Computerized tomography (CT) is the most often used imaging modality in the evaluation of acute clinical stroke. However, the rapidity with which CT density changes occur after acute, severe, focal ischemia cannot be determined clinically. Even if the time of symptom onset is known, clinical stroke severity is highly variable. We studied the time course of CT density change after severe, rapid onset, acute, focal ischemia as documented by stable xenon CT cerebral blood flow (CBF) in monkeys. Eight monkeys (Macaca mulatta) were subjected to transorbital occlusion of the left posterior cerebral, anterior, middle, and internal carotid arteries to induce focal ischemia. CT density Hounsfield units (HU), CBF by stable xenon CT, arterial blood pressure, and blood gases were measured before occlusion, immediately after occlusion, at 30 min, and hourly for up to 6 h. Occlusion of the cerebral arteries decreased CBF to 8 ± 5 ml/100 g/ min within 15 min postocclusion. At 6 h, CBF was unchanged at 9 ± 4 ml/100 g/ min. CT density within the ischemic core fell from 42 to 38 HU immediately after occlusion (P < 0.05), rose transiently, then fell at 2 h (P < 0.01) and plateaued at 36 ± 5 HU for a total decrease of 4-5 HU between 4 and 6 h poststroke. Changes in CT density lag severe focal ischemia by 2 h. Thus, when CT hypodensity is seen in acute stroke, it is likely 2 h old. It also provides an explanation for the phenomenon of clinical CT mismatch with clinical deficits and normal CT.