A synthetic laboratory population of the diamondback moth, Plutella xylostella (L.), was used to test the F2 screen developed for detecting the frequency of rare resistance alleles to Cry1Ac and Cry1C toxins of Bacillus thuringiensis (Bt). Of the 120 single-pair matings set up, 106 produced enough F2 families for screening of Cry1Ac or Cry1C resistance alleles using both transgenic broccoli and an artificial diet overlay assay with a diagnostic dose. When using Bt broccoli plants as the F2 screen method, only one F2 family was detected for Cry1Ac resistance and no family was detected for Cry1C resistance. Six families were detected for either Cry1Ac or Cry1C resistance using the diet assay. The survivors in the diagnostic diet assay were crossed with the resistant individuals to confirm their resistance genotypes. Four F2 families were confirmed to contain one copy of an allele resistant to Cry1Ac in the original single-pairs and four other F2 families contained an allele resistant to Cry1C. Our results suggest that using transgenic plants expressing a high level of a Bt toxin in an F2 screen may underestimate the frequency of resistance alleles with high false negatives, or fail to detect true resistance alleles. The diagnostic diet assay was a better F2 screen method to detect alleles, especially for the Cry1Ac resistance with monogenic inheritance in the diamondback moth. The estimated probabilities of false positives and false negatives were 33 and 1%, respectively, for detecting Cry1Ac resistance at the allele frequency of 0.012 using the diagnostic diet assay. Careful validation of the screening method for each insect-crop system is necessary before the F2 screen can be used to detect rare Bt resistance alleles in field populations.