One risk of growing Bacillus thuringiensis (Bt) crops is the potential nontarget effects which are likely related to the environmental behavior of crystal (Cry) toxins. Bt rice residues left in field after harvest constitute a main source of Cry toxins entering the environment. To our knowledge, very few studies have simultaneously evaluated the persistence of Cry toxins in Bt rice residues under field conditions using different methods. Here, we established a bioassay method with a target insect: the striped stem borer (SSB), Chilo suppressalis Walker. The reaction limit of the SSB to Cry toxins ranged from 5.4 to 12.7 ng g-1 in artificial diet, indicating that the detection limit of the bioassay ranged from 54 to 127 ng g-1 rice residues. A field decomposition experiment lasting for 210 d was conducted with the straw of two Bt rice lines transformed with either cry1Ab/1Ac or cry2A. Enzyme-linked immunosorbent assays (ELISAs) revealed that the Cry toxins in the Bt rice residues experienced rapid degradation to below 25% of the initial level in the first 42 d, and then decreased to below 100 ng g-1 rice residues within 100 to 140 d. Flooded conditions accelerated the degradation in the beginning compared with buried conditions. The Cry toxins were still detectable by ELISA, although at levels below 10 ng g-1 rice residues (<0.3% of the initial level) 210 d after harvest. However, the bioassay revealed that the SSB no longer had a significant reaction to Bt rice residues added into artificial diets 16 to 18 d after harvest under both conditions, which indicated that the level of bioactive Cry toxins had declined to below the detection limit. Our results suggest that ELISA overestimate the persistence of Cry toxins and that the potential risks mediated by Cry toxins may be much smaller than originally expected.