The object of this investigation was to study the effects of intermittent, low dose mannitol therapy on conscious cats after acute left middle cerebral artery (MCA) occlusion. A simple implanted device was applied to the proximal left MCA of 40 adult cats using microsurgical techniques. In the acute experiments, 10 cats were untreated and 10 cats received mannitol (0.5 g/kg intravenously) immediately before occlusion and again 3, 6, and 9 hours later. They subsequently underwent intra-arterial perfusion with colloidal carbon and buffered paraformaldehyde 12 hours after occlusion. The plasma osmolality immediately before perfusion was 316 ± 2 (SD) milliosmoles in untreated cats and 331 ± 5 milliosmoles in treated cats. Gross swelling, impaired carbon filling, and breakdown of the blood-brain barrier (BBB) to fluorescein were seen in the left MCA territory of 8 untreated cats and 1 treated cat. The mean percentage of gray matter cross sectional area where severe ischemic neuronal alterations predominated was 45 ± 12% in untreated and 14 ± 16% in treated cats (p < 0.01). The mean capillary luminal diameter in the left sylvian cortex was 4.5 ± 1.0 μ in untreated cats (control, 6.5 ± 1.0 μ) and 5.5 ± 1.0 μ in treated cats. In the subacute experiments, 10 cats were not treated and 10 cats received mannitol as in the acute experiments. The cats were killed with a large bolus of sodium pentobarbital 48 hours after left MCA occlusion. Gross swelling and breakdown of the BBB were less severe in treated cats. The mean cross sectional area of infarcted tissue was 55 ± 12% in untreated cats and 33 ± 21% in treated cats (p < 1.0). The findings of this study indicate that intermittent, low dose mannitol therapy delays the onset of ischemic cerebral injury and may reduce the size of the eventual infarct or convert a potential infarct into a so-called “transient ischemic attack.”