Pollution-induced community tolerance (PICT) as an ecotoxicological test system has been claimed to detect pollutant effects highly specifically and sensitively. However, the specificity might be limited by the occurrence of cotolerance. Another limitation of the application of any ecotoxicological test system lies in variation of the measured responses. We tested the variation and the occurrence of cotolerance experimentally, using antibiotics as toxicants, soil microcosms as microbial communities, and tolerance determination in Biolog plates as PICT detection test. Bacteria have been discussed as being prone to multiple tolerances due to the possible accumulation of multiple resistance genes on mobile genetic elements. However, in our experiments, cotolerance occurred only between antibiotics of the same group (oxytetracycline and tetracycline), as expected from their identical mode of action. Cotolerance between oxytetracycline and tylosin in soil microcosms exposed to oxytetracycline was low, as was cotolerance to oxytetracycline in tylosin-exposed microcosms. We conclude that tolerance development to antibiotics in soils reflects the actual selection pressure rather than a general pattern of multiple resistances. Concerning variation, the PICT effect of tetracycline was well reproducible in two consecutive years. The response variation linked to PICT experiments in controlled microcosms was comparable to that of ecotoxicological test systems of equivalent complexity. In conclusion, our results support an application of the PICT methodology as an effective means to study the soil ecotoxicology of antibiotics.