To develop an easy, reproducible experimental model of cerebral infarction (CI) without craniotomy in New Zealand white rabbits, a silicone rubber cylinder embedded in a nylon suture was delivered to the middle cerebral arteries through the internal carotid artery in anesthetized animals. Rabbits were sacrificed 0.5-5 h after embolization. CI size and location were ascertained by the triphenyl-2H-tetrazolium chloride (TTC) staining method; cerebral blood flow (CBF) was measured prior to and after embolization. PCO2, temperature and blood pressure were monitored and kept constant. CI occurred in all rabbits after 4 h of ischemia, in 50% after 3 h and only in 33% after 2.5 h. CI did not occur within less than 2.5 h of ischemia. No correlation was found between size and location of CI and occlusion time. CBF was maximally reduced in the right MCA territory but was also reduced in both anterior cerebral arteries and left MCA territories. This model is technically easy and the retrievable embolus allows the study of reperfusion by pulling on the nylon suture. It is suitable for studying chemical and molecular changes of the ischemic cells and/or for studying neuroimage changes after ischemic stroke.