Abstract: Previous research has reported on the presence of Bt Cry toxins in herbivores and the subsequent effect on predators in tritrophic study systems. Further investigation has revealed that effects observed on the predator in Bt treatments were likely due to a nutritional deficiency caused by the suboptimal, Bt-sensitive prey (indirect effects) when predators have no alternative prey choices. Therefore, the question arises whether predators prefer optimal prey that is not affected by the Cry toxin uptake when given a choice. To answer this question, choice and no-choice experiments were performed using the generalist predator, Harmonia axyridis, feeding upon the fall armyworm, Spodoptera frugiperda, a major insect pest of maize. Harmonia axyridis larvae were fed with 1-5 day-old S. frugiperda larvae fed Bt or non-Bt maize expressing Cry1Ab. Spodoptera frugiperda larvae are sensitive to Cry1Ab; larvae feeding on Bt maize are smaller and development is delayed. In choice-experiments, H. axyridis ate more Bt-suppressed S. frugiperda larvae than larvae of the same age that had fed on non-Bt maize. Harmonia axyridis was not likely choosing prey based on whether it had fed Bt or non-Bt maize, but rather on the size of the larvae. These results may have implications for insect resistance management (IRM) for Bt crops. Larvae of pests harboring various levels of resistance against plant-expressed Bt toxins could be smaller than the sensitive larvae developing on a non-Bt crop. No differences between H. axyridis fed on Bt or non-Bt-fed S. frugiperda were observed in length of development and body weight. If Bt-fed S. frugiperda larvae are more likely to be preyed upon by H. axyridis, it is possible that such larvae, having experienced Bt selection, may be preferentially removed from local field populations. This scenario is now more likely to be played out under field conditions in the U.S. where the “refuge in the bag” approach, consisting of seed blends of Bt and non-Bt hybrid seed, randomly mixed within commercial fields (e.g., blends of 5 or 10% non-Bt seed placed in each Bt seed bag). The frequency of this phenomenon should be studied further under field conditions.