The mode of action of the entomopathogenic bacterium Bacillus thuringiensis (Bt) remains a matter of debate. Recent reports have claimed that aseptic lepidopteran hosts were not susceptible to Bt and that inoculation with mid-gut bacteria restores pathogenicity. These claims are controversial because larvae were rendered aseptic by consuming antibiotics, although the effect of these antibiotics on Bt was not examined. We tested the generality of the mid-gut bacteria hypothesis in the diamondback moth, Plutella xylostella using properly controlled experiments that investigated the effect of antibiotic consumption and absence of gut microbiota separately. We found that purified Bt toxin and spore/toxin mixtures were fully pathogenic to larvae reared aseptically. Persistence of antibiotics in larval tissues was implicated in reducing host mortality because larval consumption of the antibiotic rifampicin reduced the pathogenicity of rifampicin-sensitive Bt strains but not rifampicin-resistant strains. Inoculating larvae with Enterobacter sp. Mn2 reduced the mortality of larvae feeding on Bt HD-1 and the presence of a culturable gut microbiota also reduced the pathogenicity of the Bt toxin Cry1Ac, in agreement with other studies indicating that an intestinal microbiota can protect taxonomically diverse hosts from pathogen attack. As ingestion of antibiotics suppresses host mortality the vegetative growth of Bt in the host must be important for its pathogenicity. Furthermore, claims that aseptic larvae are not susceptible to Bt must be supported by experiments that control for the effect of administering antibiotics.