Nineteen Mediterranean natural soils with a wide range of properties and the Organisation for Economic Co-operation and Development (OECD) artificial soil were used to assess the influence of soil properties on the results of avoidance and reproduction tests carried out with the soil collembolan species Folsomia candida. Compared to natural soils, the OECD soil was mostly rejected by individuals when a natural soil was offered in avoidance tests, and the number of offspring produced was generally lower than the one obtained in natural soils. None of the soil properties assessed showed a significant influence on the avoidance behavior. More precisely, only soil moisture was included in the model explaining the avoidance response (avoidance increased with increasing differences in moisture), but its contribution was marginally not significant. The model derived explained only 16% of the variance in avoidance response. On the contrary, several soil properties significantly influenced reproduction (number of offspring increased with increasing moisture content, increasing coarse texture, and decreasing nitrogen content). In this case, the model explained 45% of the variance in reproduction. These results, together with the fact that most of the selected soils fulfilled the validity criteria in both avoidance and reproduction tests, confirm the literature experience showing that this species is relatively insensitive to soil properties and hence highly suitable to be used in ecotoxicological tests with natural soils. In addition, our study highlights the need for accuracy in soil moisture adjustment in soil ecotoxicological tests with this species. Otherwise, results of both avoidance and reproduction tests might be biased. © 2011 SETAC.
- Behavioral toxicology
- Soil properties
- Terrestrial invertebrate toxicology