Retention interval effects are seen in taste-aversion learning when single-element aversions are significantly weaker 24 h after conditioning compared with tests at later intervals. This report contains three experiments which suggest that the source of the increased drinking at the 1-day interval is nonassociative interference produced by the novel conditioning episode. In Experiment 1, a parametric analysis demonstrated that aversion strength increased monotonically over a 30-h period following conditioning, and that by 48 h after conditioning it was stabilized. In Experiment 2, a single US preexposure was used to reduce the novelty of the US prior to conditioning. As a result, animals preexposed to the US had stronger taste aversions than did non-preexposed controls at a 1-day retention interval; however, no differences were seen at a 5-day interval. Experiment 3 investigated whether the counterintuitive outcome of Experiment 2 was due to the summation of environment-illness and taste-illness associations at the 1-day test. The results ruled out the summation argument; the US preexposure did not need to be presented in the conditioning context to strengthen the aversion at the 1-day interval. Collectively, these results suggest that the presentation of a surprising US can interfere with the retrieval of the taste-illness association for a short period after conditioning, and that this contributes to the retention interval effect.