Results of life-cycle toxicity experiments are supposed to be indicative for long-term effects of exposure to toxicants. Several studies, however, have shown that adaptation or extinction of populations exposed for several generations may occur. The aim of this study was therefore to determine if the effects of the PAH phenanthrene on survival and reproduction of the springtail Folsomia candida exposed for 10 consecutive generations to contaminated soil would progressively increase, or, alternatively, if adaptation of the test organisms to the toxicant would occur. LC50 values for the first four generations were similar (171-215 micromol/kg dry soil), as expected for a narcotic compound. In the fourth generation, springtails exposed to a concentration similarto the EC50 for one generation (163 micromol/kg dry soil) showed internal phenanthrene concentrations in the range known to cause mortality; no reproduction took place, and the population went extinct. From the fifth generation onwards, survival and reproduction were not affected by the remaining exposure concentrations. Apparently, up to a certain threshold concentration (above 77 and below 163 micromol/kg dry soil), the springtails were able to metabolize phenanthrene, as shown by the lack of adverse effects and the lack of adaptation. During multigeneration exposure, the graded concentration-response relationship changed into an all-or-nothing response with a defined threshold concentration.Together with the worsening of effects, this raises concerns about the use of single-generation studies to tackle long-term population effects of environmental toxicants.