The slow-growth-high-mortality (SG-HM) hypothesis states that arthropods fed on suboptimal diets have prolonged developmental time, which will expose the arthropods to higher mortality from their natural enemies because the window of vulnerability to natural enemies is increased. In this study, SG-HM was tested with cotton plants, Gossypium hirsutum, beet armyworm, Spodoptera exigua, and the parasitoid, Cotesia marginiventris. Differential growth rates of beet armyworm were generated by rearing beet armyworm larvae on cotton plants receiving 42 (low) and 196 (high) ppm nitrogen (N). The results demonstrated that there was a defined range of sizes (from ca. 0.08-8 mg/larva) below and above which S. exigua caterpillars were not suitable for C. marginiventris development. Spodoptera exigua larvae reared on low N plants reached the thresholds more slowly than larvae reared on high N plants. However, no significant parasitism difference was observed when S. exigua larvae of the two N treatments were exposed to C. marginiventris females both in no-choice and cage choice tests. The SG-HM failed partly because in early S. exigua developmental stages (3-d-old and younger), more larvae from high N plants were parasitized, while in later stages more larvae from low N plants were parasitized. Therefore, the timing of vulnerability of S. exigua grown on host plants with different N levels to C. marginiventris shifted, but the size of the window of vulnerability remained the same.