PURPOSE: This study was designed to determine the embolic effect of PVA particles of various sizes on the cerebral artery of a cat and to determine the appropriate particle size for embolization. MATERIALS AND METHODS: A total of 21 cats were divided into three groups according to the PVA particle size injected: group I(n=7), embolized with 45-150 micrometer PVA; group II(n=7), with 150-250 micrometer PVA; and group III(n=7), with 350-500 micrometer PVA. PVA particles were slowly injected into the left common carotid artery of each cat, and T2-weighted coronal MR images were obtained 24 hours after injection. During histologic examination of brain sections we analyzed the size, number of occluded vessels, and the ischemic changes caused by the particles. RESULTS: On T2 weighted images, areas of high signal intensity (infarction) were observed in four of the seven cats (57%) in group Iand in two of the seven (29%) in group II. High signal intensity was not found in group III. The mean percentage of areas of high signal intensity was 11.86 +/-1 . 37 % in group Iand 5.18 +/-1 . 77% in group II( P <0.05). During histologic examination, occlusion of the distal branches of the anterior cerebral (ACA) and/or the middle cerebral arteries(MCA) by PVA particles was observed in all seven cats (100%) in group I, and in four of the seven cats (57%) in group II. No group IIIcat showed occlusion of the distal branches of the ACA and/or MCA. The mean caliber of occluded vessels was 175 micrometer in Group Iand 258 micrometer in Group II. The mean number of occluded vessels seen on all slide sections was 14 in Group I and 5 in Group II. CONCLUSION: Small PVA particles had a greater cerebral embolic effect than did those which were medium or large. For the induction of embolic infarction in cat brain, PVA particles 45-150 micrometer in size are appropriate.