Effects of different soil types on the collembolans Folsomia candida and Hypogastrura assimilis using the herbicide Phenmedipham

Soil ecotoxicology studies are usually performed in standard soils such as Organization for Economic Cooperation and Development artificial soil or LUFA ST. 2.2, a natural soil. When assessing the toxic effects in the environment, soil properties are often different from those in standard soils, which might lead to a different exposure situation for the test species and therefore to misleading conclusions. Selected to cover a broad range of properties and based on the Euro-Soils concept, 17 different soils were studied regarding their suitability to two test species: Folsomia candida and Hypogastrura assimilis (Collembola). In reproduction tests, the test species reacted differently to the soils. F. candida was less affected by soil properties: 500 to 1200 juveniles/vessel were found in untreated soils (i.e., controls). These differences can be attributed to normal interindividual variability. H. assimilis showed a significant correlation with maximum water-holding capacity and also a tendency to lower the reproductive output in soils with a low pH (< 4). Therefore, some soils were revealed to be inappropriate for tests with H. assimilis. In the main tests, the effect of the reference test substance Phenmedipham ( formulation Betosyp) was studied in those soils where sufficient reproduction was determined beforehand. Clearly, the chronic end point was more sensitive than survival when testing Phenmedipham. In H. assimilis, because of high variability and low effects in the tested dosages, no conclusions could be drawn. In F. candida, different soils caused different toxic effects: Juveniles preferred soils with high C-to-N ratios. Higher microbial activity might support a quicker metabolization of the test substance. In general, the toxic response is caused by a synergistic action of several soil properties with each of them exerting an effect too small to be clarified with the available set of data.