Some considerations on the transfer of laboratory data to the field

The problem of obtaining reproducible data in field experiments is discussed. If a stringent statistical definition is followed reproducibility of field experiments in soil microbiology does not exist. Thus, a direct transfer of laboratory data to the field situation is not possible. In this paper the necessity of establishing causal links between soil properties and the toxicity of chemicals to the microflora is emphasized. Rather than continuing to monitor the side‐effects of pesticides on the soil microflora, the ultimate approach should be prediction of the toxicity of pesticides, under any environmental condition, by mathematical modelling.

Key words:

• Pesticides
• microbial activity
• hazard assessment
• side‐effect testing
• reproducibility of data
• data evaluation