to realize that an in-depth understanding of
the physical and biological processes regulating populations is necessary before
effects of contaminants can be predicted. For the most part, current population
models cannot effectively deal with all of the physical and biological processes
affecting populations. As an endpoint, population dynamics are extremely complex
and absolutely species-specific. There are no formal methodologies for using
changes in population dynamics as a risk assessment tool; however, knowledge of
the dynamics of a population is necessary to any assessment of stress to a
population.
2.4 METHODS FOR SELECTING APPROPRIATE POPULATIONS TO MONITOR
Evaluation of individual species forms an important part of any ecosystem
study. Individual species effects can provide readily measurable indicators of
ecosystem health. They pinpoint some of the factors responsible for changes in
population density. However, individual species do not exist in ecological
vacuums; interspecific interactions, such as competition or predation, have strong
59
-------
effects on population dynamics. These interspecies interactions may be very
sensitive to effects of pollutants.
An important initial step in assessing population-level ecotoxicological
endpoints is the selection of the most appropriate populations for study. The
methods for assessing population-level changes range from individual evaluation of
all species to the evaluation of certain representative species of the community.
There are benefits to using each of these methods. They are discussed in detail
below.
2.4.1 Total Population
The total-population approach involves evaluation of the total number of
organisms present in the area of concern. Measurements are made using one of
three major field sampling techniques. The first is the nearest neighbor method.
In this method, a geographic point is randomly selected and then is searched in
ever widening circles until an organism is found. The searching continues until a
second member of the same species is found. The density of individuals per unit
area is determined by using simple mathematical expressions based on the distance
between the nearest neighbors (Southwood, 1978). This method and several of
its variations have been used with some success in making rough estimates of
plant and insect populations present in very small areas (such as the organisms
present on one plant). This method is most useful for immobile, sedentary, or
territorial organisms; it is less useful for mobile animals (Tanner, 1978).
60
-------
More accurate measures of determining a total population include marking
techniques or taking absolute samples of a unit area. Techniques for marking
animals, particularly mark and recapture, have been used on large or slow animals
that can be readily marked, such as snails and many vertebrates. An advantage
of mark/recapture techniques is that they can provide estimates of birth and
death rates in addition to population size.
The total-population approach is quite comprehensive and is useful for
population studies. This approach samples quadrats that are representative of the
system of interest. Sedentary organisms, such as plants or barnacles, may be
censused nondestructively. But in many other cases, samples must be taken of
the organisms. present in air, soil and litter, water, or sediments (if appropriate).
This is a complicated and labor-intensive process requiring biota sampling
equipment, such as suction traps for air samples, corers for soil samples, and nets
and traps for water and land samples (Southwood, 1978). The organisms in the
soil cores must be extracted, and the organisms from all of the samples must be
identified and counted.
Single measurements of abundance obtained by one of these methods do not
provide much interpretable information, but repeated measures over time can be
used to estimate important population parameters, such as rates of recruitment,
mortality, and population growth. However, estimation of population parameters
does not necessarily identify the causes for many of the changes in populations,
which may include interactions with other members of the community or with the
physical environment.
61
-------
2.4.2 Species Dominance
The concept of species dominance relates to the importance of an individual
species relative to others in the ecosystem. The dominant species is variously
defined as the one with the greatest abundance, the greatest productivity, the
largest individuals, the most occupied space, or the greatest impact on community
dynamics (Clapham, 1973; Sheehan, 1984b).
Shifts in dominance may be used as measures of pollutant effects on
community structure. For example, in a bay stressed by pulp and paper effluent,
yellow perch replaced white sucker as the dominant fish species in the area near
the mill discharge. The shift was attributed to pollutant-induced decreases in the
abundance of benthic invertebrates, the primary food source of the white sucker
(Kelso, 1977). Dominance shifts were also noted in a study of macroinvertebrate
communities of streams stressed by heavy metal pollution (Winner et al., 1975
cited in Sheehan, 1984b). In this particular case, reductions in the numbers of
sensitive taxa leads to the dominance of metal-tolerant chironomid species. A
number of abundance-based dominance indices have been devised to provide more
rigorous methods of assessing this measure. These, however, have not been
widely used in pollution studies (Sheehan, 1984b).
Clapham (1973) suggests that there are a number of problems associated with
the practical use of the dominance concept. For instance, the most commonly
used measure of dominance is species abundance. However, a large-bodied species
may be more influential than a far more abundant smaller species. Similar
problems are associated with other measures of dominance. Dominance may
62
-------
reflect lack of competition, i.e., pollutant-sensitive organisms may die or migrate
away from the study area. Invertebrates are very opportunistic in this way—with
no competition, there are dramatic increases in numbers (Folmar, personal
communication). Another consideration is that dominance within a single
community may be best assessed using different measures at different trophic
levels. Thus, a dominant plant species might be best defined based on ground
cover, while a dominant carnivore species might be best defined based on biomass
or net productivity. Although dominance implies a position of advantage in
community interactions, it does not necessarily correlate with tolerance to
chemical stress. Abundance (dominance) under pollution stress often depends on
the opportunities and life-history strategy of the species (see section on Life
Histories and Response to Contaminants). Nevertheless, in some ecosystems
dominance can supply a clear simplifying concept that is useful for assessing
pollutant impact (Clapham, 1973).
2.4.3 Indicator Species
The indicator-species method of evaluating ecosystem stress has been used
for most of this century. The method is based on the notion that the continued
presence of certain species indicates acceptable environmental conditions, whereas
their absence would indicate the lack of appropriate environmental conditions
(Sheehan, 1984b). The indicator method is useful as an evaluative tool, and it can
be employed to determine if damage to a system has already occurred.
Sheehan (1984b) points out problems in the general definition of indicator
species. For example, the absence of a species in a system does not always occur
63
-------
as a result of poor environmental conditions. Absence can also result from a lack
of dispersal (e.g., changes in wind patterns resulting in altered seed distribution
in the study area), sampling of an inappropriate seasonal lifestage (particularly for
invertebrates), or biotic influences such as introduction of a new predator. It is
clear that merely observing the presence or absence of a species does not prove a
cause-and-effect relationship between the disturbance being investigated and the
occurrence of the species.
The indicator concept has been modified somewhat to recognize the ability
of indicators to reflect the subtle as well as the gross effects of pollution
(Sheehan, 1984b). Effects that occur prior to the total disappearance of a
population from a system must be measured. To accomplish this, the species
selected must evoke characteristic toxicological responses.
One of the problems associated with the indicator method involves the
selection of the indicator. Because there are no consistently reliable algorithms
for selecting appropriate species, the determination of an indicator is subjective
(Levin et al., 1984). The methodology for determining appropriate indicators is
based on two different techniques. The first is subjective selection of the
organism (Levin et al., 1984; Limburg et al., 1984), involving identification of
species that possess characteristics such as:
o Amenability to laboratory handling and testing;
o Characteristic toxicological responses;
o Economic, recreational, and/or ecological importance;
o Universal distribution;
64
-------
o Susceptibility to the toxicants; and
o Availability and abundance.
The second basis for selecting species is mathematical. Statistical analysis
can be performed based on the log-normal distribution of individuals among
species. A method has been proposed that selects species of moderate abundance
(16 to 63 individuals per species) to be appropriate as indicators (Gray and
Pearson, 1982; Pearson et al., 1983). A problem with this method arises because
it does not account for practical matters, such as exhibiting a toxicological
response to a contaminant or being easy to handle in a laboratory.
In conclusion, the indicator method can be a useful endpoint for ecological
risk assessment in both aquatic and terrestrial systems; however, it must be used
with caution. The selection of the particular species to monitor is the key. The
indicator species must be chosen by scientists with a strong expertise concerning
the particular environment being studied.
2.4.4 Keystone Species
Keystone species provide a valuable endpoint in ecological risk assessment.
They can be a means of bridging the gap between single-species toxicity testing
and community level evaluations. Single-species tests are quick and easy, but
their use can be limited if the species is a laboratory strain that may not
represent the sensitivity of species present in the actual environment under
examination. Community level evaluations are often complicated, costly, and time-
consuming. However, evaluations of keystone species offer the advantages of a
65
-------
single-species approach, which yields insights as to whether a stress will affect
community level endpoints.
"A few species by their size, form, abundance, or activity may exclude or
promote other macro species, may provide new niches of smaller organisms, or
may modify the physical environment in which they live." (Lewis, 1978, cited in
Bowmer et al., 1986). This is the currently held definition for keystone species.
It is a more general definition than the original, which focused on predatory
control of species densities. According to Paine (1966), keystone species are
"predators that keep the population densities of the prey below levels where
resources become limiting." This prevents a high level of competition for
resources whereupon only a few species would become dominant and species
diversity would decline. By definition, keystone species modify the community
and/or the environment. Therefore, any disturbances to the keystone species
would lead to a disturbance to the community and/or environment. The process
of keystone species selection is more objective than selection of indicators.
Furthermore, there has been experimental demonstration that keystone species
strongly influence abundances of other species in the community.
The keystone relationship is illustrated in a study of the starfish Pisaster
ochraceus and mussel Mytilus californianus conducted by Paine (1974). Paine
observed a natural population of mussels in a well-defined band. The principal
predator for the mussel larvae was the starfish Pisaster ochraceus. Most of
these predators were removed manually from the area on a monthly or twice-
monthly basis for 5 years; monitoring of the area continued for 6 years. Under
the conditions of a predator-free environment, the mussels excluded many other
66
-------
species that were in competition for the limited resource of space. Paine
concluded that removal of the starfish resulted in "the local elimination of at
least twenty-five macroscopic species that otherwise would have existed on or
immediately associated with the primary space."
Hixon and Brostoff (1983) discuss a keystone relationship that is different
from the previous examples. In the Paine study, predation by the keystone
species was found to maintain a high diversity of prey by preventing competitive
exclusion among the prey. Hixon and Brostoff observed a relationship whereby
the predator maintained a high diversity of prey by reducing the overall intensity
of predation. The predator was the territorial damselfish. Damselfish are one of
several herbivorous fish that feed on algae in tropical reefs. In the study, plates
for growing algae were placed in three different areas: one set was placed
outside the damselfish territories; one set was placed within some of the
damselfish territories; and the final set was placed in fish-exclusion cages within
the damselfish territories. Plates were sampled over the course of 1 year.
Initially, colonization of the plates was comparable in the territories and the
cages, although colonization was somewhat slower outside the territories.
Throughout the experiment, biomass and diversity remained lowest where there
were no damselfish to moderate the amount of grazing. Within the territories,
diversity was highest on the caged plates at the 6-month mark. By the 1-year
mark, diversity was greatest on the plates inside the territories, but biomass was
higher on the caged plates. This study shows that damselfish can maintain a high
diversity of prey by excluding other herbivorous fish from their territories, thus
reducing the overall intensity of predation. This study supports the "intermediate
disturbance hypothesis," which holds that diversity is maximized by an
67
-------
intermediate level of predation or other disturbance and declines at both low and
high predation intensities (Connell, 1978).
Many criticisms of this method are identical to criticisms of species-level
approaches. However, keystone species do have a demonstrable influence on
community structure. It is important to identify the types of ecosystems in which
key species exert influence. In monitoring keystone species, it is important to
note that systems have mechanisms, such as functional redundancy, that can
minimize general structural or functional changes that could take place (Levin et
al., 1984). For example, predators may switch their food choice from a toxicant-
sensitive prey species to a more resistant one without changing the food web
structure. Pollutant-sensitive species may be replaced by competitors without
changing system biomass or productivity (Sheehan, 1984c). If compensatory
mechanisms are not present, keystone species removal may have wide-ranging
indirect effects on community structure that would not be evident from species-
level approaches. For example, removal of sea otters from Alaskan islands led to
increases in grazing sea urchins, decimation of kelp beds, and exposure of
shorelines to greater wave damage (Estes and Palmisano, 1974). This example
illustrates that interactions among species and between species and their
environment must be understood before effects of contaminants can be predicted.
Thus, these species-level effects may not adequately serve as indicators of whole
ecosystem-level effects and should be used in combination with other measures in
an ecosystem risk assessment.
68
-------
2.4.5 Representative and Important Species Approach
A multiple-species approach based on an examination of the Representative
and Important Species (RIS) concept was presented by Limburg et al. (1984). This
is a regulatory method that is used in the Clean Water Act and the National
Environmental Policy Act. It combines several of the species-level effect
measures (like keystone organisms or indicator species) with practicality.
Limburg et al. (1984) recommended general criteria for selecting a RIS. The
species should have some of the following characteristics:
o Representative of the ecological community most exposed to an impact;
o Direct commercial, recreational, or aesthetic value;
o Critical to maintaining the integrity of an ecological community whether
by structure (e.g., coral reefs) or function (e.g., primary producers,
keystone predators);
o Characteristically predominant in a habitat either by virtue of numbers or
biomass (e.g., Spartina in a salt marsh);
o Particularly sensitive to the perturbation of concern; usually the
protection of the most sensitive species provides a "safety window" for
the rest of the organisms in the area; and
o Able to sequester chemical substances with known or suspected toxic
effects, either because of physiology or life history.
Several states have modified these criteria for particular locations. For
example, Maryland Regulation 10.50.01.13 defines the following criteria for
developing RIS lists for estuaries and other coastal zones:
69
-------
o Consider only those species normally present in the local salinity regime;
o Determine the spatial and temporal distribution of resident and migratory
species with respect to their various life stages; and
o Select at least one fin fish, mollusc, and arthropod, and one other species
for intensive study, using as selection criteria species abundance,
commercial or recreational importance, and sensitivity per life stage to
facility operations. In fresh water studies, one insect species shall also
be selected. Additional R1S may be selected from a list of important
taxa, which includes three species of waterfowl, one mammal, eight
molluscs, one crab, three insects, and twenty-eight species of fish.
This approach to evaluating ecosystem health is a valuable one. It can be
made site-specific, and it solves some of the main criticisms of laboratory
toxicological testing. Examination of a set of species provides some information
on interspecies sensitivity. The RIS species selected are members of the natural
system, and looking at species like keystone predators predicts some of the
effects on ecological processes and structure (Levin et al., 1984). One objection
to the uses to which the RIS concept has been applied is that little attention has
been paid to microorganisms, (such as bacteria, fungi, microscopic algae, or
microinvertebrates), which may be extremely important in ecosystem structure and
function.
2.5 CONCLUSION
Species-level methods are valuable indicators of ecosystem effects, but they
do have drawbacks. They do not accurately account for changes in community or
ecosystem structure that are caused by interactions between species or with other
environmental factors. Also, changes in population levels are not necessarily
correlated with adverse effects on the system due to buffers (replacement species)
70
-------
that minimize disruption of system function. Substitutions can occur without
causing noticeable modification in ecosystem processes (Levin et al., 1984).
Single-species measures provide methods of monitoring changes that are
easier to define and measure than measurements at the community level. When
used in conjunction with community-level endpoints, populations can provide
necessary information on ecosystem health. For example, keystone species modify
the community, and thus can provide insights into changes at the community
level. Similarly the representative and important species approach provides a
more integrative indicator of ecosystem stress. It selects species for examination
based on practical, economic, and scientific considerations. Used properly, this
approach (with particular emphasis on keystone species) is extremely valuable for
ecological risk assessment.
2.6 SPECIES INTERACTIONS
Species interactions are defined as relationships between two or more species
(Odum, 1971). Because communities are composed of groups of interacting
populations, impairment of the ability of a population to function normally is
likely to lead to adverse impacts on the structure and function of the community.
Two types of species interactions have been used extensively as endpoints for
determining stress to an ecosystem; these are alterations to predator-prey
interactions and competition (Sheehan, 1984a).
71
-------
2.6.1 Predator-Prey Interactions
Predator-prey interactions have been studied extensively, particularly with
regard to pesticide effects. Complex behavioral patterns may be utilized in
avoidance of predators. For example, schooling, surface swimming, formation of a
circular aggregation, and hiding under natural vegetation are used by fathead
minnows to avoid their principal predator, the largemouth bass (Sullivan and
Atchison, 1978). Any alteration in prey vulnerability to predation may strongly
affect population dynamics of both prey and predator.
The usefulness of predator-prey interactions as a measure of ecosystem
stress was illustrated in a study conducted by Tagatz (1976). He observed1 the
effect of a pesticide (Mirex) on predation by pinfish on grass shrimp. Sublethal
concentrations of Mirex significantly increased the susceptibility of grass shrimp
to pinfish predation. This type of behavior alteration was also observed by Kania
and O'Hara (1974). Sublethal mercury concentrations of 0.1, 0.05, and 0.01 ppm
reduced the ability of mosquitofish to avoid predation by largemouth bass. In this
situation, the effect on the mosquitofish had a potentially more serious effect on
the ecosystem, as large amounts of mercury were ingested from contaminated prey
by largemouth bass. It is likely that bass prey selectively upon the most heavily
contaminated mosquitofish, as they are most likely to have reduced capacity to
avoid capture. This would lead to increased bioaccumulation and biomagnification
of mercury up the food chain.
Sylvester (1972) examined the effects of thermal stress on predator
avoidance in sockeye salmon. In a laboratory study, tanks containing coho salmon
72
-------
(predator) and sockeye salmon fry (prey) were monitored. Elevated temperatures
significantly increased predation rates on sockeye salmon fry. Industrial processes
using cooling water that is discharged into rivers and streams may thus have a
high impact on predator-prey relations.
Toxic stress can also have an effect on the selection of prey species. Farr
(1978) exposed three aquatic species to sublethal doses of methyl parathion under
laboratory conditions. The predator was gulf killifish and the prey were grass
shrimp and sheepshead minnow. Only the grass shrimp exhibited any toxic effects
from the pesticide. As a result, methyl parathion increased the preference of
killifish for grass shrimp. Increased predation on grass shrimp could affect the
ecosystem-level processes of decomposition and nutrient recycling. Grass shrimp
are important in the conversion of organic detritus to forms more readily usable
by other organisms. Replacement species may functionally compensate for the
reduction of grass shrimp, but many of these species are also susceptible to
methyl parathion poisoning.
Another way that pollutants can alter community structure is through
predator removal. Hurlbert (1975) noted that pesticide-induced density increases
of prey populations can occur due to either removal of a predator or increased
food supply. He cites several cases where rotenone and other pesticides
eradicated fish populations. After removal of fish, reduced predation allowed
many types of invertebrates and frogs to substantially increase in number.
Similarly, pesticide-induced depletions of crustacean zooplankton have led to algal
blooms and increases in abundance of herbivorous rotifers (both prey and
competitors with crustaceans).
73
-------
Most predator-prey studies are performed using aquatic organisms, largely
due to the ease of laboratory handling of fish, the intensity of predator-prey
interactions in aquatic systems, and the preponderance of data on the behavioral
effects of contaminants such as pesticides on fish. There are several methods of
performing these types of studies. Microcosms are a logical choice if adequate
time and space are available. A common laboratory technique used to detect the
effects of stress on aquatic predator-prey interactions is described by Goodyear
(1972). He recommends the use of two controls: one with' the toxicant present
but without the predator, the other without the toxicant but with the presence of
the predator. In general, the tank should functionally represent the natural
habitat (e.g., some predator species require hiding areas from which to pounce,
and prey generally require a refuge). Goodyear further notes that the distractive
presence of investigators can affect the behavior of certain species; therefore,
knowledge of the test species is critical prior to beginning an investigation.
2.6.2 Interspecies Competition
A less commonly used measure of stress to the ecosystem is alteration of
competition. Interspecific competition is defined as two or more species making
simultaneous demands on one resource that is present in limited amounts.
Interspecific competition may result in either the emergence of one dominant
population or coexistence of multiple species (McNaughton and Wolf, 1979;
Moriarty, 1983).
Competition studies have not been used as extensively as predation studies
(Sheehan, 1984a); however, they can serve as an endpoint. For example, toxic
74
-------
stress can alter competition intensity in a variety of ways. As discussed in the
preceding section, Hurlbert (1975) noted that pesticide-induced mortality of
predator populations can result in increased abundance of prey. This reduction of
the predator population could actually have an adverse effect on competitively
inferior prey species, as release from predation may lead to competitive exclusion.
Even if local extinction of one or more populations does not occur, release from
predation will likely alter community function and structure (Moriarity, 1983;
Sheehan, 1984a). A possible example of community effects of reductions of a
dominant competitor has been illustrated by Schmidt (1986). Coyotes are
territorial and are the dominant member of a guild of large canids that prey on
domestic livestock. Population reduction aimed at coyotes had serious effects on
the structure of the guild. As the coyote population declined, other large
carnivore populations increased. There is insufficient information about whether
the shift in guild structure affected prey populations; however, it does appear
that the reduction of coyotes has resulted in an increase in the competitively
inferior smaller canids.
Pollutant stress could alter interspecific competition by increasing the
amount of a limiting resource. For example, phosphates discharged into aquatic
systems often result in algal blooms (Reynolds and Walsby, 1975; Barica and Mur,
1979). Under these conditions, blue-green algae often replace other species,
apparently by competitive elimination (Smith, 1983).
75
-------
2.7 CONCLUSION
These examples show how pollutant stress can affect interactions among
populations. Examination of population interactions is an integral part of
determining the effects of toxicants on ecosystems. Communities are groups of
interacting populations, and serious changes in the functional capabilities of
populations are likely to lead to adverse impacts on the community structure and
function. These effects of pollutants on behavioral interactions among organisms
cannot be determined from single species toxicity tests. The limitations of single
species endpoints suggest that assessment of effects of pollutants must incorporate
population, community, and ecosystem level ecotoxicological endpoints.
76
-------
CHAPTER 3
ECOSYSTEM-LEVEL ENDPOINTS
3.0 INTRODUCTION
An ecosystem is a highly complex structure composed of a diversity of
interacting biotic and abiotic components (Odum, 1971). The stress responses of
the system reflect, in part, the integrated responses of these components. In an
ecosystem risk assessment, effects on ecosystem components can be evaluated
using organism- or population-level measures. Such measures may provide
information on dominant, keystone, or indicator species; these may then serve as
indicator signals of ecosystem-level stress. Organism-level measures can also be
used to evaluate the status of economically or aesthetically important species
(Sheehan, 1984a; Levin et al., 1984). Organism- or population-level effects,
however, are not always indicative of ecosystem-level changes, because the system
often responds in ways that are independent of the responses of specific
components (Sheehan, 1984c). For example, individual populations can fluctuate or
die off, yet the system, through a variety of internal buffering mechanisms--
predator flexibility, replacement species—can survive. Thus, monitoring of only
organism- or population-level changes may not provide an adequate picture of the
responses of the system.
Ecosystem-level endpoints, such as primary production and nutrient cycling,
provide important indicators of ecosystem stress. Although these endpoints have
not received much attention as measures of pollution-induced stress, the literature
is now expanding concerning the use of endpoint changes in both ecosystem
77
-------
function and structure. The ecosystem-level work done to date has not been as
comprehensive as that done at the individual organism level. The research does,
however, identify tested, measurable, ecosystem-level endpoints such as primary
production, leaching of nutrients from soil, species richness, and similarity that
provide quantitative data for risk assessment.
3.1 ECOSYSTEM FUNCTION
Functional components of a system include the processes involved in the
movement and transformation of chemicals and energy. These processes in turn
provide the basic support for the system's structural components. As a result,
the maintenance of population and community structure is dependent on the
functional integrity of the ecosystem (Sheehan, 1984c). Societal goals for
environmental protection (which are embodied in the statutes mandating ecological
risk assessment) focus on protection of population and ecosystem structure, such
as maintenance of species diversity or protection of endangered species. The
dependence of these structural components on ecosystem functional processes
make functional endpoints important elements of ecological risk assessment.
The processes involved in chemical and energy flow are revealing endpoints
for an ecosystem risk assessment. Measurable processes include primary
production, photosynthetic rates, respiration, decomposition, nutrient levels, and
nutrient leaching from soil. These processes are often highly sensitive to low
levels of ecosystem stress and are thus valuable indicators of ecosystem stress.
78
-------
Despite their importance, functional endpoints have not been widely used in
ecosystem pollution studies. There is, therefore, a lack of extensive baseline data
establishing levels at which changes in ecosystem function will adversely affect
ecosystem structure. More work using functional endpoints to monitor pollutant
effects on ecosystems must be done before these endpoints can provide fully
predictive measurements. In the following section, the current state of
knowledge concerning these functional endpoints and methods of their
measurement will be reviewed.
3.2 ENERGY FLOW
3.2.1 Primary Productivity
Production, or the "process of energy input and storage in an ecosystem"
(McNaughton and Wolf, 1979), is one of the most tested endpoints for evaluating
ecosystem stress in both aquatic and terrestrial systems. Studies usually focus on
either gross or net primary production. Gross primary production has been
variously defined as the "energy fixed in photosynthesis" (Krebs, 1985) or the
"total assimilation of organic matter by the plant community" (Odum, 1971). Net
primary production is defined as gross primary production less the energy lost
through respiration (McNaughton and Wolf, 1979). Net primary production is thus
the total amount of photosynthetic input that could be available to other trophic
levels (Beadle and Long, 1985).
Photosynthesis is the mechanism through which solar energy is transformed
into biomass (Sheehan, 1984c). It is, therefore, the ultimate source of all
79
-------
production in an ecosystem, and the various measures of photosynthesis—CO2
uptake, O2 production, 14C uptake, chlorophyll levels—can serve as indicators of
primary production.
As the energy base for the ecosystem, primary productivity is a fundamental
factor in ecosystem health. It is critical to both autotrophic maintenance and
heterotrophic development, and it determines the amount of living tissue that an
ecosystem can support (Woodwell, 1970). Measurements of change in primary
productivity provide important information about ecosystem health.
One of the measures of the rate of photosynthesis is often used as an
endpoint of ecosystem toxicity. For example, SO2 and a number of other
phytotoxic pollutants act to suppress photosynthesis rates. This suppression has
been investigated in lichens and a variety of other producers. In one such study
conducted by Bennett and Hill (1974), alterations in the rate of photosynthesis
were measured as changes in CO2 uptake. These investigators observed a
reversible suppression in photosynthesis in crop plants chronically exposed to
pollutants such as CI2, O3, SO2, and NO2. In another study, Carlson (1979)
observed the effects of SO2 and O3 on the rate of photosynthesis. Measured as
CO2 uptake, photosynthesis was significantly reduced in exposed maple and ash
leaves. Other studies using photosynthesis as an endpoint are described in
Sheehan (1984c).
Although primary productivity is clearly an important measure of ecosystem
function, its usefulness as an endpoint for ecological risk assessment is dependent
on the nature of both the pollutant and ecosystem being studied. Effects of
80
-------
different pollutants on primary productivity in different ecosystems are highly
variable. SO2 and acid rain, for example, consistently reduce productivity in
terrestrial systems, and yet they may either decrease or increase productivity in
different aquatic systems. Similarly, low levels of heavy metals have been shown
to produce rapid, readily measurable changes in net primary productivity in
aquatic systems, while intermediate levels of metal pollution produce potential, but
difficult to measure, changes in productivity of forest ecosystems (Sheehan,
1984c). Primary productivity therefore serves as an appropriate measure of
ecosystem stress in some situations but not others.
When productivity is used as a measure, changes should be evaluated in
terms of the status of the system being monitored. Natural systems normally
undergo fluctuations in productivity as a result of changes in grazing and in
available resources such as CO2, nitrogen, phosphorous, and light. If resources
flow into and out of the system at the same rate, the system is considered to be
in a steady state. A system may naturally shift from one steady state, and
therefore one level of productivity, to another, depending on resource availability
(Odum, 1971). Thus, any measure of pollutant effects on a system's productivity
can only be fully assessed with prior knowledge of the steady-state status of the
system. In the absence of information on steady-state productivity levels, it may
be possible to estimate them by using appropriate models or by comparison to
similar, unimpacted systems. If ecosystems are not in a steady state, however,
even prior knowledge of system productivity will not be sufficient to assess
pollutant impact (Hurlbert, 1984; Stewart-Oaten et al., 1986).
81
-------
Productivity changes must also be assessed in terms of the seasonal and
successional stages of an ecosystem. Ecosystems in different stages of seasonal
development demonstrate different sensitivities to pollutants. For example,
terrestrial systems tend to be more sensitive to pollutant-induced stress during
the season when plants are in the germinal stages of growth (Sheehan, 1984c).
Similarly, pollutant-induced changes in nutrient flux are most likely to affect
plant productivity in late-succession terrestrial systems, where nutrient cycles are
stabilized (Vitousek and Reiners, 1975, cited in Sheehan, 1984c).
Another factor that must be evaluated is the transience of productivity
changes. Long-term changes of large magnitude are clearly suggestive of
ecosystem damage. Transient changes, however, may be suggestive of ecosystem
stability, when in fact other kinds of functional and structural changes are taking
place. Studies on the effects of hydrocarbons in lake systems have shown that
while algal production drops initially in response to pollutants, the algae show
rapid recovery and growth, returning to normal productivity levels, even as
nutrient cycles and other parts of the system are breaking down (Hobbie, personal
communication).
Evidence from eutrophication studies suggests that recovery of algal
production may involve a change in species composition, which can then affect
other components of the ecosystem. Eutrophication commonly causes increases in
primary production that are accompanied by a shift in algal species composition to
dominance by blue-green algae (Reynolds and Walsby, 1975; Barica and Mur, 1979).
Because blue-green algae are often toxic or of poor nutritional value to algal
82
-------
consumers (Fulton and Paerl, 1987), the transfer of primary production to higher
trophic levels may be reduced following eutrophication.
Pesticides, such as DDT, also induce transient changes in productivity.
Pesticides cause an initial increase in crop productivity due to a reduction in pest
species numbers; after a number of years, however, productivity drops, due to
pesticide effects on other components of the system (Sheehan, 1984c). Clearly,
changes in productivity must be evaluated with knowledge of changes that could
be taking place in other parts of the system.
3.2.2 Respiration
Respiration, defined as "any energy-yielding biotic oxidation" (Odum, 1971),
provides a measure of the rate at which organic matter is oxidized. As such, it
serves as an indicator of community metabolism (Cooper and Copeland, 1973). For
example, soil respiration can be used as an indicator of the rate of decomposition
of organic matter (Garten et al., 1985). In addition, respiration can be used with
measures of net production to determine gross production. Methods for
measuring respiration include determinations of nighttime C>2 uptake and CO2
release.
A number of studies illustrate the use of respiration as an independent
endpoint. In one study, conducted by Baddeley et al. (1973), respiration rates,
measured as O2 uptake, decreased when lichens were exposed to SO2. Cooper
and Copeland (1973) also observed a depression in respiration, measured as CO2
production, in estuarine ecosystems stressed with industrial effluent.
83
-------
3.2.3 Photosynthesis/Respiration Ratio
The photosynthesis/respiration (P/R) ratio provides an integrative measure of
ecosystem metabolism. It was originally proposed by Odum (1971) as a method for
classifying ecosystems and has since developed into a useful method of measuring
ecosystem stress. Odum suggested that, in autotrophic systems where energy is
obtained principally from the sun (McNaughton and Wolf, 1979), the rate of
photosynthesis tends to exceed the rate of community respiration. In such a
system, biomass tends to increase and therefore the P/R ratio is greater than 1.
In heterotrophic systems, energy is obtained primarily from preformed sources of
organic energy (McNaughton and Wolf, 1979) and the rate of respiration tends to
exceed the rate of photosynthesis, yielding a P/R ratio of less than 1. During
the process of succession, the P/R ratios of both autotrophic and heterotrophic
systems approach 1, and energy cost is in balance with energy fixation (Odum,
1971).
Studies have shown that the P/R ratio can be a sensitive indicator of
ecosystem stress. In mature systems, the presence of toxicants and other
stressors will cause deviations from the predicted P/R ratio of 1 (Giddings and
Eddlemon, 1978). Such changes have been observed in a number of micro-
ecosystems. In a study conducted by Maki and Johnson (1976), changes in
measurement of O2 levels were used to determine primary production and
community respiration in 3-trifluoromethyl-r-nitrophenol (TFM) poisoned model
stream communities. Exposure to TFM was found to cause a significant
depression in gross primary production and P/R ratios. The P/R ratios were
found to be very sensitive indicators of TFM influence on the community.
84
-------
Similarly, Giddings and Eddlemon (1978) found that P/R ratios, determined
from changes in 02 concentrations, declined in arsenic-stressed pond microcosms.
The ratio decline was correlated with the arsenic concentrations of the individual
systems. This study indicated that the P/R ratio could be useful when monitoring
changes in ecosystems at varying distances from the pollutant source.
Although it provides a useful endpoint, the P/R ratio should be used with
some caution. In situations where a toxic chemical reduces both primary
production and respiration, the P/R ratio may exhibit little change, even though
both the total energy base for the ecosystem and biomass have been reduced
(Sheehan, personal communication).
3.2.4 Methods of Measurement of Primary Production
A variety of methods have been used to measure primary production. These
methods all monitor some aspect of the energy-transforming photosynthetic
process, and the general photosynthetic equation.
6C02 + 12H20 673 kilocalories C6H1206 + 602 + 6H20
can be used with these methods to calculate production. Measures of gross
production additionally take into account energy lost through respiration.
In making net and gross productivity determinations, it has generally been
assumed that net production can be determined during the light period of the diel
cycle, while respiration rates for calculating gross production can be determined
85
-------
al., 1963). Thus, using nighttime respiration rates to estimate daytime, i.e., total,
plant community respiration, will probably lead to an underestimate of gross
primary production. Because actual quantitative measures of daytime respiratory
rates are lacking, however, daytime respiration is normally assumed to be equal to
nighttime respiration (Cooper and Copeland, 1973).
*4C UPTAKE: uptake is the most commonly used measure of primary
production in aquatic systems. It is currently considered the most sensitive
method of measuring aquatic primary production in ponds, lakes, and oceans. This
method was introduced by Steeman Nielson in 1952, and it provides an estimate of
the amount of carbon taken up by a system's plants and converted through
photosynthesis to organic compounds (Peterson, 1980). A simple light-dark bottle
technique is used to measure ^CC>2 uptake in aquatic systems. ^CC>2 may also
be used to measure production in terrestrial systems by introducing it into a
transparent chamber containing representative plants. The plants are then
harvested, and whole-plant levels are determined (Krebs, 1985).
Problems with methodology have been detailed by a number of
investigators. Peterson (1980) suggests that uptake normally provides no
direct estimate of respiration, as dark bottle uptake measures both active dark
uptake of CO2 and abiotic formation of labelled particulate carbon. Thus, this
method does not measure either gross or net production. Other problems
associated with this technique include:
0 Potential contamination of samples with metals or other toxic
materials from containers used during collection and incubation.
0 Potential loss of during inoculation, incubation, and sample
86
-------
o Potential contamination of samples with metals or other toxic
materials from containers used during collection and incubation.
o Potential loss of '^C during inoculation, incubation, and sample
preparation.
o Uptake and release of 14C by bacteria and zooplankton in natural
plankton communities.
o Underestimation of primary production when compared to other methods
in both oligotrophic and certain highly productive ecosystems (Peterson,
1980; Bemer et al., 1986; Fitzwater et al., 1982; Sakamoto et al., 1984).
The method has been carefully analyzed and refined over the years and
remains the best, most sensitive method currently available for measuring primary
production in lakes, ponds, and oceans (Hobbie, personal communication). It is
evident, however, that this method requires careful monitoring of sampling and
counting techniques.
O; measurements: Another way to use primary production as an endpoint in
an ecological risk assessment is through measurement of O2 production. This is a
widely used method of measuring primary production in marine and freshwater
environments. Light-dark bottle techniques, which were pioneered by Gaarder and
Gran (1927), involve a setup similar to that already described for measuring
uptake. Oxygen production in incubated samples is measured using an oxygen
meter, or by chemical analysis. Data obtained from light bottles provide a
measure of net community production (Odum, 1971). Data obtained from dark
bottles represent the normal respiratory consumption of oxygen and thus can be
added to data from light bottles to provide a measure of gross production
(Kormondy, 1969), which cannot be measured using (Peterson, 1980).
87
-------
Oxygen production has been used in a variety of ecosystem pollution studies
as an indication of changes in primary production. In one such study conducted
by Patil et al. (1985), the effects of DDT contamination in a pond ecosystem were
monitored. Primary production and respiration were measured, using the light and
dark bottle method, as changes in Oj. It was found that DDT caused decreases
in both gross production and net primary production, as well as increases in
respiration.
Oxygen production provides a good measure of aquatic primary production
because it is simple to monitor and it produces data on both production and
respiration (Peterson, 1980). While 14C is currently considered the most sensitive
measure, recent studies suggest that oxygen may provide a more accurate measure
of primary productivity than 14C in certain highly productive ecosystems, e.g.,
hypertrophic ponds (Berner et al., 1986). Oxygen measurements are also useful
for estimating production in flowing water systems (Hobbie, personal
communication).
COj assimilation: Monitoring of CO2 assimilation using an infrared gas
analyzer is a useful method of measuring photosynthesis, respiration, and primary
production in terrestrial ecosystems. Measurements can be made with great speed
and precision to + 1 ppm, a level of accuracy not achievable for field measures of
O2 (Woodwell and Botkin, 1970). The ease of measurement has made this method
one of the most widely used measures of terrestrial production.
CO2 assimilation is determined by enclosing all or part of an ecosystem in
an airtight or monitored, open-flow, light-transmitting chamber, and measuring
88
-------
CC>2 levels with an infrared gas analyzer. Daytime uptake of CO2 serves as an
indicator of photosynthetic uptake minus respiratory release of CO2 and provides
a measure of net community production, while nighttime production of CO2 is a
measure of respiration. These can be added together to provide a measure of
gross primary production (Brinson, et al., 1981; Carlson, 1979; Kormondy, 1969).
CO2 has also been measured in open systems using an aerodynamic method
developed by Huber (Odum, 1971). This method involves measurement of the
vertical gradient of CO2 from the ground up. Vertical gradients result from
uptake of CO2 by vegetation and release of CO2 through soil and litter
respiration. Total community respiration can be estimated from nighttime
gradients. In using this method, however, account must be taken of incidental
factors such as mass air movements and soil CO2 evolution that could alter CO2
concentrations. Aerodynamic measurement methods have been used on crops,
grasslands, and even forest communities (Odum, 1971). Refinements in monitoring
techniques may make this an extremely useful way of measuring CO2 in the
future.
CO2 assimilation can be used to monitor primary production in a variety of
terrestrial systems following toxicant insult. For example, Bennett and Hill (1974)
measured CO2 uptake rates of barley and oat canopies, following 2-hour
fumigations with environmental pollutants. Similarly, Carlson (1979) used the rate
of CO2 exchange to measure photosynthetic rates in maple, ash, and oak branches
exposed in an experimental chamber to sulphur dioxide and ozone.
89
-------
Brinson et al. (1981) cite several problems associated with measurement of
primary production by CO2 assimilation, including: the potential for changes in
temperature and air flow to alter rates of CO2 assimilation; the high cost of
using sophisticated monitoring equipment to obtain long-term measures in the
field; and the potential for error in large, complex systems where parts must be
used to estimate whole ecosystem function. The latter problem, however, becomes
insignificant in systems such as nonforested wetlands, where the whole system can
be enclosed.
dH method: Because the pH of water is a function of dissolved CO2, pH
changes in aquatic ecosystems can be used as an index of production. To use pH
as a measure, it is necessary to create a calibration curve that incorporates the
buffering capacity of the system. Methods of preparing a calibration curve are
described in Beyers (1963).
pH measurements are useful in that they do not involve any disturbance to
the community, which could potentially skew any conclusions regarding
environmental hazards. This method has been used primarily to determine gross
community production in laboratory microecosystems (Odum, 1971). Another
advantage of using the pH method is that, in contrast to O2 and CO2, the H+
ions and OH- ions that largely govern the pH in natural waters are not readily
lost to the atmosphere. pH measurements, additionally, are useful for pollutant
studies because they are not generally influenced by the presence of foreign
substances (Cooper and Copeland, 1973).
90
-------
Gorden et al. (1969) point out several problems associated with pH-derived
measures of community production. They note that excretion of organic acids and
ammonium by bacteria and algae can result in pH changes. Furthermore,
heterotrophic CO2 uptake may affect measures of both photosynthetic uptake and
respiratory production. Cooper and Copeland (1973), however, suggest that the
error caused by heterotrophic uptake is likely to be minor. Another potential
problem is that acidity produced by nitrification can reduce pH (Fenchel and
Blackburn, 1979).
Chlorophyll: A determination of chlorophyll levels or chlorophyll
fluorescence is primarily used as a measure of plant biomass, but is sometimes
used as an index of primary production or photosynthesis. Odum (1971) suggests
that chlorophyll concentrations can be used to estimate gross production when
both the assimilation ratio (rate of O2 production per unit biomass of chlorophyll)
and the available light levels are known. The assumption of a constant
assimilation ratio may be false when production is limited by factors other than
light, such as nutrient availability, toxicity, or other stresses.
Measurement of chlorophyll levels requires extraction of chlorophyll from
phytoplankton using an organic solvent and determination of concentrations using
a spectrophotometer (Kormondy, 1969). Chlorophyll extraction has proved to be a
relatively speedy and inexpensive alternative to '^C and O2 methods for
measuring production in large bodies of water. In a study conducted by Ryther
and Yentsch (1957, as cited in Odum, 1971), production rates of marine phyto-
plankton proved to be similar when the chlorophyll and O2 light-dark bottle
methods were compared.
91
-------
Biomass change: Biomass can be defined as the total weight of living
organisms or the total stored energy content of an ecosystem (McNaughton and
Wolf, 1979). Estimates of biomass change over time is useful as an endpoint for
ecological risk assessment. Historically one of the oldest methods of measuring
production, the harvest method of biomass measurement is the most commonly
used.
The harvest method, in which plant material is removed and weighed at
periodic intervals, can be employed in areas where herbivores are insignificant and
the ecosystem does not approach a steady-state condition. Typically, farmers use
this method when measuring production of cultivated crops. It is also useful in
early successional systems, where small numbers of annual plants predominate and
little herbivore consumption takes place before plants are fully grown. In these
situations, a determination of peak standing crop can represent net production.
Although some investigators consider biomass production a measure of primary
production, technically it is a measure of net community production, as it does
not account for biomass loss through herbivory and plant respiration (Odum,
1971).
In diverse environments where all species do not mature simultaneously, peak
standing crop does not account for plant mortality prior to the time of
measurement or plant growth following the time of measurement, nor does it
account for different species attaining peak standing crop at different times. To
deal with these problems, Wiegert and Evans (1964) developed a method of
determining net production by combined measurements of the annual disappearance
of dead material from both green and dead standing crops. Lomnicki et al. (1968)
92
-------
simplified the Wiegert and Evans method by using the collection of green and
dead material only. It was subsequently shown that, in a grassland community,
this modified method produces results comparable to those of the original method
(Lomnicki et al., 1968).
In forests, biomass change can be used to determine production by trees,
shrubs, and ground flora. Biomass change is determined by summing the mass of
a number of components, including flowers, fruit, buds, and leaves. Their mass
can be estimated using traps to collect litter fall, measuring branch mass by
destructive sampling (i.e., selective tree-falling), and measuring root biomass
through root sampling. Repeated measures can then be used to determine biomass
change (Newbould, 1967).
The measurement of biomass offers a relatively simple approach to the
determination of primary production. There are, however, a number of problems
associated with this form of measurement that hinder its utility for use in an
ecological risk assessment. These problems (summarized in Beadle and Long, 1985)
include:
o A tendency by investigators to not consider inorganic (ash) content when
determining biomass;
o Difficulties in identifying living material in senescent leaves and in
separating out live and dead roots;
o Difficulty in obtaining representative samples in heterogeneous plant
communities;
o Errors in the use of regression estimates to determine forest biomass
based on small samples;
o Difficulties in measuring mass losses due to grazing; and
93
-------
o Failure to account for below-ground biomass, including both roots and
root exudates. This factor may pose a particular problem in wetland
communities with large amounts of underground biomass.
Despite accuracy problems, the simplicity of the biomass measure has made it
a standard technique for measuring production, particularly in terrestrial
ecosystems. Biomass is also used as a measure of community structure. This use
is described in a later section.
3.3 BIOGEOCHEMICAL CYCLING
Biogeochemical cycles are defined as "the patterns of flow of chemical
elements through biological organisms and their geological (abiotic) environment"
(Kormondy, 1969). Elements are moved through the action of biochemical,
chemical, and physical forces. During the process, the chemicals undergo a series
of physical and chemical transformations making them more or less available to
the system (Atlas and Bartha, 1981).
The cyclic movement of chemical elements through an ecosystem is
fundamental to the health of the system. Organisms are dependent on the
constant availability of some 20 elements, which are required for all life
processes. The major elemental components of living organisms include C, H, O,
N, P, and S; minor and trace elements include Mg, K, Na, Fe, Mn, Ca, Al, B, Co,
Cr, Cu, Mo, Ni, Se, Si, V, and Zn. These elements are cycled to varying extents
through the biosphere. Toxicant effects on the availability of these elements in
an ecosystem ultimately translate into effects on many other aspects of ecosystem
94
-------
structure and function. Various aspects of nutrient cycles thus provide important
endpoints for use in an ecological risk assessment.
Nutrient cycling can be monitored using a variety of measures. Information
can be obtained on nutrient concentrations in soil and living organisms, and on
levels of nutrient leaching in soil. Serious changes in any of these measures can
provide good evidence of ecosystem stress. Measures of excessive nutrient export
in a mature, normally nutrient-conservative terrestrial system, for example, are
highly suggestive of a serious breakdown in nutrient cycling processes (Sheehan,
1984c).
3.3.1 Nutrient Analysis
Nutrient analyses must be performed with an understanding of the numerous
processes that affect nutrient concentrations in different parts of a system. In
aquatic systems, the molecular forms of carbon, nitrogen, phosphorus, and sulfur
are continuously being altered by physical and biological processes. These
processes include continental erosion and dissolution of gases in atmospheric
water droplets. The nutrient concentrations in aquatic systems are further
affected by factors such as inputs from watersheds and from groundwater,
chemical composition of rivers, stratification and vertical mixing of the water
column, photosynthesis in the near-surface euphotic zone, settlement of organic
matter to deep waters, and the microbial decomposition of organic matter to its
component nutrients.
95
-------
In terrestrial systems, nutrient concentrations may also be affected by
processes such as volcanic eruption, glacier melting, aerial transport of soil and
seeds, precipitation, erosion and runoff, germination and growth of seeds into
plants, photosynthesis, nitrogen fixation, chemical and physical decomposition of
dead plants and organic matter by soil microorganisms, and microbial storage of
nutrients.
Nutrients can be analyzed using standardized laboratory techniques for
element analysis. For example, ammonia-nitrogen can be analyzed by
nesslerization, phenate, titration, or ammonia-selective electrode methods; nitrate-
nitrogen can be analyzed by cadmium reduction or chromotropic acid procedures;
nitrite-nitrogen can be analyzed by the formation of a reddish-purple azo dye;
and sulfate can be analyzed by gravimetric or turbidimetric methods (American
Public Health Assoc., 1985). These standard methods are well defined,
inexpensive, and readily performed in an established laboratory. They may,
however, lack sensitivity for detection of trace elements.
Atomic absorption spectroscopy is another commonly used method for
nutrient analysis. This method is highly sensitive and has the ability to measure
elements in the ppb range. The substance to be analyzed is converted to an
atomic vapor and absorbance at a selected wavelength is measured and compared
with that of a reference substance. The usefulness of the method for ecological
risk assessment, however, is limited by its cost and technical complexity.
96
-------
An autoanalyzer system is a fast and accurate method for measuring levels
of nutrients such as nitrate, phosphate, and silicate. Like atomic absorption
spectroscopy, however, autoanalysis tends to be expensive (Warfar et al., 1983).
Nutrient levels clearly provide useful endpoints for monitoring nutrient
cycles. They are readily measurable, using well standardized and often highly
sensitive methodologies. Nutrient levels alone, however, are sometimes inadequate
measures of pollutant-induced changes in nutrient cycles. During productive
periods in aquatic systems, for example, macronutrient levels are often low and
recycling processes are well developed. Disruptions to nutrient cycles in this case
will only be obvious as changes in nutrient flux rates, rates of transformation,
and associated biological activities (Sheehan, personal communication).
3.3.2 Chlorophyll Content
Analysis of chlorophyll content is a rapid method for estimating the biomass
of photosynthetic algae in aquatic systems. Because algal productivity, and thus
biomass, increases in nutrient-rich systems, chlorophyll measurements provide
indicators of nutrient conditions in aquatic ecosystems. Methods for analysis of
chlorophyll are described in the section on primary productivity.
3.3.3 Leaching
Leaching is the process through which percolating water removes soluble
substances from the soil. The analysis of soil leachate provides a well-tested,
97
-------
sensitive method of monitoring changes in nutrient cycling because it measures
the rate at which nutrients leave the system.
Measurements of the rate of leaching of a chemical indicate the length of
time that a chemical is retained in the topsoil, where it is most subject to
degradation, dissipation, or plant uptake (Hamaker, 1975). In determining the rate
of nutrient disappearance, it is important to consider not only leaching, but also
the rate of degradation. There are few studies in the literature, however, that
deal with the problem of simultaneous degradation and leaching (King and
McCarty, 1968).
The rate of the nutrient loss through leaching is a natural candidate for an
endpoint for an ecosystem risk assessment. Changes in the rate of leaching can
reflect the breakdown of any of a number of nutrient cycling processes. In
addition, a leachate study can detect small changes in nutrient content that are
not measurable using other methods. It can thus provide a sensitive, early
measure of detrimental changes to systems. It is a particularly useful endpoint in
situations where it is difficult to predict the exact mode of action of a new
pollutant (O'Neill et al., 1977).
Soil leaching analysis is generally performed using a lysimeter. Lysimeters
were first used by De LaHire in 1703 and Ebermayer in 1897. Since that time,
they have been modified and improved, and are employed to study water
movements in agricultural and forest systems (Cole et al., 1961). In pollution
impact studies, the lysimeter may be used to assess nutrient leaching, as well as
the movement of toxic chemicals through the soil into the groundwater.
98
-------
Leachate studies can be performed in the laboratory by adding water, at
seasonal rainfall averages, to soil samples and analyzing the leachate for metals.
Although this procedure may oversimplify the leaching process, neglecting many
factors that affect leaching rates in the natural habitat, it can be useful for
making comparisons of leaching rates from different soil types.
Numerous studies illustrate the usefulness of soil leachate measures in
studying pollutant effects on ecosystems. In one study, Nieboer et al. (1980, cited
in Sheehan, 1984c) found that K.+ efflux from lichens correlated with SO2 levels
and the acidity of precipitation. Overrein (1972, cited in Sheehan, 1984c) found a
similar increase in leaching of calcium from forest soils exposed to acid rain.
In 1961, Cole et al. noted that a modified lysimeter proved to be a sensitive
tool when used to monitor the effects of fertilizer containing potassium (added as
KC1) and nitrogen (added as (NH^SC^), on ion movement in the soil. Data from
the lysimeter indicated that both potassium and nitrogen acted as mass ions,
stimulating increased leaching of Ca and other elements.
In a soil leachate study conducted by O'Neill et al. (1977), it was found that
arsenic and lead caused significant increases in losses of calcium and nitrate from
soil microcosms, even though there were no measurable changes in population or
community parameters. These results led O'Neill et al. to conclude that soil
nutrient leaching can provide a sensitive, early-warning indicator of environmental
stress.
99
-------
3.3.4 Determination of a Nutrient Budget
Nutrient budget determination is a "measurement of the total input minus
total output of nutrients in a system." The nutrient budget provides a good
measure of the dynamics of ecosystem function. It includes information on many
nutrients, thus providing a comprehensive overview of ecosystem function. It can
further provide information on a variety of channels of nutrient movement
including sources of inflow (precipitation, dust, and weathering), recycling, and
outflow (runoff, erosion, leaching, and volatilization). This information is critical
in assessing the retention or loss of nutrients that can potentially be stored in
the biomass of the ecosystem for long periods of time (Kormondy, 1969).
A complete nutrient budget study includes determinations of standing stocks
of nutrients, rates of movement of essential nutrients between different ecosystem
compartments, and inputs and outflows of nutrients through the system (Westman,
1985). As part of this study, measurements are made of meteorologic inputs and
geologic outputs. Meteorologic input to the ecosystem is measured through
chemical analysis of precipitation and calculated for each element in terms of
mass flux per unit area. Geological output can be measured through analysis of
chemicals dissolved in water. This is combined with data on the flow rate and
volume of drainage water to obtain an estimate of nutrient loss as grams of
element lost per hectare watershed (Bormann and Likens, 1970).
Likens et al. (1970) were able to obtain a substantial amount of information
on nutrient cycling when they used the nutrient-budget method to study the
effects of forest cutting and herbicide treatment in the Hubbard Brook watershed-
100
-------
ecosystem. Comparison of the streamwater nutrient concentrations in the
deforested and undisturbed areas of the watershed showed that budgetary net
losses from deforested watershed were up to 20 times higher than in the adjacent,
undisturbed watershed. Thus, it appeared that alterations of the nutrient cycles
resulted in an increased loss of dissolved nutrients from the deforested ecosystem.
Although changes in the nutrient budget would seem to be excellent
endpoints for ecosystem risk assessment, in general the costs and time required
for such a study are prohibitive. A study of the effects of a contaminant on
nutrient budgets requires continuous long-term measurements of both the
undisturbed system and stressed ecosystem. In addition, measurements of nutrient
budgets are hampered by the existence of nearly unmeasurable pathways of
nutrient losses such as deep seepage, groundwater circulation, and wind (Bormann
and Likens, 1967), rendering nutrient budget determination infeasible in many
systems. Alternative approaches for nutrient budget estimation may include
microcosm experiments or restricting attention to certain transformations in
nutrient cycles.
3.3.5 Nutrient Cycling: The Nitrogen Cycle
Various phases of nutrient cycles provide readily measurable endpoints for
ecosystem risk assessment. Portions of the nitrogen cycle, in particular, have
frequently been used as endpoints of toxicity. Measurements of alterations in the
processing of nitrogen can serve as sensitive indicators of environmental change
in ecosystems (Sheehan, 1984c; Bollag and Barabasz, 1979). As a critical
constituent of living biomass, and a frequently limiting element in terrestrial and
101
-------
aquatic ecosystems, nitrogen is an appropriate endpoint for measurement
(Westman, 1985; Cook, 1984). The nitrogen cycle includes four major steps:
nitrogen fixation, mineralization, nitrification, and denitrification (Kormondy,
1969).
Though high concentrations of nitrogen are present in the atmosphere, most
organisms are unable to use atmospheric nitrogen. Nitrogen becomes available to
organisms after it has been fixed into nitrate or other inorganic nitrogen
compounds that can be utilized by other organisms (McNaughton and Wolf, 1979).
In aquatic systems, nitrogen fixation is carried out by blue-green algae and
bacteria. In terrestrial systems bacteria are the primary nitrogen fixers.
Nitrogen fixation takes place through reduction of dinitrogen (N2), and the
resulting ammonia may be assimilated by algae and larger aquatic or terrestrial
plants. The reduced inorganic nitrogen is assimilated by various organisms into
proteins and nucleic acids and is eventually released as metabolic waste or as
protoplasm in dead organisms. Many heterotrophic bacteria, actinomycetes, and
fungi convert the organic nitrogen and release it as ammonia. The process is
referred to as mineralization or ammonification.
Ammonia and ammonium salts are converted by nitrifying bacteria to nitrite,
which in turn is converted to nitrate in a pH-dependent process known as
nitrification. Both the nitrate and ammonia may be rapidly taken up by plants or
microbes.
Under certain conditions, particularly anaerobic ones, nitrate is reduced by
denitrifying bacteria to nitrite, ammonia, and dinitrogen. Denitrification occurs
102
-------
primarily under anaerobic conditions in the presence of large amounts of decaying
organic matter. Dinitrogen gas produced during denitrification is released back
into the atmosphere (Kormondy, 1969).
3.3.6 Methods for Analysis
The nitrogen cycle can be monitored through four general processes: soil
nitrogen availability, soil denitrification, nitrification, and nitrogen fixation.
Soil nitrogen availability: There is no generally accepted method for
measuring available nitrogen in soil. Problems arise because quantities of
available nitrogen are difficult to determine, in part because 97 to 99 percent of
soil nitrogen is present in organic forms that are not directly available to plants.
Only after mineralization occurs can nitrogen be used. The amount of nitrogen
that would be mineralized is difficult to predict, as mineralization depends on
numerous environmental factors. In addition, once nitrogen has been mineralized,
it is subject to losses through leaching, denitrification, and microbial conversion
to organic forms (Goh and Haynes, 1986).
Two techniques commonly used for measuring soil nitrogen availability are
the incubation and the chemical indices methods. A number of different
incubation procedures have been proposed. These include short-term (1-6 week)
aerobic incubations and anaerobic incubations. Proposed procedures have been
reviewed by Keeney (1982), Stanford (1982), Sahrawat (1983), and Goh and Haynes
(1986).
103
-------
Incubation methods are usually unsuitable for routine soil testing because
they are both space and time consuming. In addition, it is likely that they
provide only relative estimates of available soil nitrogen. Other problems
associated with the use of incubation methods are discussed by Keeney (1982) and
Goh and Haynes (1986). They include water loss, improper aeration, and improper
sample pretreatment. These can be alleviated through careful handling of samples.
Nonetheless, measurements made using incubation methods are highly
correlated with nitrogen uptake by plants and have been used in a number of
studies. In one such study, anaerobic incubation was used by Shumway and
Atkinson (1977, in Keeney, 1980) to measure NH4-N production. The results
indicated that NH4-N production was correlated with increased diameter growths
in Douglas fir trees.
Chemical indices have been proposed as speedy, precise, and convenient
alternatives to biological incubation procedures. Chemical indices involve
measurements of ammonium or total nitrogen from treated soil samples. Chemical
index methods have been reviewed by Keeney (1982), Stanford (1982), Sahrawat
(1983), and Goh and Haynes (1986).
Soil denitrification: Soil denitrification is generally analyzed by measuring
the products and conditions of anaerobic incubation of soil samples.
Denitrification has been measured under field conditions by disappearance of
applied (as 15^ and ^N20), and by measuring N2O fluxes in the presence of
an N2O reduction inhibitor, acetylene (Rolston et al., 1982; Colbourn et al., 1984;
Haynes and Sherlock, 1986).
104
-------
Soil denitrification was measured in a study assessing the effects of
pollutant metals performed by Bollag and Barabasz (1979). The study, which used
laboratory incubations, involved measurements of denitrification in liquid growth
medium, autoclaved soil, and native soil. The study showed that, depending on
the medium used, cadmium, copper, and zinc all inhibited denitrification to varying
extents.
Nitrification: Nitrification is the process by which ammonium is oxidized via
nitrite to nitrate. Inhibition of nitrification by toxic pollutants can lead to
accumulation of toxic levels of ammonia and nitrites (Liang and Tabatabai, 1978).
Various methods have been used to measure nitrification rates. One
commonly used method involves the incubation of samples in the presence of
chlorate, which inhibits nitrite oxidation, the second step in nitrification. Under
these conditions, the rate of nitrite accumulation provides an estimate of nitrifier
activity (Belser and Mays, 1982). However, this method must be interpreted
cautiously, as chlorate can also inhibit oxidation of ammonia in some cases (Hynes
and Knowles, 1983). Other inhibitors used in measurement of rates of
nitrification include nitrapyrin and allylthiourea (Hall, 1984). In another method,
a tracer amount of '^N-nitrate is added to the sample and nitrification is
measured as the rate of dilution of '^N-nitrate during incubation. Nitrification
rates have also been estimated as changes in nitrate and ammonia concentrations
during long-term incubations (Koike and Hattori, 1978).
Nitrogen fixation: Nitrogen fixation in terrestrial systems can be studied
through the analysis of symbiotic legumes. Fixation processes are sensitive to
105
-------
nonlethal doses of pollutants, which impair the survival of organisms dependent on
nitrogen fixation. Methods of analyzing legume fixation performance include
measures of ethylene formation and nodulation performance, as well as nitrogen
mass balance and isotope analyses.
The ethylene formation (acetylene reduction) method is based on the fact
that the nitrogenase enzyme involved in nitrogen fixation also reduces acetylene
to ethylene. The amount of ethylene produced during incubation can be easily
measured by gas chromatography. The ethylene production rate is simple,
sensitive, and inexpensive to measure, and is thus frequently used as a measure of
nitrogen fixation.
The nodulation performance method involves a quantification of nitrogen-
fixing root nodules on leguminous plants. Simple nodule quantification, however,
is not an adequate measure of nitrogen fixation, as not all nodules are active.
Further analysis must be performed to test for nitrogenase activity in individual
nodules.
It is possible to quantitatively estimate rates and amounts of nitrogen fixed
by nitrogen mass balance techniques. Kjeldahl digestion of the sample will
provide a measure of total organic nitrogen (American Public Health Assoc., 1985).
To use mass balance as a measure of nitrogen fixation, it would be necessary to
account for other sources of available nitrogen, and for losses of nitrogen
subsequent to fixation (Silvester, 1983).
106
-------
Nitrogen fixation has also been measured as the uptake of N1^ determined
with a mass spectrometer. Although this method is quite sensitive, it is also
time-consuming and expensive. This and other methods for analysis of nitrogen
fixation are discussed in detail in Bergersen (1980) and Silvester (1983).
Numerous studies have demonstrated pollutant-induced inhibition of nitrogen
fixation. Home and Goldman (1974) observed the suppression of nitrogen fixation
by blue-green algae with the addition of low levels of copper. Similarly, Francies
et al. (1980, cited in Sheehan, 1984c) observed that nitrogen fixation by free-
living bacteria was substantially reduced at pH levels below 6.4. Alexander (1980,
cited in Sheehan, 1984c) additionally noted that the root nodulation process is
highly sensitive to soil acidity.
Various phases of the nitrogen cycle, and nutrient cycles in general, provide
sensitive endpoints for ecological risk assessment. As nutrient cycles are critical
functional processes for ecosystem survival, measurements indicating disruptions of
these cycles can have far reaching implications concerning the health of an
ecosystem. In addition, many nutrient-related measures such as rates of nutrient
soil-leaching and nitrogen fixation are highly sensitive to low levels of pollutant
stress. These endpoints are also readily measurable using a variety of techniques.
The specific choice of a nutrient measurement endpoint is best made based on
knowledge of the limiting factors in a particular ecosystem.
107
-------
3.3.7 Decomposition
Decomposition is the process through which organic material is degraded
and organically bound nutrients are released into the ecosystem (Levin et al.,
1984). Through its vital role in the immobilization and release of nutrients,
decomposition links primary productivity and nutrient cycling. The process of
degradation involves the movement of energy and carbon through a series of
decomposer trophic levels. Microorganisms initiate the process by assimilating the
proteins and carbohydrates from detritus (dead organisms and excreta). Bacteria,
actinomycetes, and fungi further degradation, while cellulose-, hemicellulose-, and
chitin-digesting organisms complete the process. Detritivorous animals strongly
affect rates of decomposition by mechanical breakdown of detritus, and ingestion
of detritus and decomposing microorganisms. Measures of decomposition thus
provide good universal endpoints for monitoring pollution stress at the ecosystem
level (Sheehan, 1984c).
The rate of decomposition is affected by both the quantity and quality of
available substrate, as well as the physical, chemical, and biological status of the
ecosystem. Pollutants may alter decomposition by causing changes in the chemical
status of the environment. Toxic organic pollutants (oil, industrial organic
effluents, pesticides) have also been observed to affect the other controlling
variables.
108
-------
3.3.8 Methods of Measurement
Methods commonly used for monitoring changes in the rates of decomposition
include measures of litter bag decomposition rates and populations of litter
decomposers, as well as indices of microbial activity.
Litter bag analysis: Litter bag analysis is the most commonly used method
for studying decomposition. The method makes use of litter bags of a
predetermined pore size, which are filled with known weights of dried substrate
samples and placed under a suitable amount of soil surface litter or suspended in
the water of an aquatic system. Decomposition rates are measured as changes in
dry-weight biomass over time (Gorden, 1972).
Populations of litter decomposers: Populations of litter decomposers such as
bacteria and macroinvertebrates can be monitored to assess potential effects of
pollutants on decomposition. Macroinvertebrates can be measured using estimates
of species abundance and biomass. Microfungal population densities can be
estimated by using dilution plate count methods. Macroinvertebrate decomposer
analysis can also be achieved by using species lists or by counting their numbers.
Measures of decomposer populations are often combined with litter bag analysis to
provide a more complete picture of changes in the decomposition process.
Indices of microbial activity: The evolution of carbon dioxide from soil has
often been used as a measure of microbial activity in soil. Unfortunately, the
many types of organisms present in soil make it an invalid measure of microbial
activity. It may instead be considered a measure of community respiration from
109
-------
the soil. The measurement of soil respiration by CO2 flux can be used as an
index of litter breakdown. Estimates of CO2 flux can be made on litter samples
using KOH to absorb respired CO2 (Gorden, 1972).
Most naturally occurring compounds are degradable by microorganisms, and
the degradation of organic compounds will stimulate microbial biomass, which may
be estimated by a number of indirect methods. These methods include
measurement of ATP levels and monitoring of soil enzyme activity (Visser et al.,
1984).
Although a large number of studies have monitored the effects of pollutants
on decomposition, most of these involve laboratory, microcosm, and short-term,
small-scale field investigations. Long-term effects on larger-scale natural
ecosystems are thus not well understood.
One field study of decomposition effects was conducted by Freedman and
Hutchinson (1980b). The effects of smelter pollution were examined using analysis
of forest litter decomposition. Measurements were made of litter standing crop
and rates of litter input, and populations of litter-decomposers. Additional
measures were also made of soil CO2 flux and soil enzyme activity. It was found
that there was a reduction in the rate of litter decomposition in forested areas
close to the smelter. This reduction was correlated with decreases in soil CO2
flux and acid phosphatase activity. In addition, reductions were noted in
populations of soil microfungi and microarthropods at the contaminated sites
(Freedman and Hutchinson, 1980b).
110
-------
In another decomposition study, Forbes and Magnuson (1980) monitored
decomposition and microbial colonization of leaves in a stream stressed by coal
ash effluent. ATP content proved to be a sensitive measure of pollutant-induced
effects on leaf decomposition. After 27 and 96 days of exposure, the ATP
content of leaves placed in the effluent-exposed stream was found to be
significantly lower than that of leaves placed in a reference stream. In addition,
there was a lack of normal macroinvertebrate colonization in the exposed leaf
packs. This was correlated with reduced colonization and decomposition by fungi.
The ash effluent appeared to indirectly affect macroinvertebrates through its
interference with leaf decomposition and thus with food availability.
Studies to date indicate that decomposition provides a good, readily
measurable endpoint for an ecological risk assessment. Reliable, sensitive
measures are available for monitoring changes in both the decomposition process
and in decomposing organisms. More long-term field work must be done,
however, to fully assess the usefulness of decomposition measures for monitoring
long-term stress effects on an ecosystem.
3.4 CONCLUSION
The functional characteristics of an ecosystem provide important endpoints
for determining pollutant effects on that ecosystem. Primary production, nutrient
cycling, and decomposition are critical measures of energy and nutrient flow
through the system. As these processes form the foundation for ecosystem
structure, they can serve as critical indicators of ecosystem survival. In addition,
111
-------
some measures, such as nutrient leaching from soil, are extremely sensitive to low
levels of stress and may be particularly useful as sentinels of ecosystem decline.
Decisions about which characteristics to use for a particular risk assessment
should be made on a case-specific basis, depending on the nature of both the
pollutants that are affecting a system and the system itself. Thus, while primary
productivity might be an appropriate measure to use in a heavy-metal-polluted
aquatic system, it might not be a very sensitive measure in a similarly polluted
terrestrial system. In general, several kinds of functional measures are required
to fully assess the responses of a system.
In addition to their importance as endpoints, functional measures can also
provide insights into the causes of changes in ecosystem structure. The cause of
a change in species biomass, for example, might be revealed through functional
measures as alterations in nutrient levels and/or primary productivity. When used
in combination with measures of ecosystem structure and species level effects,
functional measures can thus provide information on the dynamics involved in an
ecosystem's response to stress.
3.5 ECOSYSTEM OR COMMUNITY STRUCTURE
The structure of an ecosystem is defined by the communities of organisms
that comprise it. Structural features include the abundance, biomass, diversity,
and spatial distribution of populations, as well as the taxonomic, functional, and
trophic organization of the community. While structural and functional aspects of
an ecosystem are clearly linked to each other, they may be affected by stress in
112
-------
totally independent ways. For example, pollutant-induced changes in community
diversity may not result in concurrent changes in ecosystem productivity.
Similarly, changes in productivity can occur without any substantial change in
diversity. Aspects of ecosystem response to stress may thus be missed, if only
one endpoint is examined. Measures of both ecosystem structure and function
must therefore be included in an ecosystem risk assessment (Sheehan, 1984b).
Any analysis of stress-induced change in community structure has certain
inherent limitations. Baseline data that would provide a description of the pre-
exposure structure of a community is rarely available. Such baseline data is of
limited value for communities that are not at equilibrium. An alternative
procedure is to compare the exposed community with a similar unexposed site.
This makes a questionable assumption that the two sites differ only in the
presence of the pollutant (Hurlbert, 1984; Stewart-Oaten et al., 1986). Studies of
structure are also limited by the ability of the investigator to identify all species
and species interactions in the community. At best, a representative sample of
the community must be used to deduce the state of interactions of the whole
(Herricks and Cairns, 1982).
Indicators of community structure range from simple, descriptive measures of
abundance and biomass to more analytical, multivariate analyses of ecological
similarity. The type of indicator used depends on the nature of the community to
be assessed and the availability of time and resources.
113
-------
3.6 ABUNDANCE AND BIOMASS
Biomass is defined as the total weight (Clapman, 1973) and abundance as the
absolute numbers of living organisms (McNaughton and Wolf, 1979) in a
community. Both abundance and biomass provide simple, gross measures of
community structure, and may be specified in terms of trophic or taxonomic units,
thus providing comparative measures of changes in trophic structure. Abundance
or biomass are often expressed in terms of population per unit of space (density).
When absolute abundance measures are not feasible, often measures of relative
abundance are obtained, such as number encountered per unit time, or proportion
of individuals captured that are of a particular type.
While these measures are useful in terms of their simplicity, they lack
sensitivity and provide little information about the overall character of the
system. They also display both seasonal and temporal variations. In addition,
both measures carry an inherent bias. Measures of abundance tend to exaggerate
the importance of small, abundant species, while measures of biomass tend to
overemphasize the importance of large, nonabundant species. These measures are
best used in conjunction with other measures of community structure (Sheehan,
1984b).
Chlorophyll content can be used for estimating biomass of photosynthetic
organisms. In aquatic systems chlorophyll estimates are a rapid and useful way to
estimate phytoplankton biomass.
114
-------
A number of pollution studies have demonstrated variable effects of
pollutants on ecosystem abundance and biomass. In a study of acidified
Norwegian lakes, conducted by Leivestad et al. (1976), it was found that both the
biomass and the density of benthic invertebrates declined with decreasing pH.
Hendrey et al. (1976) also studied acidified lakes, however, and observed an
increase in periphyton biomass accompanying a decrease in abundance of
zooplankton and fish. In addition, Stokes (1986) reviewed studies of aquatic
communities of phytoplankton, periphyton, and macrophytes and found that while
acidification caused little change in biomass and productivity, it caused consistent
reductions in species richness and composition.
Sensitivity problems associated with abundance measures are illustrated in a
study conducted by Winner et al. (1975, in Sheehan, 1984b). In this study, several
measures of structure were used to examine a community of benthic
macroinvertebrates dispersed along a pollutant gradient in a copper-contaminated
stream. Although abundance did show an inverse relationship to copper
concentrations, large variations in abundance obscured differences at median
concentration levels. A clear, graded response was observable, however, using
measures of species richness (see description below) (Sheehan, 1984b).
Problems associated with accurate determinations of biomass are detailed in
the section on Primary Productivity (Section 3.2.1).
115
-------
3.7 SPECIES LISTS
The listing of species is a straightforward indicator of community structure
that is commonly used in studies of ecosystem stress. Species lists from samples
at varying distances from a pollutant source can be compared to provide an
indication of pollutant effects (Sheehan, 1984b) Species lists may also comprise
the first step in determining other measures, such as species richness, and species
dominance, which can in turn be used as components of biological diversity and
similarity indices (Herricks and Cairns, 1982). In ecosystem-level studies of
effects of pollutants on experimental lakes, among the most sensitive responses
were species composition of phytoplankton and disappearances of sensitive aquatic
species (Schindler, 1987).
3.8 BIOLOGICAL INDICES (POLLUTION INDICES)
Biological indices are measures developed from observations of responses of
groups of indicator species (see Indicator Species, Section 2.3.3) to pollutant
stress in aquatic systems (Hellawell, 1977). They provide numerical rankings of
species and species assemblages. These indices were first developed in the early
1900's, primarily as a method of evaluating the effects of municipal sewage or
organic wastes on aquatic systems (Sheehan, 1984b).
Numerous biotic indices have been proposed over the years. Some of the
simpler indices compare numerical differences between groups of species or
individual species, which are classed as being tolerant or intolerant to pollution.
Many of these simple indices look only at species numbers and make no allowance
116
-------
for species abundance. For example, in a pollution index developed by Beck
(1954, in Hellawell, 1977), the index is: I = 2C\-C2, where Cj represents the
number of intolerant and C2 represents the number of tolerant macroinvertebrate
species.
Some of the more complex biotic indices integrate measures of key species
abundance, their known pollution tolerance, and their reliability as indicators. In
one such index, developed by Chandler (1970, in Hellawell, 1977), scores are
determined for both species abundance and pollutant tolerance. The resultant
values are then combined to produce a total biotic score.
One of the oldest indicator systems is the saprobic approach (Kolkwitz and
Marsson, 1902, 1908, 1909, cited in Sladecek, 1963). This system was designed for
use in running water. It is based on the assumption that the presence of
particular organisms living in running water bodies permits estimation of both the
level of contamination in the water and the trend of the general conditions (for
example, deterioration or self-purification). Water quality is predicted based on
the number of specific microorganisms present in 1 ml of test water as well as
chemical properties of the water such as DO, H2S content, and BOD (Sladecek,
1963). The data on the quantity of microorganisms and the chemical parameters
are fit into rigid categories that describe water quality. Sladecek (1963) lists
criticisms of the saprobic system including force fitting of data, consideration of
only processes related to bacterial decomposition, and no validation of the system.
Overall, saprobity is not considered a very valuable method for most situations.
117
-------
The usefulness of biotic indices in evaluating ecosystem stress was illustrated
in a study conducted by Solbe (1977). In this investigation, two biotic indices,
two diversity indices (see section on Diversity Indices) and a similarity index (see
section on Similarity Indices) were used to evaluate the invertebrate communities
in a zinc-polluted stream. It was found that both the Trent Biotic Index and the
Chandler Biotic Score served as good indicators of pollutant effects, with the
Chandler Biotic Score being most sensitive to low levels of pollution. MargalePs
diversity indices and a cluster analysis that measured similarity, however, proved
to be less sensitive than the biotic indices.
While biotic indices provide a semi-quantitative measure for evaluating
ecosystem stress, their usefulness for studying the effects of toxic chemicals has
not been well tested (Sheehan, 1984b). In addition, many of the indices require
subjective determinations of organism tolerance or indicator values (Herrick and
Cairns, 1982). Scores produced, furthermore, may not clearly distinguish between
different combinations of evaluated factors. Thus, the same score may be
obtained both for a few individuals of a sensitive, pollution-intolerant species and
a ubiquitous, pollution-indifferent species (Hellawell, 1977). As Solbe's study
indicates, however, biotic indices can provide sensitive indicators for certain well-
defined systems.
3.9 SPECIES RICHNESS
Species richness can be defined as the number of species present in a system
(McNaughton and Wolf, 1979). As it would be virtually impossible to determine
the total number of species in a natural community, richness has been quantified
118
-------
as the number of species per fixed number of individuals, or the number of
species per unit area (Peet, 1974).
Species richness is widely used as an endpoint in measurements of polluted
ecosystems. It is also a component of measures of species diversity (Levin et al.,
1984). When determined by making direct counts of species numbers in a sample,
it provides a simple, practical, objective measure of community structure. It can
also be used to compare species diversity between communities, when the
relationships of species importance are similar (Peet, 1974).
A problem in the measurement of species richness is its inherent dependence
on sample size. A series of simple richness indices have been developed that
assume a consistent relationship between species numbers and sample size. If this
assumption is not satisfied, these richness indices will be biased in an unknown
manner. One such index, developed by Margalef (1951, cited in Peet, 1974), is
described by the logarithmic relationship: Rj = (S-l)/Log N, where S is the
number of species and N is the number of individuals in a sample.
Another well-known example of a richness index that assumes a known
relationship between species numbers and sample size is Preston's analysis.
Preston's relationship calculates the expected number of species in a total sample,
based on the proposition that a log-normal distribution best describes species
abundance data (Peet, 1974). This log-normal distribution arises from the fact
that a multiplicity of factors causes populations to increase in a geometric, rather
than arithmetic, pattern. In a large sample, individuals in a population thus tend
to be distributed in log-normal pattern (Gray, 1979). Patrick (1949, in Levin et
119
-------
al., 1984) was able to use this index to demonstrate a loss of diversity in diatom
assemblages of stream communities polluted with sewage and organic outputs
(Levin et al., 1984).
Studies have shown that species richness is a good measure of pollutant
stress. It is sensitive to gradients of pollutant concentration and thus provides a
"concentration-response relationship" useful for monitoring stress in ecosystems
(Sheehan, 1984b). For example, in a retrospective study conducted by Stokes
(1986), the effects of acidification on phytoplankton, periphyton, and macrophytes
in a number of aquatic systems were compared. The results showed that species
richness declined consistently with declining pH. In contrast, biomass and
productivity did not change substantially.
Varied effectiveness of species richness measures have been observed in
terrestrial studies. Freedman and Hutchinson (1980a) found that species richness
provided a sensitive monitor of distance from a pollutant source in a forest
community. In this study, richness of the ground flora vegetation was found to
correlate well with distance from a metal smelter. In measures of the overstory
or tree canopy, however, biomass-related measures were more sensitive than
measures of either richness or diversity. The lack of correlation in the overstory
results was attributed to the fact that only a small number of species were
present in the initial community.
While species richness can provide a useful measure of pollutant stress and
community-level effects along a gradient, it does not provide a complete picture
of ecosystem dynamics. As the Stokes (1986) study indicates, changes in species
120
-------
richness do not necessarily complement changes in biomass or productivity. It is
evident that in a system with functional redundancy, one species might readily
compensate for the loss of another (Sheehan, 1984c; Peet, 1974). Thus, richness
alone may not be a complete measure but is a useful component of a complex of
measures of ecosystem function.
3.10 DIVERSITY
Although the concept of diversity has been broadly explored in the
literature, there is still no universally accepted definition of the term (Sheehan,
1984b). A generally held definition, however, involves the dual-component concept
introduced by Simpson (1949, cited in Peet, 1974). According to this definition,
diversity incorporates both species richness and equitability (the relative evenness
of abundance of species distribution) (Peet, 1974). The concept thus provides a
measure of the "frequency of species occurrence in a community" (Herrick and
Cairns, 1982).
Diversity indices are widely used measures of community structure. The
theory behind these indices is that diversity increases as a community ages and
higher diversity is associated with community stability (Odum, 1971). Caswell
(1976, in Gray, 1979), however, has shown that diversity decreases in the final
stages of succession. Additionally, the relationship between diversity and stability
has come under serious question. Nonetheless, there is evidence to suggest that
pollution can produce a decline in ecosystem diversity (Sheehan, 1984b).
121
-------
Diversity calculations can additionally provide a useful method of
summarizing large amounts of data (Herrick and Cairns, 1982), and they permit
comparative measures, evaluating the degree to which species abundances in
stressed environments deviate from predicted values (Hellawell, 1977). Because of
these advantages, large numbers of diversity indices have been developed over the
years. Some of these emphasize the richness component of diversity, some
emphasize the equitability component, and others incorporate both components
(heterogeneity).
A variety of equitability-based indices have also been developed. These
generally scale a heterogeneity measure to the maximum possible value for a fixed
species number and sample size (Peet, 1974). Equitability indices are based on the
assumption that if species numbers are equal, communities with the most equitable
distribution of individuals among species can be considered to be the most diverse
(Sheehan, 1984b).
Some of the heterogeneity indices are derived from information theory.
Heterogeneity, in these cases, is based on the degree of uncertainty associated
with the species of an individual randomly selected from a population. As the
number of species and evenness of the sample increases, so does the uncertainty
of selecting an individual of a particular species. Thus, increased uncertainty is
correlated with increased diversity.
The Shannon-Weaver (Wiener) formulation is one of the most commonly used
information-based indices. Pielou (1966) suggests that this method is useful for
estimating diversity in a large collection of organisms in which all the species
122
-------
cannot be identified. She further demonstrates how different formulas for
estimating Shannon's measure of information can be used to estimate diversity in
different kinds of communities.
The Shannon-Weaver and other diversity indices have produced mixed results
in pollution studies. Bechtel and Copeland (1970) found that fish species and
biomass diversity, determined using the Shannon-Weaver formula, provided useful
indicators of environmental pollution in Galveston Bay. Similarly, in a second
study by the same authors, the diversity of phytoplankton, zooplankton, and
nekton provided a reliable indicator of pollutant levels (Copeland and Bechtel,
1971, cited in Sheehan, 1984b).
On the other hand, in the previously mentioned study by Freedman and
Hutchinson (1980a), the Shannon-Weaver formula did not prove to be a powerful
indicator of pollutant-induced stress. It was less sensitive than measures of
either biomass or total cover to changes in a forest community along a gradient
of smelter pollution. Similarly, in a study of a deciduous forest exposed to air
pollution, McClenahen (1978) found that diversity, as measured by the Shannon-
Weaver index, was inadequate to describe changes in the pollutant-exposed shrub
layer. In this situation, richness and evenness varied inversely and thus obscured
any changes that might have been evident from measures of diversity. In another
study conducted by Eisele and Hartund (1976, cited in Sheehan, 1984b), it was
found that, in a methoxyclor-contaminated stream community, invertebrate
populations declined while diversity remained unaffected.
123
-------
Because of studies like those described above, numerous authors have offered
alternative methods for measuring structure. Gray (1979) has suggested, for
example, that measures such as the Shannon-Wiener (Weaver) index and the
rarefaction method of Sanders can be difficult to interpret. Additionally, they are
not very sensitive to pollutant-induced changes in community structure. Gray
suggests that a log-normal distribution analysis, an analysis that measures the
departure from the log-normal distribution of individuals among species, is a much
more sensitive measure of ecosystem stress. Sheehan (1984b), however, notes that
this technique has not been widely used in ecotoxicology studies.
Standard diversity indices have been criticized in the literature for their
inherent biases. Shannon's formula and similar indices tend to overemphasize the
importance of rare species at low population densities (Sheehan, 1984b). On the
other hand, Simpson's and related indices tend to be most sensitive to changes in
the most common species (Peet, 1974). Another problem associated with use of
diversity indices is the difficulty of interpreting their results (Hellawell, 1977).
Green (1975, cited in Read et al., 1978) noted that diversity indices could measure
changes in community structures, but that they provide little information on the
nature of the change. This point is further emphasized by Herricks and Cairns
(1982), who suggest that information about the community structure is lost when
large quantities of data are summarized in a single index.
Green (1979) argues against the use of diversity indices. He indicates that
the concept of diversity is vague, combining two components that may vary
independently. He further notes that diversity indices are "often uncritically
124
-------
applied, without regard to the assumptions implicit in the various diversity
formula and the biases in their estimation."
Problems with sensitivity, bias, and information content of diversity indices
may render them inadequate measures for assessment of pollutant effects in many
ecosystems. Sheehan (personal communication) has suggested that, while these
indices might be useful as measures of gross changes in pollutant-stressed
systems, they are not particularly sensitive to low level effects.
3.11 COMPARATIVE INDICES (SIMILARITY INDICES)
Comparative indices are designed to evaluate two or more parameters of
community structure over space and time. These methods were initially designed
by plant ecologists to evaluate plant communities. They are useful for monitoring
the effects of pollutants at varying distances from a source or for detecting
ecosystem recovery or deterioration over time (Hellawell, 1977). They also
provide a means of comparing stressed and unstressed communities (Brock, 1977,
cited in Sheehan, 1984b).
Comparative indices identify the similarities or differences between samples.
Some of these indices compare species lists or species abundance data. Those
which make use of species lists include the Jaccard index, a measure that is
particularly useful for distinguishing between similar samples. The Dice and
Ochiai coefficients, on the other hand, are species list indices that are most
sensitive for very dissimilar samples.
125
-------
Other similarity indices make use of quantitative abundance tabulations or
distance measures between communities. These measures, which include percent
similarity, coefficient of community, the Czekanowski index, and the Bray-Curtis
coefficient, may require modification to account for overestimates of the
importance of abundant species (Herricks and Cairns, 1982).
Species ranking provides another method of comparing communities. In these
methods, species are ranked by their relative importance in a community. The
rankings in different communities are then compared (Hellawell, 1977). Ranking
indices include Spearman's rank correlation coefficient and Kendall's coefficient.
These measures have been criticized for a number of different reasons. Problems
include methodological difficulties involved in evaluating species ties and species
absence (Herricks and Carins, 1982) and similarities of rankings that can arise in
structurally different communities (Hellawell, 1977).
Finally, correlation coefficients have been used to identify relationships
between communities based on species distributions or species presence or
absence. These measures include the point correlation coefficient and the product
moment correlation.
Studies have shown that similarity indices provide sensitive measures of low-
level stress in polluted ecosystems. These measures appear to be more sensitive
than those obtained by diversity indices (Sheehan, 1984b). For example, in a
study conducted by Marshall and Mellinger (1980, cited in Sheehan, 1984b)
changes in a Lake Michigan zooplankton community in response to cadmium
pollution were assessed with percentage similarity and coefficients of community
126
-------
indices. Using these measures, significant changes in community structure could
be identified at cadmium levels that produced minimal effects on diversity indices.
The coefficient of community was also found to be a useful measure by
McClenahen (1978) in his study of air pollutant effects on a deciduous forest. As
previously mentioned, McClenahen found that changes in community structure at
low pollutant levels were difficult to identify using the Shannon diversity index.
He noted, however, that the coefficient of community could identify significant
correlations between community structure and air pollutant exposure along a
decreasing pollutant gradient.
3.12 MULTIVARIATE ANALYSIS
A more complex approach to similarity measures used in ecological risk
assessment is multivariate analysis. Multivariate methods permit associations to
be made between samples or groups of samples based on many different variables
(Herricks and Cairns, 1982). Green (1979) describes such techniques as being
powerful measures that are sensitive to subtle patterns of difference and provide
a visual representation of relationships between samples.
Multivariate measures of similarity can be broadly grouped into two
catagories, cluster analysis and ordination. Clustering methods involve grouping
samples with similar characteristics into clusters and fitting the clusters into a
two-dimensional display of complex resemblance patterns. The display may then
be tested using an indicator of accuracy.
127
-------
Ordination techniques involve the organization of samples along one or more
axes or gradients, using statistical comparisons of variables such as species
abundances or similarity coefficients,. Strong discontinuities between communities
are represented by a steeply sloped gradient (Odum, 1971). These techniques have
been found to be useful in determination of the causes of associations between
communities (Herrick and Cairns, 1982).
3.13 TROPHIC ORGANIZATION
Trophic structure may serve as a framework within which other measures of
community structure or function can be monitored. Determinations may be made
of productivity, dominance, biomass, and other endpoints at different trophic
levels.
Energy from the sun is stored in chemical bonds during photosynthesis and
subsequently passes through a series of organisms in an ecosystem. This
movement of energy through the system is characterized as a food chain. Food
energy passes through the food chain in steps, as one organism feeds on another
and is eaten itself. Feeding steps or levels are defined as trophic levels
(McNaughton and Wolf, 1979; Odum, 1971).
Organisms in a particular system are functionally grouped into different
trophic levels. Green plants are classified as producers, herbivores as primary
consumers, carnivores as secondary consumers, and secondary carnivores as
tertiary consumers (Odum, 1971). Trophic relationships may also be described in
terms of feeding mechanisms. Cummins (1974) categorized aquatic insects as
128
-------
shredders, collectors, scrapers, and predators. Functional feeding groups have
also been termed guilds (Johnson, 1981).
Changes in trophic organization may be measured in terms of changes in
species numbers, species abundance, or species diversity at different trophic
levels. They may also be quantified in terms of shifts in trophic dominance.
Trophic shifts were observed in a study by Hall and Likens (1980, cited in
Sheehan, 1984c). They found that in acidified streams, the ratio of consumers to
producers declined. Another example of trophic shifts in aquatic systems are
blooms of phytoplankton primary producers following insecticide-induced removal
of crustacean herbivores (Hurlbert, 1975).
3.14 SPATIAL STRUCTURE
The spatial structure of an ecosystem involves both the vertical and
horizontal patterns of species organization within the community. This spatial
organization is readily quantifiable for terrestrial producers. It is difficult to
define, however, for mobile terrestrial consumers or in aquatic communities where
many species are mobile.
Spatial structure has been used as a measure of pollution-induced changes in
ecosystem structure. Sheehan (1984b) describes a number of terrestrial studies in
which simplification of structure was observed to result from environmental stress.
Changes in plant distribution were also observed by Dawson and Nash (1980) in a
study of the effects of copper smelter effluent on a forest ecosystem. Reductions
in the abundance of annuals and grasses were observed in the area near the
129
-------
smelter. Large shrubs, however, remained abundant. The pattern of distribution
was attributed to the presence of a thin layer of copper-contaminated, acidified
soil in the area near the smelter, which killed plants such as annuals that have
shallow root systems.
A small number of studies thus suggest that spatial structure offers a
potentially useful endpoint for measures of toxicant effects on the structure of
terrestrial systems.
3.15 GUILDS
The guild concept, though still in the early stages of research and
development, provides a potentially useful organizing structure for the
performance of a risk assessment. A guild was originally defined by Root in 1967
as a "group of species that exploit the same class of environmental resources in a
similar way" and was designed as a method of grouping species based on niche
overlap rather than taxonomy (Root, 1967). In general, this concept has been
used to group animals based on feeding or nesting strategies. Thus, animals might
be grouped into the same guild if they all feed on a common set of flowering
plants, forage on the ground for insects, or nest in tree holes. A community can
then be conceived of as containing a complex of potentially interacting guilds
(Krebs, 1985) which can be organized into a matrix. A feeding and nesting guild
matrix for an avian community in a terrestrial habitat is illustrated in Figure 3-1.
Since its original development, the usefulness of the guild concept as a
general research tool and as a component of the risk assessment process has been
130
-------
extensively analyzed in the literature. It has been suggested that the concept
simplifies the analysis of community structure by reducing the number of
components in a community to a manageable number. In addition, use of the
guild concept might diminish the economic costs of a risk assessment by reducing
the amount of long-term input required to monitor the status of a stressed system
(Landres, 1983; Severinghaus, 1981). The study of a guild containing a small
group of ecologically related species may thus provide a practical alternative to
the single-species or community analysis approaches to risk assessment (Petersen,
1986).
One of the main problems associated with the use of the guild concept arises
from the lack of an objective mechanism of delineating guilds in a particular
system. Guilds are often defined by an investigator, based on a subjective
classification of resource use (Karr, 1980, cited in Landres, 1983). They may also
be defined somewhat more objectively using statistical techniques such as cluster
analysis, principal components analysis, and canonical correlation (Landres, 1983).
For example, statistical methods were employed by Poysa (1983) to study guilds in
a waterfowl community. Both cluster analysis and principal component analysis
were used to identify guilds based on feeding habitat and feeding method. These
analyses required investigator-determined classifications of feeding habitats and
feeding methods, thus introducing a subjective component to even these forms of
analysis.
131
-------
Figure 3-1. Guilds matrix with primary
feeding and nesting zones. (Adapted from Verner, 1984)
ai
z
o
N
o
z
2
UJ
HI
u.
>
tr
<
2
E
Q.
vaaw
Snags
ACWO
COHA
$8HA
Am.
ANHU
BTM
Tree canopies
«.n
WCKI
YRWA
WBK
SOON
PHAI
HIM
NOOA
OCJU
NUWO
YSSA
W»NU
Tree boles
and limbs
WBSH
HOWR
90
HfTH
CATM
Btwn
RCKI
WW
Shrubs
rov
UM
TUVU
AOAO
AMXI
PTTHA
A**0
CAQU
SHCO
SCOW
MOOO
MTO
COMA
LI 00
con.
OHOW
Ground
WCMf
ACV
WTO
star
LfOW
ftCJA
HOfl
©
Jr
Ground
Shrubs
Tree boles
and limbs
Tree canopies
Snags
1
»
9
Jfi
«
M
¦D
s
Ł
PRIMARY NESTING ZONE
132
-------
Because of the subjectivity of guild determination, there is a general lack of
consistency in the ways in which different investigators define guilds. Yerner
(1984) found this problem to be particularly acute in the literature on avian
guilds. In some instances, different investigators identified different sets of
guilds among avian species in similar habitats. It appears that the ability to
classify organisms into guilds is, in part, based on the experience and
understanding of the ecologist performing the analysis (Gauch and Whittaker,
1981, cited in Landres, 1983). Thus, the usefulness of the guild concept may vary
with investigator competence.
Recently, a number of investigators have attempted to define an approach to
the use of guilds that would be useful in the performance of a risk assessment.
One such approach involves the use of "guild indicators" and is based on the idea
that all members of a guild will react similarly to environmental change. The
responses of any one species of the guild can thus be used as an indicator of the
responses of the others (Severinghaus, 1981). While this method might provide an
economic approach for conducting a risk assessment, it has been criticized by a
number of investigators. Landres (1983) notes that while all members of a guild
may exploit a common resource, they may not necessarily be similar in all aspects
of their life history and thus may not be equally affected by a change in that
resource. Some species, for example, may be readily able to shift to a new
resource while others cannot (Landres, 1983; Verner, 1984). It would thus be
wrong to assume that a single species can automatically serve as an indicator for
the rest of the guild.
133
-------
The whole guild approach is another method of using guilds as a tool in risk
assessment. This approach, as described by Verner (1984), uses the entire guild
as an indicator of habitat quality. Guild categories are based on "zones of
habitat that are likely to respond in similar ways to various sorts of
perturbations." The guild then provides a basis for obtaining information on
individual guild species and on the ability of the habitat to support them. Verner
used this approach to categorize bird populations in a pine-oak woodland. His
guilds were organized into a grid based on primary feeding and nesting zones (see
Figure 3-1). Intraguild comparisons were made of differences in species richness
on grazed and ungrazed plots of land. It could then be determined how well an
ungrazed plot would support a guild such as ground-feeding birds.
Szaro (1986) took a third approach to using guilds by placing groups of bird
species in a ponderosa pine forest into response guilds. These guilds were made
up of bird species that demonstrated similar changes in densities, in particular
habitat-zones, over time. Typical response guilds included:
o Species that were absent in one year on most or all study plots and
showed no preference for any forested site;
o Species that had their highest densities on the medium-cut and light-cut
plots; and
o Species that had their highest densities on untreated, light-cut, and
medium-cut plots.
Szaro found that there was a strong correlation between species density and
response guild density. Szaro further noted that this correlation was much
weaker for functional guilds (habitat-zone guilds) as defined by Verner (1984). He
134
-------
concludes that response guilds revealed relationships that would not be seen using
other types of guild matrices.
In conclusion, changes in guilds are potentially useful endpoints for
ecosystem risk assessments. They facilitate analyses of species responses to
functional or structural changes in specific habitat zones of an ecosystem.
Furthermore, they reduce the numbers of community components and thus simplify
community structure analyses. Unfortunately, while there are numerous studies
that make use of guilds to analyze ecosystems, little work has been done to apply
the guild concept to systems exposed to pollutant stress. In addition, there is
still no consistently agreed-upon approach for defining guilds. Further research
and testing will be required to determine if an approach such as response guilds
would be useful in an ecosystem risk assessment.
3.16 CONCLUSION
Measures of changes in ecosystem structure can provide important
components of an ecosystem risk assessment. Measures such as abundance,
biomass, , and species lists provide relatively simple, gross measures of ecosystem
stress. Species richness appears to be a good, sensitive measure of pollutant-
induced stress, and can provide a partial picture of changes in community
structure. Similarity indices and multivariate analyses further provide important
measures of differences in pollutant-effects between systems or along gradients.
They also seem to be sensitive measures of pollutant effects.
135
-------
Other methods may prove to be less sensitive or less desirable for use as
endpoints of toxicity. Diversity indices, for example, produce mixed results in
ecosystem studies and often prove to be less sensitive than richness or similarity
indices. Biotic indices are often confounded by the need for subjective
determinations of organism tolerance or value as indicators. In general, diversity
and biotic indices are not strongly recommended as measures for use in ecosystem
risk assessments. It is important to note that measures of ecosystem structure,
such as species richness or diversity, are influenced by numerous poorly
understood factors, including regional diversity, geographical dispersal,
competition, predation, adaptation, and environmental variation (e.g., Ricklefs,
1987). All of these factors may vary among local sites. Therefore, in comparing
contaminated with uncontaminated sites, it is difficult to conclude with certainty
that any differences noted in ecosystem structure are due solely to the
contaminant. Whole ecosystem experimentation may be the only way to determine
the effects of a particular chemical on an ecosystem (Levin et al., 1984).
3.17 STABILITY
One of the fundamental characteristics of an ecosystem is its degree of
stability. The concept of ecosystem stability is associated with a wide spectrum
of ecosystem traits such as inertia, elasticity, and amplitude. These traits
generally refer to the tendency of a system to be upset by, and to recover from,
the effects of perturbations. Elements of stability are clearly important factors
to consider in any evaluation of an ecosystem's response to pollutant-induced
stress.
136
-------
A large body of literature exists on the various aspects of ecosystem
stability and recovery. While the theoretical aspects of these concepts have been
debated for years, their practical use in the monitoring of pollutant-induced
ecosystem stress is still in its early stages (Sheehan, 1984b). Because stability-
related concepts have been detailed in another document (Ecological Risk
Assessment Issues, Volume II. Ecosystem Stability/Recovery Report, TRI, 1988d),
they will be briefly summarized here.
Stability can be defined as the ability of a system to return to an
equilibrium state after a temporary disturbance. The equilibrium state applies to
aspects of both ecosystem structure and ecosystem function (McNaughton and
Wolf, 1979). The stability definition implies that the less a system fluctuates, the
more stable it is (Holling, 1973). Many systems, however, naturally experience
repeated cyclic fluctuations. These systems would still be considered stable if
they displayed a tendency to return to their original patterns of fluctuation
(Sheehan, 1984b).
Considerable debate has focused on the relationship between stability and
diversity. The original hypothesis stated that species diversity stabilizes
ecosystem functional properties (McNaughton, 1977). While some investigators
provide evidence to support this concept (Van Voris et al., 1980), mathematical
models suggest that under certain conditions increased diversity has a
destabilizing effect (May, 1973, cited in McNaughton, 1977). Further theoretical
and empirical research indicates that there is not a simple relationship between
stability and diversity. Stability may not decrease as diversity increases if there
is a decrease in connectance or strength of food-web interactions (Pimm, 1984;
137
-------
Moore and Hunt, 1988). An alternative view of the diversity-stability relationship
is that diversity is a function of environmental stability. If, as theory predicts,
complex systems tend to be dynamically fragile, they may be expected to occur in
more predictable habitats (May, 1986). Some empirical evidence suggests that
diversity is maximized at intermediate levels of environmental disturbance
(Connell, 1978).
Methods of determining stability involve examination of changes in both
structural and functional characteristics representative of total ecosystem
responses. In a study by Van Voris et al. (1980), loss of calcium in leachate was
used as an index of ecosystem stability. Other studies have focused on
fluctuations in diversity and population density. Sheehan (1984b) suggests that
measures of similarity, species richness, total biomass, and primary productivity
would also be suitable for describing ecosystem fluctuations.
A large variety of terms have emerged from the concept of ecosystem
stability. These include: inertia, elasticity, amplitude, resilience, hysteresis,
malleability, constancy, and persistance (Sheehan, 1984b). A few of these are
briefly summarized below.
3.17.1 Resilience
Resilience can be defined as the ability of a system to absorb perturbations
and return to a stable configuration (Holling, 1973). The resiliency of a
particular ecosystem is dependent on a number of different factors. In general,
high resiliency is found in moderately stable systems, located in temperate areas,
138
-------
which have high productivity and large niche sizes, and which experience large
environmental fluctuations. On the other hand, resilience tends to be low in
communities that are highly stable, such as tropical rain forests, or highly
unstable, such as the tundra. These systems are either too specialized to adapt
to large perturbations or too low in productivity to readily compensate for the
stress-induced changes (Clapham, 1973).
3.17.2 Amplitude
Amplitude is one component of resilience. The amplitude of a system refers
to the amount of perturbation that the system can absorb and still retain the
ability to recover (Westman, 1978, cited in Sheehan, 1984b). It describes the
recovery threshold for the system.
Amplitude can be ascertained by monitoring recovery along a pollutant
gradient or comparing pollutant effects at varying levels of exposure. In one
such amplitude study conducted by Baker (1971, cited in Sheehan, 1984b), it was
found that salt marsh grass loses its ability to recover from crude oil applications
after 12 successive oilings.
3.17.3 Elasticity
Another component of resilience is elasticity. Elasticity is defined as the
speed of recovery to a stable state and has been measured using an index of
elasticity (Stauffer and Hocutt, 1980). Factors that influence this index were
listed by Cairns and Dickson (1977). They include:
139
-------
o The availability of nearby epicenters of replacement organisms for the
damaged system;
o The mobility of organisms in the system;
o The management capabilities for rapid control of the damaged system;
o The condition of the system after perturbation; and
o The extent to which residual toxicants remain in the system.
One of the problems involved in estimating elasticity following pollutant
stress is the fact that recovery may not be a linear process, and thus may be
difficult to predict early on. In addition, different features of a system may
recover at different rates, making it difficult to produce one comprehensive
prediction of time-to-recovery (Sheehan, 1984b).
3.17.4 Inertia
Inertia (alternatively termed resistance) represents the ability of a system to
resist stress-induced changes (Cairns and Dickson, 1977; Sheehan, 1984b). It
appears to be related, like resilience, to the degree of specialization and the
amount of exposure to environmental fluctuations that are normally experienced
by a system (Boesch, 1974, cited in Sheehan, 1984). According to Sheehan
(1984b), while inertia may be quantified as the amount of pollutant per unit area
per unit time that will cause a specified amount of ecosystem damage, resilience
would be quantified as the maximum level of pollutant exposure that would still
permit ecosystem recovery.
An index for the prediction of the inertia of streams has been developed by
Cairns and Dickson (1977) and has been used to classify stream fish communities
by Stauffer and Hocutt (1980). The index of inertia of a stream may be
calculated using the following parameters:
o The tendency of the system to experience variable environmental
conditions;
o The functional and structural redundancy of the system (in terms of
species numbers and trophic interactions);
140
-------
o The dependability of the stream flow, its flushing capacity, etc.;
o The detoxification capacity of the water;
o The proximity of the system to major ecological transition areas; and
o The water quality management facilities of the area (Cairns and Dickson,
1977; Stauffer and Hocutt, 1980).
3.18 CONCLUSION
Various aspects of stability have been briefly reviewed. One measure of
stability, inertia, may be potentially useful for predicting effects of pollutants on
ecosystems. Measures of stability in general, however, require long-term
monitoring of numerous parameters. Thus, stability measures are not useful for
rapid determinations of effects. In addition, these measures have not been widely
used in the field and require more testing to establish their applicability.
141
-------
CHAPTER 4
SUMMARY
4.0 INTRODUCTION
A variety of ecotoxicological endpoints have been proposed to assess the
effects of pollutants on ecological systems. Potential endpoints occur at the level
of the individual organism, the population, and the ecosystem. In general,
endpoints at lower levels of organization (organism or suborganism levels) have
been more widely used because they are simpler, more rapidly and inexpensively
assessed, and are most useful in determining the mechanisms of toxicological
effects. Endpoints at the population or ecosystem levels of organization are more
complex and difficult to interpret, but are probably ecologically more realistic
because they incorporate the complexity of interactions among organisms, and
between organisms and their abiotic environment. A major unresolved question is
to what extent endpoints at lower levels of organization can be used to predict
pollutant impacts at higher levels of organization.
It is important to note that the ecosystem- and population-level
ecotoxicological endpoints, and to a lesser extent the organismal-level endpoints,
are often generalized indicators of stress or disturbance, rather than specific
responses to toxic pollutants. There are many sources of stress or disturbance in
natural, unpolluted ecosystems (e.g., Sousa, 1984), so evidence of disturbance in an
ecosystem is not necessarily an indication of significant effects of pollution.
Ideally, measurements would be made before and after pollutant discharge to
provide direct evidence of pollutant effects (Stewart-Oaten et al., 1986).
142
-------
4.1 ORGANISM-LEVEL ENDPOINTS
Physiological endpoints most closely related to individual fitness are acute
mortality, growth and development, and reproductive success. Acute lethality
testing is widely used to provide minimal estimates of toxicity. However, such
testing is not sufficiently sensitive to assess sublethal or chronic effects, which
occur at lower toxicant concentrations and may be of considerable ecological
importance.
Biochemical response endpoints may provide information on the mechanism of
toxic action. Since biochemical processes are in general particularly sensitive to
pollutants, biochemical response endpoints may provide early warning of potential
impacts on the individual. However, most biochemical processes respond to
conditions other than pollutant stress and the response of these endpoints may be
adjusted with acclimation of the individual to the stress. Correlations between
biochemical response endpoints and individual success need to be established to
enhance the value of these sensitive endpoints as predictors of higher level
impacts.
Osmoregulatory activity is an appropriate endpoint for assessing impacts on
certain freshwater and estuarine fish and invertebrates. Again, the ability of the
individuals to acclimate to osmoregulatory stress must be considered in
interpreting osmoregulatory response data.
143
-------
Musculoskeletal endpoints have been used to monitor stress responses in fish.
Correlations between abnormalities and the ecological success of deformed fish
need to be established.
Respiratory activity has been used as a response endpoint for a number of
species. However, it is difficult to generalize about the patterns of respiratory
response to stress. Respiration rates may be elevated or inhibited by pollutants
and exposed individuals may adjust ventilation rates with acclimation time.
Behavioral alterations are appropriate endpoints for impact assessments if
they act either to protect the individual from harm, as in avoidance behavior, or
make the individual more vulnerable to the stress, as in the loss of antipredator
behavior. Although behavioral responses are not easy to demonstrate in the
laboratory or in the field, these endpoints, if demonstrated, may be easily
extrapolated to predict potential population-level effects.
Genotoxicity and carcinogenicity are endpoints that provide early warning of
stress. Data must be gathered on the natural incidence of mutations and tumors
to aid in interpreting the importance of chemically induced mutation and tumor
incidence rates.
Endpoints measuring growth, development, and reproductive success of
individuals are of most obvious utility in predicting population-level impacts.
Because these endpoints are directly related to population success, their use in
impact studies where single-species test data is extrapolated to predict population-
level impacts is recommended. These endpoints have been less frequently used
144
-------
because of the time and expense required to conduct full-life-cycle chronic
toxicity tests. However, the more frequently used short-term physiological and
biochemical endpoints cannot be recommended until their relationships to
organismal growth and reproductive success are determined.
A number of studies have documented interactions between effects of
pollutants and abiotic and biotic factors in the environment. These examples
illustrate the inadequacy of laboratory single-species, single-factor testing to
estimate all ecological effects of contaminants, and point to the necessity of
relating ecotoxicologic effects on individual organisms to population- and
ecosystem-level effects of pollutants.
4.2 POPULATION-LEVEL ENDPOINTS
At the population level, stress response may be monitored in terms of
changes in the abundance, distribution, age structure, or gene makeup of exposed
populations. The first three endpoints can be quite clearly related to the overall
success of the exposed population. Changes in the gene pool may be related to
future adaptability of the population to similar types of stress.
Also at question is the selection of an appropriate population or populations
to be monitored in an impact assessment. Quite clearly, monitoring the effects on
commercially or aesthetically valuable species is important to predict impacts on
those species. More valuable to predicting higher level impacts is population-
response data on representative and ecologically important species within exposed
communities. Included within this category are keystone species, which strongly
145
-------
influence the structure of the communities or the functioning of the ecosystem.
If there is interest in extrapolating population response to predict ecosystem-level
impacts, emphasis should be placed on gathering data on populations from major
functional groups, including primary producers, primary and tertiary consumers,
and decomposers.
A problem in the use of population-level endpoints as indicators of pollution
is that the numerous factors regulating population structure are, as yet, poorly
understood. This makes it difficult to discriminate pollutant effects from
naturally occurring processes. As population structure is influenced by
interactions among population members, with other populations, and with the
abiotic environment, it becomes necessary to examine the effects of pollutants at
the ecosystem level.
4.3 ECOSYSTEM FUNCTION ENDPOINTS
The analysis of functional response endpoints can provide data on energy
flow and nutrient cycles. The functional capability of the ecosystem is, in fact,
the ultimate criterion of ecological success. The effective use of endpoints in
describing impacts is dependent upon a theoretical and practical knowledge of
ecosystems for proper interpretation, and on the collection of sufficient baseline
data to establish normal process rates. A longer history of measuring functional
response variables will be necessary to establish threshold values for unacceptable
reductions in functional capability.
146
-------
Primary productivity provides the energy for the base of the food web. This
process has been shown to be sensitive to a variety of pollutants and other forms
of stress. Reductions in primary productivity that are of substantial magnitude
and of long duration are unquestionably detrimental to energy processing in
exposed ecosystems.
Disruptions in material cycles will be critical if the effects on cycling
processes indirectly inhibit ecosystem production. Material cycles can be upset by
pollutant inhibition of the decomposition process, interference with the functional
links in specific nutrient cycles, or disruption of nutrient conservation
mechanisms. The effects on decomposition can be measured in terrestrial and
aquatic ecosystems and changes in decomposition rate and the completeness of
mineralization can be related to the level of pollution stress. At present, little
data are available on the long-term impacts of reduced decomposition on
ecosystem production.
Specific nutrient cycling processes are key to the production efficiency of
ecosystems. Identification of the critical cycles in specific ecosystems will be
necessary for the selection of appropriate monitoring points.
Nutrient conservation is exceedingly important in terrestrial ecosystems.
Evidence of excessive leaching of essential nutrients is a sign of stress. Leaching
loss of nutrients has been correlated with reduced nutrient availability in the
plant-root zone and reduced plant growth in nutrient-deficient soil.
147
-------
A problem in the use of ecosystem-function endpoints is their relative
insensitivity to ecosystem structure. Shifts in species composition to more
pollutant-resistant species may or may not result in changes in such functional
processes as productivity or nutrient cycling. Thus, an assessment of pollutant
effects at the ecosystem level should include both structure and function
endpoints.
Because the factors controlling ecosystem structure and function are
numerous and poorly understood, it is difficult to distinguish ecosystem-level
effects of pollutants from naturally occurring processes. Many of the ecosystem-
level endpoints depend on the questionable assumption that unpolluted ecosystems
are at a stable, undisturbed state.
4.4 ECOSYSTEM STRUCTURE ENDPOINTS
Measures of ecosystem structure can provide important data for ecosystem
risk assessments. Structural changes in stressed ecological communities may be
visualized as an information network reflecting environmental conditions, but not
demonstrating the external mechanisms or internal interactions that brought about
the reorganization in species composition or dominance patterns.
Structural endpoints such as abundance and biomass of communities provide
relatively simple, gross measurements of ecosystem stress. Species richness has
been shown to be sensitive to the level of stress and can provide a partial
picture of changes in community composition accompanying stress.
148
-------
Combined numerical indices such as similarity and ordination measures may
be used to track changes in community structure with changes in pollutant
concentrations. Although diversity indices have been widely used in hazard
assessment studies, these integrated measures are often insensitive to stress and
provide data that are difficult to interpret. The use of numerical indices
exclusive of the biological data from which they are calculated should be
discouraged.
4.5 CONCLUSIONS
A multilevel ecological risk assessment, which makes use of a combination of
organism, population, and ecosystem-level endpoints, provides the most effective
approach to examining ecosystem stress. Multilevel testing would both enhance
the sensitivity of a risk assessment and broaden its focus to more complex levels
of ecological organization. In contrast, the traditional approach of using only
single-species testing is generally inadequate to account for the pollutant-induced
effects on the complex organization of the ecosystem. Supplementation of
organism-level endpoints with carefully selected population- and ecosystem-level
endpoints may greatly improve the accuracy of ecological risk assessments.
It is difficult to precisely specify the most appropriate endpoints for use in
an ecological risk assessment because the choice depends partially on both the
ecosystem being stressed and the nature of the pollutant stressor. A variety of
organism-, population-, and ecosystem-level endpoints are available to choose
from. Among the factors that should be considered in the choice of endpoints
are sensitivity to low pollutant levels, predictability of long-term stress to
149
-------
individual species or to ecosystems, and ease of measurement. Endpoints that
appear to be of generally high value in assessing the ecological risk of pollutants
include primary productivity, nutrient cycling, trophic structure, and growth and
reproduction, particularly for keystone species. Further guidance in choosing
endpoints is given in a related document, "Ecological Endpoint Selection Criteria"
(TRI, 1988a). There is still, however, an inadequacy of field data documenting the
usefulness of organism-, population-, and ecosystem-level endpoints in predicting
ecological risk. Future research in this area would facilitate the development of
the multilevel risk assessment approach.
150
-------
REFERENCES
Addison, R.F., M.E. Zinck, and D.E. Willis. 1981. Time- and dose-response of
hepatic mixed function oxidase activity in brook trout Salvelinus fontinalis on
polychlorinated biphenyl residues: Implications for biological effects monitoring.
Environ. Pollut. Ser. A. 25: 211-218.
Alink, G.M., E.M.H. Frederix-Wolters, M.A. van der Gaag, J.F.J, van de Kerkhoff,
and C.L.M. Poels. 1980. Induction of sister-chromatid exchange in fish exposed
to Rhine water. Mutat. Res. 78: 369-374.
American Public Health Association, Inc. 1985. Standard Methods for the
Examination of Water and Wastewater Including Bottom Sediments and Sludges.
Sixteenth Edition. 1268 pp.
Anderson, J.W., J.M. Neff, and S.R. Petrocelli. 1974. Sublethal effects of oil,
heavy metal and PCBs on marine organisms, pp. 83-121 in Khan, M.A.Q. and J.P.
Bederka, eds., Survival in Toxic Environments. Academic Press, New York.
Andrewartha, H.G., and L.C. Birch. 1954. The Distribution and Abundance of
Animals. Univ. of Chicago Press.
Antonovics, J., A.D. Bradshaw, and R.G. Turner. 1971. Heavy metal tolerance in
plants. Adv. Ecol. Res. 7: 1-85.
Atlas, R.M., and R. Bartha. 1981. Microbial ecology: Fundamentals and
Applications. Addison-Wesley Publishing Company, Reading, MA.
Babich, H., and G. Stotzky. 1980. Physicochemical factors that affect the
toxicity of heavy metals in microbes in aquatic habitats, pp. 181-203 in Colwell,
R.R., and J. Foster, eds., Aquatic Microbial Ecology. Proceedings of the
Conference of the Amer. Soc. Microbiology. Maryland Sea Grant Publ. Univ. of
Maryland, College Park, MD.
Baddeley, M.S., B.W. Ferry, and E.I. Finegan. 1973. Sulphur dioxide and
respiration in lichens, pp. 299-313 in Ferry, B.W., M.S. Baddeley, and D.L.
Hawksworth, eds., Air Pollution and Lichens. Univ. of Toronto.
Barica, J., and L.R. Mur, eds. 1979. Hypertrophic Ecosystems. Developments in
Hydrobiology 2. Junk, The Hague.
Barnthouse, L.W., R.V. O'Neill, S.M. Bartell, and G.W. Suter. 1986. Population
and ecosystem theory in ecological risk assessment, pp. 82-96 in Poston, T.M. and
R. Purdy, eds., Aquatic Toxicology and Environmental Fate, 9th Symposium. STP
921. Am. Soc. Testing Materials, Philadelphia.
Bayne, B.L., J. Anderson, D. Engel, E. Gilfillan, D. Hoss, R. Loyd, and F.
Thurberg. 1980. Physiological techniques of measuring the biological effects of
pollution in the sea, in Mclntyre, A.D. and J.B. Pierce, eds., Biological Effects of
Marine Pollution and the Problems of Monitoring., Rapp, P.-V. Reun, Cons. Int.
Explor. Mer. 179: 88-89.
151
-------
Beadle, C.L., and S.P. Long. 1985. Photosynthesis - is it limiting to biomass
production? Biomass. 8: 119-168.
Bechtel, T.J., and B.J. Copeland. 1970. Fish species diversity indices as
indicators of pollution in Galveston Bay, Texas. Contrib. Mar. Sci. 15: 103-132.
Beijer, K., and A. Jernelov. 1978. Ecological aspects of mercury selenium
interactions in the marine environment. Environ. Health Perspect. 25: 43-45.
Belser, L.W., and E.L. Mays. 1982. Use of nitrifier, activity measurements to
estimate the efficiency of viable nitrifier counts in soils and sediments. Appl.
Environ. Microbiol. 43: 945-948.
Bengtsson, B.E. 1979. Biological variables, especially skeletal deformities in fish,
for monitoring marine pollution. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 286:
457-464.
Bengtsson, B.E., A. Bengtsson and M. Himberg. 1985. Fish deformities and
pollution in some Swedish waters. Ambio. 14: 32-35.
Bennett, J.H., and A.C. Hill. 1974. Acute inhibition of apparent photosynthesis
by phytotoxic air pollutants, pp. 115-127 in Dugger, M., ed., Air Pollutant Effects
on Plant Growth. ACS Symposium Series 3. Amer. Chem. Soc.
Bergersen, F.J. 1980. Methods for Evaluating Biological Nitrogen Fixation. John
Wiley and Sons, New York.
Berner, T., Z. Dubinsky, F. Schanz, U.J. Grobbelaar, H. Rai, U. Uehlinger, and
P.G. Falkowski. 1986. The measurement of primary productivity in a high-rate
oxidation pond (HROP). J. Plankton Res. 8(4): 659-672.
Beyers, R.J. 1963. The metabolism of twelve aquatic laboratory microecosystems.
Ecol. Monogr. 33: 281-306.
Billiard, R. 1978. Effects of heat pollution and organo-chlorinated pesticides on
fish reproduction, pp. 265-267 in Final Reports on Research Sponsored Under the
First Environmental Research Program, Commission of the European Communities,
Publ. by CEC Brussels, IBN 99-825-0/85-C, Brussels, Belgium. (Cited in: Effects of
Pollutants at the Ecosystem Level, P.J. Sheehan, D.R. Miller, G.C. Butler and P.
Bourdeau, eds., John Wiley and Sons, New York, pp. 23-50, 1984.)
Birge, W.J., J.A. Black, and R.A. Kuehne. 1980. Effects of organic compounds on
amphibian reproduction. University of Kentucky, Water Resources Research
Institute, Research Report 121, Lexington, Kentucky. 39 pp. (Cited in: Effects of
Pollutants at the Ecosystem Level, P.J. Sheehan, D.R. Miller, G.C. Butler and P.
Bourdeau, eds. John Wiley and Sons, New York, pp. 23-50, 1984.)
Birge, W.J., J.A. Black, and D.M. Stern. 1979. Toxicity of organic chemicals to
embryo-larval stages of fish. U.S. EPA-560-11-79-007; PB80-101637.
152
-------
Bishop, W.E., and L.C. Valentine. 1982. Use of the central mudminnow (Umbra
limi) in the development and evaluation of a sister chromatid exchange test for
detecting mutagens in vivo, pp. 99-108 in Pearson, J.C., R.B. Foster, and W.E.
Bishop, eds., Aquatic Toxicology and Hazard Assessment: Fifth Symposium. STP
766. Am. Soc. Testing Materials, Philadelphia.
Blus, L.J., C.D. Gish, A.A. Belisle and R.M. Prouty. 1972. Logarithmic
relationship of DDE residues to eggshell thinning. Nature. 235: 376-377.
Bollag, J.M., and W. Barabasz. 1979. Effect of heavy metals on the
denitrification process in soil. J. Environ. Qual. 8(2): 196-201.
Borgmann, U., R. Cove, and C. Loveridge. 1980. Effect of metals on the biomass
production kinetics of freshwater copepods. Can. J. Fish. Aquat. Sci. 37: 567-575.
Bormann, F.H., and G.E. Likens. 1967. Nutrient cycling. Science. 155: 424-428.
Bormann, F.H., and G.E. Likens. 1970. The nutrient cycles of an ecosystem.
Scientific American 223(4): 92-101.
Bowmer, T., R.G.V. Boelens, B.F. Keegan, and J.O. O'Neill. 1986. The use of
marine benthic "key" species in ecotoxicological testing: Amphiura filiformis.
Aquat. Toxicol. (NY) 8: 93-109.
Boyce, M.S. 1977. Population growth and stochastic fluctuations in the life table.
Theor. Popul. Biol. 12: 366-373.
Brinson, M.M., A.E. Lugo, and S. Brown. 1981. Primary productivity
decomposition and consumer activity in freshwater wetlands. Annu. Rev. Ecol.
Syst. 12: 123-161.
Bull, C., and J. Mclnemey. 1974. Behavior of juvenile coho salmon Oncorhvnchus
kisutch exposed to Sumithion (fenitrothion) an organophosphate insecticide. J.
Fish Res. Board Can. 31: 1867-1872.
Busby, D.G., P.A. Pearce, and N.R. Garrity. 1981. Brain cholinesterase response
in songbirds exposed to experimental fenitrothion spraying in New Brunswick,
Canada. Bull. Environ. Contam. Toxicol. 26: 401-406.
Cairns, J., ed. 1986. Community Toxicity Testing. STP 920, Am. Soc. Testing
Materials, Philadelphia.
Cairns, J., and K.L. Dickson. 1977. Recovery of streams and spills of hazardous
materials. pp. 24-44 in Cairns, J., K.L. Dickson, and E.E. Henricks, eds.,
Recovery and Restoration of Damaged Ecosystems. Univ. of Virginia Press.
Call, D.J., L.T. Brooke, M.L. Knuth, S.H. Poirier, and M.D. Hoglund. 1985. Fish
subchronic toxicity prediction model for industrial organic chemicals that produce
narcosis. Environ. Toxicol. Chem. 4: 335-341.
Campbell, P.G.C., and P.M. Stokes. 1985. Acidification and toxicity of metals to
aquatic biota. Can. J. Fish Aquat. Sci. 42: 2034-2049.
153
-------
Carlson, R.W. 1979. Reduction in the photosynthetic rate of Acer auercus and
fraxinus species caused by sulphur dioxide and ozone. Environ. Pollut. 18: 159-
170.
Christensen, E.R., and N. Nyholm. 1984. Ecotoxicological assays with algae:
Weibull dose-response curves. Environ. Sci. Technol. 18: 713-718.
Clapham, W.B. 1973. Natural Ecosystems. Case Western Reserve Univ.,
Cleveland, OH.
Colbourn, P., M.M. Iqbal, and I.W. Harper. 1984. Estimation of the total gaseous
nitrogen losses from clay soils under laboratory and field conditions. J. Soil. Sci.
35: 11-22.
Cole, D.W., S.P. Gessel, and E.E. Held. 1961. Tension lysimeter studies of ion
and moisture movement in glacial till and coral atoll soils. Soil Sci. Soc. Am.
Proc. 25:321-325.
Connell, J.H. 1978. Diversity in tropical rain forests and coral reefs. Science
199: 1302-1310.
Constantinidou, H.A., and T.T. Kozlowski. 1979. Effects of sulfur dioxide and
ozone on Ulmus americana seedlings. I. Visible injury and growth. Can. J. Bot.
57: 170-175.
Cook, R.B. 1984. Man and the biogeochemical cycles: Interacting with the
elements. Environ. 26(7): 11-40.
Cooke, A.S. 1972. The effects of DDT, dieldrin and 2,4,-D on amphibian spawn
and tadpoles. Environ. Pollut. 3: 51-68.
Cooke, A.S. 1973. Shell thinning in avian eggs by environmental pollutants.
Environ. Pollut. 4: 85-152.
Cooke, A.S. 1981. Tadpoles as indicators of harmful levels of pollution in the
field. Environ. Pollut. Ser. A. Ecol. Biol. 25: 123-133.
Cooper, D.C., and B.J. Copeland. 1973. Responses of continuous-series estuarine
microecosystems to point-source input variations. Ecol. Monographs 43: 213-236.
Couch, J.A., and J.C. Harshbarger. 1985. Effects of carcinogenic agents on
aquatic animals: An environmental and experimental overview. Environ. Carcin.
Revs. 3: 63-105.
Cummins, K.W. 1974. Structure and function of stream ecosystems. Bioscience.
24: 631-641.
Czuba, M., and D.P. Ormrod. 1984. Cadmium concentration in cress shoots in
relation to cadmium-enhanced ozone phytotoxicity. Environ. Pollut. 25: 67-76.
154
-------
Davey, F.B., H. Kleerekoper, and P. Gensler. 1972. Effect of exposure to
sublethal DDT on the locomotor behavior of the goldfish (Carassius auratus). J.
Fish Res Board Can. 29: 1333-1336.
Dawson, J.L., and T.H. Nash. 1980. Effects of air pollution from copper smelters
on a desert grassland community. Environ. Exp. Bot. 20: 61-72.
Donaldson, E.M. 1981. The pituitary-interrenal axis and as indicator of stress in
fish. pp. 11-17 in Pickering, A.D., ed., Stress and Fish. Academic Press, New
York.
Donaldson, E.M., and H.M. Dye. 1975. Corticosteroid concentration in sockeye
salmon (Oncorhvnchus nerka) exposed to low concentration of copper. J. Fish
Res. Board Can. 32: 533-539.
Drummond, R.A., W.A. Spoor, and G.F. Olson. 1973. Some short-term indicators
of sublethal effects of copper on brook trout, Salvelinus fontinalis. J. Fish Res.
Board Can. 30: 698-701.
Edwards, R.H., and R.M. Overstreet. 1976. Mesenchymal tumors of some
estuarine fishes of the northern Gulf of Mexico. I. Subcutaneous tumors, probably
fibrosarcomas, in the striped mullet, Mueil cephalus. Bull. Mar. Sci. 26(1): 33-40.
Eisenberg, J.F. 1980. The density and biomass of tropical mammals, pp. 35-55 in
Soule, M.E. and B.A. Wilcox, eds., Conservation Biology, an Evolutionary-
Ecological Perspective. Sinauer Assoc. Publ., Sunderland, MA.
Elton, C. 1933. The Ecology of Animals. Methuen, London.
Estes, J.A., and J.F. Palmisano. 1974. Sea otters: Their role in structuring
nearshore communities. Science 185: 1058-1060.
Farr, J.A. 1978. The effect of methyl parathion on predator choice of two
estuarine prey species. Trans. Am. Fish Soc. 107: 87-91.
Feder, W.A. 1973. Cumulative effects of chronic exposure of plants to low levels
of air pollutants, in Naegele, J.A., ed., Air Pollution Damage to Vegetation. Adv.
Chem. Ser. 122: 21-30.
Fenchel, T., and T.H. Blackburn. 1979. The nitrogen cycle. pp. 101-126 in
Fenchel, T., and T.H. Blackburn, eds., Bacteria and Mineral Cycling. Academic
Press, New York.
Fitzwater, S.E., G.A. Knauer, and J.H. Martin. 1982. Metal contamination and its
effects of primary production measurements. Limnol. Oceanogr. 27(3): 544-551.
Fleming, W.J., M.A.R. McLane, and E. Cromartie. 1982. Endrin decreases screech
owl productivity. J. Wildl. Manage. 46: 462-468.
Folmar, L.C. 1976. Overt avoidance reaction of rainbow trout fry to nine
herbicides. Bull. Environ. Contam. Toxicol. 15: 509-514.
155
-------
Forbes, A.M., and J.J. Magnuson. 1980. Decomposition and microbial colonization
and leaves in a stream modified by coal ash effluent. Hydrobiology 76: 263-267.
Fowler, C.W. 1981. Density dependence as related to life history strategy.
Ecology 62(3): 602-610.
Fox, G.A., A.P. Gilman, D.B. Peakall, and F.W. Anderka. 1978. Behaviorial
abnormalities of nesting Lake Ontario herring gulls. J. Wildl. Manage. 42: 477-483.
Franson, J.C., J.W. Spann, G.H. Heinz, C. Bunck, and T. Lamont. 1983. Effects of
dietary ABATE on reproductive success, duckling survival, behavioral, and clinical
pathology in game-farm mallards. Arch. Environ. Contam. Toxicol. 12: 529-534.
Freedman, B., and T.C. Hutchinson. 1980a. Long-term effects of smelter pollution
at Sudbury, Ontario on forest community composition. Can, J. Bot. 58: 2123-2140.
Freedman, B., and T.C. Hutchinson. 1980b. Smelter pollution near Sudbury,
Ontario, Canada, and effects on forest litter decomposition. pp. 395-434 in
Hutchinson, T.C., and M. Havas, eds., Effects of Acid Precipitation on Terrestrial
Ecosystems. Plenum Publ., New York.
Fulton, R.S., and H.W. Paerl. 1987. Toxic and inhibitory effects of the blue-
green alga Microcystis aeruginosa to herbivorous zooplankton. J. Plankton Res. 9:
837-855.
Gaarder, T., and H.H. Gran. 1927. Production of plankton in the Oslo Fjord.
Rapp. Proc. Verb. Cons. Perm. Int. Explor. Mer. 42: 1-48.
Garten, C.T., Jr., G.W. Suter II, and B.G. Blaylock. 1985. Development and
evaluation of multispecies test protocols for assessing chemical toxicity,
ORNL/TM-9225. Martin Marietta, Oak Ridge National Laboratory, Oak Ridge, TN.
Gentile, J.H., S.M. Gentile, N.G. Hairston, Jr., and B.K. Sullivan. 1982. The use
of life-tables for evaluating the chronic toxicity of pollutants to Mvsidopsis bahia.
Hydrobiologia 93: 179-187.
Giddings, J., and G.K. Eddlemon. 1978. Photosynthesis respiration ratios in
aquatic microcosms under arsenic stress. Water, Air and Soil Pollut. 9: 207-212.
Goh, K.M., and R.J. Haynes. 1986. Nitrogen and agronomic practice, pp. 379-468
in Haynes, R.J., ed., Mineral Nitrogen in the Plant System. Acad. Press, Orlando,
FL.
Goodyear, P.C. 1972. A simple technique for detecting effects of toxicants or
other stresses on a predator-prey interaction. Trans. Am. Fish Soc. 101: 367-370.
Gorden, R.W. 1972. Field and Laboratory Microbial Ecology. Wm. C. Brown Co.,
Publishers, Dubuque, IA.
Gorden, R.W, R.J. Beyers, E.P. Odum, and R.G. Eagon. 1969. Studies of a simple
laboratory microecosystem: Bacterial activities in a heterotrophic succession.
Ecology 50: 86-100.
156
-------
Gray, J.S. 1979. Pollution-induced changes in populations. Philos. Trans. R. Soc.
Lond. 286: 545-561.
Gray, J.S., and T.H. Pearson. 1982. Objective selection of sensitive species
indicative of pollution-induced change in benthic communities. I. Comparative
methodology. Mar. Ecol. Prog. Ser. 9: 111-119.
Green, R.H. 1979. Sampling design and Statistical Methods for Environmental
Biologists. John Wiley and Sons, New York.
Grime, J.P. 1977. Evidence for the existence of three primary strategies in
plants and relevance to ecological and evolutionary theory. Am. Natl. Ill: 1169-
1194.
Haegele, M.A., and R.H. Hudson. 1977. Reduction of courtship behavior induced
by DDE in male ringed turtle doves. Wilson Bull. 89: 593-601.
Hall, G.H. 1984. Measurement of nitrification rates in lake sediments:
Comparison of the nitrification inhibitors nitrapyrin and allylthiourea. Microb.
Ecol. 10: 25-36.
Hamaker, J.W. 1975. The interpretation of soil leaching experiments in Haque,
pp. 115-133 in Haque, R., and V.H. Freed, eds., Environmental Science Research,
Vol. 6. Environmental Dynamics of Pesticides. Plenum Press, New York.
Harrison, F.L., and I.M. Jones. 1982. An in vivo sister-chromatid exchange assay
in the larvae of the mussel Mvtilus edulis: Response to 3 mutagens. Mutat. Res.
105: 235-242.
Hatakeyama, S., and M. Yasuno. 1981. Effects of cadmium on the periodicity of
parturition and blood size Moina macrocopa (Cladocera). Environ. Pollut. Ser. A.
Ecol. Biol. 26: 111-120.
Havas, M. 1985. Aluminum bioaccumulation and toxicity to Daohnia magna in
soft water at low pH. Can. J. Fish. Aquat. Sci. 42: 1741-1748.
Havas, M., and T.C. Hutchinson. 1982. Aquatic invertebrates from the Smoking
Hills, N.W.T.: Effect of pH and metals on mortality. Can. J. Fish Aquat. Sci. 39:
890-903.
Haynes, R.J., and R.R. Sherlock. 1986. Gaseous losses of nitrogen, pp. 242-302 in
Haynes, R.J., ed., Mineral Nitrogen in the Plant-Soil System. Academic Press,
Orlando, FL
Heath, R.L. 1975. Ozone, pp. 23-55 in Mudd, J.B. and T.T. Kozlowski, eds.,
Response of Plants to Air Pollution. Academic Press, New York.
Heinz, G.H. 1976. Behavior of mallard ducklings from parents fed 3 ppm DDE.
Bull. Environ. Contam. Toxicol. 16: 640-645.
Heinz, G.H. 1979. Methylmercury: Reproductive and behaviorial effects on three
generations of mallard ducks. J. Wildl. Manage. 43(2): 394-401.
157
-------
Heinz, G.H., S.D. Haseltine, W.L. Reichel, and G.L. Hensler. 1983. Relationships
of environmental contaminants to reproductive success in red-breasted mergansers
Mergus serrator from Lake Michigan. Environ. Pollut. Ser. A. Ecol. and Biol. 32:
211-232.
Hellawell, J.M. 1977. Change in natural and managed ecosystems: Detection,
measurement and assessment. Proc. R. Soc. B. Biol. Sci. 197: 31-56.
Hendry, G.R., K.. Baalsrud, T.S. Traaen, M. Laake, and G. Raddum. 1976. Acid
precipitation: Some hydrobiological changes. Ambio 5: 224-227.
Herricks, E.E., and J. Cairns, Jr. 1982. Biological monitoring. Part Ill-Receiving
system methodology based on community structure. Water Res. 16: 141-153.
Hetrick, F.M., M.D. Knittel, and J.L. Fryer. 1979. Increased susceptibility of
rainbrow trout to infectious hematopoietic necrosis virus after exposure to copper.
Appl. Environ. Microbiol. 37: 198-201.
Hixon, M.A., and W.N. Brostoff. 1983. Damselfish as keystone species in reverse:
intermediate disturbance and diverity of reef algae. Science 220: 511-513.
Hoch, G., O.V.H. Owens, and B. Kok. 1963. Photosynthesis and respiration.
Arch. Biochem. Biophys. 101: 171-180.
Hoffman, D.J., O.H. Pattee, S.N. Wiemeyer, and B. Mulhern. 1981. Effects of lead
shot ingestion on delta-aminolevulinic acid dehydratase activity, hemoglobin
concentration, and serum chemistry in bald eagles. J. Wildl. Dis. 17: 423-431.
Hoffman, D.J., L. Sileoa, and H.C. Murray. 1984. Subchronic organophosphorus
ester-induced delayed neurotoxicity in mallards. Toxicol. Appl. Pharmacol. 75:
128-136.
Holcombe, G.W., D.A. Benoit, E.N. Leonard, and J.M. McKim. 1976. Long-term
effects of lead exposure on three generations of brook trout (Salvelinus
fontinalis). J. Fish Res. Board Can. 33: 1731-1741.
Holling, C.S. 1973. Resilence and stability of ecological systems. Annu. Rev.
Ecol. Syst. 4: 1-23.
Home, A.J. and C.R. Goldman. 1974. Suppression of nitrogen fixation by blue-
green algae in a eutrophic lake with trace additions of copper. Science 83: 409-
411.
Hose, J.E., J. Hannah, H. Puffer, and M.L. Landolt. 1984. Histologic and skeletal
abnormalities in benzo(a)pyrene-treated rainbow trout alevins. Arch. Environ.
Contam. Toxicol. 13: 675-684.
Hughes, P.R., J.E. Potter, and L.H. Weinstein. 1982. Effects of air pollution on
plant/insect interactions: Increased susceptibility of greenhouse grown soybeans
to the Mexican bean beetle following plant exposure to SO. Environ. Entmol. 11:
173-176.
158
-------
Hurlbert, S.H. 1975. Secondary effects of pesticides on aquatic ecosystems.
Residue Rev. 57: 81-148.
Hurlbert, S.H. 1984. Pseudoreplication and the design of ecological field
experiments. Ecol. Monogr. 54: 187-211.
Hutton, M. 1980. Metal contamination of feral pigeons Columbia livia from the
London area: Part 2. Biological effects of lead exposure. Environ. Pollut. Ser.
A. Ecol. Biol. 22: 281-293.
Hynes, R.K., and R. Knowles. 1983. Inhibition of chemoautotrophic nitrification
by sodium chlorate and sodium chlorite: a reexamination. Appl. Environ.
Microbiol. 45: 1178-1182.
Jackim, R. 1973. Influence of lead and other metals on fish delta-aminolevulinate
dehydrase activity. J. Fish Res. Board Can. 30: 560-562.
Jensen, A.L., and J.S. Marshall. 1983. Toxicant-induced fecundity compensation:
A model of population responses. Environ. Manage. 7(2): 171-175.
Johnson, R.A. 1981. Application of the guild concept to environmental impact
analysis of terrestrial vegetation. J. Environ. Management 13: 205-222.
Jones, M.B. 1975. Synergistic effects of salinity, temperature and heavy metals
on mortality and osmoregulation in marine and estuarine isopods (Crustacea).
Mar. Biol. 30: 13-20.
Kania, H.J., and J. O'Hara. 1974. Behaviorial alterations in a simple predator-
prey system due to sublethal exposure to mercury. Trans. Am. Fish. Soc. 103:
134-136.
Keeney, D.R. 1980. Prediction of soil nitrogen availability in forest ecosystems:
A literature review. Forest Sci. 26: 159-171.
Keeney, D.R. 1982. Nitrogen-availability indices, pp. 71 1-733 in Page, C.A.L., ed.,
Methods of Soil Analysis. Part 2. Chemical and Microbiology Properties. Am.
Soc. Agron., Madison, WI.
Kelso, J.R.M. 1977. Density, distribution and movement of Nipigon Bay Fishes in
relation to a pulp and paper mill effluent. J. Fish Res. Board Can. 34: 879-885.
King, K.A., D.H. White, and C.A. Mitchell. 1984. Nest defense behavior and
reproductive success of laughing gulls sublethally dosed with parathion. Bull.
Environ. Contam. Toxicol. 33: 499-504.
King, P.H., and P.L. McCarty. 1968. A chromatographic model for predicting
pesticide migration in soil. Soil Sci. 106: 248.
Koenig, C.C., R.J. Livingston, and C.R. Cripe. 1976. Blue crab mortality:
Interaction of temperature and DDT residues. Arch. Environ. Contam. Toxicol. 4:
119-128.
159
-------
Koike, I., and A. Hattori. 1978. Simultaneous determinations of nitrification and
IS
nitrate reduction in coastal sediments by a JJN dilution technique. Appl. Environ.
Microbiol. 35: 853-857.
Kormondy, E.J. 1969. Concepts of Ecology. Prentice-Hall Biological Science
Series. Prentice-Hall, Inc., Englewood Cliffs, NJ.
Krebs, C.J. 1985. Ecology: The Experimental Analysis of Distribution and
Abundance. 3rd Edition. Harper and Row, New York.
Kynard, B. 1974. Avoidance behavior of insecticide susceptible and resistant
populations of mosquitofish to four insecticides. Trans. Am. Fish Soc. 103: 557-
561.
Lack, D. 1954. The Natural Regulation of Animal Numbers. Clarendon Press,
Oxford.
Landres, P.B. 1983. Use of the guild concept in environmental impact
assessment. Environ. Manage. 7(5): 393-398.
Leivestad, H., G. Hendry, I.P. Muniz, and E. Snekvik. 1976. Effects of acid
precipitation on freshwater organisms, pp. 87-111 in Braekke F.H., ed., Impact of
Acid Precipitation on Forest and Freshwater Ecosystems in Norway. Research
Rep. 6/76. The SNSF Project.
Levin, S.A., K.D. Kimball, W.H. McDowell, and S.F. Kimball. 1984. New
perspectives in ecotoxicology. Environ. Manage. 8(5): 375-442.
Liang, C.N., and M.A. Tabatabai. 1978. Effects of trace elements on nitrification
in soils. J. Environ. Qual. 7(2): 291-293.
Lighthart, B. 1980. Effects of certain cadmium species on pure and litter
populations of microorganisms. Antonie van Leeuwenhoek. 46: 161-167.
Likens, G.E., F.H. Bormann, N.M. Johnson, D.W. Fisher, and R.S. Pierce. 1970.
Effects of forest cutting and herbicide treatment on nutrient budgets in the
Hubbard Brook watershed ecosystem. Ecol. Monogr. 40(1): 23-47.
Limburg, K.E., C.C. Harwell, and S.A. Levin. 1984. Principles for estuarine
impact assessment: Lessons learned from the Hudson River and other estuarine
experiences. Ecosystem Res. Center. Cornell Univ., Ithaca, NY. ECR-024. 75 pp.
Livingston, R.J. 1977. Review of current literature concerning the acute and
chronic effects of pesticides on aquatic organisms. CRC Crit. Rev. Environ.
Control. 7: 325-351.
Lomnicki, A., E. Bandola, and K. Kankowska. 1968. Modification of the Wiegert-
Evans method for estimation of net primary production. Ecology. 49(1): 147-149.
Lorz, H.W., and B.R. McPherson. 1976. Effects of copper or zinc in fresh water
on the adaption to sea water and ATPase activity, and the effects of copper on
160
-------
migratory disposition of Coho salmon (Qncorhvnchus kisutch). J. Fish Res. Board
Can. 33: 2023-2030.
Lotti, M., and M.K. Johnson. 1978. Neurotoxicity of organophosphorus pesticides:
Predictions can be based on in vitro studies with hen and human enzyme. Arch.
Toxicol. 41: 215-221.
Ludke, J.L., E.F. Hill, and M.P. Dieter. 1975. Cholinesterase (ChE) response and
related mortality among birds fed ChE inhibitors. Arch. Environ. Contam. Toxicol.
3: 1-21.
Macek, K., W. Birge, F.L. Mayer, A.L. Buikema, Jr., and A.W. Maki. 1978.
Discussion session synposis, pp. 27-32 in Cairns, J., K.L. Dickson, and A.W. Maki,
eds., Estimating the Hazard of Chemical Substances to Aquatic Life. Am. Soc.
Testing Materials, Philadelphia, PA.
Maki, A.W. 1979. Respiratory activity of fish as a predictor of chronic fish
toxicity values for surfactants, pp. 77-95 in Marking, L.L., and R.A. Kimmerle,
eds., Aquatic Toxicology. ASTM STP 667. Amer. Soc. for Testing Materials,
Philadelphia, PA.
Maki, A.W., and H.E. Johnson. 1976. Evaluation of a toxicant on the metabolism
of model stream communities. J. Fish. Res. Board Can. 33: 2740-2746.
Martin, A. 1983. Assessment of the effects of chemicals on the reproductive
functions of reptiles and amphibians, pp. 405-413 in Vouk, V.B., and P.J. Sheehan,
eds., Methods for Assessing the Effects of Chemicals on Reproductive Functions.
John Wiley and Sons, New York.
Mauck, W., P. Mehrle, and F.L. Mayer. 1978. Effects of polychlorinated biphenyl
Aroclor 1254 on growth, survival and bone development in brook trout (Salvelinus
fontinalisl. J. Fish Res. Board Can. 35: 1084-1088.
May, R.M. 1986. The search for pattern in the balance of nature: Advances and
retreats. Ecology 67: 1115-1126.
McCain, B.B., K.V. Pierce, S.R. Willings, and B.S. Miller. 1977. Hepatomas in
marine fish from an urban estuary. Bull. Environ. Contam. Toxicol. 18: 1-2.
McClenahen, J.R. 1978. Community changes in a deciduous forest exposed to air
pollution. Can. J. For. Res. 8: 432-438.
McGreer, E.R. 1979. Sublethal effects of heavy metal contaminated sediments on
the bivalve Macoma balthica (L.). Mar. Pollut. Mar. 10: 259-262.
McKim, J.M. 1985. Early life stage toxicity tests, pp. 58-95 in Rand, G.M., and
S.R. Petrocelli, eds., Fundamentals of Aquatic Toxicology. Hemisphere Publ. Co.,
Washington, DC.
McKim, J.M., and D.A. Benoit. 1971. Effects of long-term exposures to copper
on survival, growth, and reproduction of brook trout (Salvelinus fontinalis). J.
Fish Res. Board Can. 28: 655-662.
161
-------
McLaughlin, S., and O.U. Braker. 1985. Methods for evaluating and predicting
forest growth responses to air pollution. Experientia 41: 310-319.
McLaughlin, S.B. 1985. Effects of air pollution on forests. A critical review. J.
Air Pollut. Control Assoc. 35: 512-534.
McLenahen, J.R. 1978. Community changes in a deciduous forest exposed to air
pollution. Can. J. For. Res. 8: 432-438.
McNaughton, S.J. 1977. Diversity and stability of ecological communities. A
comment on the role of empiricism in ecology. Amer. Nat. Ill: 515-525.
McNaughton, S.J., and L.L. Wolf. 1979. General Ecology. Holt, Rinehart and
Winston, New York. pp. 1-702.
Mearns, A.J., and M.J. Sherwood. 1974. Environmental aspects of fin erosion and
tumors in southern California dover sole. Trans. Am. Fish Soc. 103: 799-810.
Meyers, T.R., and J.D. Hendricks. 1982. A summary of tissue lesions in aquatic
animals induced by controlled exposures to environmental contaminants,
chemotherapeutic agents, and potential carcinogens. Mar. Fish Rev. 44: 1-17.
Miles, L.J., and G.R. Parker. 1980. Effects of cadmium and a one-time drought
stress on survival, growth, and yield of native plant species. J. Environ. Qual. 9:
278-282.
Miller, J.E., J.J. Hassett, and D.E. Koeppe. 1977. Interaction of lead and cadmium
on metal uptake and growth of plant species. J. Environ. Qual. 6: 18-20.
Monk, D.C. 1983. The uses and abuses of ecotoxicology. Mar. Pollut. Bull. 14(8):
284-288.
Moore, J.C., and H.W. Hunt. 1988. Resource compartmentation and the stability
of real ecosystems. Nature 333: 261-263.
Moore, M.N., and A.R.D. Stebbing. 1976. The quanitative cytochemical effects of
three metals on a lysosomal hydrolase of a hydroid. J. Mar. Biol. Ass. U.K. 56:
995-1005.
Moriarty, F. 1983. Ecotoxicology. Academic Press, New York.
Murray, B.G. 1979. Population Dynamics: Alternative Models. Academic Press,
New York.
Newbould, P.J. 1967. Methods for Estimating the Primary Production of Forests.
IBP Handbook No. 2. Blackwell Scientific Publications. Oxford.
Nur, N. 1984. The consequences of brood size for breeding blue tits. II.
Nestling weight, offspring survival and optimal brood size. J. Anim. Ecol. 53:
497-517.
162
-------
O'Neill, R.V., B.S. Ausmus, D.R. Jackson, R.I. Van Hook, P. Van Voris, C.
Washburne, and A.P. Watson. 1977. Monitoring terrestrial ecosystems by analysis
of nutrient export. Water Air Soil Pollut. 8: 271-277.
Odum, E.P. 1971. Fundamentals of Ecology. W.B. Saunders Co., Philadelphia, PA.
pp. 1-574.
Pagano, G., M. Cipollaro, G. Corsale, A. Esposito, E. Ragucci, G.G. Giordano, and
N.M. Trieff. 1986. The sea urchin: Bioassay for the assessment of damage from
environment contaminants, pp. 66-91 in Cairns, J., ed., Community Toxicity
Testing, A Symposium Sponsored by ASTM Committee D-19 on Water. Colorado
Springs, CO. 6-7 May 1985. STP 920. Am. Soc. Testing Materials, Philadelphia,
PA.
Paine, R.T. 1966. Food web complexity and species diversity. Am. Nat. 100(910):
65-75.
Paine, R.T. 1974. Intertidal community structure: experimental studies on the
relationship between a dominant competitor and its principal predator. Oecologia.
15: 93-120.
Patil, S.G., D.K. Harshey, and D.F. Singh. 1985. The effects of DDT on primary
production in a pond ecosystem. Comp. Physiol. Ecol. 10: 139-140.
Peakall, D.B. 1983. Methods for assessment of the effects of pollutants on avian
reproduction, pp. 345-363 in Vouk, V.B., and P.J. Sheehan, eds., Methods of
Assessing the Effects of Chemicals on Reproductive Functions. SCOPE 20. John
Wiley and Sons, New York.
Pearson, T.H., J.S. Gray, and P.J. Johannessen. 1983. Objective selection of
sensitive species indicative of pollution-induced change in benthic communities. 2.
Data analyses. Mar. Ecol. Prog. Ser. 12: 237-255.
Peet, R.K. 1974. The measurement of species diversity. Ann. Rev. Ecol. Syst. 5:
285-307.
Pesch, G.G., and C.E. Pesch. 1980. Neanthes arenaceodentata (Polychaeta:
Annelida), a proposed model for marine genetic toxicology. Can. J. Fish Aquat.
Sci. 37: 1225-1228.
Petersen, R.C. 1986. Population and guild analysis for interpretation of heavy
metal pollution streams, pp. 180-198 in Cairns, J. ed., Community Toxicity Testing.
STP 920. Am. Soc. Testing Materials, Philadelphia, PA.
Peterson, B.J. 1980. Aquatic primary productivity and the ^C-C02 method: A
history of the productivity problem. Annu. Rev. Ecol. Syst. 11: 359-385.
Pielou, E.C. 1966. The measurement of diversity in different types of biological
collections. J. Theor. Biol. 13: 131-144.
Pimm, S.L. 1984. The complexity and stability of ecosystems. Nature 307: 321 -
326.
163
-------
Poysa, H. 1983. Resource utilization pattern and guild structure in a waterfowl
community. Oikos. 40: 295-307.
Rattner, B.A., V.P. Eroschenko, G.A. Fox, D.M. Fry, and J. Gorsline. 1984. Avian
endocrine response to environmental pollutants. J. Exp. Zool. 232: 683-689.
Read, P.A., T. Renshaw, and K.J. Anderson. 1978. Pollution effects on intertidal
macrobenthic communities. J. Appl. Ecol. 15: 15-31.
Reinert, R.A., and H.W. Spurr. 1972. Differential effects of fungicides on ozone
injury and brown spot disease of tobacco. J. Environ. Qual. 1: 450-452.
Renfro, J.L., B. Schmidt-Nielsen, D. Miller, D. Benos, and J. Allen. 1974. Methyl
mercury and inorganic mercury: Uptake, distribution and effect on osmoregulatory
mechanisms in fishes, pp. 101-122 in Vernberg, F.J., and W.B. Vernberg, eds.,
Pollution and Physiology of Marine Organisms. Academic Press, New York.
Reynolds, C.S., and A.E. Walsby. 1975. Water-blooms. Biol. Rev. 50: 437-481.
Ricklefs, R.E. 1973. Ecology. Chiron Press, New York.
Ricklefs, R.E. 1987. Community diversity: Relative roles of local and regional
processes. Science 235: 167-171.
Rodsaether, C.M., J.A. Olafsen, J. Raa, K. Myhre, and J.B. Steen. 1977. Copper
as initiating factor of vibriosis (Vibrio aneuillarum) in eel (Anguilla anguilla). J.
Fish Biol. 10: 17-21.
Rolston, D.E., A.N. Sharpley, O.W. Toy, and F.E. Broadbent. 1982. Field
measurements of denitrification. III. Rates during irrigation cycles. Soil Sci.
Soc. Am. J. 42: 863-869.
Root, R.B. 1967. The niche exploitation pattern of the blue-gray gnatcatcher.
Ecol. Monogr. 37(4): 318-350.
Rose, F.L., and J.C. Harshbarger. 1977. Neoplastic and possible related skin
lesions in neotenic tiger salamanders from a sewage lagoon. Science 196: 315-317.
Rosenberg, R., and J.D. Costlow. 1976. Synergistic effects of cadmium and
salinity with constant and cycling temperatures on the larval development of two
estuarine crab species. Mar. Biol. 38: 291-303.
Royce, W.F. 1972. Introduction to the Fishery Sciences. Academic Press, New
York.
Sahrawat, K.L. 1983. Nitrogen availability indexes for submerged rice soils. Adv.
Agron. 36: 415-451.
Sakamoto, M., M.M. Tilzer, R. Gachter, H. Rai, Y. Collos, P. Tschumi, P. Berner,
D. Zbaren, J. Zbaren, M. Dokulil, P. Bossard, U. Uehlinger, and E.A. Nusch. 1984.
Joint field experiments for comparisons of measuring methods of photosynthetic
production. J. Plankton Res. 6(2): 365-383.
164
-------
Saunders, R.L., and J.B. Sprague. 1967. Effects of copper-zinc mining pollution
on a spawning migration of Altantic salmon (Salmo salar). Water Res. 1: 419-432.
Schindler, D.W. 1987. Detecting ecosystem responses to anthropogenic stress.
Can. J. Fish Aquat. Sci. 44(Suppl. 1): 6-25.
Schindler, D.W., K.H. Mills, D.F. Malley, D.L. Findlay, J.A. Shearer, I.J. Davis,
M.A. Turner, G.A. Linsey, and D.R. Cruikshank. 1985. Long-term ecosystem
stress: the effects of years of experimental acidification on a small lake.
Science 228: 1395-1401.
Schmidt, R.H. 1986. Community-level effects of coyote population reduction, pp.
49-65 in Cairns, J., Jr., ed., Community Toxicity Testing. STP 920. Am. Soc.
Testing Materials, Philadelphia, PA.
Schofield, C.L. 1976. Acid precipitation effects on fish. Ambio 5: 228-230.
Schreck, C.B. 1981. Stress and compensation in teleostean fishes: response to
social and physical factors, pp. 295-321 in Pickering, A.D., ed., Stress and Fish.
Academic Press, New York.
Severinghaus, W.D. 1981. Guild theory development as a mechanism for assessing
environment impact. Environ. Manage. 5(3): 187-190.
Sheehan, P.J. 1984a. Effects on individuals and populations, pp. 23-50 in
Sheehan, P.J., D.R. Miller, G.C. Butler, and P. Bourdeau, eds., Effects of Pollutants
at the Ecosystem Level. John Wiley and Sons, New York.
Sheehan, P.J. 1984b. Effects on community and ecosystem structure and
dynamics, pp. 51-99 in Sheehan, P.J., D.R. Miller, G.C. Butler and P. Bourdeau,
eds., Effects of Pollutants at the Ecosystem Level. John Wiley and Sons, New
York.
Sheehan, P.J. 1984c. Functional changes in the ecosystem, pp. 101-145 in
Sheehan, P.J., D.R. Miller, G.C. Butler, and P. Bourdeau, eds., Effects of Pollutants
at the Ecosystem Level. John Wiley and Sons, New York.
Silvester, W.B. 1983. Analysis of nitrogen fixation, pp. 173-212 in Gordon, J.C.,
and C.T. Wheeler, eds., Biological Nitrogen Fixation in Forest Ecosytems:
Foundations and Applications. Nijhoff/Junk Publ., The Hague.
Sindermann, C.T. 1980. The use of pathological effects of pollutants in marine
environment monitoring programs in Mclntyre, A.D., and J.B. Pearce, eds.,
Biological Effects of Marine Pollution and the Problems of Monitoring. Rapp. P.-
V. Reun. Cons. Int. Explor. Mer. 179: 135-151.
Sladecek, V. 1963. The future of the saprobity system. Hydrobiologia 25: 518-
537.
Smith, V.H. 1983. Low nitrogen to phosphorus ratios favor dominance by blue-
green algae in lake phytoplankton. Science 221: 669-671.
165
-------
Solbe, J.F. 1977. Water quality, fish and invertebrates in a zinc polluted stream,
pp. 97-105 in Alabaster, J.S., ed., Biological monitoring of Inland Fisheries.
Applied Science Publ., London.
Sousa, W.P. 1984. The role of disturbance in natural communities. Annu. Rev.
Ecol. Syst. 15: 353-391.
Southwood, T.R.E. 1978. Ecological Methods. Chapman and Hall.
Sprague, J.B., P.F. Elson, and R.L. Saunders. 1965. Sublethal copper-zinc
pollution in a salmon river - a field and laboratory study, pp. 65-82 in Jaag, O.,
ed., Advances in Water Pollution Research. Pergamon Press, New York. (Cited in:
Sheehan, P.J., D.R. Miller, G.C. Butler, and P. Bordeau, eds., 1984. Effects of
Pollutants at the Ecosystem Level. John Wiley and Sons, New York.
Stanford, G. 1982. Assessment of soil nitrogen availability, pp. 651-688 in
Stevenson, F.J., ed., Nitrogen in Agricultural Soils. Am. Soc. Agron., Madison, WI.
Stauffer, J.R., and C.H. Hocutt. 1980. Inertia and recovery: An approach to
stream classification and stress evaluation. Water Res. Bull. 16(1): 72-78.
Stegeman, J.J. 1980. Mixed-function oxygenase studies in monitoring for effects
of organic pollution, in Mclntyre, A.D., and J.B. Pearce, eds., Biological Effects of
Marine Pollution and Problems of Monitoring. Rapp. P-V Reun. Cons. Int. Explor.
Mer. 179: 33-38.
Stewart-Oaten, A., W.W. Murdoch, and K.R. Parker. 1986. Environmental impact
assessment: "Pseudoreplication" in time? Ecology 67: 929-940.
Stokes, P.M. 1986. Ecological effects of acidification on primary producers in
aquatic systems. Water Air Soil Pollut. 30(1-2): 421-438.
Stoner, A.W., and R.J. Livingston. 1978. Respiration, growth and food conversion
efficiency and pinfish (Laeodon rhomboides) exposed to sublethal concentrations of
bleached kraft mill effluent. Environ. Pollut. 17: 207-217.
Sullivan, J.F., and G.J. Atchison. 1978. Predator-prey behaviour of fathead
minnows, Pimephales promelas and largemouth bass, Micropterus salmoides in a
model ecosystem. J. Fish Biol. 13: 249-253.
Sylvester, J.R. 1972. Effect of thermal stress on predator avoidance in sockeye
salmon. J. Fish Res. Board Can. 29: 601-603.
Szaro, R.C. 1986. Guild management: An evaluation of avian guilds as a
predictive tool. Environ. Manage. 10(5): 681-688.
Tagatz, M.E. 1976. Effect of mirex on predator-prey interaction in an
experimental estuarine ecosystem. Trans. Am. Fish Soc. 105: 546-549.
Tanner, J.T. 1978. Guide to the Study of Animals Populations. Univ. Tenn.
Press, Knoxville, TN.
166
-------
Tingey, D.T., R.A. Reinert, J.A. Dunning, and W.W. Heck. 1973. Foliar injury
responses of eleven plant species to ozone/sulfur dioxide mixtures. Atmos.
Environ. 7: 201-208.
TRI. 1988a. Ecological endpoint selection criteria, Office of Research and
Development, U.S. EPA, September, 1988, Washington, D.C.
TRI. 1988b. Ecological model selection criteria. Office of Research and
Development, U.S. EPA, September 1988, Washington, D.C.
TRI. 1988c. Ecological risk assessment issues Volume I - Ecological
measurements used in risk assessment: State of the science. Office of Research
and Development. U.S. EPA, September 1988, Washington, D.C.
TRI. 1988d. Ecological risk assessment issues Volume II - Ecosystem
stability/recovery report. Office of Research and Development, U.S. EPA,
September 1988, Washington, DC.
Usher, M.B. 1972. Developments in the Leslie matrix model, pp. 29-60 in Jeffers,
J.N.R, ed., Mathematical Models in Ecology. Blackwell Scientific Publ. Oxford.
Van Voris, P., R.V. O'Neill, W.R. Emanuel, and H.H. Shugart. 1980. Functional
complexity and ecosystem stability. Ecology 61(6): 1352-1360.
Verma, S.R., I.P. Tonk A.K. Gupta, and R.C. Dalela. 1981. In vivo enzymatic
alterations in certain tissues of Saccobranchus fossils following exposure to four
toxic substances. Environ. Pollut. Ser. A. Ecol. Biol. 26: 121-127.
Verner, J. 1984. The guild concept applied to management of bird populations.
Environ. Manage. 8(1): 1-13.
Visser, S., J. Fujikawa, C.L. Griffiths, and D. Parkinson. 1984. Effect of topsoil
storage on microbial activity, primary production and decomposition potential.
Plant and Soil 82: 41-50.
Walley, K.A., M.S.I. Khan, and A.D. Bradshaw. 1974. The potential for evolution
and heavy metal tolerance in plants. I. Copper and zinc tolerance in Aerostis
tenuis. Heredity 32: 309-319.
Warfar, W.V.M., P. LeCorre, and J.L. Birrien. 1983. Nutrients and primary
production in permanently well-mixed temperate coastal waters. Est. Cstl. Shelf
Sci. 17: 431-446.
Warren, C.E. 1971. Biology and Water Pollution Control. W.B. Saunders Co.,
Philadelphia, PA.
Weis, J., and P. Weis. 1987. Pollutants as developmental toxicants in aquatic
organisms. Environ. Health Perspect. 71: 77-85.
Wellings, S.R., C.E. Alpers, B.B. McCain, and B.S. Miller. 1976. Fin erosion
disease of the starry flounder (Platichthvs stellatusl in the estuary of the
Duwamish River, Seattle, Washington. J. Fish Res. Board Can. 33: 2577-2586.
167
-------
Wentsel, R., A. Mcintosh, W.P. McCafferty, G. Atchinson, and V. Anderson. 1977.
Avoidance response of midge larvae (Chironomus tentans't to sediment containing
heavy metals. Hydrobiology. 55: 171-175.
Westman, W.E. 1985. Ecology, Impact Assessment, and Environmental Planning.
John Wiley and Sons, New York. pp. 1-532.
Wiegert, R.G., and F.C. Evans. 1964. Primary production and the disappearance
of dead vegetation on an old field in southeastern Michigan. Ecology 45(1): 49-
63.
Winner, R.W., T. Kelling, R. Yeager, and M.P. Farrell. 1977. Effect of food type
on the acute and chronic toxicity of copper to Daphnia magna. Freshwater Biol.
7: 343-349.
Woodwell, G.M. 1970. The energy cycle of the biosphere. Sci. Amer. 223: 64-
74.
Woodwell, G.M., and D.B. Botkin. 1970. Metabolism of terrestrial ecosystems by
gas exchange techniques: The Brookhaven approach, pp. 73-85 in Reichle, D.E.
ed., Analysis of Temperate Forest Ecosystems. Springer, NY.
Wu, R.S.S. and C.D. Levings. 1980. Mortality, growth and fecundity of
transplanted mussell and barnacle populations near a pulp mill outfall. Mar.
Pollut. Bull. 11:11-15.
168
------- |