Ring-Testing and Field-Validation of a Terrestrial

Ecotoxicology, 13, 105–118, 2004  2004 Kluwer Academic Publishers. Manufactured in The Netherlands.

Ring-Testing and Field-Validation of a Terrestrial Model Ecosystem (TME) – An Instrument for Testing Potentially Harmful Substances: Effects of Carbendazim on Earthworms JO¨RG RO¨MBKE,1,* CORNELIS A.M. VAN GESTEL,2 SUSAN E. JONES,3 JOSE´E E. KOOLHAAS,2 JOSE´ M.L. RODRIGUES4 AND THOMAS MOSER1 1 ECT Oekotoxikologie GmbH, Bo¨ttgerstr. 2-14, D-65439 Flo¨rsheim am Main, Germany 2 Institute of Ecological Science, Vrije Universiteit, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands 3 University of Wales, School of Agricultural and Forestry Sciences, Bangor, Gwynedd LL57 2UW, Wales, UK 4 Departamento de Biologia, Universidade de Aveiro, Aveiro, Portugal Accepted 24 October 2002 Abstract. The effects of the fungicide carbendazim (applied in the formulation Derosal
) on earthworms (Lumbricidae) was determined in Terrestrial Model Ecosystem (TME) tests and field-validation studies. TMEs consisted of intact soil columns (diameter 17.5 cm; length 40 cm) taken from a grassland or, in one case, from an arable site. The TMEs were taken from the same site where the respective field-validation study was performed. The tests were performed in Amsterdam (The Netherlands), Bangor (Wales, UK), Coimbra (Portugal) and Flo¨rsheim (Germany). The sites selected had an earthworm coenosis representative of the different land use types and regions. In addition, the differences between the coenosis found in the TMEs and the respective field sites were in general low. A high variability was found between the replicate samples, which reduces the probability of determining significant differences by the statistical evaluation of the data. Similar effects of the chemical treatment were observed on abundance as well as on biomass. Effects were most pronounced 16 weeks after application of the test chemical. The observed effects on earthworm abundance and biomass did not differ between the TME tests and the respective field-validation studies. Effects on earthworm diversity were difficult to assess since the number of individuals per species was low in general. However, the genus Lumbricus and in particular L. terrestris and L. rubellus seemed to be more affected by the chemical treatment than others. NOEC and EC50-values derived from the TME pre-test, the TME ring-test and the field-validation study indicate that the TMEs of the different partners delivered comparable results although different soils were used. Due to the high variability NOECs could often not be determined. The EC50-values for the effect of carbendazim on earthworm abundance ranged between 2.04 and 48.8 kg a.i./ha (2.71–65.2 mg/kg soil) and on earthworm biomass from 1.02 to 34.6 kg a.i./ha (1.36–46.0 mg/kg soil). These results indicate that the abundance and biomass of earthworms are suitable endpoints in ecotoxicological studies with TMEs. Keywords: carbendazim; Lumbricidae; earthworms; soil mesocosms; community effects

*To whom correspondence should be addressed: Tel.: +49-6145-9564-20; Fax: +49-6145-9564-99; E-mail: [email protected]

106 Ro¨mbke et al. Introduction Risk assessment of chemicals is usually based on the results of single species toxicity tests, which are performed in the laboratory (Van Leeuwen and Hermens, 1995). It is, however, realised that such single species tests may not be sufficient to predict effects on the level of ecosystem structure and functioning (Cairns, 1984). In the past, such effects have been studied in the field but the results are difficult to assess and the efforts are high. Therefore, mesocosm tests have been advocated, which may close the gap between laboratory and field studies. For the soil environment, this has resulted in the development of several types of model ecosystems (Morgan and Knacker, 1994; Sheppard, 1997; Edwards et al., 1998; Knacker et al., 2004). Such model ecosystems may be applied in a higher tier of risk assessment, when results of single species laboratory toxicity tests provide reasons for concern with regard to the potential risk of a chemical in the soil environment (see Weyers et al., 2004). In this paper, results on earthworm abundance, biomass and diversity gained by tests using Terrestrial Model Ecosystems (TMEs) are described. TMEs are defined as controlled, reproducible systems that attempt to simulate the processes and interactions of components in a portion of the terrestrial environment (Gillett and Witt, 1980; Sheppard, 1997). These tests were performed within the framework of the TME-project sponsored by the European Union (Contract No.: ENV4-CT97-0470). The aim of this project was the improvement and validation of the TME test system, first described by Van Voris et al. (1985) and used in studies reported by Knacker et al. (1989), Frederickson et al. (1991) and Chekai et al. (1993). For that purpose, the TME tests were performed at four different European sites with different soils, but using similar equipment, test chemical, test design, and endpoints. The tests were designed in a way which should allow a comparison of NOEC- and EC50-values for several endpoints. In order to investigate the ecological realism of this laboratory TME test system, a fieldvalidation study was performed. Carbendazim was chosen as the model chemical for the TME tests and the field-validation study. Carbendazim is a fungicide that is used on a large scale in agriculture

throughout Europe (Cuppen et al., 2000; Frampton and Wratten, 2000). This chemical was also chosen as a reference substance in the earthworm chronic reproduction test (ISO 1998; OECD 2001), since it is known for its high toxicity to earthworms.

Materials and methods Experimental set-up Three types of tests were performed: in the first year a TME pre-test, in the second year, based on the experience gained during the pre-test, the TME ring-test and in parallel to the ring-test a fieldvalidation study. The TME-project was conducted at four sites throughout Europe by the following project partners: ECT Oekotoxikologie GmbH, Flo¨rsheim, Germany (1), Vrije Universiteit Amsterdam (Institute of Ecological Science, The Netherlands) (2), University of Wales, Bangor (School of Agricultural and Forestry Sciences, UK) (3) and Universidade de Coimbra (Instituto Ambiente e Vida, Portugal) (4). In Coimbra, TME tests and the field-validation study were performed at an arable site, whereas the respective work at the three other sites was done on a grassland. The properties of the soils from these four sites were measured using ISO guidelines (ISO 1993; ISO 1994; ISO 1995; Table 1). The TME tests started with the extraction of the TME soil cores. TMEs were taken by means of a soil core extractor, containing a HDPE tube (diameter 17.5 cm; length 40 cm), which served as a soil core encasement. Grass was cut just before TME extraction. The TMEs were either placed in temperaturecontrolled carts in a climatic chamber (Amsterdam, Flo¨rsheim, Coimbra) or in a greenhouse (Bangor). TMEs were irrigated up to three times per week using artificial rainwater slightly modified according to Velthorst (1993). The model chemical carbendazim was applied after an acclimatisation period of two to four weeks. For the field-validation study, 30 field plots, each 25 m2, were marked out by each partner at the same site where the TMEs were extracted. Six plots served as controls, and 24 plots were treated with the model chemical (four plots for each of the six treatment levels; the plots were completely

TME – Effects of Carbendazim on Earthworms 107 Table 1. Soil properties and site use of the four experimental fields from which the soil cores for the TME tests were extracted and the respective field-validation studies were performed (for details see Knacker et al. (2004); OM = organic matter Participant/ location

Country code

Texture

Clay (%)

OM (%)

pH (KCl)

Land use

Amsterdam Bangor Coimbra Flo¨rsheim

NL UK P D

Silty loam Loam Silty loam Silty clay loam

7.9 20.5 24.7 36.5

4.5 6.1 3.4 5.2

4.8–5.1 5.8–6.6 6.4–7.1 5.3–5.9

Meadow Pasture Arable field Meadow

randomised). Before spraying the model chemical the grass cover was cut on the grassland sites (Amsterdam, Bangor, Flo¨rsheim) or the soil was ploughed at the arable site (Coimbra). Carbendazim was applied to the TMEs and field plots as Derosal
, containing 360 g carbendazim/l. In the TME pre-tests, carbendazim was applied at dosages of 0 (T0), 0.36 (T1), 2.16 (T2), 13.0 (T3) and 77.8 (T4) kg a.i./ha. In the TME ring-test and the field-validation study at dosages of 0 (T0), 0.36 (T1), 1.08 (T2), 3.24 (T3), 9.72 (T4), 29.2 (T5) and 87.5 (T6) kg a.i./ha. In the laboratory, carbendazim (dissolved in 50 ml of water per soil core surface (=222 cm2)) was applied using a pipette. The TMEs were irrigated immediately after treatment. In the field-validation study, carbendazim was applied by a plot sprayer, commonly used in agricultural practice, using a 3 l spray solution (6 l for highest dosage) per plot (@1200 l/ha). Immediately after spraying, a volume of 30 l of water was sprayed onto the plot to wash off the Derosal
from the plants onto the soil. The control plots were also sprayed with 30 l of water. Various measurement endpoints were chosen to determine the fate and effect of the model chemical as well as the structure and function of the terrestrial compartment. The fate endpoints included the measurement of residues of the model chemical in the upper soil layer (0–5 cm), in the soil layer from 5 to 15 cm, and in the leachate. The effect endpoints were classified into functional endpoints (nutrients in leachate and soil which describe aspects of nutrient cycling; soil enzyme activity; microbial substrate induced respiration; bacterial growth; feeding activity of soil organisms; organic matter decomposition) and structural endpoints (abundance as well as diversity and community structure of microarthropods, nematodes, enchytraeids, lumbricids and plant biomass. For a de-

tailed description of the set-up of the TMEproject, see Knacker et al. (2004). Earthworm sampling In the TME pre-test, earthworm samples were taken 1, 8 and 16 weeks after application. After taking the samples for all other endpoints, the remaining soil of the TME cores was sorted by hand for earthworms. At each sampling point six untreated TMEs (control) and three TMEs for each treatment level were sampled. In the TME ringtest, samples were taken 1, 8 and 16 weeks after application of the test chemical. After 1 week, only the control, the lowest (T1) and the highest (T6) treatment were sampled. Six control TMEs and four TMEs per treatment level were used in the TME ring-test. In the field, sampling was done only at the end of the test; i.e. after 16 weeks. In the field a combination of methods was used to achieve a complete assessment of the earthworm biocoenosis (Dunger and Fiedler 1997): firstly, a square of 50
50 cm (0.25 m2) soil was dug out up to a depth of 15 cm. This excavated topsoil layer was sorted by hand. Additionally, 10 l of a 0.5% formaldehyde solution was sprinkled in the same plot from which the topsoil was removed to extract the remaining worms from the soil. Earthworms were col1ected in 70% ethanol, then fixed in 4% formaldehyde for 1–2 weeks. Afterwards, they were finally preserved in 70% ethanol. Shortly after fixation the adult worms were counted and determined to the species level while juveniles were identified at the genus level (Graff 1953; StøpBowitz 1969; Bouche´ 1972; Sims and Gerard 1985). Site-specific keys for three sites (Amsterdam, Bangor and Flo¨rsheim) were developed by Partner (1). Directly after determination the biomass was measured as wet mass and converted by

108 Ro¨mbke et al. a factor of ten to dry mass. All values are given per square meter (i.e. the conversion factor between individuals per TME soil core and per square meter was 41.5).

test with those in the respective field-validation study a U-test according to Mann–Whitney (2sided, p £ 0.05) was used (Sparks, 2000; Sachs, 1999; Norusis, 1998).

Statistical evaluation Results NOEC/LOEC and EC50-values were determined for earthworm abundance and biomass. NOEC/ LOEC-values were determined from single replicate values. Firstly, the data were tested for homogeneity by applying Cochran’s test. In case of variance homogeneity, NOEC/LOEC-values were determined by analysis of variance (ANOVA) followed by a Dunnett’s t-test (l-sided; p £ 0.05). In case of inhomogeneity, NOEC/LOEC-values were determined using a Bonferroni-U-test according to Holm (l-sided; p £ 0.05). The EC50values were calculated applying a logistic model according to Haanstra et al. (1985) using the treatment mean values. To compare the earthworm abundance and biomass in the TME ring-

800

800

Bangor

In the TME pre-test (Fig. 1) the mean number (sd, n ¼ 4–6) of earthworms in the controls was 218 ± 179 ind/m2 in Amsterdam, 382 ± 123 ind/ m2 in Bangor and 374 ± 243 ind/m2 in Flo¨rsheim after 1 week. In Coimbra no earthworm sampling took place. After 8 weeks, the mean number of earthworms decreased in Amsterdam and increased slightly in Bangor and Flo¨rsheim. Towards the end of the TME pre-test (after 16 weeks) an increase in the number of earthworms was observed in Amsterdam whereas in Bangor and Flo¨rsheim it returned to the level observed after 1 weeks.

800

Pre-test

Coimbra

800

Pre-test

600

600

600

600

400

400

400

400

200

200

200

200

0

Abundance [ind/m²]

Amsterdam Pre-test

Earthworm abundance and biomass: controls

800

+1

+8

+16

0 800

Ring-test

+1

+8

+16

0 800

Ring-test

+1

+8

+16

0 800

Ring-test

600

600

600

600

400

400

400

400

200

200

200

200

0 800

+1

+8

+16

0 800

Field

+1

+8

+16

0 800

Field

+1

+8

+16

0 800

Field

600

600

600

600

400

400

400

400

200

200

200

200

0

+1

+8

+16

0

+1

+8

+16

0

+1

+8

+16

0

Flörsheim Pre-test

+1

+8

+16

Ring-test

+1

+8

+16

+8

+16

Field

+1

Sampling Point Figure 1. Abundance of earthworms in control TMEs and field plot, 1, 8 and 16 after application. Data are given for the TME pre-test, TME ring test and field-validation study at Amsterdam, Bangor, Coimbra and Flo¨rsheim.

TME – Effects of Carbendazim on Earthworms 109 The earthworm biomass in the TME pre-test was lowest in Amsterdam (4.2 ± 3.3 g dw/m2), followed by Bangor with 7.7 ± 2.8 g dw/m2 and highest in Flo¨rsheim (18.44 ± 9.37 g dw/m2). The change in biomass with time showed a similar pattern as the abundance. In the TME ring-test after 1 week, the mean number of earthworms in the controls was in the same order of magnitude for all sites (between 280 ± 58.8 and 334 ± 68.4 ind/m2) except for Coimbra, where the mean number of earthworms was low (72.2 ± 60.5 ind/m2) (Fig. 1). With time, the number of earthworms at all sites remained more constant than in the TME pre-test. Earthworm biomass was slightly higher in Amsterdam (13.1 ± 7.7 g dw/m2) than in Bangor (8.9 ± 1.8 g dw/m2), but it was highest for the Flo¨rsheim TMEs (18.6 ± 9.6 g dw/m2). In the Coimbra TME ring-test the earthworm biomass was low (2.2 ± 1.8 g dw/m2) corresponding to the observed low earthworm numbers. The fluctuation of earthworm biomass with time was low and similar to the pattern observed for the number of earthworms (data not shown). In the field-validation study, the mean number of earthworms found in the control plots in Amsterdam, Bangor and Flo¨rsheim after 16 weeks did not differ markedly from those in the respective TMEs in the ring-test at the same sampling point (Fig. 1). In Coimbra the mean number of earthworms was nearly twice as high in the field than in the TMEs. The differences between earthworm numbers found in the field-validation study and in the respective TMEs at all sites were not statistically significant. In general, a high variability was found between the replicate samples. That was observed for all partners at all sampling points in the TME pre-test and the TME ring-test. In the field-validation study, variability was still high but in all cases lower than in the respective TME ringtests. In the field-validation study control earthworm biomass did not differ from that in the TMEs in Amsterdam, whereas for Bangor and Coimbra a significantly higher biomass was found in the field compared to the TMEs. In Flo¨rsheim earthworm biomass in the control field plots was more than 50% below that in the TMEs. This difference was not statistically significant, probably due to the high variability between the samples.

Earthworm abundance and biomass: effects of the model chemical carbendazim One week after carbendazim treatment, slight to moderate (up to 50%) reductions in earthworm numbers were seen at the three highest (Flo¨rsheim TME pre-test), two highest (Amsterdam TME pretest) or highest treatment level only (Amsterdam TME ring-test, Bangor TME pre-test and TME ring-test). These effects never differed significantly from the control. In Coimbra no earthworms were collected in the TME pre-test. In the TME ringtest only a very low number of earthworms was found. The data of the TME ring-test after 1 week were not evaluated statistically since only T1 and T6 were sampled. Eight weeks after carbendazim treatment, mean earthworm numbers in the Amsterdam TME pretest were significantly reduced compared to the control at the two highest treatment levels (Fig. 2). In the TME ring-test, earthworm numbers were reduced at several treatment levels, but this reduction was only statistically significant at the highest treatment level. In the Bangor TME pretest and TME ring-test, earthworm abundance decreased at the two or three highest treatment levels, but effects found were not statistically significant in comparison to the controls. In Coimbra, a reduction in earthworm numbers was found at all treatment levels in the TME ring-test, but due to the low number of earthworms, a statistical evaluation of the data was not applicable. In the Flo¨rsheim TME pre-test, earthworm numbers showed a dose-related reduction, which was statistically significant at the highest treatment level. In the TME ring-test, earthworm numbers were reduced at several treatment levels, but effects were not statistically significant in comparison to the controls. The mean number of earthworms was significantly reduced at the two highest treatment levels in the Amsterdam and Flo¨rsheim pre-tests 16 weeks after treatment with carbendazim. (Fig. 3). In the Bangor TME pre-test, a similar trend was seen, but due to the high variability between the replicate samples and the absence of earthworms at the highest treatment level, statistical evaluation of the data was not applicable. In the TME ring-test, a dose-related decrease in earthworm numbers was found, which was

110 Ro¨mbke et al. Amsterdam

800

Bangor

800

Pre-test

800

Pre-test

+8

Coimbra

800

Pre-test

+8

+8

+8

Flörsheim Pre-test

600

600

600

600

400

400

400

400

200

200

200

200

Abundance [ind./m²]

*

0

800

T0

T1

T2

*

*

T3

T4

0

800

Ring-test

T0

T1

T2

T3

T4

0

800

Ring-test +8

+8

T0

T1

T2

T3

T4

Ring-test

0

800

+8

600

600

600

400

400

400

400

200

200

200

0

*

T0 T1 T2 T3 T4 T5 T6

0

T0 T1 T2 T3 T4 T5 T6

0

T0 T1 T2 T3 T4 T5 T6

T1

T2

T3

T4

Ring-test +8

600

200

T0

0

T0 T1 T2 T3 T4 T5 T6

Treatment Level Figure 2. Effect of carbendazim on the abundance of earthworms (ind/m2). Data are given for the TME pre-test and the TME ring test, after 8 weeks, performed in Amsterdam, Bangor, Coimbra and Flo¨rsheim. Significant differences compared to the control are indicated by an asterisk. Treatment levels: TME pre-test: T0 ¼ Control, T1 ¼ 0.36, T2 ¼ 2.16, T3 ¼ 12.96, T4 ¼ 77.76 kg a.i./ha; TME ring test: T0 ¼ Control, T1 ¼ 0.36, T2 ¼ 1.08, T3 ¼ 3.24, T4 ¼ 9.72, T5 ¼ 29,16, T6 ¼ 87.48 kg a.i./ha.

statistically significant at the three highest treatment levels in Amsterdam (at the highest treatment level no earthworms were found) and Flo¨rsheim, but not in Bangor. In the Coimbra TME ring-test, no effects on earthworm numbers were seen. In the field-validation studies a doserelated decrease in earthworm numbers was found, which was statistically significant at the three highest treatment levels in Amsterdam, at the two highest treatment levels in Bangor and at the highest treatment level only in Flo¨rsheim (Fig. 3). In the field-validation study of Coimbra, earthworm numbers were reduced at the four highest treatment levels, but effects were not statistically significant in comparison to the controls. Effects of carbendazim on earthworms biomass showed a similar pattern as on the earthworm abundance (Fig. 4). This was observed at all sampling points in the TME pre-test, the TME ring-test and the field-validation study in Amsterdam, Bangor and Coimbra. In Flo¨rsheim effects on biomass seemed to be more pronounced than

effects on abundance in particular in the ring-test and the field-validation study.

NOEC and EC50-values Due to the high variability between the replicate samples of each treatment level and the control, it was not possible in many cases to detect significant differences between the different treatment levels and the control by parametric or non-parametric tests although an effect on earthworm abundance or biomass seems to be obvious. Therefore, very often the NOEC was greater or equal to the highest treatment level (Tables 2 and 4). Despite the high variability, the calculation of EC50-values was possible, but the 95% confidence intervals were sometimes high (Tables 3 and 5). The NOECs for the effect of carbendazim on earthworm abundance ranged from 2.16 to ‡87.5 kg a.i./ha and for earthworm biomass from 1.08 to ‡87.5 kg a.i./ha. The EC50-values ranged from

TME – Effects of Carbendazim on Earthworms 111 800

Amsterdam

800

Pre-test

Bangor

800

Pre-test

Coimbra

800

Pre-test

600 +16

600 +16

600 +16

600

400

400

400

400

200

200

200

200

Flörsheim Pre-test +16

*

Abundance [ind/m²]

0 800

T0

T1

T2

*

*

T3

T4

* 0 800

Ring-test

T0

T1

T2

T3

T4

0

T0

T1

T2

T3

T4

800

Ring-test

0 800

Ring-test

T0

T1

T2

T3

T4

Ring-test

600 +16

600 +16

600 +16

600 +16

400

400

400

400

200

200

200

* *

200

*

*

* 0 T0 T1 T2 T3 T4 T5 T6 800

Field

200

T0 T1 T2 T3 T4 T5 T6

* *

*

Field

600

400 * *

0 800

600 +16

600 +16 400

0 800

T0 T1 T2 T3 T4 T5 T6

Field +16

0

*

T0 T1 T2 T3 T4 T5 T6

800

Field 600 +16

400

400

200

200

*

200

* 0

T0 T1 T2 T3 T4 T5 T6

0

T0 T1 T2 T3 T4 T5 T6

0

T0 T1 T2 T3 T4 T5 T6

0

T0 T1 T2 T3 T4 T5 T6

Treatment Level Figure 3. Effect of carbendazim on the abundance of earthworms (ind/m2). Data are given for the TME pre-test and the TME ring test and the field-validation study, after 16 weeks, performed in Amsterdam, Bangor, Coimbra and Flo¨rsheim. Significant differences compared to the control are indicated by an asterisk. Treatment levels: see Fig. 2.

2.04 to 49.0 (earthworm abundance) and from 1.02 to 34.6 kg a.i./ha (earthworm biomass). Interestingly enough the EC50-values for the TME pre-test and TME ring-test are often below the respective NOECs, whereas the EC50-values for the fieldvalidation study (with a lower variability compared to the TME tests) were higher than the NOECs. A comparison of the EC50-values indicated that the variation between the results of the different partners was low. Additionally, the EC50values derived from the TME tests are comparable to those derived from the field-validation study although those from the TME tests seemed to be somewhat lower.

Earthworm diversity: controls and chemical treatments The earthworm biocoenosis (Table 6) was most divers in Flo¨rsheim and Coimbra with 7 species determined, followed by Bangor (6 species) and

Amsterdam (4 species). The most common species Aporrectodea caliginosa was found at all sites. The Amsterdam earthworm community was dominated by A. caliginosa and Lumbricus rubellus. L. castaneus and Allolobophora chlorotica were only found in one or two TME or field samples. In Bangor three Aporrectodea species (A. caliginosa, A. longa and A. rosea) were most abundant. In Flo¨rsheim earthworm community was dominated by juvenile Aporrectodea individuals (probably mainly A. caliginosa) A. antipae, Octolasion sp. and Lumbricus terrestris, which occurred only in the soils of Bangor and Flo¨rsheim. Murchieona minuscula was only found in Bangor and Aporrectodea antipae only in Flo¨rsheim. The earthworm community of Coimbra differed clearly from the other investigated sites because 3 of the 7 species found in Coimbra were determined only for this partner. One of these 3 species did occur only once in the samples. Probably it is Allolobophora fernandae (Graff, 1957), but the taxonomic status of this rare species is not clear (Zicsi, 1976).

112 Ro¨mbke et al.

30

30

Bangor

30

Pre-test

Coimbra

30

Pre-test

Flörsheim Pre-test

25 +16 20

25 20

25 +16 20

25 +16 20

15

15

15

15

10

10

10

10

5

5

5

5

0

0

0

Biomass [g dw/m²]

Amsterdam Pre-test

30 25

T0

T1

T2

* T4

T3

0 30

Ring-test

+16

* T0

T1

T2

* T4

T3

30

Ring-test

T0

T1

T2

T3

T4

Ring-test

30

25

20

25 +16 20

20

20

15

15

15

15

10

10

10

5

5

5

0

0

0

+16

10

*

5

* * 0 T0 T1 T2 T3 T4 T5 T6 30 25

25

T0 T1 T2 T3 T4 T5 T6

30

Field

T0 T1 T2 T3 T4 T5 T6

Field

25 +16

25

20

20

20

15

15

15

15

10

10

5

5

5

0

0

* *

10

0

*

10 *

5

T1

T2

T3

T4

Ring-test +16

*

*

*

T0 T1 T2 T3 T4 T5 T6

*

T0 T1 T2 T3 T4 T5 T6

T0 T1 T2 T3 T4 T5 T6

T0 T1 T2 T3 T4 T5 T6

30

25 +16 20

+16

T0

* 30

Field

+16

*

0

Field +16

*

*

*

T0 T1 T2 T3 T4 T5 T6

Treatment Level Figure 4. Effect of carbendazim on the biomass of earthworms (g dw/m2). Data are given for the TME pre-test, the TME ring test and the field-validation study, after 16 weeks, performed in Amsterdam, Bangor, Coimbra and Flo¨rsheim. Significant (Dunnett-t-test, lsided; p £ 0.05) differences compared to the control are indicated by an asterisk. Treatment levels: see Fig. 2.

Table 2. NOEC-values (in kg a.i./ha) for the effect of carbendazim on earthworm abundance in the TME pre-test, the TME ring test and the field-validation studie´d, 8 and 16 weeks after application of the model chemical in Amsterdam (A), Bangor (B), Coimbra (C) and Flo¨rsheim (F) NOEC

TME TME TME TME Field

pre-test ring test pre-test ring test

Weeks

A

B

C

F

þ8

2.16 29.2 2.16 3.24 3.24

‡77.8 29.2

–

a

–

‡87.5 9.72

a

13.0 ‡87.5 2.16 3.24 29.2

þ16

a

‡87.5

bricus was more affected by the chemical treatment than other genera. In particular in Flo¨rsheim, and to a lesser extend in Bangor the number of L. terrestris decreased clearly with the treatment, while the number of individuals belonging to the genus Aporrectodea remained constant. A similar observation was made in Amsterdam, where the species L. rubellus was affected more than the two Aporrectodea species.

Discussion

– = No sampling. a Not applicable.

Controls

Despite the fact that effects of carbendazim on the number of earthworm species were visible, the absolute number of specimen belonging to one particular species was too low (except for Flo¨rsheim) to use this parameter as a measurement endpoint. However, it seems that the genus Lum-

The abundance and biomass of the earthworms determined for the grasslands in Amsterdam, Bangor and Flo¨rsheim, as well as for the arable land in Coimbra, corresponded to the values found for similar field types referred to in the literature (Satchell, 1983; Ro¨mbke et al., 1997). Fluctuations between the sampling points and differences be-

TME – Effects of Carbendazim on Earthworms 113 Table 3. EC50-values (in kg a.i./ha) and the 95% confidence limits (CL) calculated for the effect of carbendazim on earthworm abundance in the TME pre-test, the TME ring test and the field-validation study, 8 and 16 weeks after application of the model chemical at Amsterdam (A), Bangor (B), Coimbra (C) and Flo¨rsheim (F) EC50 lower–upper 95% CL

TME TME TME TME Field

pre-test ring test pre-test ring test

Time, week

A

þ8

10.9 12.9 2.80 3.00 11.5

þ16

B a

1.4–117 0.2–42.7 1.5–5.9 0.4–324

C

2.04 6.02 2.35 3.91 48.8

0.02–230 0.5–67.7 1.2–4.6 1.1 13.3–179

F

–

–

b

b

–

–

b

b

b

b

4.01 26.5 7.41 4.33 49.0

0.2–83.8 0.8–858 1.5–37.7 1.3–14.4 30.1–79.6

– = no sampling. a Lower CL < 10)3 · EC50 and upper CL > 103 · EC50; bNot applicable.

Table 4. NOEC (in kg a.i./ha) for the effect of carbendazim on earthworm biomass in the TME pre-test, the TME ring test and the field-validation study, 8 and 16 weeks after application of the model chemical in Amsterdam (A), Bangor (B), Coimbra (C) and Flo¨rsheim (F) NOEC

TME TME TME TME Field

pre-test ring test pre-test ring test

Week

A

B

C

F

þ8

‡77.8 ‡87.5 13.0 3.24 3.24

‡77.8 ‡87.5 2.16 ‡87.5 9.72

–

13.0 9.72 13.0 1.08 3.24

þ16

a

– a

‡87.5

– = No sampling. a Not applicable.

Table 5. EC50-values (in kg a.i./ha) and the 95% confidence limits (CL) calculated for the effect of carbendazim on earthworm biomass in the TME pre-test, the TME ring test and the field-validation study, 8 and 16 weeks after application of the model chemical in Amsterdam (A), Bangor (B), Coimbra (C) and Flo¨rsheim (F) Time, week EC50 lower–upper 95% CL A TME TME TME TME Field

pre-test ring test pre-test ring test

þ8 þ16

B

C

F

a

a

a

a

–

–

a

a

3.44 3.29 4.42 8.69

b

7.83 2.16 8.57 34.6

0.3–230.0

a

a

b

b

–

–

b

a

a

4.8–248.4

9.84

b

3.16 1.02 1.07 9.13

0.6–18.3 1.8–10.6 0.02–4842

b

0.7–1.6 b

– = No sampling. a Lower CL < 10)3 · EC50 and upper CL > 103 · EC50; b Not applicable.

tween the TME pre-test, TME ring-test and fieldvalidation study were within the ranges expected from publications cited above. The variability between replicate control samples is well known from many field studies since earthworms are not evenly distributed in soil (Edwards and Bohlen, 1996).

Soil moisture content and the amount of food available are the most important factors influencing distribution patterns. In addition, Satchell (1955) assumed that aggregations might occur when earthworms are reproducing more rapidly than the offspring can disperse from the breeding

114 Ro¨mbke et al. Table 6. Earthworm species found in the TME pre-test, the TME ring test and the field-validation study at Amsterdam, Bangor, Coimbra and Flo¨rsheim Amsterdam

Bangor

Coimbra

Flo¨rsheim

Allolobophora chlorotica Aporrectodea caliginosa Lumbricus castaneus Lumbricus rubellus

Aporrectodea caliginosa Aporrectodea longa Aporrectodea rosea Lumbricus rubellus Lumbricus terrestris Murchieona minuscula

Allolobophora chlorotica Allolobophora fernandaea Aporrectodea caliginosa Aporrectodea rosea Criodrilus lacuum Eiseniella tetraeda Octolasion lacteum

Aporrectodea caliginosa Aporrectodea rosea Aporrectodea antipae Lumbricus castaneus Lumbricus terrestris Octolasion cyaneum Octolasion lacteum

a

Taxonomical revision necessary.

site. The inhomogeneity of the earthworm distribution in the field seems to have been realistically reflected by the TMEs. The large sampling area of each field sample (0.25 m2) compared to the small TMEs (approx. 0.024 m2) can explain the generally lower variability obtained for worm numbers and biomass in the field. The low earthworm biomass in the TMEs of Bangor compared to the earthworm biomass in the field, can be explained by the absence of the species Lumbricus terrestris in the TMEs. As L. terrestris is the biggest middle European earthworm species, the presence of a single specimen can contribute considerably to the overall biomass, in particular if the total number of earthworms is low. In Coimbra the field was ploughed before extracting the soil cores. Therefore, the initial earthworm population was already disturbed when the experiment started. As a consequence the number of earthworms found was very low, making the interpretation of the Coimbra data difficult. Compared to the field the biomass of earthworms was high in the TMEs of Flo¨rsheim. This was probably due to a long period of drought in the field before sampling. Drought causes migration of L. terrestris to deep soil layers. From there L. terrestris could either not be extracted; because the formaldehyde solution did not penetrate deep enough into the soil, or the earthworms were in a state of quiescence and thus not reacting to formaldehyde (Edwards and Bohlen, 1996). Effects of carbendazim Carbendazim strongly adsorbs to soil organic material and remains in the soil for up to 3 years (World Health Organization, 1993). From various

laboratory studies it is known to be highly toxic to earthworms (Adema et al., 1985; Vonk et al. 1986, Van Gestel 1992). Acute effects (mainly mortality) have been found at soil concentrations between 0.9 and 5.7 mg carbendazim/kg, using different earthworm species, test durations and test substrates. Chronic effects occurred at concentrations between 0.6 and 1.9 mg carbendazim/kg. The toxicity of carbendazim to earthworms is confirmed by the results of the presented work. At all sites, independent of soil properties, effects on earthworms have been observed. These effects seemed to be most pronounced 16 weeks after application of the model chemical (Fig. 3). At this sampling point a clear dose–response relationship was observed in the TME pre-test, the TME ringtest and the field-validation study at Amsterdam, Bangor and Flo¨rsheim. Effects on single species could not be statistically evaluated since the absolute numbers were too low at the different sites and sampling points. However, at least in Flo¨rsheim L. terrestris and in Amsterdam L. rubellus seemed to be more affected than e.g. A. caliginosa. This was also reported by other authors (Federschmidt, 1994; Lofs-Holmin, 1981). One reason for the more sensitive reaction of L. terrestris and L. rubellus might be the feeding behaviour of these species, which feed on the soil surface where carbendazim is adsorbed. In addition, referring to the results of laboratory studies with carbendazim (Federschmidt, 1994) and benomyl (Heimbach, 1988), the LC50 values for L. terrestris are lower in comparison to those for other species such as Eisenia fetida. The NOEC and EC50-values derived from the TME pre-test, TME ring-test and field-validation study (Tables 2–5) indicate that the TME tests of

TME – Effects of Carbendazim on Earthworms 115 the different partners delivered comparable results, although different soils from different sites were used. Additionally, the NOEC and EC50-values derived from the TME tests are comparable to those derived from the field-validation study. Since it was often not possible to determine NOEC-values due to the high variability of the data, only the EC50-values will be discussed in detail here (Landis et al., 1997). However, when calculating the EC50values, this variability results in high 95% confidence intervals. The EC50-values were transformed from kg carbendazim/ha to mg carbendazim/kg by a conversion factor of 1.33 according to EPPO (2001). The converted EC50-values for earthworm abundance appear to range from 2.71 to 65.2 mg carbendazim/kg and for earthworm biomass from 1.36 to 46.2 mg carbendazim/kg for all TME tests and the field-validation study. The lower range of the biomass EC50-values can be explained by the reaction of L. terrestris, which is more affected than other species (Federschmidt, 1994). This difference between the two measurement endpoints was only observed in Bangor and Flo¨rsheim (Figs. 3 and 4). The range of effect concentrations is in agreement with data from the literature; for example, LC50values for E. fetida or E. andrei of 5.7 and 9.3 mg carbendazim/kg (Vonk et al., 1986, Van Gestel et al., 1992) or an EC50-value of 2.9 mg carbendazim/kg for E. andrei (cocoon production) (Van Gestel et al., 1992) have been reported. Federschmidt (1994) determined LC50-values for L. terrestris of 0.9 mg carbendazim/kg after 14 days and 2.6 mg carbendazim/kg after 28 days. Carbendazim applied in various field studies has caused effects on earthworms independently from site use and soil characteristics at concentrations ranging between 0.4 and 1.6 mg carbendazim/kg (Van Gestel, 1992). Under field conditions species specific increases or decreases in the number of earthworms can be induced by carbendazim (LofsHolmin 1981), which may result in effects on litter decomposition (Cook and Swait, 1975; Keogh and Whitehead, 1975). On a grassland near Frankfurt/ Main, Germany, which showed similar soil characteristics as the field site of Flo¨rsheim, the number of earthworms decreased and litter decomposition was delayed at concentrations of 0.48 and 4.8 mg carbendazim/kg (Knacker et al., 1994). In the latter study carbendazim was sprayed six times bimonthly, to achieve an almost constant con-

centration level in the uppermost soil layer for about 1 year. In summary, effects of carbendazim on the earthworms on European meadow and crop sites were found starting at estimated soil concentrations of approximately 0.5 mg a.i./kg. Independently from the site and soil characteristics as well as the respective earthworm biocoenosis, direct and indirect effects were found in narrow concentration ranges (for earthworm biomass mainly between 1 and 15 mg carbendazim/kg). The comparison of the EC50-values indicated that the reproducibility was higher than, for example, in the laboratory ring-test of the acute earthworm test (Edwards, 1984). It is recommended that earthworm abundance and biomass can be used as the main endpoints. However, if large species, especially L. terrestris, occur at a specific site, biomass is preferred, since it is more sensitive than the abundance of earthworms. No distinct differences between the three investigation levels (laboratory, TME and field) were found. Thus, theoretically the investigation of one level would have been sufficient for the model chemical carbendazim, but the ecological relevance of the results gained in laboratory tests could only be proven by performing TME or field tests. Earthworm diversity The species composition in the Bangor and Flo¨rsheim sites is as expected (Sims and Gerard, 1985); only the relatively rarely found species M. minuscula is exceptional for Bangor. Probably this worm is more widely distributed than documented in the literature because it is difficult to identify due to external taxonomic characteristics, which are often not well developed. With only four species (and in fact only two dominant species), the site near Amsterdam showed the lowest diversity of all sites investigated. Compared to the other sites, the soil of Amsterdam had the highest content of sand and the lowest water holding capacity, which might be the reason for the low number of earthworm species. Maybe the former use of the site as an agricultural field is also partly responsible for this result. The diversity at the site of Coimbra is difficult to assess since similar sites have rarely been studied in Portugal (Trigo et al., 1988). Eiseniella tetraeda and Criodrilus lacuum, found at the

116 Ro¨mbke et al. Coimbra site, are a clear indication that the respective field site is a retention area of the neighbouring river, since these species normally occur at riverbanks or in the sediment (Bouche´, 1972), while A. chlorotica is a typical inhabitant of moist, often anthropogenically disturbed sites, including agricultural land (Graff, 1953). It is remarkable that the diversity at this site was high, in spite of the fact that the field was ploughed before the start of the experiment. Only one individual, probably A. fernandae (Graff, 1957), occurred in a TME sample at Coimbra. It is known so far only from three specimens found near Setubal in a soil with a high content of organic matter. This species, which needs taxonomic revision, seems to be endemic to Central Portugal. In the past it has been confused with Eophila asconensis, a species usually found in Northern Italy (Zicsi, 1965; Omodeo, 1988). Overall conclusion The results presented here indicate that the abundance and biomass of earthworms are suitable endpoints for the assessment of effects of chemicals in TMEs. Despite different soil properties, climatic conditions and earthworm communities, the toxicity of the model chemical carbendazim on earthworms was comparable at the four sites investigated. However, predictability of the biomass results derived from TMEs is restricted if the number of large earthworms is high. At sites where the abundance of earthworms is low, data interpretation might become difficult. Due to the important ecological role of earthworms in many soil ecosystems and in the light of the results presented here the evaluation of these organisms should be incorporated when using TMEs for the environmental risk assessment of chemicals in soil.

Acknowledgements This research was financially supported by the European Union (Project No. ENV4-CT97-0470).

References Adema, D.M., Barug, D. and Vonk, J.W. (1985). Comparison of the effects of several chemicals on microorganisms, higher

plants and earthworms. In Actes Symposium Intl. ‘‘Ecotoxicologie terrestre’’. pp. 199–214. Bouche´, M. (1972). Lombriciens de France. Ecologie et Systematique. Paris, France: INRA Publ. 72-2, Institut National de Recherches Agriculturelles, 671 pp. Cairns, J. (1984). Are single species toxicity tests alone adequate for estimating environmental hazard? Environ. Monit. Assessm. 4, 259–73. Checkai, R.T., Wentsel, R.S., Phillips, C.T. and Yon, R.L. (1993). Controlled environment soil-core microcosm unit for investigating fate, migration, and transformation of chemicals in soils. J. Soil Contamination 2, 229–43. Cook, M.E. and Swait, A.A.J. (1975). Effects of some fungicide treatments on earthworm populations and leaf removal in apple orchards. J. hort. Sci. 50, 495–9. Cuppen, J.G.M., Van den Brink, P.J., Camps, E., Uil, K.F. and Brock, T.C.M. (2000). Impact of the fungicide carbendazim on freshwater microcosms. I. Water quality, breakdown of particulate organic matter and responses of macro invertebrates. Aquat. Toxicol. 48, 233–50. Dunger, W. and Fiedler, H.J. (1997). Methoden der Bodenbiologie, Jena: G. Fischer Verlag. 539 pp. Edwards, C.A. (1984). Report of the second stage in development of a standardized laboratory method for assessing the toxicity of chemical substances to earthworms. Environment and quality of life, Report EUR 9360 EN. Commission of the European Communities, Luxembourg, 99 pp. Edwards, C.A. and Bohlen, P.J. (1996). Biology end Ecology of Earthworms, 3rd edn., 426 pp. London: Chapman and Hall. Edwards, C.A., Knacker, T. and Pokarzhevskii, A. (1998). The prediction of the fate and effects of pesticides in the environment using tiered laboratory soil microcosms. The 1998 Brighton Conf. – Pests Diseases 4C-1, 267–72. EPPO (2001). Environmental risk assessment for plant protection products. Chapter 7: Soil organisms and functions. EPPO Standards, draft version 2001-02-12. Organisation Europe´enne et Me´diterrane´enne pour la protection des Plantes 1, Paris. Federschmidt, A. (1994). Die Oligocha¨tenfauna zweier O¨kosysteme auf Lo¨ßlehm unter Beru¨cksichtigung der Auswirkungen von Chemikalienstreß. Dissertation Universita¨t Frankfurt. Fredrickson, J.K., Van Voris, P., Bentjen, S.A. and Bolton, H. Jr. (1991). Terrestrial microcosm for evaluating the environmental fate and risks associated with the release of chemicals or genetically engineered micro-organisms to the environment. Toxic Subst. J. 11, 65–110. Frampton, G.K. and Wratten, S.D. (2000). Effects of benzimidazole and triazole fungicide use on epigeic species of Collembola in wheat. Ecotoxicol. Environ. Safety 46, 64– 72. Gillett, J.W. and Witt, J.M. (1980). Chemical evaluation: projected application of terrestrial microcosm technology. In J.P. Giesy, (ed). Microcosms in Ecological Research, pp. 1008–33. Springfield, Virginia, USA: Technical Information Centre US Department of Energy. Graff, O. (1953). Die Regenwu¨rmer Deutschlands. Schriftenreihe Forschungsinstitut Landwirtschaft 7, 1–70. Graff, O. (1957). De lumbricidis quibusdam in Lusitania habitantibus. Agronomia Lusitana 19, 299–304.

TME – Effects of Carbendazim on Earthworms 117 Haanstra, L., Doelman, P. and Oude Voshaar, J.H. (1985). The use of sigmoidal dose response curves in soil ecotoxicological research. Plant Soil 84, 293–7. Heimbach, F. (1988). A comparison of laboratory methods for toxicity testing with earthworms. In C.A. Edwards and E.F. Neuhauser, (eds). Earthworms in Waste and Environmental Management, pp. 329–35. The Hague: SPB Academic Publ. ISO (1993). Soil quality – Determination of particle size distribution in mineral soil material – Method by sieving and sedimentation following removal of soluble salts, organic matter and carbonates. Guideline 11277, International Organisation for Standardisation, Geneva. ISO (1994). Soil Quality – Determination of pH. Guideline 10390, International Organisation for Standardisation, Geneva. ISO (1995). Soil quality – Determination of organic carbon and total carbon after dry combustion. Guideline 10694, International Organisation for Standardisation, Geneva. ISO (1998). Soil quality Effects of pollutants on earthworms (Eisenia fetida). Part 2: Determination of effects on reproduction. Guideline 11268-2, International Organisation for Standardisation, Geneva. Keogh, R.G. and Whitehead, P.H. (1975). Observations on some effects of pasture spraying with benomyl and carbendazim on earthworm activity and litter removal from pasture. N. Z. J. Exp. Agric. 3, 103–4. Knacker, T., Van Gestel, C.A.M., Jones, S.E., Sousa, P., Schallnass, H.-J., Fo¨rster, B. and Edwards, C.A. (2004). Ring-testing and field-validation of a Terrestrial Model Ecosystem (TME) – an instrument for testing potentially harmful substances: Conceptual approach and study design. Ecotoxicology 13, 9–27. Knacker, T., Ro¨mbke, J., Eder, M., Federschmidt, A., Fo¨rster, B. and Schallnass, H.-J. (1994). Arbeiten zur U¨bertragbarkeit und Pra¨zisierung der Wirkungsmechanismen chemischer Belastung in verschiedenen O¨kosystemen. TP 2: Zoologische und mikrobiologische Parameter zum Streuabbau. Bericht des Battelle-Instituts fu¨r das Bundesministerium fu¨r Bildung und Forschung Nr. 0339302B, 536 pp. Knacker, T., Marcinkowski, A., Fo¨rster, B., Arthur, M.F. and Tolle, D.A. (1989). Refinement of terrestrial microcosms for evaluating fate and effects of chemicals. Brighton Crop Protection Conference – Weeds, 259–65. Landis, W.G., Matthews, R.A. and Matthews, G.B. (1997). Design and analysis of multispecies toxicity tests for pesticide registration. Ecol. Appl. 7, 1111–6. Lofs-Holmin, A. (1981). Influence in field experiments of benomyl and carbendazim on earthworms (Lumbricidae) in relation to soil texture. Swed. J. Agric. Res. 11, 141–7. Morgan, E. and Knacker, T. (1994). The role of laboratory terrestrial model ecosystems in the testing of potentially harmful substances. Ecotoxicology 3, 213–33. Norusis, M.J. (1998). SPSS
8.0 Guide to data analysis, p. 563. Upper Saddle River. OECD (2001). Guideline for Testing of Chemicals (Draft). Earthworm Reproduction Test (Eisenia fetida/andrei). Organisation for Economic Co-Operation and Development, Paris. Omodeo, P. (1988). The genus Eophila (Lumbricidae, Oligochaeta). Boll. Zool. 55, 73–84.

Ro¨mbke, J., Beck, L., Fo¨rster, B., Fru¨nd, C-H., Horak, F., Ruf, A., Rosciczewski, K., Scheurig, M. and Woas, S. (1997). Boden als Lebensraum fu¨r Bodenorganismen und bodenbiologische Standortklassifikation – Literaturstudie. Texte und Berichte zum Bodenschutz 4/97. Landesanstalt fu¨r Umweltschutz Baden- Wu¨rttemberg (Karlsruhe). 390 pp. and appendix. Sachs, L. (1999). Angewandte Statistik. Anwendung statistischer Methoden, 881 pp. Berlin. Satchell, J.E. (1983). Earthworm Ecology. From Darwin to Vermiculture, 495 pp. London: Chapman and Hall. Satchell, J.E. (1955). Some aspects of earthworm zoology. In D.K. Mc. E. Kevan, (ed) Soil Zoology, pp. 180–201. London: Butterworths. Sheppard, S.C. (1997). Toxicity testing using microcosms. In: Tarradellas, J., Bitton, G. and Rossel, D. (eds). Soil Ecotoxicology, Lewis Publ., Boca Raton, pp. 345–73. Sims, R.W. and Gerard, B.M. (1985). Earthworms. In: Kermack, D.M. and Bames, R.S.K. (eds). Synopses of the British Fauna (New Series) No. 31. 171 pp. London: E.J. Brill/Dr. W. Backhuys. Sparks, T. (2000). Statistics in Ecotoxicology. 320 pp. Chichester: Wiley. Stop-Bøwitz, C. (1969). A contribution to our knowledge of the systematics & zoogeography of Norwegian earthworms. Nytt Magazin Zoologie 17, 169–280. Trigo, D., Mascato, R., Mato, S. and Diaz Cosin, D. (1988). Biogeographical divisions of continental Portugal as regards earthworm fauna. Boll. Zool. 55, 85–92. Van Gestel, C.A.M. (1992). Validation of earthworm toxicity tests by comparison with field studies: A review of benomyl, carbendazim, carbofuran, and carbaryl. Ecotoxol. Environ. Safety 23, 221–36. Van Gestel, C.A.M., Dirven-Van Breemen, E.M., Baerselman, R., Emans, J.J.B., Janssen, J.A.M., Postuma, R. and Van Vliet, P.J.M. (1992). Comparison of sublethal and lethal criteria for nine different chemicals in standardized toxicity tests using the earthworm Eisenia andrei. Ecotoxicol. Environ. Safety 23, 206–20. Van Leeuwen, C.J. and Hermens, J.L.M. (eds). (1995). Risk Assessment of Chemicals. An Introduction, 374 pp. Dordrecht: Kluwer Academic Publishers. Van Voris, P., Tolle, D.A. and Arthur, M.F. (1985). Experimental Terrestrial Soil-Core Microcosm Test Protocol. United States Environmental Protection Agency, Washington, EPA/600/3-85/047 PNL-5450, UC-11. Velthorst, E.J. (1993). Manual for chemical water analysis. Department of Soil Science and Geology, Agricultural University, Wageningen, The Netherlands. Vonk, J.W., Adema, D.M.M. and Barug, D. (1986). Comparison of the effects of several chemicals on microorganisms, higher plants and earthworms. In J.W. Assink, and W.J. Van den Brink (eds). Contaminated Soil, pp. 191–202. Dordrecht, The Netherlands: Martinus Nijhoff Publishers. Weyers, A., Sokull-Klu¨ttgen, B., Knacker, T., Martin, S. and Van Gestel, C.A.M. (2004). Use of Terrestrial Model Ecosystem data in environmental risk assessment for industrial chemicals, biocides and plant protection products in the EU. Ecotoxicology 13, 163–176.

118 Ro¨mbke et al. World Health Organisation (1993). Carbendazim. Environmental Health Criteria 149. Published under the joint sponsorship of the United Nations Environment Programme, the International Labour Organisation, and the World Health Organisation, Geneva, Switzerland, pp. 1–125.

Zicsi, A. (1976). Weitere Angaben zur Regenwurm-Fauna (Oligochaeta: Lumbricidae) des Tessins (Schweiz). Revue Suisse Zool. 83, 515–20. Zicsi, A. (1965). Eine neue Regenwurm-Art aus Portugal (Oligochaeta: Lumbricidae). Acta Zool. 11, 217–25.