Taste and odor have different properties in toxiphobic conditioning. When each is used alone, taste becomes aversive when followed by immediate or delayed poison, while odor becomes aversive only if followed by immediate poison. However, if odor and taste are presented as a compound and followed by delayed poison, then odor does become aversive when tested alone. It is as if taste has potentiated the odor signal. Several experiments assessed the role of the amygdala in this potentiation effect by anesthetizing the amygdala with 10% novocaine. Novocaine applied 30 min before presentation (Pre-CS) of an odor-taste compound disrupted the potentiated odor aversion but not the taste aversion. In contrast, novocaine applied 1 min after the compound odor-taste or 1 min prior to LiCl poison did not dissociate odor and taste aversions; both odor and taste aversions were facilitated. Novocaine applied 30 min before an odor alone also disrupted an odor aversion induced by immediate LiCl. But identical treatment did not disrupt odor avoidance conditioned by immediate foot-shock, suggesting that amygdala anesthesia does not simply produce anosmia. Pre-CS novocaine treatment also disrupted flavor neophobia prior to conditioning. The results suggest that novocaine applied to the amygdala disrupts the integration of odor with taste and illness during toxiphobic conditioning.