&NA;Anesthetized lambs were subjected to epidural cord compression at T13 by means of an epidural balloon distended to 200 mmHg for 80 minutes. Determinations of spinal cord blood flow (SCBF) were made by labelled microspheres prior to and during compression, and then ½, 1½, and 2½ hours after compression. Twelve control animals received saline (80 ml/h). Nine animals received hetastarch (a 20 ml/kg bolus followed by an 80 ml/h infusion) and 8 animals received mannitol (a 1 g/kg bolus followed by 1 g/kg/hr). An additional 10 animals received phenylephrine to raise mean arterial pressure by 20 to 40%. Somatosensory evoked potentials (SEPs) were recorded following each determination of SCBF. Animals were killed following determination of the 2½ hour postcompressive flow. The animals treated with either hetastarch or phenylephrine had a postcompressive mean arterial pressure that was significantly greater than that of the controls (P< 0.01). During treatment, cardiac output in the hetastarch and mannitol animals was significantly greater and hematocrit significantly less than in the controls (P< 0.05). In spite of these changes, postcompressive SCBF was significantly increased only in those animals treated with phenylephrine (P< 0.05). Statistical analysis of the relationship between mean arterial pressure and postcompressive SCBF revealed that autoregulation is indeed lost with this experimental model of spinal cord injury. Histological examination of the injured site failed to demonstrate any difference in the area of hemorrhagic necrosis among the four groups. Electrical conduction across the injured site, as measured with SEPs, was lost 10 minutes following balloon inflation and did not recover during the 2½ hour duration of this experiment. (Neurosurgery24:228‐234, 1989)