The effects of pesticides on soil microorganisms are studied for risk assessment because of the importance of microbes in soil processes. These studies often comprise laboratory studies, such as soil respiration and mineralization inhibition tests, or small-scale field studies. Many of the field studies have been carried out using pesticides or other substances that are no longer used for crop protection in the Nordic countries. Furthermore, the effects of whole-plant protection programs have been studied in only a few cases. Jones et al. (1992) studied the effects of many years of plant protection programs on soil microbes in connection with wheat cultivation. They observed higher microbial activity after being supervised and reduced than after conventional full-insurance pesticide use. However, the possible detrimental effects on soil microbes have not been very drastic or long lasting.

Harmful effects of a pesticide can often be observed in laboratory studies with soil types that allow high bioavailability of the chemical. These "worst case" studies are needed for the initial risk assessment of the pesticide. The bioavailability criteria are met by the soil that was selected by an OECD working group for the assessment of effects of chemicals on soil microorganisms (70% sand and only 0.5–1.5% organic carbon) (OECD, 2000). Since in this soil the adsorption of the chemical is minimal and its bio-availability to microbes is maximal, tests with other soils are generally regarded as unnecessary. In field studies, natural soils with clay and high organic carbon content may have such a high adsorption capacity that no effects of the same pesticide on biota are detected. In Finland, most of cereal crop production is located in areas with soils of high clay and organic carbon content. This probably affects the bioavailability of the pesticides used.

In field or laboratory studies, pesticide effects on microbial biota are often assessed by measuring microbial activity relative to carbon and nitrogen metabolism. These measurements such as soil respiration (Nordgren, 1988) and nitrification ( Müller et al., 1981) were adopted in the EPPO/Coe pesticide risk assessment guidelines ( OEPP/EPPO, 1994). Corresponding test guidelines were also adopted by OECD for testing chemical effects on soil carbon and nitrogen mineralization ( OECD, 2000). The soil microbial biomass can be assessed using several methods. One is measurement of soil ATP content, indicating the amount of active microbes ( Jenkinson et al., 1979; West et al., 1986). The sum parameters of microbial activity and biomass can indicate a severe detrimental effects on the total microbial community, but fail to indicate the possible harmful changes in microbial community structure. The effects of a pesticide may be very different on fungi and bacterial ( Jones et al., 1992; Malkomes, 1992). Because of the microbial community changes in tolerance and species composition in the stressed environment, the measurable changes can be masked and are difficult to prove statistically ( Giller et al., 1998).

The potential toxicity of chemicals can be assessed with single-species bacterial toxicity tests (Bitton and Dutka, 1986). Often these tests are more sensitive than multispecies tests in which the different responses of different microbial species can mask the effects on the most sensitive ones. These simple tests can also be used for assessment of the bioavailability of the chemical.

The inhibition of light production by luminescent Vibrio fischeri indicates a disturbance in the energy metabolism of this heterotrophic bacterium. Because the luminescence pathway is a direct branch of the electron transport chain, the luminescence measurement is a measure of the metabolic status of this bacterium (Hastings, 1978). Hence the reduction in bacterial luminescence when these bacteria are exposed to chemicals, soil, or soil extracts can be used as an indicator of potential toxicity. The bacterium Pseudomonas putida represents a common heterotrophic microorganisms. P. putida cells are cultured under specified conditions in a defined medium with different concentrations of chemical over several generations. During this cultivation and exposure, toxic substances may inhibit multiplication of the bacteria (Brinkmann and Kühn, 1977). These two simple bacterial tests measuring different aspects—energy metabolism and cell multiplication—could be used to rank the potential harmful effects of pesticides.

The main aim of this study was to assess the effects of two pesticide regimens (full insurance compared with supervised and reduced pesticide use) and two cultivation techniques (conventional tillage and fertilization vs no tillage and less fertilization) on soil microbes in field studies. Another aim was to evaluate the field results with different laboratory toxicity and bioavailability tests

2. MATERIALS AND METHODS

2.1. Field study design and sampling

Field studies were carried out near Jokioinen in southern Finland (60°8′N, 23°5′E). The size of the six replicate study fields (plots) for each treatment was 54×120 m². They were set in the area to perform a randomized block experiment. There was some heterogeneity between the study fields in organic matter (2.6–4.1%) and clay content (>70%), but all the soil in the area would be classified as clay soil with a pH in water of 5.6 (spring) to 6.15 (autumn). The soil had not been treated with any pesticide during the 7 preceding years.

During two barley growing seasons, different herbicides, fungicides, and insecticides were used according to the general "full-insurance" guidance calendar (treatment C) or the actual monitored need for "supervised" plant protection (treatment D). The herbicides used were chlorsulfuron, MCPA, bentazone, the fungicides carboxin–imazalin and propiconazole, and the insecticides dimethoate and pirimicarb (Table 1).

 

 

Table 1. Pesticide treatments in conventional full insurance (AC, BC) and supervised (AD, BD) treatment regimens during the two growing seasons
 

 

The effects of two different cultivation techniques were also studied. These were normal tillage and fertilizer use (treatment A) and cultivation without tillage and with reduced use of fertilizers (treatment B). Hence the four different treatment compilations were (AC) conventional full-insurance pesticide use, normal tillage; (AD) reduced pesticide use, normal tillage; (BC) conventional full-insurance pesticide use, no tillage, less fertilizer; and (BD) reduced pesticide use, no tillage, less fertilizer.

During both years the sampling was performed before and after pesticide treatments, and the soil was also sampled twice at the end of the growing season to assess the acute and long-term effects after crop harvest. From each of the six replicate study fields (plots), 30 surface soil (0–5 cm core) samples were pooled together. These sieved samples were kept at +4°C for a maximum of 4 days before nitrification potential and ATP-content analyses. Samples for soil respiration measurements were frozen immediately at −20°C for later analysis.

2.2. Soil respiration measurement

For the soil basal respiration rate measurements, 60-g (dry weight) samples of thawed soil samples were placed in respirometer cuvettes. Soil moisture content was adjusted to 50% of the water holding capacity (WHC) and samples were stored for 1 week at 20°C to stabilize the respiration level. The cuvettes were placed in the respirometer, which captured CO₂ in KOH solution and measured the change in conductivity (Nordgren, 1988). Soil respiration was measured for 1 week at 20°C.

2.3. Soil nitrification potential measurement

Soil nitrification potential was assessed as the capacity of the soil to reduce ammonium ((NH₄)₂SO₄) to nitrate (–NO₃) (Müller et al., 1981). The moisture content of 50-g (dry weight) fresh soil samples was adjusted to 50% WHC, and 50 mL of 20 mM (NH₄)₂SO₄ solution was added to the soil in 250-mL Erlenmeyer flasks. The flasks were incubated at 20°C for 20 days and the concentration of NO₃⁻ ions was measured using a nitrate electrode (Orion 920 pH-meter). The nitrification potential NO₃–N mg/g was calculated on the basis of the dry weight of soil samples.

2.4. Soil ATP content analysis

For the soil ATP analysis, samples of 10 g fresh wt of sieved soil were placed in 100-mL vials containing 10 mL 20% trichloroacetic acid and 10 mL 8 mM EDTA and shaken vigorously for 30 min (Vanhala and Ahtiainen, 1994). Suspended solids were removed by filtration (Schleicher & Schuell 604 paper filter). Aliquots of 0.5 mL of this filtrate were mixed with 0.5 mL 0.1 M Tris+2 mM EDTA buffer in Eppendorf tubes. These tubes were kept on ice. The samples were further diluted 50 times in the same buffer for the measurement. ATP concentrations were measured with a BioOrbit luminometric assay kit in a BioOrbit 1251 luminometer (BioOrbit, Turku, Finland). The ATP content of the samples was then calculated on the basis of the dry weight of the soil samples.

2.5. Microbiological toxicity and bioavailability testing of chemicals

To assess and score potential harmful effects on microbes, the pesticides (products used in field experiments) were tested with two standardized bacterial toxicity tests in aqueous solutions. The soil respiration inhibition potential in the same soil as in the field studies was measured for four selected pesticides and the bioavailability of the pesticides in this soil was assessed by a solid-phase modification of the luminescent bacteria test.

Toxicity of the pesticides was assessed according to the standardized luminescence bacteria test procedure (ISO, 1998). The luminescence inhibition test was accomplished by combing different dilutions of the pesticide with V. fischeri DSM 7151 (identical to NRRL B-1117). The test tubes were incubated in a 15°C water bath. After 30 min incubation luminescence was measured with a luminometer (BioOrbit 1253). EC₅₀ values of the pesticides were estimated by luminescence inhibition in different dilutions compared with a deionized water control.

Pesticides were also tested with an automated modification of the standard P. putida growth inhibition test (ISO, 1995) using the Bioscreen C analyzer (Labsystems). In this test, P. putida MIGULA (DSM 50026) bacteria were grown in liquid medium in special cuvettes and the turbidity due to bacterial growth was measured by vertical photometry. EC₅₀ values of the pesticides were then estimated by growth inhibition (%) in different dilutions compared with a deionized water control.

Soil basal respiration inhibition by dimethoate, propiconazole, and chlorsulfuron was measured using an automated respirometer (Nordgren, 1988). The same clay soil as in the field studies was supplemented with different amounts of pesticides as commercial formulations (10, 50, 100, 500 mg/kg). Soil respiration was measured after intervals of 0–2, 2–4, 19–21, and 23–25 days during 1 month of incubation at 20°C. Soil moisture content was adjusted to 50% of the WHC at the beginning and end of each week.

The bioavailability of the pesticides in the soil was assessed by a solid-phase modification of the luminescent bacteria tests (Brouwer et al., 1990). Different amounts of pesticides formulations (10, 50, 100, 500 mg/kg soil) were added to the soil, and the toxicity of the soil was tested with the soil-contact luminescent bacteria test at the beginning of the exposure, after 2 and 24 h, and after 12 days. Homogenized soil subsamples of 5 g from different treatments were weighed into 50-mL centrifuge tubes and diluted with 16 mL deionized water. This suspension was supplemented with 2 mL 20% NaCl solution and with a 2-mL inoculum of V. fischeri (strain NRRL B 1117) luminescent bacteria. Tubes were incubated for 15 min at 15°C in a water bath. After incubation the tubes were quickly centrifuged (5 min, 1660 rpm, 300g) to separate the solids from the bacteria. The luminescence of a 1-mL sample of this supernatant was measured with a luminometer (BioOrbit 1253). The percentage of luminescence inhibition in pesticide-treated soil compared with nontreated soil was measured and calculated for different exposure periods.

2.6. Data analysis

Field data on microbial activity and biomass of six replicate study fields of each treatment on six sampling occasions, separately during the two growing seasons, were evaluated with PROC UNIVARIATE (SAS Institute Inc., 1998) for normal distribution (a log transformation was applied as needed) for the analysis of variance (ANOVA) to compare the treatments. The ANOVA was performed for each sampling of each treatment (SAS Institute Inc., 1998) and if any significance was found at a risk level of 5% or lower (P<0.05), pairwise t tests were performed.

Additionally, possible effects of other environmental factors (soil water content, organic carbon content, NO₃⁻ and NH₄⁺ concentrations) were also statistically evaluated by stepwise least-squares regression analysis (LSA) (SAS Institute Inc., 1998).

3. RESULTS

3.1. Microbiological activity and biomass measurements

Seasonal succession of soil respiration, nitrification potential, and soil ATP content in four different pesticide and cultivation treatments (AC, normal pesticides and normal cultivation; AD, reduced pesticides and normal cultivation; BC, normal pesticides and lighter cultivation; BD, reduced pesticides and lighter cultivation) during the first summer is illustrated in Fig. 1, and the results during the next summer in Fig. 2. In particular, the beginning of the first summer was very dry, with only occasional rainfall (Fig. 3). This was strongly reflected in soil nitrification potential and in soil respiration level when comparing the two summers.

 

Fig. 1. Seasonal succession of soil respiration (a), nitrification potential (b), and soil ATP content (c) as means of six replicate plots and the SEM (bars) in different pesticide and cultivation treatments (AC, AD, BC, BD) during the summer of 1992 (S=sowing, P1 and P2=pesticide spraying, H=harvest).

 

 

Fig. 2. Seasonal succession of soil respiration (a), nitrification potential (b), and soil ATP content (c) as means of six replicates and the SEM (bars) in different pesticide and cultivation treatments (AC, AD, BC, BD) during the summer of 1993 (S=sowing, P1 and P2=pesticide spraying, H=harvest).

 

 

Fig. 3. Weather conditions during both growing seasons 1992 (a) and 1993 (b) as pentadi values: mean temperature (broken line), effective temperature sum, above +5°C (histogram), and rainfall (black bars). The time lines under the figures indicate the timing of sowing (P), pesticide treatments (T1=pesticide–insecticide, T2=fungicide–growth regulator), crop harvesting (H), and soil sampling (I–VI).

 

At the beginning of both growing seasons, before the first pesticide spraying, no acute effects of the seed disinfection were observed (ANOVA). The immediate effects of the first herbicide and insecticide treatment could be seen at the third sampling. At this sampling (III) the nitrification potential in herbicide-treated C plots was significantly higher (NOVA: P<0.02) compared with that in nontreated D plots (Fig. 1b). However, at the next sampling (IV) the difference was no longer statistically significant. This increase in nitrification after herbicide treatment was not observed during the second year ( Fig. 2b) although the treatment was identical.

The acute effects of the second fungicide and growth regulator treatment should have been observed at the fourth sampling (IV) in both years. During the summer of 1992 the ATP content of the soil (Fig. 1c) was higher in fields with conventional cultivation techniques (AC, AD) probably due to the greater use of fertilizers. However, the difference was not statistically significant (ANOVA, P=0.167).

The last two sampling occasions (V and VI) were analyzed to assess the longer-term effects of whole-plant treatment programs on soil microbes. However, no statistically significant (ANOVA) effects of the treatments were observed at the end of the summer.

Because the pesticide treatments did not generally have strong or long-lasting significant effects on soil microbial activity and biomass, the possible effects of other environmental factors (soil water content, organic carbon content, NO₃⁻ and NH₄⁺ concentrations) were also statistically evaluated by stepwise LSA. The model used chose factors having at least P=0.15 explanatory power. However, the explanatory power was weak (R²<20%) and inconsistent and no adequate statistical explanations could be identified using these tested factors.

3.2. Microbial toxicity and bioavailability testing of the chemicals

In the laboratory studies the toxicity of the fungicide, in particular, and of certain herbicides was clearly detected by the bacterial toxicity tests (Table 2). The luminescent bacteria test, in particular, was very sensitive to propiconazole and MCPA (EC₅₀=1.25 g/L). In soil respiration inhibition studies (Table 2) with tested clay soil, inhibition was observed only at concentrations much higher than those used in the field. The lowest effective concentration (LOEC) for propiconazole was 125 mg/kg.

 

 

Table 2. Toxicity of pesticides in bacterial toxicity tests (EC₅₀) and in the soil respiration inhibition test (LOEC) in the same clay soil as in the field studies
 

 

In the bioavailability studies with the direct contact modification of the luminescent bacteria test, it was observed that the toxicity of the soil was already reduced 24 h after pesticide treatment (Fig. 4). This was clearly demonstrated by the tests with propiconazole (Tilt 250 EC). The reduction in toxicity was faster with dimethoate (Roxion) an chlorsulfuron (Glean 20 DF), although the results were less consistent.

 

Fig. 4. Toxicity of soil at different times after three pesticide treatments—Glean 20 DF herbicide (a), Roxion insecticide (b), and Tilt 250 EC fungicide (c)—in various concentrations (50, 100, and 500 mg/kg product).

4. DISCUSSION

No strong or long-lasting significant effects (P<0.05) of the pesticide treatments were detected in the filed studies. The only observed effect (P=0.02) was the contemporary elevated nitrification potential after chlorsulfuron treatment in the first summer. Microbial activity and biomass mainly followed the weather conditions. The study strategy simulated realistically normal crop production condition, and the sampling program and analyses of field samples were planned to be adequate to assess possible effects. However, because of the large scale (temporal and spatial) of the field experiment the variability of the environmental factors (aerial soil heterogeneity, weather) was too high to enable strong statistical conclusions. The observed changes tended not to be statistically significant (ANOVA). Hence it was not possible to conclude that the reduced and supervised pesticide use (AD, BD) was less harmful to soil microbes than conventional use (AC, BC), as did Jones et al. (1992). With respect to pesticide use, the supervised AD and BD treatments can be regarded almost as zero controls when only one insecticide treatment was needed during the first growing season ( Table 1). The chosen variables of microbial activity and biomass were on certain occasions sufficiently sensitive to detect some differences caused by the pesticide treatments, cultivation practices, or other environmental factors. This could also be seen statistically in the elevated nitrification potential immediately after the herbicide treatment during the first season and in biomass differences (not statistically significant) caused by cultivation techniques including fertilization and tillage. There was no positive control such as a very heavy pesticide treatment to validate the sensitivity of the field analysis. In this respect the field study differed from standardized toxicity tests under laboratory conditions. This kind of control treatment, known to have an effect on microbes, could have provided more information on the variability of the study plots and on method sensitivity.

Many pesticides have been assessed for their side effects on carbon and nitrogen transformation in soils. Anderson (1992) reported studies in which the effects of 68 different pesticides as normal 5-fold and 10-fold doses on soil carbon and nitrogen mineralization processes were assessed. Only 5 of 68 pesticides caused a decrease of 15% in nitrification potential at the normal dose. Moreover, only 15 of the pesticides studied had some effects even at the highest doses. The pesticides were not named and hardly any of them were the same as those used in this survey.

In the laboratory soil respiration inhibition tests, some effects of propiconazole would have been expected on the basis of the toxicity testing of water solutions (EC₅₀ was 1.25 mg/L in the luminescent bacteria test, Table 2). However, the only inhibitive effect on soil respiration was observed at the highest concentration tested (125 mg/kg propiconazole) in the same soil as used in the field experiments. The high clay (>70%) and organic matter (2.6–4.1%) contents of the studied soil appeared to have an effect on the bioavailability of the chemicals, hence reducing the exposure of organisms to the pesticides studied. The reduction in soil toxicity to luminescent bacteria could clearly be seen in the bioavailability assays (Fig. 4). This toxicity change during 1 day could not be explained by the rapid degradation of the chemicals. The reduction was probably due to the rapid adsorption to soil particles and hence reduced bioavailability of the chemicals. Similar results have been observed in soil animal studies in which the degradation of dimethoate has also been documented ( Martikainen et al., 1998).

During the field experiments propiconazole was found in the surface soil generally at a maximum concentration of 2 mg/kg (data not shown). Hence the lack of effects in the field was reasonable. The fungicide propiconazole selectively affects fungi, and it has been reported that it decreased the length of fungal hyphae in a field test (Emholt, 1992). However, agricultural soils tend to have bacterium-based energy channels rather than fungal mechanisms as in forest soils ( Moore and Hunt, 1988). In earlier laboratory studies propiconazole has caused dose–response effects on soil respiration only at concentrations higher than those used in the field ( Emholt, 1992).

In the field studies and in the laboratory assessments (inhibition of soil respiration, nitrogen and carbon mineralization processes), a natural heterogenic population of soil microbes is exposed to the chemical for various periods. If there are sufficient resistant species or taxonomic groups in the exposed community, the activity of these microbes can camouflage the possible effects on other, more sensitive microbes. This is not the case with single-species tests, which can be seen as more sensitive but which are considered to be ecologically less relevant. In agricultural soils the key microbial processes in nutrient circulation should be assessed and it should be understood that one intention in cultivation is to control the diversity of the microbial community for more favorable crop production conditions.

The study design and the selection of analytical methods are very important in large field experiments. Varying environmental conditions can interfere with the treatment responses being studied. However, in the case of more severe soil contamination due to heavy metal deposition, the effects on soil microbes have been assessed by measurements similar to those used in this study (Vanhala and Ahtiainen, 1994).

Boreal environmental conditions and poor bioavailability of chemicals in soil may lead to poor biological degradation of substances and possible accumulation of pesticide residues in soil. Whether this is a risk, especially when the field is turned to forest or pasture, and the possible changes in soil structure and pH that could affect bioavailability should also be carefully assessed.

5. CONCLUSION

In the field studies, statistically significant effects of pesticides on soil microbes were difficult to detect. This could be due mainly to other environmental factors such as soil characteristics, pesticide bioavailability, and weather conditions. In the laboratory tests, toxicity was observed in aqueous solutions but not in the same soil as in the field experiment. The bioavailability testing in this soil revealed toxicity reduction (rather rapid with propiconazole) probably due to the adsorption of three tested chemicals to the soil matrix.

ACKNOWLEDGEMENTS

The authors thank Miia Aalto and Kirsi Puisto for skillful technical assistance.

REFERENCES

Anderson (1992). Anderson, J.P.E., 1992. Side-effects of pesticides on carbon and nitrogen transformations in soils. In: Anderson, J.P.E., Arnold, D.J., Lewis, F., Torstensson, L. (Eds.), Proceedings of the International Symposium on Environmental Aspects of Pesticide Microbiology, 17–21 August 1992, Sigtuna, Sweden.

Bitton and Dutka (1986). G. Bitton and B.J. Dutka , Introduction and review of microbial and biochemical toxicity screening problem. In: G. Bitton and B.J. Dutka, Editors, Toxicity Testing using Microorganisms Vol. 1, CRC Press, Boca Raton, FL (1986), pp. 1–8.

Brinkmann and Kühn (1977). G. Brinkmann and R. Kühn , Limiting values for the damaging action of water pollutants to bacteria Pseudomonas putida and green algae Scenedesmus quadricauda in cell multiplication inhibition test. Z. Wasser Abwasser-Forsch. 10 (1977), pp. 87–98.

Brouwer et al (1990). H. Brouwer, T. Murphy and L. McArdle , A sediment-contact bioassay with Photobacterium phosphoreum. Environ. Toxicol. Chem. 9 (1990), pp. 1353–1358.

Emholt (1992). S. Emholt , Effect of propiconazole on substrate amended soil respiration following laboratory and field application. Pestic. Sci. 34 (1992), pp. 139–146.

Giller et al (1998). K.E. Giller, E. Witter and S.P. McGrath , Toxicity of heavy metals to micro-organisms and microbial processes in agricultural soils: a review. Soil Biol. Biochem. 30 (1998), pp. 1389–1414.

Hastings (1978). J.W. Hastings , Bacterial luminescence: an overview. In: M. Deluca, Editor, Methods in Enzymology Vol. 57, Academic Press, New York (1978).

International Organisation for Standardisation (ISO) (1995). International Organisation for Standardisation (ISO), 1995. ISO 10712: water quality: Pseudomonas putida growth inhibition test (Pseudomonas cell multiplication inhibition test). International standard.

International Organisation for Standardisation (ISO) (1998). International Organisation for Standardisation (ISO), 1998. ISO 11348: water quality: determination of the inhibitory effect of water samples on the light emission of Vibrio fischeri (luminescent bacteria test). International standard.

Jenkinson et al (1979). D.S. Jenkinson, S.A. Davidson and D.S. Powlson , Adenosine triphosphate and microbial biomass in soil. Soil Biol. Biochem. 11 (1979), pp. 521–527.

Jones et al (1992). Jones, S.E., Jones, A.U., Johnson, D.B., 1992. Effects of differential pesticide inputs on the size and the composition of soil microbial biomass: results from the Boxworth and SCARAB projects. In: Anderson, J.P.E., Arnold, D.J., Lewis, F., Torstensson, L. (Eds.), Proceedings of the International Symposium on Environmental Aspects of Pesticide Microbiology, 17–21 August 1992, Sigtuna, Sweden.

Malkomes (1992). Malkomes, H.-P., 1992. Interactions between pesticides and soil biocenosis. In: Anderson, J.P.E., Arnold, D.J., Lewis, F., Torstensson, L. (Eds.), Proceedings of the International Symposium on Environmental Aspects of Pesticide Microbiology, 17–21 August 1992, Sigtuna, Sweden.

Martikainen et al (1998). E. Martikainen, J. Haimi and J. Ahtiainen , Effects of dimethoate and benomyl on soil organisms and soil processes: a microcosm study. Appl. Soil Ecol. 9 (1998), pp. 381–387.

Moore and Hunt (1988). J.C. Moore and H.W. Hunt , Resource compartmentation and stability of real ecosystems. Nature 333 (1988), pp. 261–263.

Müller et al (1981). M.M. Müller, C. Rosenberg, H. Siltanen and T. Wartiovaara , Fate of glyphosate and its influence on nitrogen cycling in two Finnish soils. Bull. Environ. Contam. Toxicol. 27 (1981), pp. 724–730.

Nordgren (1988). A. Nordgren , Apparatus for the continuous, long-term monitoring of soil respiration rate in large numbers of samples. Soil. Boil. Biochem. 20 (1988), pp. 955–957.

OEPP/EPPO (1994). OEPP/EPPO, 1994. Decision-making scheme for the environmental risk assessment of plant protection products. Soil microflora. OEPP/EPPO Bulletin, Vol. 24 (Chapter 7). Blackwell Scientific, Oxford, UK.

Organization for Economic Cooperation and Development (OECD) (2000). Organization for Economic Cooperation and Development (OECD), 2000. Test Guidelines for Testing Chemicals. 11th Addendum. OECD, Paris.

SAS Institute Inc. (1988). SAS Institute Inc., 1988. SAS Procedures Guide, 6.03 Edition. SAS Institute Inc., Cary, NC.

Vanhala and Ahtiainen (1994). P. Vanhala and J.H. Ahtiainen , Soil respiration, ATP content and Photobacterium phosphoreum toxicity test as indicators of metal pollution in soil. Environ. Toxicol. Water Qual. 9 (1994), pp. 115–121.

West et al (1986). A.W. West, G.P. Sparling and W.D. Grant , Correlations between four methods to estimate total microbial biomass in stored, air-dried and glucose amended soils. Soil Biol. Biochem. 18 (1986), pp. 569–576.

(order Full Text from publisher)