One conditioned taste aversion experiment with rats assessed the impact of extinguishing a target conditioned stimulus (CS), S, in compound with a second CS, A, upon conditioned responding elicited by CS S when presented alone at test. Following initial conditioning treatment with CSs A and S, the experiment manipulated number of extinction trials with CS A alone (i.e., 0, 5, or 10 trials) prior to AS compound treatment. In addition, two control groups received either extinction trials with S alone or no extinction treatment with S. Conditions receiving either 0 or 10 extinction trials with CS A prior to nonreinforced exposures to AS showed results indicating that aversion elicited by CS S was protected from extinction, whereas a condition receiving 5 extinction trials with CS A prior to AS trials showed unprotected extinction of aversion elicited by CS S. Current associative models are challenged in accounting for this pattern of results. (Contains 1 table, 2 figures, and 1 footnote.)