SCOPE 53 - Methods to Assess the Effects of Chemicals On Ecosystems

6.1 INTRODUCTION

The sources and types of chemical pollution to which freshwater systems can be exposed are many and varied. The origins of pollutants entering these systems range from those that are predominantly point sources (i.e., from sewage treatment works, fish farms, and industry) to those that are mainly diffuse, such as agricultural run-off and acid deposition. Some of these are effectively continuous, whereas others are intermittent or episodic.

Pollutants can also be characterized by their quality: general organic loading, specific organic toxicants, inorganic toxicants, and acids. (Holdgate, 1979). While most of the sources listed above produce mixtures of these characteristics, particular components are enhanced. Thus, sewage treatment works, fish farms, and farmyards produce general organic loading, but may also be the source of pesticides, and organic and inorganic toxicants. Industrial effluents generally are the source of specific toxicants, but might also lead to general organic loading.

Some consideration should also be given to the types of system exposed to these pollutants. Most freshwater systems flow, but some more rapidly than others; that is, lakes and ponds have a slow throughput, whereas that for rivers and streams is rapid. Because they flow rapidly, the so-called lotic systems have been used for the transport of materials, such as removing industrial pollutants from factories and organic effluents from sewage works (Hellawell, 1986).

A view exists that lotic systems that depend substantially on organic inputs from terrestrial ecosystems as a basis for their economy should have the capacity for self-cleansing of at least organic pollutants (Calow et al., 1990). However, this perception has been challenged recently (Royal Commission on Environmental Pollution, 1992). Of course, "still" or so-called lentic waters may also be important repositories of pollutants, and are certainly exposed to diffuse inputs from agricultural land leading to eutrophication (Harper, 1992).

Thus, the challenge for aquatic ecotoxicology is to develop methods that can not only assess but also estimate the ecological impact of chemicals on a variety of complex ecosystems under diverse and complicated circumstances (e.g., periodic perturbations or steady trickles, and mixtures of varying complexity). The predictive approach is important in risk assessment used in the regulation of chemical pollutants. Assessment, at times referred to as a "retrospective approach," is used to determine whether particular contaminants have an impact on specified natural ecosystems. This chapter reviews current methods, first predictive and then retrospective approaches, to address such problems. Initially, one must define what is to be measured, for it has relevance to both approaches.

6.2 WHAT TO MEASURE

Effects of contaminants can be identified across the entire biological hierarchy from subcellular systems to ecosystems. Moving down in scale, effects are usually measured more quickly and with more experimental rigour and control. Moreover, in this direction, systems are more general; most organisms contain DNA, but most ecosystems differ considerably in species composition. Since ecotoxicology is concerned with protecting ecological systems, this generally means consideration of populations and communities that compose ecosystems.

Considerable debate continues about the ecological relevance of observations in individual organisms and of the processes functioning within them (Calow, 1989). One view is that ecological processes are driven "bottom up," so that specific functional links can be made between effects observed at the organismic and suborganismic levels and those at the population-community levels. Thus, any impact on the physiological materials and energy flows in organisms can be linked with consequences for developmental rates, survival chances, and reproductive output; these then can be linked to consequences in population dynamics and, alternatively, to the role of a species or population within the community (Calow and Sibly, 1990). By contrast, the impact of a pollutant on keystone predators (Paine, 1988) is likely to alter the interaction between prey species in a way that could not have been predicted from ecotoxicological observations on the species individually.

The relative values of "bottom-up" and "top-down" regulation may vary from one ecosystem to another, and may depend upon the extent to which particular ecosystems are structured. Thus flowing-water systems that may be relatively unstructured, because of their temporal dynamism (Peckarsky, 1984; Hildrew and Townsend, 1987), might be subject predominantly to "bottom-up" control. This outcome could be a justification for use of organismic-suborganismic test systems in defining the hazards associated with chemicals to which the organisms are likely to be exposed. More research is needed to clarify these relationships.

Even for ecosystems difficult questions have to be considered to decide which attributes should be measured. Unlike for organisms (Calow, 1992), ecosystem "goal states" are unlikely, although states related to ecosystem stability and resilience are known to exist (Westman, 1978). Traditionally in ecotoxicology, measurements of effects have been made of physicochemical variability, species composition and diversity (i.e. Crossland and Wolff, 1985), and, less commonly, functional features (Odum, 1985; Cairns and Pratt, 1986). Yet food-web structure might be as important (Hildrew, 1992), as might be the analysis of functional components and their ability to maintain balance of energy flow and of material cycles within systems. Protecting species deemed important for humans might also be important.

6.3 PREDICTIVE TESTS

These theoretical considerations notwithstanding, single-species tests have dominated studies in aquatic ecotoxicology for the past 20 years (Maltby and Calow, 1989). Moreover, certain taxa, particularly the daphnids, have dominated these single-species studies (Maltby and Calow, 1989), because single species tests can be accomplished more quickly and easily and, in principle, daphnids are easily maintained in laboratory culture for such tests.

6.3.1 SINGLE-SPECIES TESTS

These tests can be classified according to either (1) the form of the response measured: discontinuous or quantal (i.e., death or cessation of a specified behaviour) or continuous (e.g., a reduction in growth or reproduction) or (2) the relationship between the time-course of the response and the generation time (GT) of the test organism, may be short or long relative to GT. Tests are usually referred to as acute (quantal and short term) or chronic (continuous or long term).

Most tests in aquatic ecotoxicology have been performed with a single species and of acute duration. Both acute and chronic test systems have been incorporated into legislation concerned with the assessment of ecological hazards of both new and existing chemicals and detailed descriptions of these standard test systems are provided by Persoone and Janssen (1993) and Solbe (1993). Their use has arisen largely from the convenience of some post hoc justification of relevance. Thus, the use of daphnids as test organisms can be justified by (1) the key roles that they and their taxonomic associates play in freshwater ecosystems such as zooplankton (interestingly, not often the case for lotic systems) and (2) their proven sensitivity relative to other freshwater species (Baudo, 1987).

Convenience has been a criterion in the choice of test systems, because it enables tests to be executed on a routine basis and to be standardized. The latter is paramount, since without the ability to get the same result from the same test system in different laboratories, something not always possible even for daphnids (Baird et al., 1989), the whole legal and scientific basis of the test is undermined (Calow, 1993).

For the following reasons, attention must also be paid to the ecological relevance of the test. First, demonstrating that the test species is pivotal to estimating broad environmental consequences, its ecological legitimacy would be strengthened. However, often too little understanding exists of the structure of freshwater ecosystems to know if keystone species exist at all (Hildrew, 1992). Second, ecological relevance might be claimed if the test species were known to be more sensitive than most other species. The daphnids appear to be quite sensitive, but much variation in sensitivity to numerous toxicants is apparent even between genotypes of Daphnia magna (Baird et al., 1991). The concept of the more sensitive indicator species seems to have been undermined by Cairns (1986). Much more needs to be known about the distribution of sensitivities to toxicants both within and between species in communities (Kooijman, 1987; Van Straalan and Denneman, 1989).

Moreover, the emphasis on pelagic organisms also needs careful consideration. Epibenthic and sediment systems may be as important, if not more so, in freshwaters as are pelagic ones; yet, because of difficulties in manipulation, sediments have been relatively neglected. Sediment ecotoxicity is nevertheless developing, as indicated by useful information provided by Reynoldson and Day (1993).

Given the predominance of acute tests and the interest in using them to set priorities and environmental standards through the application of safety factors, the relationships between acute and chronic responses must be well understood. Pioneering work on this topic reported by Sloof et al. (1986) and carefully performed studies reported by ECETOC (1993) support the general view that chronic effects are approximately equal to LC₅₀/LC₄₀. Since this finding is based on a correlation, no guarantee exists that this relationship is likely to apply to any particular chemical, particularly given differences in activation of toxicants (Maltby and Calow, 1989). To increase scientific confidence, the mechanistic basis of the relationship between acute short-term and chronic longer-term responses needs to be supported by additional data.

6.3.2 MULTI-SPECIES TESTS

Two reasons that multi-species systems might bring more ecological relevance to ecotoxicity tests include:

1. The test results may expose crucial emergent effects that arise out of species interactions.
2. The test findings from testing in complex milieu may provide insight into the distribution (fate) of toxicants between various compartments of the systems (i.e., sediment and water column) and hence on bioavailability.

If a divergence occurs between single- and multi-species responses, knowing whether it arises from basis l or 2 above would prove useful.

Multi-species tests can certainly be conducted for a variety of scales and using a number of alternative microcosms and macrocosms. However, no generally accepted distinction exists among systems based solely on size (SETAC Europe, 1991; SETAC, 1992). They can be based indoors or outdoors, and they can be closed or open. Closed systems are observed more frequently than are open ones, largely because of the complexity of the "plumbing" needed for open systems; however, both indoor and outdoor artificial streams have been constructed and used (Cairns and Cherry, 1993).

Mesocosms have been used in a legislative context: for pesticide registration in the US where the USEP A can require a mesocosm analysis if information from other tests is insufficient to "negate a presumption of risk" or indicates that significant biological populations will be at risk. The best practice in this context has been evaluated in Workshops organized by the Society of Environmental Toxicology and Chemistry for the use of both microcosms (SETAC, 1992) and mesocosms (SETAC Europe, 1991). Standard multi-species ecotoxicological test systems have been described, and reviewed by Cairns and Cherry (1993).

For the effort and expense involved in establishing and managing these systems, careful consideration needs to be given to their usefulness and relevance. Three major problems are apparent: (1) because of the size and complexity of these systems, data obtained from them are difficult to replicate, so rigorous statistical analysis is not always possible; (2) methods of construction differ sufficiently to increase the difficulty in reproducing results; and (3) uncertainty exists over exactly which properties should be measured, as it does for natural ecosystems.

Multi-species systems can either be set up by (1) sampling from natural systems (e.g. enclosing a sample of plankton in a closed flask; Leffler, 1981) or (2) colonization (e.g., leaving the pond or stream open for colonisation; Eaton et al., 1985; Crossland and Wolff, 1985); or (3) by deliberate construction (Taub and Read, 1982). Techniques (1) and (2) cannot guarantee reproducibility, and the enclosing effects in (1) and (3) could lead to peculiar dynamics that interact with or potentially mask other perturbations. These effects need to be thoroughly investigated in control systems. These features mean that standardization is probably difficult to achieve, and ecological relevance may be illusory.

6.4 APPLICATION OF TOXICANTS

Most of the single- and multi-species test systems apply the toxicant continuously; however, episodic events are also common in freshwater systems. A few studies have been conducted in which the toxicant was applied as a pulse. Thurston et al. (1981) studied the effects of fluctuating ammonium on trout, and concluded that, over a 96-hour period, they could withstand a fixed concentration of ammonia better than fluctuating concentrations with the same mean. By contrast, Brown et al. (1969) found that, with either zinc or mixtures of zinc and ammonia, no significant difference in survival was observed for trout exposed either continuously or alternately to low and high concentrations.

To distinguish between standard toxicity tests and those including post-exposure mortality, Abel (1980) proposed the use of median lethal exposure time (i.e., time required to kill half the population within a predetermined post-exposure period). A similar index, median-post-exposure-lethal time (i.e., time from end of exposure period by which 50 percent of the animals are dead), has been proposed by Pascoe and Shazili (1986).

Within the context of episodic events, more careful consideration needs to be given to the effects of amplitude, duration, and frequency of pulse of the dose on populations and communities, not only in terms of survivorship but also of developmental rates and reproductive output.

A further complication to be addressed is that test substances are most often applied singly to test systems, because complex mixtures' are more difficult to handle and because regulations focus predominantly on single substances. All the most appropriate mixtures and conditions in the field for carrying out tests would be impossible to imagine, let alone construct experimentally. Yet interest in the effects of complex effluents on natural systems has directed attention to this problem (USEPA, 1986; de March, 1987a, 1987b; McCarty et al., 1992). Many mixtures of organic chemicals appear to act primarily by simple, sometimes joint, actions so that the contribution of the components to the total mixture are additive (Broderius and Kahl, 1985; Deneer et al., 1988); however, not all toxicants produce this effect (Marking, 1985).
Understanding of basic mechanisms of ecotoxicological action needs to be extended to ascertain how toxicants exert their primary effects and how toxicants might interfere with or enhance each other's actions.

6.5 RETROSPECTIVE ASSESSMENT

Assessments can be carried out on different scales: (1) surveying the general state of fresh waters regionally, nationally (National Rivers Authority, 1991) and, locally; (2) assessing impacts of contamination at a particular site; and (3) monitoring effluent on a site on a continuous basis to assess quality. In principle, all these activities can be carried out using effects at all biological levels described previously; often, however, organismic and suborganismic responses are convenient to use in limited studies, whereas community and ecosystem properties are more convenient in the more extensive ones. In all, ecological relevance remains a critical determinant.

Chemical analyses on their own are rarely adequate in assessments, because the dynamic complexity of inputs means that samples taken at specific times (the norm for chemical analysis) may be misleading. Also, the qualitative complexity of inputs implies extreme difficulty in assessing ecological impact from a database comprised of only those substances present. Biological systems, however, integrate effects over time of complex mixtures of contaminants.

Two key issues in designing and interpreting retrospective biological assessments are: (1) what to measure and (2) how to identify cause from correlation (related to problems of design of sampling programmes).

6.6 THE MEASUREMENT PROBLEM REVISITED

Are there characteristics of healthy ecosystems that can be defined and against which divergences can be gauged? As noted previously, while "goal states" are not appropriately applied to ecosystems, steady or stable states might be. One approach based on an attempt to define a characteristic assemblage of benthic invertebrates for a particular river habitat is the River Invertebrate Predictive And Classification System (RIVPACS) (Armitage et al., 1987). This system defines relationships between physicochemical variables and species assemblages on the basis of an initial survey of clean sites, and then uses the model based on the results to estimate expected assemblages at particular sites from observation of the physico-chemical variables. The deviation of observed from predicted findings is considered an index of disturbance that can be used for both surveys and impact assessments. However, this and similar approaches developed for other taxa (i.e., fish) are based on correlations between what is presumed to be dependent and independent variables (Bayley and Li, 1992); thus, deviations from expectation that correlate with an extra variable (i.e., presence of a contaminant) need not be causally related.

Another approach focuses not so much on the details of species occurrence as on the general properties of ecological diversity (Wilhm, 1972). Disputes on how this property should be measured remain (Magurran, 1988), and certainly no a priori theory that predicts what levels of diversity might be expected in particular habitats exists. How diversity should respond¾either positively or negatively¾to stress is simply unclear; that is, diversity of plankton appears to diminish continuously with organic enrichment, and to increase and then diminish for benthic invertebrates (Calow, 1984).

The expectations for the functional organization of ecosystems and how it might change under stress are defined relatively easily: for any stable ecosystem, the input of energy (either autochthonous or allochthonous) should balance output. Organic loading can destabilize the balance by causing input to exceed output; alternatively, toxic stress can destabilize this process by causing output to exceed input, resulting from the increased metabolic work needed to counteract this kind of pollution. This situation leads to changes in ecosystem functional ratios involving production, biomass, and respiration (Odum, 1985). Such changes will also be closely linked to the cycling and spiralling of specific nutrients, a topic beginning increasingly to attract attention (Newbold, 1992).

The extent to which structure and function are coupled in ecosystems is problematic (Cairns and Pratt, 1986), so that the one probably cannot be used as a surrogate for the other. This situation means that a choice has to be made either to measure either structural or functional responses, but not both, or to combine structural and functional indices as a basis for inferences about divergences from the norm (Karr, 1991). For the latter approach, care has to be taken to ensure that the different indices are not autocorrelated (Barbour et al., 1992).

A close coupling exists between functioning and trophic structure embodied in food webs. The properties of these are coming under increased scrutiny. Structural-functional attributes are likely to be identified for use in defining relative stability as well as resilience to disturbance in ecosystems (De Angelis, 1992). Benthic invertebrates have been classified into trophic functional groups (Cummins and Klug, 1979), and predictable shifts in the relative abundance of these along rivers (from shredders to collectors) have been claimed as part of the River Continuum Concept (Cummins, 1992). Thus, deviations from these could signal disturbances. Two caveats persist: (1) predictions are not yet precisely defined, and the RCC theory is not universally accepted (Townsend, 1989); and (2) causation should not be mistaken for correlation.

If ecosystem properties and processes are difficult to measure, an alternative is to use alterations in individual species populations as indicators of change. Searching for key species may be tempting; however, "key" species may have several definitions: sensitivity, functional role, or impact on the dynamics of other species. Approaches relying on indicator species (involving presence or absence of a single species; ratios of sensitive to tolerant species; and biotic scores) are based largely on the construct of "sensitive" species, a philosophy whose difficulties have already been noted (Metcalfe, 1989). Moreover, separated populations within species can display considerable variation in response to pollutants based possibly on differential local conditions and selection pressures (Maltby et al., 1987). Shifts from sensitive to tolerant forms of organisms (both inter- and intra-species) have been used independently as indicators of stress as is the case for the Pollution Induced Community Tolerance (PICT) system (Blanck et al., 1988).

6.7 CAUSE OR ONLY CORRELATION?

In complex field situations, defining confidently causal links from direct observations is always difficult. Usually a change is observed to be correlated with a variable of interest (e.g., a reduction in diversity or density downstream of an effluent), but to say that one caused the other is unjustified without further research, because important contributing variables may be present but hidden from view.

However, without some a priori yardstick against which to compare observed systems, divergences from normal have to be judged comparatively by reference to the state of the system before the presumed disturbances or with similar systems in habitats not exposed to the pollutants of interest (e.g., upstream and downstream comparisons). Some profound problems are associated, however, with being able to distinguish between natural variation through space and time and that caused by the putative disturbance¾the so-called sampling problem of pseudodesign (Norris et al., 1992).

The only way to identify causal links with more certainty is to associate the field observations with an experimental programme. Testing ecological hypotheses derived from field observations in more rigorously controlled experimental circumstances can be used widely in ecology, and ought to be used more in ecotoxicology. For example, conclusions about putative interactions between contaminants and organisms in field situations could be tested using classical dose- response experiments. The reverse is sometimes undertaken; that is, quantitative dose-response comparisons can result from simple laboratory tests using the more complex microcosms, mesocosms, and with true ecosystems (Crossland et al., 1992).

6.7.1 PLANTED SYSTEMS

Planted systems are often used to monitor effluents or even whole rivers in which physiological or behavioral features are used to signal change in systems that range from caged fish to microbes (Pascoe and Edwards, 1984).

Planted systems can also be used in upstream-downstream, before-after studies as in situ bioassays. They avoid problems of pseudodesign, because in principle all replicates start in the same condition. Of course, these systems raise questions of relevance. Careful choice of a system and of effects may help to overcome this difficulty. The Gammarus Scope for Growth Bioassay (Naylor et al., 1989) serves to illustrate this point: the organism, G. pulex was chosen as the test animal because of its relative sensitivity to a wide range of toxicants and because of its important role in the decomposer food chain is of considerable importance in the economy of flowing-water systems. Furthermore, a physiological response involving the net energy available for somatic and reproductive products was chosen, because it is sensitive to contaminants, can be linked causally to population dynamics (Calow and Sibly, 1990), and is obviously related to important ecosystem processes in the decomposer food chain.

6.8 REFERENCES

Abel, P.D. (1980) A new method for assessing the lethal impact of short term high level discharges of pollutants on aquatic animals. Progr. Water Technol. 13, 347-352.

Armitage, P.D., Gunn, R.J.M., Furse, M.J., Wright, J.F., and Moss, D. (1987) The use of prediction to assess macroinvertebrate response to river regulation. Hydrobiologia 144, 25-32.

Baird, D.J., Barber, I., Bradley, M., Calow, P., and Soares, A.M.V.M. (1989) The Daphnia bioassay: A critique. Hydrobiologia 188/189, 403-406.

Baird, D.J., Barber, I., Bradley, M., Soares, A.M.V.M., and Calow, P. (1991) a comparative study of genotype sensitivity to acute toxic stress using clones of Daphnia magna Straus. Ecotoxicol. Environ. Safety 21, 257-265.

Barbour, M.T., Plafkin, J.L., Bradley, B.P., Graves, C., and Wisseman, R.W. (1992) Evaluation of EPA's rapid bioassessment benthic metrics: metric redundancy and variability among reference sites. Environ. Toxicol. Chem. II, 437-449.

Baudo, R. (1987) Ecotoxicological testing with Daphnia. In Peters, R.H., and de Bernardi, R. (Eds.) Daphnia, Vol. 45, pp. 461-482. Mem. Ist. Ital. Idrobiol.

Bayley, P.B., and Li, H.W. (1992) Riverine fishes. In: Calow, P., and Petts, G. (Eds.) Rivers Handbook, Vol. 1, pp. 251-281. Blackwell Scientific, Oxford. ,

Blanck, H., Wängberg, S.A., and Molander, S. (1988) Pollution-induced community tolerance¾a new ecotoxicological tool. In: Cairns, J., and Pratt, J.R. (Eds.) Functional Testing of Aquatic Biota for Estimating Hazard of Chemicals, pp. 219-230. AST STP 988. American Society for Testing and Materials, Philadelphia.

Broderius, S., and Kahl, M. (1985) Acute toxicity of organic chemical mixtures to the fathead minnow. Aquat. Toxicol. 6, 307-322. .

Brown, V.M., Jordan, D.H.M., and Tiller, B.A. (1969) The acute toxicity of rainbow trout in fluctuating concentrations of mixtures of ammonia, phenol and zinc. J. Fish Biol. 1, 1-9.

Cairns, J. (1986) The myth of the most sensitive species. Bioscience 36, 670-672.

Cairns, J., and Cherry, D.S. (1993) Freshwater multispecies test systems. In: Calow, P. (Ed.) Handbook of Ecotoxicology, Vol. 1. Blackwell Scientific, Oxford.

Cairns, J., and Pratt, J.R. (1986) On the relation between structural and functional analysis of ecosystems. Environ. Toxicol. Chem. 5, 785-786.

Calow, P. (1984) Factors affecting the diversity of littoral invertebrates: A hypothesis. Ing. Sanit. 30, 41-46.

Calow, P. (1989) The choice and implementation of environmental bioassays. Hydrobiologia 188/189, 61-64.

Calow, P. ( 1992) Can ecosystems be healthy? Critical consideration of concepts. J. Aquatic Ecosys. Health 1, 1-5.

Calow, P. (1993) Seeking standardization in ecotoxicology. In: Soares, A.M.V.M., and Calow, P. (Eds.) Progress in Standardization of Aquatic Toxicity Tests. Lewis, Boca Raton, Florida.

Calow, P., and Sibly, R.M. (1990) A physiological basis of population processes: ecotoxicological implications. Funct. Ecol. 4, 283-285.

Calow, P., Armitage, P., Boon, P., Chave, P., Cox, E., Hildrew, A., Learner, M., Maltby, L., Morris, G., Seager, J., and Whitton, B. (1990) River Water Quality, British Ecological Society Ecological Issues No.1. Field Studies Council, Montford Bridge, Shrewsbury .

Crossland, N.O., and Wolff, C.J.M. (1985) Fate and biological effects of pentachlorophenol in outdoor ponds. Environ. Contam. Toxicol. 4, 73-86.

Crossland, N.O., Mitchell, G.C., and Dorn, P.B. (1992) Use of outdoor artificial streams to determine threshold toxicity concentrations for a petrochemical effluent. Environ. Toxicol. Chem. 11, 49-59.

Cummins, K.W. (1992) Invertebrates. In: Calow, P., and Petts, G. (Eds.) Rivers Handbook, Vol. 1, pp. 234-250. Blackwell Scientific, Oxford.

Cummins, K. W., and Klug, M.J. (1979) Feeding ecology of stream invertebrates. Ann. Rev. Ecol. Sys. 10, 147-172.

De Angelis, D.L. (1992) Dynamics of Nutrient Cycling and Food Webs. Chapman and Hall, London.

de March, B.G.E. (1987a) Mixture toxicity indices in acute lethal toxicity tests. Arch. Environ. Contam. Toxicol. 16, 33-37.

de March, B.G.E. (1987b) Simple similiar action and independent joint action¾two similar models for the joint effects of toxicants applied as mixtures. Aquat. Toxicol. 9, 291-304.

Deneer, J.W., Seinen, W., and Hermens, J.L.M. (1988) Growth of Daphnia magna exposed to mixtures of chemicals with diverse modes of action. Ecotoxicol. Environ. Safety 15, 72-77.

Eaton, J., Arthur, J., Hermanutz, R., Kiefer, R., Mueller, L., Anderson, R., Erickson, R., Nordling, B., Rogers, J., and Pritchard, H. (1985). Biological effects of continuous and intermittent dosing of outdoor experimental streams with chlorphyrifos. In: Aquatic Toxicology and Hazard Assessment. Eighth Symp. ASTM STP 891. American Society for Testing and Materials, Philadelphia.

ECETOC (1993) Environmental Hazard Assessment of Substances. Technical Report No.51. ECETOC, Brussels.

Harper, D. (1992) Eutrophication of Freshwaters. Chapman and Hall, London.

Hellawell, J.M. (1986) Biological Indicators of Freshwater Pollution and Environmental Management. Elsevier, London.

Hildrew, A.G. (1992) Food webs and species interaction. In: Calow, P., and Petts, G. (Eds.) Rivers Handbook, Vol. 1. Blackwell Scientific, Oxford.

Hildrew, A.G., and Townsend, C.R. (1987) Organization in freshwater benthic communities. In: Gee, J.H.R., and Giller, P.S. (Eds.) Organisation of Communities: Past and Present, pp. 347-371. 27^th Symposium of the British Ecological Society. Blackwell Scientific, Oxford and London.

Holdgate, M.W. (1979) A Perspective of Environmental Pollution. Cambridge University Press, Cambridge.

Karr, J.R. (1991) Biological integrity: a long neglected aspect of water resource management. Ecol. Appl. 1, 26-35.

Kooijman, S.A.L. (1987) A safety factor for LC₅₀ values allowing for differences among species. Water Res. 21, 269-276.

Leffler, J.W. (1981) Aquatic Microcosms and Stress Criteria for Assessing Environmental Impact of Organic Chemicals. Office of Pesticides and Toxic Substances, US Environmental Protection Agency, Washington, D.C., pp. 50.

McCarty, L.S., Ozburn, G.W., Smith, A.D., and Dixon, D.G. (1992) Toxicokinetic modelling of mixtures of organic chemicals. Environ. Toxicol. Chem. 11, 1037-1047.

Magurran, A.E. (1988) Ecological Diversity and Its Measurement. Croom Helm, London. Maltby, L., Calow, P., Cosgrove, M., and Pindar, L. (1987) Adaptation to acidification in aquatic invertebrates: Speculation and preliminary observations. Ann. Soc. Roy. Zool. Belgique 117, 105-115.

Maltby, L., and Calow, P. (1989) The application of bioassays in the resolution of environmental problems; past, present and future. Hydrobiologia 188/189, 65-76.

Marking, L.L. (1985) Toxicity of chemical mixtures. In: Rand, G.M., and Petrocelli, S.R. (Eds.) Fundamentals in Aquatic Toxicology, pp. l64-176. Hemisphere, Washington, D.C.

Metcalfe, J.L. (1989) Biological water quality assessment of running waters based on macroinvertebrate communities: history and present status in Europe. Environ. Pollut. 60, 101-139.

National Rivers Authoity (1991) The Quality of Rivers, Canals and Estuaries in England and Wales. Report of the 1990 Survey. National Rivers Authority, Cambridge.

Naylor, C., Maltby, L., and Calow, P. (1989) Scope for growth in Gammarus pulex, a freshwater benthic detritivore. Hydrobiologia 188/189, 517-523.

Newbold, J.D. (1992) Cycles and spirals of nutrients. In: Calow, P., and Petts, G. (Eds.) Rivers Handbook, Vol. 1, pp. 379-408. Blackwell Scientific, Oxford.

Norris, R.H., McElravy, E.P., and Resh, V.H. (1992) The sampling problem. In: Calow, P., and Petts, G. (Eds.) Rivers Handbook, Vol. 1, pp. 282-306. Blackwell Scientific, Oxford.

Odum, E.P. (1985) Trends expressed in stressed ecosystems. Bioscience 35, 419-422. 

Paine, R.T. (1988) On food webs: road maps on interaction or grist for theoretical development? Ecology 69, 1648-1654.

Pascoe, D., and Edwards, R.W. (Eds.) (1984) Freshwater Biological Monitoring. Pergamon Press, Oxford.

Pascoe, D., and Shazili, N.A.M. (1986) Episodic pollution-a comparison of brief and continuous exposure of rainbow trout to cadmium. Ecotoxicol. Environ. Safety 12, 189-198.

Peckarsky, B.L. (1984) Predator-prey interactions among aquatic insects. In: Resh, V., and Rosenberg, D.M. (Eds.) The Ecology of Aquatic Insects., pp. 196-254. Praeger, New York.

Persoone, G., and Janssen, C.R. (1993) Freshwater invertebrate toxicity tests. In: Calow, P. (Ed.) Handbook of Ecotoxicology, Vol. 1. Blackwell Scientific, Oxford.

Reynoldson, T.B., and Day, K.E. (1993) Freshwater sediments. In: Calow, P. (Ed.) Handbook of Ecotoxicology, Vol. 1. Blackwell Scientific, Oxford.

Royal Commission on Environmental Pollution (1992). Freshwater Quality. Sixteenth Report. HMSO, London.

SETAC (Society of Environmental Toxicology and Chemistry) Europe (1991) Guidance document on testing procedures for pesticides in freshwater mesocosms. In: Proceedings of a Workshop. A Meeting of Experts on Guidelines for Static Field Mesocosm Tests. SETAC Europe.

SETAC (1992) Workshop Report on Aquatic Microcosms for Ecological Assessment of Pesticides. Society of Environmental Toxicology and Chemistry, Arlington, Virginia.

Slooff, W., Van Oers, J.A.M., and De Zwart, D. (1986) Margins of uncertainty in ecotoxicological hazard assessment. Environ. Toxicol. Chem. 5, 841-852.

Solbe, J.F. de L.G. (1993) Freshwater fish. In: Calow, P. (Ed.) Handbook of Ecotoxicology, Vol. 1. Blackwell Scientific, Oxford.

Taub, F.B., and Read, P.L. (1982) Final Report and Protocol, Model Ecosystems; Design, Development, Construction and Testing. Contract No.223-80-2352. US Department of Health and Human Services, Food and Drug Administration, Washington, D.C.

Thurston, R.V., Chakoumakos, C., and Russo, R.C. (1981) Effect of fluctuating exposures on the acute toxicity of ammonia to rainbow trout (Salmo gairdneri) and cutthroat trout (S. clarki). Water Res. 15, 911-917.

Townsend, C.R. (1989) The patch dynamics of stream community ecology. J. N. Am. Benthol. Soc. 8, 36-50.

USEPA (US Environmental Protection Agency) (1986) Guidelines for the health risk assessment of chemical mixtures. Fed. Reg. 51, 34014-34025, 24 September.

Van Straalen, W.M., and Denneman, C.A.J. (1989) Ecotoxicological evaluation of soil quality criteria. Ecotoxicol. Environ. Safety 18, 241-251.

Westman, W.E. (1978) Measuring the inertia and resilience of ecosystems. Bioscience 28, 705-710.

Wilhm, J. (1972) Graphic and mathematical analysis of biotic communiies in polluted streams. Ann. Rev. Entomol. 17, 223-252.