The effect of smoking urges on cognitive performance is relatively short-lived (Zwaan, Stanfield, & Madden, 2000). We examined whether this results from the short-lived nature of the elicited urge itself, or from practice effects on the cognitive task. Smokers listened to one of two imagery scripts ("urge" vs. "neutral") and subsequently performed two cognitive tasks (math and language comprehension). Exposure to the urge script produced significantly less accurate performance at the onset of the first task than exposure to the neutral script, but there was no difference at the onset of the second task. Thus, the quick disappearance of the urge effect seems due to the transient nature of the elicited urge itself, rather than to practice effects.