Resistance to fenvalerate was examined in a laboratory-selected population of the diamondback moth, Plutella xylostella, during exposure through 10 generations. Selection results showed that LD 50 values rose from 0.003 mg per larva to 31.1 mg per larva and 36.7 mg per larva respectively, in replications 1 and 2 during those 10 generations. Resistant line moth eggs were significantly smaller than those of moths before selection. Therefore, we selected a subpopulation of resistant lines under harsh environmental conditions of low humidity and high temperature, over 10 generations (Harsh line) without insecticide exposure. We selected another subpopulation of resistant lines under optimal environmental conditions of high humidity and normal temperature through all 10 generations (Optimal line). After 10 generations, the LD 50 of fenvalerate decreased to 0.03 mg per larva and 0.02 mg per larva under harsh conditions and 0.2 mg per larva and 0.29 mg per larva under optimal conditions. The LD 50 values were lower for the Harsh lines than for the Optimal lines in all generations. The egg size of both lines increased gradually through 10 generations. Comparisons of these two lines showed that the egg size increased more rapidly in the Harsh lines than in Optimal lines in early (first-sixth) generations. In subsequent generations, egg sizes were nearly normal in both lines. Comparison of immature survivability of Harsh and Optimal lines reared under those conditions showed that the survivability of Harsh line individuals was significantly lower than for Optimal lines. Comparison of the two lines' survivability showed an opposite result from the comparison when they were reared under equivalent optimal conditions. These results suggest that susceptibility to insecticides might recover more quickly in Harsh lines than in Optimal lines because resistant insects with small eggs had lower survivability than susceptible insects with normal eggs. They were eliminated more rapidly under harsh environmental conditions.