Sixty patients, 42 with coronary disease and 18 normals, were studied to assess the impact of a delay following exercise in commencing thallium imaging on the sensitivity for detecting ischemic transient defects. Three sets of images were obtained beginning 2 min, 18 min, and 2 hr after exercise. Each patient's images were separated into two pairs of studies for analysis: 2 min-2 hr and 18 min-2 hr. Of the 42 patients with coronary disease, a greater number had transient defects detected on the 2 min-2 hr compared with the 18 min-2 hr study (22 compared with 14, p less than 0.05). False positives were not increased by earlier imaging. We conclude that a modest (18 min) delay obtaining the first set of images causes a significant reduction in the ability to detect transient thallium defects, and that imaging should begin several minutes after exercise.