A series of four experiments evaluated eyelid conditioning in the rabbit under circumstances in which the unconditioned stimulus (US) could variously be assumed to be prerepresented (primed) in short-term memory at the time of its presentation in pairing with a conditioned stimulus (CS). In both between- and within-subjects comparisons, less conditioning resulted when CS-US pairings were shortly preceded by US presentations than when not. The decremental effect of the pretrial priming US depended upon the similarity of such US to the subsequent conditioning US included in the CS-US pairing and could be removed by the interpolation of a distractor stimulus between the priming US and the CS-US pair. Finally, a similar decrement in conditioning was observed if a weak priming US was administered prior to the conditioning US but after the CS, that is, within the CS-US interval. The overall pattern of data supports Wagner's thesis that stimuli are rendered generally less effective when they are primed in short-term memory.