SCOPE 53 - Methods to Assess the Effects of Chemicals On Ecosystems, Chapter 11, Methods to Assess the Effects of Chemicals on Soils

The industrialization of our society has led to an increased production and emission of both xenobiotic and natural chemical substances. Many of these chemicals will end up in the soil. Various soil constituents have a great capacity to retain chemicals, especially those with apolar molecules or positively charged divalent and trivalent ions. Consequently, the soil is a net sink for all kinds of chemicals, and concentrations are often considerably higher than in any other environmental compartment. This situation may lead to smaller or larger impacts on the functioning of soil ecosystems. Important ecological functions of the soil are those associated with organic matter decomposition, mineralization of nutrients, and synthesis of humic substances. For that reason, an increasing need exists for methods to assess the side effects of these chemicals on soil ecosystems (OECD, 1989).

In this chapter, an overview will be provided of methods to determine the effects of chemicals on soil ecosystems. An overload of chemicals will affect both abiotic soil properties and, directly and indirectly, soil biota. This paper will, therefore, start with a description of the possible sources and consequences of chemical pollution for abiotic soil properties. Subsequently, methods are described for the determination of the effects of chemicals on soil organisms.

When considering methods to assess the effects of chemicals on soil biota, two types of tests can be distinguished. The first contributes to the prediction of the potential effects of single chemicals on soil ecosystems. For that purpose, mainly single-species laboratory tests are conducted; at times, more complex micro- ecosystem, mesocosm, or field studies are carried out. This type of testing, which aims at establishing dose-response relationships and the estimation of LC₅₀, EC₅₀, or NOEC values, may be called "prognosis."

The second type of method is aimed at assessing the potential ecological risk of a certain case of soil pollution. In such a situation, several chemicals may be involved. To determine whether a specified case of soil pollution poses a real hazard for soil biota, both laboratory and field studies may be performed. This type of testing may be called "diagnosis."

In this chapter the focus will be on terrestrial ecosystems; studies on groundwater ecosystems will not be taken into account.

11.2 QUANTIFICATION OF INPUT OF CHEMICALS IN THE SOIL

Before methods to assess the effects of chemicals on soil components can be presented, information must be provided about the various ways chemicals enter the soil and how these inputs can be quantified.

Focusing on major elements, Table 11.1 summarizes the most important processes contributing to the input of chemicals to a soil ecosystem. The relative importance of the input differs greatly among elements as well as among areas of low versus high pollution. The precision in the measurement of input quantities is mainly a question of equipment, methods, and experimental design.

This is the major process for the input of nitrogen, sulphur, and chloride, and is significant for other elements. Excluding problems such as the definition of input by mist and fog, the measurement of wet deposition is essentially the estimation of rainfall and its elemental content.

Type and design of rain gauges can be found in Tropical Soil Biology and Fertility: A Handbook of Methods (Anderson and Ingram, 1989, p.8), and their position relative to the ground surface is usually the reason for an underestimation in rainfall, depending on wind and evaporation losses in exposed areas (Eriksson, 1980). To avoid chemical changes in the sample, frequent sampling and rapid analysis are preferred to preservation.

This is defined as the direct transfer of gases and particles to different ecosystem surfaces (= receptors). Two methods are noted here:

This method is well known as the acetylene-reduction (AR) method for measuring nitrogenase activity.

Isotope techniques, such as incubation in ¹⁵N₂-containing atmosphere are attractive; but for long-running experiments, practical problems arise. Further methods are the ¹⁵N-isotope dilution technique and the classic total nitrogen difference method, based on a comparison of total N-yield in a N-fixing crop and that of a non-N fixing reference crop.

Methods to estimate current weathering rates include mass balance of watersheds, radiometric methods, and mineral bag technique.

Mass balance of watersheds or lysimeters is widely used and gives the best estimates of current weathering (e.g. Likens et al., 1977). The approach is indirect, leaving weathering as the residue in the mass balance equation:

where E(rw) is element release by weathering, E (efflux) is elemental losses mainly from leaching, E (influx) is input of elements mainly by dry and wet deposition, and E(ps) is the change in elemental storage in plants and soil.

The results of mass balance studies depend on the accuracy of estimation of all possible sources and sinks for elements in the soil. Main sinks are elemental storage in biomass and humus, but also accumulation in the soil by microbial activity, redox reactions, surface exchange processes, and formation of secondary minerals. These sinks may easily turn into sources if biological or chemical conditions are altered. For instance, present-day acid deposition depletes base cations from the exchange sites in the soil profile.

Radiometric methods using ⁸⁷Sr/ ⁸⁶Sr ratios are used to estimate calcium weathering.

For the mineral bag technique, selected soil or mineral fraction is put into a non-biodegradable mesh bag, and placed in the field for a specific period of time. The bag is then returned to the laboratory for analysis.

Major sources for the input of organic chemicals and heavy metals in soils are agriculture, industries, and traffic. Both diffuse and point sources can be identified, and some chemicals (e.g., pesticides) are applied in quite a controlled way enabling a proper prediction of the input in the soil.

To assess the deposition of pesticides that are generally applied under controlled conditions, sheets of aluminium foil or other inert substances can be placed on the soil and analysed after spraying. For other chemicals that may be released in a less controlled manner, chemical analysis of soil samples is needed to quantify the input. In all cases, rather specific analytical techniques are required to determine chemical concentrations in soils. Before analysis, complicated extraction and purification steps are often required. Generally, extraction with an organic solvent (e.g., hexane, acetonitrile, toluene or acetone) is applied, followed by analysis by HPLC, GC, or GC-MS.

To determine the soil content of heavy metals, digestion of soil samples with strong acids (e.g., HNO₃/HClO₄) is required. After that, the destruate can be analysed by atomic absorption spectrophotometry .

11.3 METHODS TO QUANTIFY EFFECTS OF CHEMICAL INPUT ON ABIOTIC SOIL CHARACTERISTICS

Recent concerns over nitrogen deficiencies have led to others concerning excess nitrogen availability and the potential for forest decline and surface water pollution. High input of N can lead to N-saturation with serious environmental impacts on soil chemistry and water quality and on fluxes of radioactively active (or "greenhouse") gases.

In Table 11.2 the characteristics of N-saturated forest soil are listed. The values for the characteristics are endpoints. To get information about the deposition level upon which these characteristics start to change, the "critical load" concept has been introduced. The definition is "the maximum deposition of elements that will not cause chemical changes leading to long-term harmful effects on ecosystem structure and function" (after Nilsson and Grennfelt, 1988). A regional assessment of critical loads is very important to formulate optimal policies for emission reductions. The generic approach to map critical loads is presented in Figure 11.1. In Table 11.3, an illustration is given of some critical chemical amounts for forest soil (water).


Characteristic	Value	Method

N cycled	25-50% NO₃^-	Anderson and Ingram, 1989;
	50-75% NH₄⁺	Faber and Verhoef, 1991

DOC concentration	Low	Complete (Nelson-Sommers)
		or partial (Walkley-Black)
		oxidation (Anderson and
		Ingram, 1989)

C/N ratio	Low	Kirsten, 1979

Ca, Mg concentration	Low	Flame atomic absorbance
		spectrophotometry

H^-, Al_i concentration	High	pH based on free protons, on
		the exchangeable fraction
		extracted with KCl, or on the
		fraction titrated with a base;
		and ICP

N₂O production	High	Lloyd, 1985; Harrison et al.,
		1990

CH₄ production	Low	Lloyd, 1985; Harrison et al.,
		1990

Sulphur is transformed in soils by processes similar to those occurring in the nitrogen cycle. Like nitrogen, sulphur can be oxidized, reduced, assimilated, or mineralised from organic matter. The major differences between the two cycles are (1) no process is equivalent to N-fixation and (2) losses of gaseous sulphur from soils are not equivalent to N-losses due to denitrification. Recent interest in soil sulphur transformations results from an increased awareness of the fertilizer value of the element and the recognition of the importance of the sulphate ion, which reaches the soil in acid rain, as the major counter-anion involved in cation leaching from soils. Sulphate adsorption by soils is an important property affecting the availability of sulphate to plants and the leaching of sulphate and associated cations. Sulphate adsorption is particularly important in soils subjected to acid precipitation, since it determines the impact of acid rain on cation mobility and leaching. Soil temperature and moisture influence sulphate adsorption, whereas desorption on waterlogging may also be an important reaction in soils exposed to atmospheric pollution.

Select receptor type
(e.g., soil ecosystems)
¯
Determine critical chemical values
¯
Select computation method (i.e., model)
¯
Quantify receptor distribution
¯
Collect input data
¯
Conduct critical load calculations
¯
Draw maps according to procedures

Figure 11.1. Flowchart to map critical loads and areas where they have been exceeded (De Vries, 1991)

Many agents have been used to extract sulphate and other sulphur ions from the soil (see Wainwright, in Harrison et al., 1990). Important is 0.01 M Ca(H₂PO₄)₂ which appears to remove sulphate from the same pool of soil sulphur that is available for plants. Sulphate can be measured by methods including gravimetrically, turbidimetrically with barium chloride, spectrophotometric ally using methylene blue, by titrimetric methods, adsorption chromatography, ion exchange chromatography, ion-selective electrodes, and thin-layer or gas chromatography (Williams, 1979).

Recent studies on the damage to ecosystems caused by oxides of sulphur and acid rain involve the use of lysimeters, and collectors to measure through-fall, stem flow, and litter deposition (Ulrich and Mayer, 1980). Unfortunately, most of these studies have omitted the microbial transformations of the element.

Recent laboratory studies have been concerned with the microbial cycling of sulphur in soils exposed to heavy atmospheric pollution from point sources. The first approach was to remove soils from the field at intervals throughout the season, and to have them analysed in the laboratory for S-ions and sulphur oxidizing micro- organisms. In the latter approach, soils were sampled from sites exposed to point-source pollution and from relatively unpolluted sites that had essentially the same soil type, vegetation, and climate as the polluted sites. They were packed into plastic tubes. Several soil columns were placed in the polluted site and several in the non-polluted site. In this way, an assessment of effects of pollution on relatively unpolluted soil could be determined as well as the time taken for heavily polluted soils to regain characteristics more typical of relatively unpolluted soil. Exposure duration was about 1.5-2 years. Similar exposure periods were found in a reciprocal transplant experiment with soil cores over a gradient of N and S input over Europe.


Criteria	Unit	Soil
[Al]	mol per m³	0.2
Al /Ca	mol per mol	1
pH	-	4.0^a
[Alk]^c	mol per m³	-0.3^a
NO₃	mol per m³	0.1^b
NH₄/K	mol per mol	5

^aFor forest top soils, pH of 3.7 and alkalinity of -0.4 are suggested.
^bRelated to vegetation changes (from De Vries, 1991).
^c[HCO₃] + [RCOO] -[H] -[Al]

Both forests and grasslands are frequently phosphorus deficient to a variable degree, and this deficiency limits their productivity. Fertilizer application is, therefore, a principal means of increasing timber and grass production.

In forestry , tree needle analysis has been used for many years as the main guide in the assessment of phosphorus fertilizer requirements. Recent publications, however, indicate that this type of analysis is unrealistic as a predictor of fertilizer responses in commercial forest trees (Axelsson, 1984; McIntosh, 1984). Analysis of the forest or grassland soils has been proposed as an alternative (Hunter et al. , 1985). Extraction methods for P_i rely on three different principles:

Methods to extract P_o have employed alkaline solutions (e.g., NaHCO₃ or NaOH) or various organic solvents (e.g., acetylacetone which dissolves organic matter). Little progress has been made towards characterising P_o extracts in terms of the mechanisms for bringing P_o into solution, binding modes in the soil and its availability to plants (Anderson and Ingram, 1989, p.113).

A new, physiologically based root bioassay has been developed, that appears to be sensitive in assessing P-deficiency in plants. The bioassay relies on the negative relationship between the rate of metabolic uptake of ³²p-labelled phosphorus by roots from a standardized solution in the laboratory and the amount of phosphorus supply in the original rooting environment. This method has been successfully applied to forest stands and grasslands (Harrison et al., 1985, 1986).

The CO₂ concentration in the atmosphere has increased by 25 percent over the past 100 years, and a consensus exists that a doubling of the concentration may occur by the middle of the next century .The largest terrestrial carbon sources and sinks influencing CO₂ fluxes are the forests, which account for approximately two-thirds of the photosynthesis. The effects of a doubling of the CO₂ concentration on the growth and development of trees is known for a few species. A general finding is an increase of the tissue density of the leaves, a change in leaf structure, and an increase in the C/N ratio of the tissues. This changed C/N ratio may reduce the decomposition rates of plant material and modify the nutrient availability (Coûteaux et al., 1991). On the other hand, increased atmospheric concentrations of CO₂, together with trace gases such as methane (CH₄), nitrous oxide (N₂O), and chlorofluorohydrocarbons (CFCs), are effecting changes in the global heat balance, resulting in significant changes in climate over the next century. Twice as much carbon is found in the top metre of soil compared to the amount in the atmosphere, and CO₂ emissions from soils will increase as organic decomposition is enhanced at higher temperatures. CO₂ emissions are particularly sensitive to temperatures between O°C and 5°C. Based on a recent model predicting global emissions from soil organic matter (Jenkinson et al., 1991), a world temperature rise of 0.3°C per decade has been estimated to result in an additional release of CO₂ from soil organic matter over the next 60 years, equivalent to about 19 percent of that released by combustion of fossil fuels if present use of fuel were to continue unabated. These calculations suggest that increased decomposition of soil organic carbon could make an important contribution to the greenhouse effect.

Methods to measure C/N ratios have already been given. Measurements of CO₂ evolution under field conditions are described and standardized (Anderson and Ingram, 1989). Methods to measure soil organic carbon are given by Anderson and Ingram (1989).

Effects of organic chemicals on the abiotic soil properties are rarely mentioned in the literature. Some chemicals, such as the herbicide paraquat, are incorporated into clay particles, but the extent to which this may influence the swelling and shrinking behaviour of the clay is unknown. Other chemicals or their degradation products are incorporated into the soil organic matter. This is a physical or a biological process, which may result, in case of chlorinated organics, in chlorination of the soil organic matter. The extent to which this process occurs and the consequences for the soil characteristics are unknown. Presently, no methods exist to determine the potential impact of organic chemicals on soil abiotic properties.

Metals generally occur in the soil solution as positively charged cations, competing for negatively charged adsorption places on the soil particles. An overload of metals will affect the ionic balance of the soil; it may also lead to a release of other, less strongly bound, metals or cations from the soil. Often this process is slow, not affecting soil abiotic properties to a great extent. Only in the case of flooding a soil with salt water, containing an excess of cations, was the swelling and shrinking properties of clays shown to be severely affected. No methods can be given to measure the impact of metals on soil abiotic properties.

11.4 METHODS TO ASSESS THE POTENTIAL RISK OF CHEMICALS FOR SOIL ORGANISMS (PROGNOSIS)

A brief description is presented of single-species laboratory tests, microcosm tests, and field tests. For an extended overview of these tests, the reader is referred to Van Straalen and Van Gestel (1992a, 1992b).

Among soil invertebrates, only earthworms have seriously been considered as test organisms during the past decade, and some standardized test methods are available. For microfauna and mesofauna, only few tests are available, although these animals are among the most numerous and species-rich groups of soil animals. Many species, however, are promising test animals, because they are easy to culture and their size allows for small-scale experimental set-ups with many replications.

Besides these soil animals, higher plants have also been considered for testing, and standardized tests with some plant species are available.

In several tests, artificial substrates (nutrient solution, agar, silica gel, filter paper) are used, the composition of which greatly affects toxicity. However, extrapolation of these test results to the field remains problematic. If concentrations in test solutions can be equated with pore water concentrations, sorption data may be used to express the toxicity per unit of soil. The validity of this extrapolation, however, still has to be investigated. The same extrapolation problems may arise for tests in which the main route of exposure is via the food. For such tests, a conversion of food concentrations to soil concentrations may be needed.

A test with higher plants has been described in an international test guideline (OECD, 1984b), while some others are under discussion. The OECD guideline 208 on higher plant toxicity testing uses several plant species, representing different agricultural crops and both monocotylodoneous and dycotylodoneous species.

Protozoans and nematodes live in the soil pore water, and the best way to test them is to use methods similar to those used in aquatic toxicology. Among the protozoans the ciliates Tetrahymena pyriformis, Colpoda cucullus, and Paramecium aurelia have been considered (Berhin et al., 1984; Nyberg and Bishop, 1983) for test animals, as have the nematode species Caenorhabditis elegans, Panagrellus silusiae, and Plectus parietinus (Sturhan, 1986; Haight et al., 1982; Van Kessel et al., 1989); however, an accepted test procedure is unavailable.

Isopods are an interesting group of animals in heavy metal research, because of their unique ability to concentrate extreme amounts of metals in their bodies (Hopkin, 1989). Their use as a test animal for soil toxicity studies, however, is restricted to a few cases, and no attempts have yet been made to arrive at standardization (Dallinger and Wieser, 1977; Van Capelleveen, 1987; Hopkin, 1990; Van Straalen and Verweij, 1991; Eijsackers, 1978b). Porcellio scaber, Oniscus asellus, and Trichoniscus pusillus are three species frequently investigated. Among these, T. pusillus seems to be the most suitable as a test species, as it has a somewhat shorter life-cycle compared to P. scaber and O. asellus. All three species are very easy to culture, and do not require special conditions.

Usually isopods are kept on a plaster substrate, and are fed with partly decomposed leaves, either intact or ground, to which chemicals can be added. Increase in growth over several weeks is observed, but is rather variable, even for one individual. Reproduction is difficult to assess, because, after mating, females may retain the sperm for a long period before producing eggs, which are carried in a brood pouch. Tests require a minimum period of four weeks.

Millipedes (Diplopoda) are another important group of saprotrophic soil invertebrates, but they have never been considered seriously as test animals. The most widely investigated species is Glomeris marginata (Hopkin et al., 1985). The species Cylindroiulus britannica is also well suited as a test animal. Test conditions for millipedes are similar to those for isopods.

A reproduction toxicity test using the parthenogenetic oribatid mite Platynothrus peltifer has been described by Denneman and Van Straalen (1991). This seems to be the only oribatid used so far in soil toxicity experiments, although oribatids comprise hundreds of species, and are usually the most numerous group of arthropods in forest soils.

In the test with P. peltifer, the animals are exposed to contaminated algae, and the number of eggs are counted. The test is very laborious, as the animals hide their eggs in small crevices; it is also a rather lengthy test (9 to 12 weeks) because of the low rate of egg production in this species and its long life-cycle (1 year), which are remarkable features for such a small animal (± 1 mm).

P. peltifer appeared to be rather resistant to cadmium, copper, and lead in terms of lethality, but very susceptible in terms of egg production. Due to their peculiar habits, species such as P. peltifer tend to be forgotten in the development of toxicity tests. It is, however, the most sensitive soil invertebrate tested so far for cadmium, while it is more sensitive than springtails for copper and lead (Denneman and Van Straalen, 1991; Van Straalen et al., 1989).

Collembola are a relatively well investigated group of soil animals. Several species have been used frequently in toxicity experiments: Onychiurus spp. (Eijsackers, 1978a; Bengtsson et al., 1985; Mola et al., 1987), Folsomia candida (Thompson and Gore, 1972; Tomlin, 1977; Iglisch, 1986), Tullbergia granulata (Subagja and Snider, 1981), and Orchesella cincta (Van Straalen et al., 1989). The first three species are parthenogenetic (thelytokous); O. cincta is sexual, sperm being transferred indirectly through spermatophores deposited on the substrate by the male. Three different exposure systems have been described: (1) through feeding on fungi grown on contaminated agar, (2) through feeding on directly contaminated food, and (3) residual exposure (treated substrate, e.g. sand, leaves, soil).

When testing Collembola with contaminated fungi, the animals are kept on a plaster of Paris substrate in a Petri dish, and fed on a piece of agar, overgrown with hyphae (e.g., Verticillium bulbillosum) (Bengtsson et al., 1983, 1985). Egg production, growth and survival are recorded regularly throughout a period of several weeks. The advantage of this system is that substances are offered in a natural way, i.e., after being taken up and possibly transformed to naturally occurring complexes. Concentration levels in the fungus, however, are difficult to maintain or to set to specific values.

When testing Collembola with directly contaminated food, chemicals are added in water or acetone solution to the food (algae, yeast, ground leaf material). Food can be offered as droplets on filter paper discs, while the animals are kept on a plaster or sand substrate. In this manner, concentrations can be manipulated easily, while growth, egg production, and survival are monitored over a period of several weeks (Van Straalen et al., 1989).

The third system of testing Collembola is to use the artificial soil medium developed for earthworm toxicity tests (Jancke, 1989; Wohlgemuth et al., 1990; ISO, 1991). Juvenile Collembola (Folsomia candida) are placed in artificial soil with dry yeast provided for food. After 28 days, the number of remaining animals and their offspring are counted after flotation extraction. The Folsomia test is very easy to carry out; it requires little attention during the test; and it gives reproducible results. Another advantage is the use of artificial soil similar to the earthworm test; thus, experimental results can be compared between earthworms and springtails. The only disadvantage of the test is that reproduction cannot be observed directly, and cannot be separated from juvenile mortality and hatching success. The Folsomia test is now undergoing the process of international standardization (ISO, 1991).

Enchytraeids can be cultured easily on agar and on (artificial) soil substrates, when fed rolled oats (Römbke and Knacker, 1989; Römbke, 1989). The toxicity tests described in the literature all use species of the genus Enchytraeus. The well-known species Cognettia sphagnetorum can also be bred easily in the laboratory, but its tendency to fragmentate upon handling makes it less suitable for toxicity tests.

Westheide et al. (1991) described a test in which the test chemical is incorporated in 1.5 percent nutrient agar. Two species are used, Enchytraeus cf. globuliferus and E. minutus. Reproduction, measured as the number of cocoons and juveniles produced, is the endpoint studied in this test. This method seems to provide an easy and reproducible test, but, because of the use of an agar substrate, results cannot be translated to real soil.

Römbke (1989) described a test with Enchytraeus albidus, using the OECD artificial soil prescribed for earthworm toxicity tests. Adult E. albidus are exposed for 28 days to different concentrations of the test chemical, mixed homogeneously through the artificial soil. Survival of adult worms and the number of juveniles produced are the endpoints studied. In this way, both acute and sublethal effects are combined in one test. The reproducibility of the method cannot be judged and, because only one chemical was tested, no conclusions can be drawn with respect to the sensitivity of the sublethal endpoint. An important positive aspect of this test is that the substrate used is the same artificial soil used in the internationally accepted earthworm toxicity tests (OECD, 1984a; EEC, 1985).

In the guidelines of OECD (1984a) and EEC (1985), Eisenia fetida and its sibling species E. andrei are recommended. Both species are commonly found in compost and dung heaps, and can be cultured easily in the laboratory on a substrate of horse dung or cow dung (OECD, 1984a). According to the existing guidelines on acute toxicity testing with earthworms (OECD, 1984a; EEC, 1985), other real soil dwelling species may also be used. Such species are, however, hard to culture in the laboratory , because they have long generation times and need large volumes of soil. So, for practical reasons, the use of the two Eisenia species is recommended.

Three acute toxicity tests exist. In the filter paper contact test (OECD, 1984a), adult earthworms of the species Eisenia spp. are exposed to filter paper wetted with a solution of the test substance. Mortality is assessed, and the 48-hour LC₅₀ value is expressed as µg per cm². The method has been shown to be easy, fast, and highly reproducible (Edwards, 1983). Several authors including Heimbach (1984) have demonstrated, however, that this test has no predictive value for the effect of chemicals on earthworms in the soil; it can only be used to rank chemicals.

In the artificial soil test (OECD, 1984a), adult earthworms of the species Eisenia spp. are exposed to the test chemical, which is mixed through an artificial soil substrate for 14 days. This artificial soil is made up by mixing (dry weight) 10 percent sphagnum peat, 20 percent kaolin clay, 70 percent quartz sand, while some CaCO₃ is added to adjust the pH to 6.0±0.5. The moisture content of the substrate is adjusted to about 55 percent (w/w) or to 40-60 percent of the water holding capacity. Mortality is the only test parameter, and LC₅₀ values are expressed as mg per kg dry soil. Van Gestel and Ma (1990) have demonstrated that results obtained in this artificial soil can easily be translated to natural soils by using sorption data. For this reason, the use of the artificial soil is acceptable, and the test can be concluded to have enough predictive value with respect to effects that occur in the field (Heimbach, 1992; Van Gestel, 1992). The test has been shown to be reproducible.

In the Artisol test (Ferrière et al., 1981), adult earthworms of the species Eisenia spp. are exposed to chemicals mixed through a substrate of amorphous silica gel (Artisol) for fourteen days. Survival is the only test parameter, and results are expressed in terms of LC₅₀ values. The silica gel substrate does not bear any resemblance to natural soil; thus for reasons of ecological realism and extrapolation towards natural soil, this test cannot be recommended.

Recently, two sublethal toxicity tests have been described. In both tests, the OECD artificial soil and the earthworm species Eisenia spp. are used. In the first test (Van Gestel et al. , 1989), chemicals are mixed homogeneously through the artificial soil, and after three weeks of exposure effects on the growth and cocoon production by adult earthworms are determined. The worms are fed by supplying a small amount of (untreated) cow dung in a small hole in the middle of the soil. By incubating cocoons produced for five weeks in untreated artificial soil, effects on hatchability (percent fertile cocoons, number of juveniles per cocoon) and the total number of offspring per adult worm can be determined.

The second method (Kokta, 1992) was developed to determine the sublethal effects of pesticides on earthworms. The pesticide is sprayed onto the soil surface, and earthworms are fed by applying about 0.5 g cow dung per animal to the soil surface once a week. Pesticides are applied in two treatment levels, corresponding with the recommended dose and a five fold dose. After six weeks incubation, adult worms are removed from the substrate and weighed. The test substrate containing cocoons and juveniles is incubated for another four weeks. Food is given when required. After ten weeks, the juveniles are extracted from the substrate by hand-sorting or heat extraction and counted. Effects on earthworm growth and on the total number of offspring produced per tray are determined. The method has been subjected to a (German) ring test, and will be revised on the basis of the results of it. The method is only applicable for pesticides, and the recovery of all juveniles from the artificial soil by hand-sorting is difficult. This hinders comparison of this method with that of Van Gestel et al. ( 1989). Furthermore, the method of pesticide application is not standardized, which may be the reason for the variability observed in the first ring test (Kokta, 1992).

In the limited number of toxicity tests using terrestrial molluscs, exposure was via the food. Russel et al. (1981) described a method for toxicity experiments with the garden snail Helix aspersa. The snails were kept in polyethylene boxes, filled with a substrate of moist quartz sand covered by a piece of woven glass towel. The snails are fed a diet of ground Purina Lab-Chow (formulation for rats, mice, and hamsters) supplemented with CaCO₃. Parameters affected include survival, reproductive behaviour, dormant state, new shell growth, and food consumption. Similar test methods using the snail Helix pomatia or the slug Arion ater have been described by other authors (Marigomez et al., 1986; Meincke and Schaller, 1974; Moser and Wieser, 1979).

Arthropods that may improve the production of agricultural products are designated as "beneficials," and commercial interest exists in designing and applying pesticides in such a way that beneficials are least affected. The working group on "Pesticides and Beneficial Organisms" of the International Organization for Biological and Integrated Control of Noxious Animals and Plants (IOBC) has contributed significantly to designing ecotoxicological test methods and decision schemes to evaluate the hazard of pesticides. These methods have been reviewed by various authors (IOBC, 1988; Croft, 1990; Samsøe-Petersen, 1990).

The hymenopteran groups Ichneumonidae, Braconidae, and Chalcidoidea contain a large number of parasitoid species. The female insect deposits an egg in or on a host (usually an insect egg or larva), which is then gradually eaten as the offspring develop. The host selection process and the life-cycle of the parasitoid are finely tuned to the host, and many species will attack only a single or a few host species. Furthermore, other hymenopteran species used in toxicity tests include Diaeretiella rapae, an internal parasite of aphids such as Myzus persicae, Phygadeuon trichops, a parasite of Delia species (bulb flies), Coccygomimus (=Pimpla) turionellae, a polyphagous parasite of Lepidoptera (Tortricidae, Geometridae, Noctuidae), and Opius sp., a parasite of leaf mining insects. The methods used for these species are similar to those described for Trichogramma and Encarsia.

Within the order of the Coleoptera, the families Carabidae (ground beetles), Staphylinidae (rove beetles), and Coccinellidae (lady birds) contain representatives that are common in agricultural fields and are recognized for their predation of pests.

Among the various arthropod groups, spiders seem to be particularly sensitive. This phenomenon often appears in field tests with pesticides, where catches of surface active spiders are reduced in a manner similar to that of predatory mites following pesticide application. The families Erigonidae and Linyphiidae (money spiders) are important groups with a great species richness.

The recommendations made by the International Commission for Plant Bee Relations (ICPBR) have been included in a guideline of the European and Mediterranean Plant Protection Organization (EPPO) to evaluate the hazards of pesticides to the honey bee Apis mellifera (OEPP/EPPO, 1991). Several countries have slightly different national guidelines to test pesticides on honey bees.

Several other beneficial arthropods have been proposed as test species (IOBC, 1988; Hassan et al., 1985): Chrysoperla carnea (Neuroptera, Chrysopidae), Anthocoris nemorum (Heteroptera, Anthocoridae), Syrphus corollae and Syrphus vitripennis (Diptera, Syrphida), and Drino inconspicua (Diptera, Tachinidae).

Single-species tests are carried out under rather artificial conditions, and disregard ecological interactions between different species. To evaluate effects of chemicals under more natural conditions, model ecosystems, microcosms or micro-ecosystems have been designed that simulate certain aspects of real ecosystems, and are yet simple enough for experimental use. Decomposing invertebrates have been considered for such systems, because their activities can be assessed conveniently in terms of system functions such as leaf litter fragmentation and nutrient conversions (Teuben and Roelofsma, 1990; Eijsackers, 1991).

Several terrestrial model ecosystems have been described (Giesy, 1980; Anderson, 1978b; Verhoef and de Goede, 1985; Hågvar, 1988; Teuben and Roelofsma, 1990; Teuben and Verhoef, 1992; Mothes-Wagner et al., 1992), without attempt to arrive at standardization. The system may either be closed or open to the ambient air, and contains intact core samples from a natural habitat (e.g., Chaney et al., 1978; Ausmus et al., 1978) or a more or less standardized soil (e.g., Bond et al., 1976). For ecotoxicological tests, the use of standardized soils seems to be most appropriate, since it allows the chemical to be mixed homogeneously through the soil, and it minimizes experimental variation between replicate units. The effects of various pretreatments, such as drying, sterilizing, inoculation, litter type, age of the litter, however, have a significant impact on the behaviour of the system and need to be investigated thoroughly (Van Wensem et al., in press).

Natural rainfall may be simulated, and leachate can be collected (e.g., Verhoef and Meintser, 1991; Bengtsson et al., 1988). Various chemical analyses of the leachate solution may indicate aspects of decomposer activity: dissolved organic carbon, NH₄, NO₃, pH, and Ca. The advantage of this procedure is that repeated sampling in time from the same soil column is possible. A disadvantage of the leaching procedure is that the humidity of soil and litter is unstable, and is difficult to standardize. Moreover, toxicants added to the system may be displaced through the column or leached out. Microbial respiration can be estimated by measuring CO₂ production in the microcosms. For that purpose, CO₂-free air (20.8 percent O₂; 79.2 percent N₂) is guided through the soil column, and the resulting CO₂ is subsequently measured by infrared gas analysis (Teuben and Roelofsma, 1990). Verhoef and Dorel (1988), Verhoef et al. (1989), and Verhoef and Meintser (1991) have used these types of microcosms, filled with pine litter, to study the effects of gaseous (NH₃) and wet ((NH₄)₂SO₄) atmospheric deposition. N-deposition eliminates the stimulation of mineral leaching by the collembolan Tomocerus minor. Neither survival nor growth of the animals are affected. Reproduction, however, is negatively influenced. In pine litter, which has been confronted with high N input for several decades, T. minor slows down mineral leaching by stimulating microbial growth.

Van Wensem (1989) and Van Wensem et al. (1991) added chemicals to poplar leaf litter, which is incubated for 4 weeks, after which some replicates are terminated to determine DOC, NH₄, NO₃, and pH. The remaining replicates are incubated for another four weeks, with eight isopods (Porcellio scaber) added to each system. Survival and growth of the isopods can be assessed, as well as particle size distribution and concentrations of minerals of the remaining litter. The organotin fungicide triphenyltin hydroxide increased the concentration of soluble ammonium in the litter, due partly to excretion by isopods and partly to stimulating effects on the microflora. In systems with isopods, the organotin decreased ammoniation in treatment levels higher than 10 µg per g, but in systems without isopods the organotin had no significant effect. The addition of isopods in this case, therefore, made the system quite sensitive, which was unexpected (triphenyltin is a fungicide), and would not have been noticed in a single-species test using isopods.

Mothes-Wagner et al. (1992) described a more complex microcosm system, consisting of 25 litres of natural or standardized soil that are inoculated with nematodes (Pelodera strongyloides) and enchytraeides (Enchytraeus coronatus) and sown with bush beans (Phaseolus vulgaris). After emergence of the beans, spider mites (Tetranychus urticae) are introduced. After a preincubation period of about six months in the laboratory, in a greenhouse, or in the field, the systems can be treated with the test chemical. Test parameters are survival, reproduction, and population growth of the introduced organisms. Furthermore, measurement of several histological and enzymatic parameters in these organisms is recommended. As no substantial test results are available, the predictive value and sensitivity of this system cannot be evaluated.

Tests on single species of isolated microorganisms in artificial substrates are not regarded as representative for the soil ecosystem, and will, therefore, not be considered here.

Based on a series of workshops held during the 1970s and 1980s, Somerville and Greaves (1987) formulated several recommended tests to assess the side effects of pesticides on the soil microflora. For all microbial tests in soil, the use of freshly sampled soil containing an active microflora was considered essential. Prolonged storage and drying of the soil should be avoided. For a proper assessment of the effect of chemicals, at least two different soil types should be used (Somerville and Greaves, 1987). A short description is given here of several tests on microbial processes related to the conversion of nutrients in soil. Unless stated otherwise, all tests are carried out in the dark at a temperature of 20±2°C, and the test chemicals are mixed homogeneously through the soil. Generally, soils are tested at a moisture content corresponding to field capacity or to 40-60 percent of the water holding capacity.

In these tests the production of CO₂ from small soil samples (£100 g) treated with the test chemical is measured continuously or semi-continuously. The tests should run for a minimum of 30 days. Tests may be performed in either unamended soil or in soil amended with a substrate. For this purpose mostly 0.5 percent (w/w) lucerne or horn meal is used (Somerville and Greaves, 1987). The disadvantage of this soil respiration test is that the activity of the total soil microflora is determined. When certain species are affected by the test chemical, this will often not be noticed, as other (less sensitive) species may take over the activity of the sensitive ones.

During the past decade some new test methods have been developed which aim to determine chemical effects on more specific groups of soil micro-organisms. One is the addition of a readily degradable substrate and the determination of the short-term respiration rate. Such a test was described by Haanstra and Doelman (1984) using glutamic acid as a substrate. Soils are amended with glutamic acid and the CO₂-production is measured. Glucose may be used as a substrate. The duration of the test is no longer than 100-120 hours. The test appeared to be quite sensitive to heavy metals. These short-term respiration tests may be combined with a biomass determination, and seem to be more sensitive than the traditional respiration tests. Another alternative may be found in the addition of more persistent substrates such as lignin or cellulose (Ljungdahl and Eriksson, 1985). Only a few soil microorganisms are capable of degrading these substrates, and, when they are affected by the test chemical, no others can take over their activity.

The disadvantage of the previously described soil respiration or substrate degradation methods is that less sensitive species of micro-organisms may grow on the substrate during the test. This results in a shift among the microflora towards more resistant species, masking the possible elimination of sensitive species (Van Beelen et al., 1991). For this reason, Van Beelen et al. (1990) developed test methods using the mineralization of low concentrations of ¹⁴C-acetate, ¹⁴C-chloroform or other labelled substrates. The amount of substrate applied is very low (1 µg per L) to ensure that no growth of the microflora will occur. This amount of substrate is added to a slurry of the test soil, prepared by mixing the homogenized soil with an equal weight of ground water. The test chemicals are added to the slurry in the desired concentration levels. Samples are incubated at 10°C. The test duration depends on the capacity of the microflora in the soil sample to mineralize the test substrate, and is chosen depending on the half-life of the acetate mineralization. Acetate mineralization is measured by determining the amount of ¹⁴CO₂ released from the sample and by determining the amount of ¹⁴C remaining in the suspension at the end of the test.

In ammoniation tests, the release of inorganic nitrogen from soil organic matter or a substrate (e.g., plant material or horn meal) is studied in a way comparable to soil respiration tests. The influence of nitrification, i.e., the conversion of ammonia into nitrate, may also be studied in these tests. Ammoniation is performed by a wide variety of soil micro-organisms, and is, therefore, relatively insensitive to perturbation. The advantages of nitrification are (1) that fewer species of micro- organisms are involved in this process and (2) that the process is considered to be of ecological and agricultural importance. Therefore, either combining these parameters in one test or running a separate test on nitrification is recommended (Somerville and Greaves, 1987). Nitrification tests can be performed in soil amended with either (NH₄)₂SO₄ or with organic substrates such as lucerne or horn meal. For this purpose, substrate equivalent to approximately 100 mg N per kg soil is added, and the disappearance of NH₄⁺ and the appearance of NO₃^- is monitored. In case the rate of NO₃^- formation does not follow the disappearance rate of the NH₄⁺, the soil should also be checked for the formation of NO₂^-. To check whether the test soil is capable of nitrification and whether the organic matter amendment is suitable for ammoniation and nitrification, studies are also recommended (Somerville and Greaves, 1987).

Tests on both symbiotic and asymbiotic nitrogen fixation can be identified. Tests on symbiotic nitrogen fixation in fact consider the unique relationship between the host plant and Rhizobium, and, therefore, include in one test effects on both the plant and the bacteria. These experiments are conducted in a soil suitable for growth of the plant. The plant, seeds, or soil can be inoculated with Rhizobium if no suitable bacteria are present in the soil (Somerville and Greaves, 1987). Effects on plant growth and the degree of nodulation should be included. In tests on asymbiotic nitrogen fixation, the degree of acetylene reduction (or formation of ethylene from acetylene) by soil samples is determined in relation to the addition of the test chemical.

Denitrification is the conversion of nitrate to atmospheric nitrogen, and will especially take place under anaerobic conditions. This process may be relevant in soil, as microsites may become anaerobic. In this test, soils are generally flooded with a layer of water, and nitrate is supplied as a substrate. Additionally, an organic substrate such as glucose is added to the soil. Since the process cannot be quantified, the formation of nitrogen gas, the disappearance of nitrate, and the formation of nitrite are measured as test parameters (Anderson, 1978a).

As in the tests on microbial processes, chemicals are also mixed homogeneously through the soil in tests on enzyme activity. Moisture content is adjusted to field capacity or to 40-60 percent of water holding capacity, and all incubations are done in the dark at 20°C. Several soil enzymes, relevant to microbial processes in soil, can be used as test parameters, as noted below.

At several intervals, small soil samples (6-7g) are taken, and incubated with 5 ml of demineralized water and 1.0 ml of a solution containing 60 mM urea. Incubation is at 35°C for 5 hours on a shaking water bath. A phenylmercury acetate solution in 2 M KCl is added to the soil samples to stop the urease reaction. After 10 minutes of shaking, the soil suspensions are filtered. The filtrates are analysed photometrically at 525 nm for urea concentrations (NEN 5796, 1989).

Soil samples (5-10g) are incubated with a solution of TTC (2,3,5-triphenyl tetrazolium chloride) in 0.1 M tris buffer solution (pH 7.6), and incubated for 24 hours at 30 or 37°C. The reduced triphenyl formazan formed is extracted with methanol and quantified by measuring the absorbance at 485 nm (Casida et al., 1964; Thalmann, 1968). Dehydrogenase reflects a broad range of microbial oxidative activities, and does not consistently correlate to microbial numbers, CO₂ evolution or O₂-consumption. Additionally, dehydrogenase activity may depend upon the nature and concentration of amended C-substrates and alternative electron acceptors (Somerville and Greaves, 1987). Rossell and Tarradellas (1991) concluded that short-term (substrate-induced) dehydrogenase activity may reflect the impact of chemicals on the physiologically active biomass of the soil microflora.

At several intervals, 0.5 g soil samples are taken, and incubated with 5 mM p-nitrophenylphosphate (p-NPP) for 1 hour in a shaker at 20°C. Phosphatase activity is measured as the amount of p-nitrophenol formed using a spectrophotometer (Tabatabai and Bremner, 1969). Phosphatase is said to bear little relation to total phosphate availability in soils (Somerville and Greaves, 1987). Its relevance for microbial activity in soil may, therefore, be questionable.

Somerville and Greaves (1987) stated that soil enzyme activities would be of little value to monitor side effects of pesticides on microflora. The main reasons for this were:

The reliability of microcosm studies in the laboratory to interpret field conditions is much debated. Microcosms differ from the field situation concerning the influence of temperature and moisture dynamics, the influence of root presence, and the composition of the soil biota community. A recent study compared microcosm studies in the laboratory with mesocosm studies and direct field measurements concerning microbial respiration, enzyme activities, and availability of macronutrients in interaction with soil animals; these soil process variables appeared to be of the same order of magnitude (Teuben and Verhoef, 1992).

The tests described in the preceding sections provide only a rough estimate of the possible hazard imposed by a chemical in the environment. In many cases, this degree of precision is sufficient. Usually the laboratory test is considered to be a "worst case" situation, since test animals are exposed to a constant concentration that is relatively available, because the test substrate is prepared freshly. Under field conditions, exposure may be lower since the chemical is not distributed unifonnly over the habitat, and bioavailability will often be lower due to various sorption processes. By contrast, the laboratory test considers the test organism under optimal conditions, without secondary stresses, such as those of food shortage, drought, and cold. The uncertainties attached to the laboratory-to-field extrapolation can be avoided by conducting experiments under semi-field or field conditions.

Various organizations have recommended test protocols for field investigations (IOBC, 1988; OEPP/EPPO, 1991). In several cases, guidelines for field tests are part of the national registration procedures for pesticides. Furthennore, considerable scientific research has been done in which side-effects of pesticides have been published (Edwards and Thompson, 1973; Eijsackers and Van de Bund, 1980; Inglesfield, 1988; Jepson, 1989). Some attempts have also been made to develop standardized procedures for field tests to assess the effects of pesticides on earthworms (Kula, 1992).

Some of the arthropods used as laboratory test species can also be exposed to chemicals under semi-field conditions, while exposed in cages. Hassan et al. (1985) lists protocols developed for Trichogramma cacoeciae, Phygadeuon trichops, Coccygomimus turionellae, Phytoseiulus persimilis, Aleochara bilineata, Chrysoperla camea, and Drino inconspicua. The usual procedure is to treat a group of plants with a spray of the chemical. A cage is then put over the plants after which test animals, with hosts or food, are introduced. The cage is installed either in a greenhouse or outdoors under a cover to provide shelter from rain and excessive sunshine. After an adequate duration of exposure, the performance of beneficials is compared with water-treated controls.

Cage tests have been described in detail for testing with pollinators (Felton et al., 1986; OEPP/EPPO, 1991). Bees (Apis mellifera) from small colonies are made to forage on a flowering crop in cages measuring minimally 2 x 2 x 3 m, with a 3 mm mesh netting. The product is applied to the plants, and not to the cage walls, by spraying. The EPPO-guideline does not require replication of the treatment. Effects are recorded at several intervals, preferably 0, 1, 2, 4, and 7 days after treatment. Observations are made on the number of dead bees, on foraging activity, and on behaviour. The results are compared with a blank control (usually water-sprayed) and a positive control (a reference product known to be hazardous to bees, e.g., parathion).

OEPP/EPPO (1991) also provides a guideline for field tests on honey bees. A chemical to be tested is applied to a plot of at least 1500 m², with the crop for which the chemical is intended, or another crop attractive to bees (rape, Phacelia), in full flower. Per treatment, three colonies of honey bees are placed in or on the edge of the plot. Test plots should be separated by at least 500 to 1000 m² to avoid bees foraging on the wrong plot. Replication of the treatment is considered desirable, but is not required in the EPPO guideline. A blank control (untreated, or treated with a reference product known to present a low hazard to bees), as well as a positive control (e.g., parathion, dimethoate) are applied to separate test plots.

After treatment, observations are made at several intervals, preferably after 0, 1, 2,4,7, and 14 days. Meteorological data are recorded during the entire period of the trial. Several parameters are estimated such as the number of foraging bees in the crop, behaviour of bees on the crop and around hives, mortality of bees (using dead bee traps), pollen collection (using pollen traps), pollen in collected honey, number of bees on frames, brood status in frames, and residues in dead bees, pollen wax, and honey. For the test to be valid, mortality in the negative control should not exceed 15 percent, while mortality in the positive control should be statistically significant.

Hassan et al. (1985) summarize the recommendations made by the IOBC Working Group Pesticides and Beneficial Organisms for full-scale field tests. The methods are suitable for a variety of crops, but have been applied mostly to winter wheat.

The trial is laid out in a replicated block design: three large fields with similar agronomic history are each divided into three treatment areas, where each treatment area covers at least 3 ha. The treatments are a spray with the product to be tested, a blank treatment (water spray), and a positive control (e.g., dimethoate).

Sampling is planned on seven occasions: 10 and 5 days before treatment, 2, 5, 10, and 20 days after treatment, and just before harvest. Sampling activities are concentrated in the central parts of each treatment area. Crop foliage fauna is collected with a suction net sampler (e.g., the Dietrich vacuum sampler). Soil surface fauna are sampled using pitfall traps, left in the field for 5 days. Visual inspection, water traps, and sticky traps can provide additional information on those arthropods sampled inefficiently by pitfalls or suction samplers.

The fauna collected are identified to at least the family level. Some groups where species can be recognized easily can be further subdivided (e.g., carabid beetles).

In addition to the large-scale experiments suggested for arthropods in winter wheat, smaller set-ups have been suggested by Edwards and Thompson (1973) and Eijsackers and Van de Bund (1980). Treatment plots of 3 x 3 m are recommended for microarthropods (Collembola, mites), while 10 x 10 m plots are suitable for studies on beetles and spiders. In all cases, however, the plots should be fenced, preferably using a polythene sheet, protruding 15 cm below ground and 40 cm above. The barrier should limit the immigration from neighbouring plots by surface-active arthropods such as beetles.

The statistical treatment of data from field experiments is not harmonized. Yet, such harmonization may be important since the probability of finding effects of the treatment will depend on the power of the statistical analysis. Stewart-Oaten et al. (1986) suggested the use of "pseudoreplication in time," to allow for a detailed evaluation of the effects of a treatment in relation to a control. In this design, also called BACI (Before After Control Impact comparison), the correlation between the observations from control plots and treatment plots before treatment is used to assess the effects in the treatment plots as deviations from the expectations made on the basis of the control plots (Everts et al., 1989; lagers op Akkerhuis and Van der Voet, 1992).

Hassan et al. (1985) summarize the standardized methods developed by the IOBC Working Group "Integrated Protection in Orchards." The methods consist of catching the fauna in collectors placed under the trees to which a chemical has just been applied. The collectors can be trays, canvas sheets, or funnels ("Steiner funnels"), with at least 0.5 m² of collecting area. In each trial, both the control (water spray) and the test treatment are followed by a "cleaning" treatment, 48 hours after the initial treatment. The "cleaning" treatment consists of dichlorvos at double the recommended dose, which will remove all beneficials present. The effectiveness of the treatment can thus be expressed in relation to the total population of beneficials present in the treated tree. The use of a reference chemical with each treatment, e.g., phosalone, is also recommended.

The design of the trial is a complete randomized block design, or a balanced incomplete block design. Each replicate is represented by one tree with one or more fauna collectors; six to eight replicates per treatment are recommended. The trees should be separated by at least one untreated tree.

The fauna are gathered from the collectors 24 hours and 48 hours after the treatment, as well as 24 hours after the "cleaning" treatment. Only the arthropods, of which there is at least an average of ten individuals per collector, are considered. The fauna are identified to the species or the family level.

Although no standardized guidelines exist yet for the study of pesticide side-effects on earthworms, recommendations for the performance of such a test have been formulated by Kula (1992).

Field studies with earthworms can be performed on arable land or on permanent grass; in both cases, a minimum number (100 individuals per m² of earthworms is required, and the relevant species (Aporrectodea caliginosa and Lumbricus terrestris) must be present. Minimum plot size should be 100 m², and at least four replicate plots should be used per treatment. A study should include a control, the highest recommended dose, and a toxic standard (benomyl). In many cases also a manyfold (e.g., fivefold) of the recommended dose should be studied. At each sampling time, at least two samples (sampling area 0.25 m² should be taken from each replicate plot. Preferably treatment should take place in the spring, and samples should be taken 1, 4 to 6, and 12 months after application. For the sampling of earthworms, the formaldehyde method and electrical sampling methods seem to be the most useful.

11.5 METHODS TO ASSESS THE IMPACT OF SOIL CONTAMINATION ON SOIL ORGANISMS (DIAGNOSIS)

The main characteristic of laboratory bioassays is that the potential toxicity of samples of field soil is studied in the laboratory using laboratory-bred test organisms. So, in such a test, well known organisms having similar characteristics are used, and the test can be performed under controlled conditions ruling out other possible disturbing influences. The advantage of bioassays is that they provide a direct indication of the toxicity of a specific soil, and integrate the effect of all substances present. The disadvantage is that the specific chemicals causing the observed effects cannot be identified. Bioassays can also be applied to determine the bioavailability of pollutants in soils as an indication of the potential risk for higher trophic levels.

Laboratory bioassays with earthworms (Eisenia fetida) and higher plants (Cyperus esculentus) have been described by Marquenie and Simmers (1988) and Van Gestel et al. (1988); the latter authors used these bioassays to study the influence of soil clean-up on the bioavailability of metals. In these methods, cylinders with a diameter of about 18 cm are filled with a 20 cm layer of soil. The cylinders have a perforated bottom and are placed in a dish filled with water. After one week of preincubation, test organisms are introduced. After 4 weeks, earthworms are sorted out of the soil; after incubation on wet filter paper for 24-48 hours to void the gut, they are analysed for metal content. Plants are harvested, and shoots are analysed. Van Gestel et al. (1993) described similar bioassays, using lettuce (Lactuca sativa) and radish (Raphanus sativus), to determine metal bioavailability in soils. For these organisms, smaller amounts of soil are needed (about 400 mg), and as in the bioassays using C. esculentus, some nutrient solution was added to the soils to stimulate plant growth.

Bioassays using the earthworm species Lumbricus terrestris and E. fetida have been described by Menzie et al. (1992). They studied survival of the earthworms after exposure to contaminated soil for 28 and 14 days, respectively, and also determined the uptake of some selected chemicals in the earthworms.

Menzie et al. (1992) and Callahan et al. (1991) studied the potential risk of contaminated soils using in situ bioassays with earthworms. Contaminated soil was placed in plastic buckets placed in the ground from which the soil was taken. The buckets were constructed to allow for free exchange of air and water. Adult earthworms of the species Lumbricus terrestris were placed in the buckets, and observations were made after 1 and 7 days. Survival and morbidity (burrowing, coiling, shortening, swelling, lesions) were the test parameters, and earthworms (including gut content) were analysed to determine bioaccumulation of the main pollutants. This bioassay proved to give a good indication of the possible risk of polluted soil.

Two observations merit comment. First, the bioassays last only 7 days, which might be too short to allow for an equilibrium in the uptake of highly lipophilic chemicals. Thus, uptake of these chemicals by the earthworms may be underestimated. Second, uptake might be misjudged, because of the presence of contaminated soil in the earthworm gut. From the authors' experience, the gut content of an earthworm may account for about 50 percent of its dry weight.

Kopezski (1992) used small enclosures (3 cm long; 4.8 cm in diameter) to study the impact of acidification on population growth and decomposition activity of the collembola species Folsomia candida and Heteromurus nitidus in a forest soil. The enclosures were filled with 1 g of hazel leaf litter and 1g of wafers, and buried into the soil. By using cheese cloth for the bottom and top ends of the enclosures, free contact with the surrounding soil was ensured. After 6 months, samples were analysed. A significant correlation appeared between animal numbers and soil pH, with H. nitidus being most sensitive. Decomposition showed a somewhat weaker correlation with soil pH.

To study the effects of N-deposition on the interactions between soil fauna and microflora and the effects on mineralization in coniferous forest soils, lysimeters have been used by Berg and Verhoef (1992). Intact soil cores with total soil fauna or absence of mesofauna are treated with three (NH₄)₂SO₄ concentrations (10, 50, and 200 kg N per ha per g). The lysimeters are defaunated by means of microwave treatment. Before installation, they are incubated with a soil-spore suspension. Faunal groups extracted from the soil are added to the different treatments.

Migration of fauna between the lysimeter and the surrounding soil can occur through holes in the lysimeter. The lysimeters are covered by a gauze lid and covered just above the top by a plexiglass roof preventing rain input. The lysimeters are watered every two weeks by hand.

Besides the bioassays using invertebrates or higher plants, special methods may be necessary to assess the potential risk of contaminated soils. Some compounds present in soil as contaminants (e.g., PAHs) are known mutagens, and their effects can be assessed by genotoxicity tests developed for drinking water, surface water, and sediments. These tests may be applicable for industrially contaminated sites, waste disposals, or sewage sludge-amended soils.

The usual approach to assess mutagenicity is to record the number of revertants in a Salmonella thyphimurium strain, plated with the test substance on a histidine deficient medium. To include those mutagens that require metabolic activation, a rat liver microsome suspension is added (Ames et al., 1975). In addition to this test, several other procedures have been proposed, which are often more sensitive than this test (Van der Gaag, 1989).

To apply mutagenicity tests to contaminated soils, an extract must be obtained. The extraction may be crucial to the validity of the results: extractions with solvents such as methanol or dimethylsulphoxide often induce a stronger mutagenic response in the Salmonella test, compared to water leachates; this difference in potency is especially true for superficial soil horizons (Kool et al., 1989; Donnelly et al., 1991).

The ecological relevance of mutagenicity test results for soils is difficult to evaluate, as this field of research is still underdeveloped. Mutagenicity is detected not only at contaminated sites but also in uncontaminated soil (Kool et al., 1989; Brown et al., 1985). and often bears no clear relationship with the levels of known chemical mutagens in soil (Donnelly et al.. 1991).

Biomonitoring studies in soil often deal with the study of the effect of chemical stress on the structure and function of entire ecosystems. or at least at the community level. For that purpose. studies are performed to determine the impact on litter decomposition in forest ecosystems or on communities of soil mesofauna. Foodweb models for soil organisms have been constructed for agro-ecosystems (Hunt et al.. 1987). and are developed for a coniferous forest soil. based on stratified litterbag experiments (Berg and Verhoef. 1992). With these models. effects of excessive N-input on the soil ecosystem can be estimated. No standardized guidelines exist for such studies.

Also certain species of organisms can be selected. and followed in time to detect certain deviations that may be due to the impact of chemical contamination, a process called biomonitoring. Isopods are for instance. excellent bioindicators for metal contamination. because of their ability to concentrate metals in their body tissues (Hopkin. 1989).

Tolsma et al. (1991) recommended inclusion of primary producers (the plant species Urtica dioica and Holcus lanatus), detritivores (earthworms and isopods), and carnivores (mice or moles) in a biomonitoring system for the terrestrial environment. These organisms were selected because of their capability to accumulate metals. P AH. and organochlorine pesticides.

11.6 CONCLUSIONS

The array of methods used in testing terrestrial invertebrates is diverse mainly because different tests have been developed with different aims. Many methods are still poorly described. especially in relation to the medium to which the chemical is applied, and the consequences for bioavailability.

The potential for standardization of a test system is important when one strives for international use of test methods. For some species, a standardized method may not be possible; this holds for the oribatid mite Platynothrus peltifer that has a very long life cycle and a low reproduction rate not allowing for a proper determination of effects on reproduction within a reasonable test duration. For others, such as the earthworms and the collembola Folsomia candida, tests have already been standardized at an international level, as is the case for many tests on beneficial arthropods.

The usefulness of a test system to derive environmental quality criteria also comprises the substrate used. When real or artificial soils are used as a test substrate, test results may be applied directly to derive soil quality criteria; however, this is not the case for tests on soil organisms using other substrates or exposure routes, such as water, agar, or nutrient solutions used in tests with protozoans and nematodes. Results of such tests cannot be translated directly to soil quality criteria. This conclusion holds also for exposure routes used in tests on beneficial arthropods; such tests can only be useful for the risk assessment of pesticides when test results can be related to natural exposure routes.

By one approach, each species should be tested under its optimal conditions, and inter-species harmonization of conditions (e.g., by using the same test substrate) is not to be recommended. The usefulness of more than one test on the same chemical is, however, very limited when the results for two species cannot be compared to each other. Standardization of the test substrate should, therefore, be considered in the further development of test methods.

Only a few methods are available to determine the potential risk of contaminated soils. Because of the complex nature of contaminated soil, in which often a mixture of chemicals is involved, such methods are urgently needed. Also more knowledge is needed on the toxicity of combinations of toxicants.

For a suitable risk assessment, a battery of tests should be available. Such a test battery should contain organisms representing different taxonomic groups as well as the soil community. For that purpose, the Health Council of the Netherlands (1991) selected 24 parameters to be applied for both diagnosis and prognosis of chemical effects in terrestrial ecosystems and sediments. A conclusion was that, especially for higher levels of organization, tests are lacking. For an initial screening of possible effects, a test system should consider higher plants, decomposition capacity of the soil, invertebrates, and vertebrates (the latter being outside the scope of this paper).

11.7 REFERENCES

Ames, B.N., McCann, J., and Yamasaki, E. (1975) Methods for detecting carcinogens and mutagens with the Salmonella/mammalian microsome mutagenicity test. Mutat. Res. 113, 173-215.

Anderson, J. (1978a) Some methods for assessing pesticide effects on non-target soil microorganisms and their activities. In: Hill, I.R., and Wright, S.J.L. (Eds.) Pesticide Microbiology. Academic Press, London.

Anderson, J.M. (1978b) Competition between two unrelated species of soil Cryptostigmata (Acari) in experimental microcosms. J. Anim. Ecol. 47, 787-803.

Anderson, J.M., and Ingram, J.S.I. (1989) Tropical Soil Biology and Fertility: A Handbook of Methods. C.A.B. International.

Ausmus, B.S., Dodson, G.J., and Jackson, D.R. (1978) Behavior of heavy metals in forest microcosms. III. Effects on litter-soil carbon metabolism. Water Air Soil Pollut. 10, 19-26.

Axelsson, B. (1984) Ultimate forest productivity: what is possible? In: Proceedings of the International Union of Forest Research Organisations Symposium on Forest Site and Continuous Productivity, pp. 61-69.

Bengtsson, G., Gullnarsson, T., and Rundgren, S. (1983) Growth changes caused by metal uptake in a population of Onychiurus armatus (Collembola) feeding on metal polluted fungi. Oikos 40, 216-225.

Bengtsson, G., Gullnarsson, T., and Rundgren, S. (1985) Influence of metals on reproduction, mortality and population growth in Onychiurus armatus (Collembola). J. Appl. Ecol. 22, 967-978.

Bengtsson, G., Berden, M., and Rundgren, S. (1988) Influence of soil animals and metals on decomposition processes: a microcosm experiment. J. Environ. Qual. 17, 113-119.

Berg, M.P., and Verhoef, H.A. (1992) Effects of nitrogen deposition on nutrient cycling and soil biota in coniferous forest soils in The Netherlands; A field approach. In: Teller, A., Mathy, P., and Jeffers, J.N.R. (Eds.) Responses of Forest Ecosystems to Environmental Changes, pp. 707-708. Elsevier, Amsterdam.

Berhin, F., Houba, C., and Remade, J. (1984) Cadmium toxicity and accumulation by Tetrahymena pyriformis in contaminated river waters. Environ. Pollut. (Series A) 35, 315-329.

Bond, H., Lighthart, B., Shimabuku, R., and Russell, L. (1976) Some effects of cadmium on coniferous forest soil and litter microcosms. Soil Science 121, 278-287.

Brown, K.W., Donnelly, K.C., Thomas, J.C., Davel, P., and Scott, B.R. (1985) Mutagenicity of three agricultural soils. Sci. Total Environ. 41, 173-186.

Callahan, C.A., Menzie, C.A., Burmaster, D.E., Wilborn, D.C., and Ernst, T. (1991) On-site methods for assessing chemical impact on the soil environment using earthworms: a case study at the Baird and McGuire superfund site, Holbrook, Massachusetts. Environ. Toxicol. Chem. 10, 817-826.

Casida, L.E., Jr, Klein, D.A., and Satoro, T. (1964) Soil dehydrogenase activity. Soil Sci. 98, 371-376.

Chaney, W.R., Kelly, J.M., and Strickland, R.C. (1978) Influence of cadmium and zinc on carbon dioxide evolution from litter and soil from a black oak forest. J. Environ. Qual. 7, 115-119.

Coûteaux, M.M., Morisseau, M., Célérier, M.L., and Bottner, P. (1991) Increased atmospheric CO₂ and litter quality: decomposition of sweet chestnut leaf litter with animal food webs of different complexities. Oikos 61, 54-64.

Croft, B.A. (1990) Arthropod Biological Control Agents and Pesticides. John Wiley & Sons, New York.

Dallinger, R., and Wieser, W. (1977) The flow of copper through a terrestrial food chain. I. Copper and nutrition in isopods. Oecologia 30, 253-264.

Denneman, C.A.J., and Van Straalen, N.M. (1991) The toxicity of lead and copper in reproduction toxicity tests using the oribatid mite Platynothrus peltifer. Pedobiologia 35, 305-311.

De Vries, W. (1991) Methodologies for the Assessment and Mapping of Critical Loads and of the Impact of Abatement Strategies on Forest Soils. Report 46. DLO The Winand Staring Centre, Wageningen. 109 pp.

Donnelly, K.C., Brown, K.W., Anderson, C.S., Thomas, J.C., and Scott, B.R. (1991) Bacterial mutagenicity and acute toxicity of solvent and aqueous extracts of soil samples from an abandoned chemical manufacturing site. Environ. Tox. Chem. 10, 1123-1131.

Edwards, C.A. (1983) Report of the Second Stage in Development of a Standardized Laboratory Method for Assessing the Toxicity of Chemical Substances for Earthworms. Commission of the European Communities, DOC XI/83/700, Brussels.

Edwards, C.A., and Thompson, A.R. (1973) Pesticides and the soil fauna. Res. Rev. 45, 1-79.

EEC (1985). EEC Directive 79/831. Annex V, Part C. Methods for determination of ecotoxicity. Level I. C(II)4: Toxicity for earthworms. Artificial soil test. DG XI/128/82.

Eijsackers, H. ( 1978a) Side effects of the herbicide 2,4,5- T affecting mobility and mortality of the springtail Onychiurus quadriocellatus Gisin (Collembola). Zeitschr. angew. Entomol. 86, 349-372.

Eijsackers, H. (1978b) Side effects of the herbicide 2,4,5-T affecting the isopod Philoscia muscorum Scopoli. Zeitschr. angew. Entomol. 87, 28-52.

Eijsackers, H. (1991) Litter fragmentation by isopods as affected by herbicide application. Neth. J. Zool. 41, 277-303.

Eijsackers, H., and Van de Bund, C.F. (1980) Effects on soil fauna. In: Hance, R.J. (Ed.) Interactions Between Herbicides and the Soil, pp. 255-305. Academic Press, London.

Eriksson, B. (1980) Statistical analysis of precipitation data. Part II: frequency analysis of monthly precipitation data. Swedish Meteorol. Hydrol. Inst. Reporter, No. RMK 17, 1-152.

Everts, J.W., Aukema, B., Hengeveld, R., and Koeman, J.H. (1989) Side-effects of pesticides on ground-dwelling predatory arthropods in arable ecosystems. Environ. Pollut. 59, 203-225.

Faber, J.H., and Verhoef, H.A. (1991) Functional differences between closely related soil arthropods with respect to decomposition processes in the presence and absence of pine tree roots. Soil BioI. Biochem. 23, 15-23.

Felton, J.C., Oomen, P.A., and Stevenson, J.H. (1986) Toxicity and hazard of pesticides to honey bees: harmonization of test methods. Bee World 67, 114-124.

Ferrière, G., Fayolle, L., and Bouché, M.B. (1981) Un nouvel outil, essentiel pour l'écophysiologie et l'écotoxicologie: l'élevage des lombriciens en sol artificiel. Pedobiologia 22, 196-201.

Giesy, J.P. Jr (1980) Microcosms in Ecological Research. Technical Information Center, US Department of Energy, Springfield.

Grennfelt, P., Larsson, S., Leyton, P., and Olsson, B. (1985) Atmospheric deposition in the Lake Gårdsjön area, S. W .Sweden. In: Anderson, F., and Olsson, B. (Eds.) Lake Gårdsjön: An Acid Forest Lake and Its Catchment. Ecological Bulletin 37, 101-108.

Haanstra, L., and Doelman, P. (1984) Glutamic acid decomposition as a sensitive measure of heavy metal pollution in soil. Soil Biol. Biochem. 16, 595-600.

Hågvar, S. (1988) Decomposition studies in an easily-constructed microcosms: effects of microarthropods and varying soil pH. Pedobiologia 31, 293-303.

Haight, M., Mudry, T., and Pasternak, J. (1982) Toxicity of seven heavy metals on Panagrellus silusiae: the efficacy of the free-living nematode as an in vivo toxicological bioassay. Nematologica 28, 1-11.

Harrison, A.F., Dighton, J., Hatton, J.C., and Smith, M.R. (1985) A phosphorus-deficiency bioassay for trees and grasses growing in low nutrient status soils. In: Proceedings of the VI^th International Colloquium for the Optimisation of Plant Nutrition 3, pp. 957-963. AIONP, Montpellier.

Harrison, A.F., Hatton, J.C., and Taylor, K. (1986) Application of a root bioassay for determination of P-deficiency in high-altitude grasslands. J. Sci. Food Agric. 37, 10-11.

Harrison, A.F., Ineson, P., and Heal, O.W. (1990) Nutrient Cycling in Terrestrial Ecosystems; Field Methods, Application and Interpretation. Elsevier Science, London. 454 pp.

Hassan, S.A., Bigler, F., Blaisinger, P., Bogenschtitz, H., Brun, J., Chiverton, P., Dickler, E., Easterbrook, M.A., Edwards, P.J., Englert, W.D., Firth, S.I., Huang, P., Inglesfield, C., Klingauf, F., Ktihner, C., Ledieu, M.S., Naton, E., Oomen, P.A., Overmeer, W.P.I., Plevoets, P., Reboulet, J.N., Rieckmann, W., Samsøe-Petersen, L., Shires, S.W., Stäubli, A., Stevenson, J., Tuset, J.J., Vanwetswinkel, G., and Van Zon, A.Q. (1985) Standard methods to test the side-effects of pesticides on natural ennemies of insects and mites developed by the IOBC/WRPS Working Group "Pesticides and Beneficial Organisms." OEPP/EPPO Bull. 15, 214-255.

Health Council of The Netherlands (1991) Quality Parameters for Terrestrial Ecosystems and Sediments. A Selection of Practical Exotoxicological Assessment Methods. Publication No. 91/17E. Health Council of the Netherlands, The Hague.

Heimbach, F. (1984) Correlation between three methods for determining the toxicity of chemicals for earthworms. Pest. Sci. 15, 605-611.

Heimbach, F. (1992) In: Becker, H., Edwards, P.J., Greig-Smith, P.W., and Heimbach, F. (Eds.) Ecotoxicology of Earthworms. Intercept Press, Andover, Hants., UK.

Hopkin, S.P. (1989) Ecophysiology of Metals in Terrestrial Invertebrates. Elsevier Applied Science, London.

Hopkin, S.P. (1990) Species-specific differences in the net assimilation of zinc, cadmium, lead, copper and iron by the terrestrial isopods Oniscus asellus and Porcellio scaber. J. Appl. Ecol. 27, 460-474.

Hopkin, S.P., Watson, K., Martin, M.H., and Mould, M.L. (1985) The assimilation of metals by Lithobius variegatus and Glomeris marginata (Chilopoda; Diplopoda). Bijdr. Dierk. 55, 88-94.

Hunt, H.W., Coleman, D.C., Ingham, E.R., Ingham, R.E., Elliott, E.T., Moore, J.C., Rose, S.L., Reid, C.P.P., and Morley, C.R. (1987) The detrital food web in a shortgrass prairie. BioI. Fertil. Soils 3, 57-68.

Hunter, I.R., Nicholson, G., and Thorn, A.J. (1985) Chemical analysis of pine litter: an alternative to foliage analysis? New Zeal. J. Forest Sci. 15, 101-109.

Iglisch, I. (1986) Hinweise zur Entwicklung von Testverfahren zum Nachweis subakuter Wirkungen von Chemikalien. Angew. Zool. 2, 199-218.

Inglesfield, C. (1988) Pyrethroids and terrestrial non-target organisms. Pestic. Sci. 27, 387-428.

IOBC (1988) Guidelines for testing the effects of pesticides on beneficials: short description of test methods. WPRS Bull 11, 1-143.

ISO (1991) Soil Quality-Effects of Soil Pollutants on Collembola: Determination of the Inhibition of Reproduction. ISO/TC 190/SC4/WG2, N26.

Jagers op Akkerhuis, G.A.J.M., and Van der Voet, H. (1992) A dose-effect relationship for the effect of deltamethrin on a linyphiid spider population in winter wheat. Arch. Environ. Contam. Toxicol. 22, 114-121.

Jancke, G. (1989) Modelversuche zur subakuten und subletalen Wirkung von Herbiziden auf Collembolen im Hinblick auf ein Testsystem fiir Umweltchemikalien. Zool. Beitr. N.F. 32, 261-299.

Jenkinson, D.S., Adams, D.E., and Wild, A. (1991) Model estimates of CO₂ emissions from soil in response to global warming. Nature 351, 304-306.

Jepson, P.C. (1989) The temporal and spatial dynamics of pesticide side-effects on non-target invertebrates. In: Jepson, P.C. (Ed.) Pesticides and Non-target Invertebrates, pp. 95-127. Intercept Ltd, Wimborne.

Kirsten, W.J. (1979) Automatic methods for the simultaneous determination of carbon, hydrogen, nitrogen, and sulphur, and for phosphor alone in organic and inorganic materials. Anal. Chem. 51,1173-1175.

Kokta, C. (1992) A laboratory test on sublethal effects of pesticides on Eisenia fetida. In: Becker, H., Edwards, P.J., Greig-Smith, P.W., and Heimbach, F. (Eds.) Ecotoxicology of Earthworms, pp. 55-62. Intercept Press, Andover, Hants.

Kool, H.J., van Kreyl, C.F., and Persad, S. (1989) Mutagenic activity in groundwater in relation to mobilization of organic mutagens in soil. Sci. Total Environ. 84, 185-199.

Kopezski, H. (1992) Versuch einer aktiven Bioindikation mit den bodenlebenden Collembolen-Arten Folsomia candida (Willem) und Heteromurus nitidus (Templeton) in einem Buchemwald-Ökosystem. Zool. Anz. 228, 82-90.

Kula, H. (1992) Measuring effects of pesticides on earthworms in the field: test design and sampling methods. In: Greig-Smith, P.W., Becker, H., Edwards, P.J., and Heimbach, F. (Eds.) Ecotoxicology of Earthworms, pp. 90-99. Intercept Press, Andover, Hants.

Likens, G.E., Bormann, F.H., Pierce, R.S., Eaton, J.S., and Johnson, N.M. (1977) Biogeochemistry of a Forested Ecosystem. Springer Verlag, New York.

Ljungdahl, L.G., and Eriksson, K.-E. (1985) Ecology of microbial cellulose degradation. In: Marshall, K.C. (Ed.) Advances in Microbial Ecology. Plenum Press, New York.

Lloyd, D. (1985) Simultaneous dissolved O₂ and redox measurements: use of polarographic, bioluminescence and mass spectrometric monitoring combined with dual wavelength spectrometry. In: Degn, H., Cox, R.P., and Toftlund, H. (Eds.), pp. 37-53. Reidel, Dordrecht, The Netherlands.

McIntosh, R. (1984) Fertiliser Experiments in Established Conifer Stands. Forestry Commission Forest Record 127. HMSO, London.

Marigomez, J.A., Angulo, E., and Saez, V. (1986) Feeding and growth responses to copper, zinc, mercury and lead in the terrestrial gastropod Arion ater (Linné). J. Moll. Stud. 52, 68-78.

Marquenie, J.M., and Simmers, J.W. (1988) A method to assess potential bioavailability of contaminants. In: Edwards, C.A., and Neuhauser, E.F. (Eds.) Earthworms in Waste and Environmental Management, pp. 367-375, SPB Acad. Publ., The Hague.

Meincke, K.-F., and Schaller, K.-H. (1974) Uber die Brauchbarkeit der Weinbergschnecke (Helix pomatia L.) im Freiland als Indikator fur die Belastung der Umwelt durch die Elemente Eisen, Zink und Blei. Oecologia (Berl.) 15, 393-398.

Menzie, C.A., Burmaster, D.E., Freshman, J.S., and Callahan, C.A. (1992) Assessment of methods for estimating ecological risk in the terrestrial component: a case study at the Baird and McGuire superfund site in Holbrook, Massachusetts. Environ. Toxicol. Chem. 11, 245-260.

Mola, L., Sabatini, M.A., Fratello, B., and Bertolani, R. (1987) Effects of atrazine on two species of Collembola (Onychiuridae) in laboratory tests. Pedobiologia 30, 145-149.

Moser, H., and Wieser, W. (1979) Copper and nutrition in Helix pomatia (L.). Oecologia (Berl.) 42, 241-251.

Mothes-Wagner, U., Reitze, H.K., and Seitz, K.-A. (1992) Terrestrial multi species toxicity testing. 1. Description of the multi species assemblage. Chemosphere 24, 1653-1667.

NEN 5796 ( 1989) Bodem-Bepaling van de urease-aktiviteit (Dutch standard for determining soil urease activity).

Nilsson, O., and Grennfelt, P. (1988). Critical Loads for Sulphur and Nitrogen. pp. 418. Miljøreport 1988: 15. Nordic Council of Ministers, København.

Nyberg, D., and Bishop, P. (1983) High levels of phenotypic variability of metal and temperature tolerance in Paramecium. Evolution 37, 341-357.

OECD (1984a) Guideline for testing of chemicals no.207. Earthworm acute toxicity tests. Adopted 4 April 1984. Organisation for Economic Co-operation and Development, Paris.

OECD (1984b) Guideline for testing of chemicals no.208. Terrestrial plant growth test. Adopted 4 April 1984. Organisation for Economic Co-operation and Development, Paris.

OECD (1989) Report of the OECD Workshop on Ecological Effects Assessment. OECD Environment Monograph No.26. Organisation for Economic Co-operation and Development, Paris.

OEPP/EPPO (1991) Draft Guideline for Evaluating the Hazards of Pesticides to Honey Bees, Apis mellifera L. European Plant Protection Organization.

Rossell, D., and Tarradellas, J. (1991) Dehydrogenase activity of soil microflora: significance in ecotoxicological tests. Environ. Toxicol. Water Qual. 6, 17-33.

Römbke, J. (1989) Enchytraeus albidus (Enchytraeidae, Oligochaeta) as a test organism in terrestrial laboratory systems. Arch. Toxicol. 13, 402-405.

Römbke, J., and Knacker, T. (1989) Aquatic toxicity test for enchytraeids. Hydrobiol. J. 180, 235-242.

Russell, L.K., DeHaven, J.I., and Botts, R.P. (1981) Toxic effects of cadmium on the garden snail (Helix aspersa). Bull. Environ. Contam. Toxicol. 26, 634-640.

Samsøe-Petersen, L. (1990) Sequences of standard methods to test effects of chemicals to terrestrial arthropods. Ecotox. Environ. Safety 19, 310-319.

Somerville, L., and Greaves, M.P. (Eds.) (1987) Pesticide Effects on Soil Microflora. Taylor & Francis, London.

Stewart-Oaten, A., Murdoch, W.W., and Parker, K.R. (1986) Environmental impact assessment: "Pseudoreplication" in time? Ecology 67: 929-940.

Sturhan, D. (1986) Influence of heavy metals and other elements on soil nematodes. Rev. Nématol. 9, 311.

Subagja, J., and Snider, R.J. (1981) The side effects of the herbicides atrazine and paraquat upon Folsomia candida and Tullbergia granulata (Insecta, Collembola). Pedobiologia 22, 141-152.

Tabatabai, M.A., and Bremner, J.M. (1969) Use of p-nitrophenyl phosphate for assay of soil phosphatase activity. Soil Biol. Biochem. 1, 301-307.

Teuben, A., and Roelofsma, T.A.P.J. (1990) Dynamic interactions between functional groups of soil arthropods and microorganisms during decomposition of coniferous litter in microcosm experiments. Biol. Fertil. Soils 9, 145-151.

Teuben, A., and Verhoef, H.A. (1992) Relevance of micro- and mesocosm experiments for studying soil ecosystem processes. Soil Biol. Biochem. 24(11), 1179-1183.

Thalmann, A., (1968) Zur Methodik der Bestimmung der Dehydrogenaseaktivität im Boden mittels Triphenyltetrazoliumchlorid (TTC). Landw. Forsch. 21, 249-258.

Thompson, A.R., and Gore, F.L. (1972) Toxicity of twenty-nine insecticides to Folsomia candida: laboratory studies. J. Econ. Entomol. 65, 1255-1260.

Tolsma, D.J., Van Hattum, B., Denneman, W.D., Aiking, H., Wegener, J.W.M., and Cofino, W.P. (1991) Aanbevelingen voor de lntegratie Vanbiotische Parameters in het RIVM-Bodemkwaliteitsmeetnet. Report No. R-9l/16. Institute for Environmental Studies, Vrije Universiteit, Amsterdam.

Tomlin, A.D. (1977) Toxicity of soil applications of the fungicide benomyl, and two analogues to three species of Collembola. Can. Ent. 109, 1619-1620.

Ulrich, B., and Mayer, R. (1980) Throughfall and stemflow: a review. In: Nicholson, I.A., Paterson, I.S., and Last, F.T. (Eds.) Methods for Studying Acid Precipitation in Forest Ecosystems, pp. 21-27. Cambridge, UK.

Van Beelen, P., Fleuren-Kemilä, A.K., Huys, M.P.A., van Mil, A.C.H.A.M., and van Vlaardingen, P.L.A. (1990) Toxic effects of pollutants on the mineralization of substrates at low experimental concentrations in soils, subsoils and sediments. In: Arendt F., Hinseveld, M., and van den Brink, W.J. (Eds.) Contaminated Soil. p. 431-438. Kluwer Academic, The Netherlands.

Van Beelen, P., Fleuren-Kemilä, A.K., Huys, M.P.A., van Montfort, A.C.P., and van Vlaardingen, P.L.A. (1991) The toxic effect of pollutants on the mineralization of acetate in subsoil microcosms. Env. Tox. Chem. 10, 775-789.

Van Capelleveen, H.E. (1987) Ecotoxicity of Heavy Metals for Terrestriallsopods. Ph.D. Thesis, Vrije Universiteit, Amsterdam.

Van der Gaag, M.A. (1989). Rapid detection of genotoxins in waste water: New perspectives with the sister-chromatid exchange assay "in vivo" with Nothobranchius rachowi. In: Løkke, H., Tyle, H., and Bro-Rasmussen, F. (Eds.) Proceedings of the First European Conference on Ecotoxicology, pp. 259-262. DIS Congress Service, Vanløse.

Van Gestel, C.A.M. (1992) The influence of soil characteristics on the toxicity of chemicals for earthworms: a review. In: Becker, H., Edwards, P.J., Greig-Smith, P.W., and Heimbach, F. (Eds.) Ecotoxicology of Earthworms, pp. 44-54. Intercept Press, Andover, Hants., UK.

Van Gestel, C.A.M., and Ma, W.-C. (1990) An approach to quantitative structure-activity relationships in terrestrial ecotoxicology: earthworm toxicity studies. Chemosphere 21, 1023-1033.

Van Gestel, C.A.M., Adema, D.M.M., De Boer, J.L.M., and De Jong, P. (1988) The influence of soil clean-up on the bioavailability of metals. In: Wolf, K., Van den Brink, W.J., and Colon, F.J. (Eds.) Contaminated Soil, pp. 63-65. Kluwer Academic, Dordrecht, The Netherlands.

Van Gestel, C.A.M., van Dis, W.A., van Breemen, E.M., and Sparenburg, P.M. (1989) Development of a standardized reproduction toxicity test with the earthworm species Eisenia fetida andrei using copper, pentachlorophenol, and 2,4-dichloroaniline. Ecotoxicol. Environ. Safety 18, 305-312.

Van Gestel, C.A.M., Dirven-van Breemen, E.M., and Kamerman,J.W. (1993) The influence of soil clean up on the bioavialability of heavy metals for earthworms and plants. In: Proceedings of the EUROSOL Conference, Maastricht, 6-12 September 1992.

Van Kessel, W.H.M., Brocades Zaalberg, R.W., and Seinen, W. (1989) Testing environmental pollutants on soil organisms: a simple assay to investigate the toxicity of environmental pollutants on soil organisms using cadmium chloride and nematodes. Ecotox. Environ. Safety 18, 181-190.

Van Straalen, N.M., and Verweij, R.A. (1991) Effects of benzo(a)pyrene on food assimilation and growth efficiency in Porcellio scaber (Isopoda). Bull. Environ. Contam. Toxicol. 46, 134-140.

Van Straalen, N.M., Schobben, J.H.M., and de Goede, R.G. (1989) Population consequences of cadmium toxicity in soil microarthropods. Ecotoxicol. Environ. Safety 17, 190-204.

Van Straalen, N.M., and Van Gestel, C.A.M. (1992a) Ecotoxicological test methods using terrestrial arthropods. Detailed Review Paper for the OECD Test Guidelines Programme.

Van Straalen, N.M., and Van Gestel, C.A.M. (1992b) Soil, Chapter 15. In: Calow, P. (Ed.) Handbook of Ecotoxicology. Blackwell Scientific, Oxford.

Van Wensem, J. (1989) A terrestrial micro-ecosystem for measuring effects of pollutants on isopod-mediated litter decomposition. Hydrobiologia 188/189, 507-516.

Van Wensem, J., Jagers op Akkerhuis, G.A.J.M., and van Straalen, N.M. (1991) Effects of the fungicide triphenyltin hydroxide on soil fauna mediated litter decomposition. Pest. Sci. 32, 307-316.

Van Wensem, J., Verhoef, H.A., and van Straalen, N.M. (in press) Litter degradation stage as a prime factor for isopod interaction with mineralization processes. Soil Biol. Biochem.

Verhoef, H.A., and Brussaard, L. (1990) Decomposition and nitrogen mineralization in natural and agroecosystems: the contribution of soil animals. Biogeochemistry 11, 175-211.

Verhoef, H.A., and de Goede, R.G.M. (1985) Effects of collembolan grazing on nitrogen dynamics in a coniferous forest. In: Fitter A.H., Atkinson, D., Read, D.J., and Usher, M.B. (Eds.) Ecological Interactions in Soil, pp. 367-377. Blackwell Scientific, Oxford.

Verhoef, H.A., and Meintser, S. (1991) The role of soil arthropods in nutrient flow and the impact of atmospheric deposition. In: Veeresh, G.K. (Ed.) Advances in Management and Conservation of Soil Fauna, pp. 497-506. Vedam Books, New Delhi.

Verhoef, H.A., and Dorel, F.G. (1988) Effects of ammonia deposition on animal-mediated nitrogen mineralization and acidity in coniferous forest soils in The Netherlands. In: Mathy, P. (Ed.) Air Pollution and Ecosystems, pp. 847-851. Reidel, Dordrecht, The Netherlands.

Verhoef, H.A., Dorel, F.G., and Zoomer, H.R. (1989) Effects of nitrogen deposition on animal-mediated nitrogen mobilization in coniferous litter. Biol. Fertil. Soils 8, 255-259.

Westheide, W., Bethke-Beilfuss, D., and Gebbe, J. (1991) Effects of benomyl on reproduction and population structure of enchytraeid oligochaetes (Annelida)-sublethal tests on agar and soil. Compar. Biochem. Physiol. 100C, 221-224.

Williams, W.J. (1979) Handbook of Anion Determination, pp. 500-609. Butterworths, London.

Wohlgemuth, D., Kratz, W., and Weigmann, G. (1990) The influence of soil characteristics on the toxicity of an environmental chemical (cadmium) on the newly developed mono-species test with the springtail Folsomia candida (Willem). In: Barcelo, J. (Ed.) Environmental Contamination, Fourth International Conference, Barcelona, October 1990, pp. 260-262. CEP Consultants Ltd, Edinburgh.

11.1 INTRODUCTION
11.2 QUANTIFICATION OF INPUT OF CHEMICALS IN THE SOIL
	11.2.1 MAJOR ELEMENTS
		11.2.1.1 Wet Deposition
		11.2.1.2 Dry Deposition
		11.2.1.3 Nitrogen Fixation
		11.2.1.4 Mineral Weathering
	11.2.2 ORGANIC CHEMICALS AND METALS
11.3 METHODS TO QUANTIFY EFFECTS OF CHEMICAL INPUT ON ABIOTIC SOIL CHARACTERISTICS
	11.3.1 NITROGEN
	11.3.2 SULPHUR
	11.3.3 PHOSPHORUS
	11.3.4 CARBON
	11.3.5 ORGANIC CHEMICALS AND METALS
11.4 METHODS TO ASSESS THE POTENTIAL RISK OF CHEMICALS FOR SOIL ORGANISMS
	(PROGNOSIS)
	11.4.1 SINGLE-SPECIES LABORATORY TOXICITY TESTS
		11.4.1.1 Higher Plants
		11.4.1.2 Protozoans and Nematodes
		11.4.1.3 Isopods and Millipedes
		11.4.1.4 Oribatid Mites
		11.4.1.5 Collembola
		11.4.1.6 Enchytraeids
		11.4.1.7 Lumbricids
		11.4.1.8 Molluscs
		11.4.1.9 Beneficial Arthropods
	11.4.2 MICROCOSM TESTS INCLUDING THOSE ON SOIL MICROFLORA
		11.4.2.1 Microcosm Tests
		11.4.2.2 Tests on Soil Microbial Processes
		11.4.2.3 Tests on Enzyme Activity in Soil
	11.4.3 FIELD TESTS
		11.4.3.1 Cage Tests Using Selected Arthropod Species
		11.4.3.2 Honey Bee Field Test
		11.4.3.3 Arthropod Fauna in Arable Crops
		11.4.3.4 Arthropod Fauna in Orchards
		11.4.3.5 Earthworm Field Tests
11.5 METHODS TO ASSESS THE IMPACT OF SOIL CONTAMINATION ON SOIL ORGANISMS
	(DIAGNOSIS)
	11.5.1 LABORATORY AND FIELD BIOASSAYS
	11.5.2 MUTAGENICITY TESTS
	11.5.3 FIELD STUDIES (BIOMONITORING)
	11.6 CONCLUSIONS
11.7 REFERENCES


	N	S	P	K	Na	Ca	Mg	Cl
Wet deposition	X	X		X	X	X	X	X
Dry deposition
Gaseous input	X	X
Particle input	X	X	?	?	X	?	X	X
N₂-fixation	X
Mineral weathering		X	X	X	X	X	X

11	Methods to Assess the Effects of Chemicals on Soils
	H. A. Verhoef and C. A. M. van Gestel
	Vrije Universiteit, The Netherlands