The present research demonstrates a conditioning order effect difference: Odor-aversion conditioning is stronger following OT+/O+ conditioning than following O+/OT+ conditioning with specific odor (O) and taste (T) cues. When a weak odor cue was used in Experiments 1A and 1B, OT+/O+ conditioning produced significantly stronger odor aversions than did either O+/OT+ or O+/O+ conditioning, which did not differ. The same design was used in Experiment 2 with a strong odor, but the order effect difference was not replicated, suggesting that the order effect difference is specific to conditions that produce taste-potentiated odor aversions. The interpretation that O+/OT+ conditioning is weaker because of the absence of a taste-odor within-compound association was not supported in Experiments 3 and 4. Notably, using stimuli that supported potentiation in earlier experiments, in Experiment 4, we found evidence of a taste-odor within-compound association in the absence of potentiation. These results confirm that previous theories of potentiation (within-compound association model, sensory-and-gate channeling model) are not sufficient to produce potentiation. Instead, these results are interpreted in terms of taste-odor configural associations.