To develop an arterial injury model for testing hemostatic devices at well-defined high and low bleeding rates. material and method. A side-hole arterial injury was created in the carotid artery of sheep. Shed blood was collected in a jugular venous reservoir and bleeding rate at the site of arterial injury was controlled by regulating outflow resistance from the venous reservoir. Two models were studied: uncontrolled exsanguinating hemorrhage and bleeding at controlled rates with blood return to maintain hemodynamic stability. Transcutaneous Duplex ultrasound was used to characterize ultrasound signatures at various bleeding rates. Results. A 2.5 mm arterial side-hole resulted in exsanguinating hemorrhage with an initial bleeding rate of 400 ml/min which, without resuscitation, decreased to below 100 ml/min in 5 minutes. After 17 minutes, bleeding from the injury site stopped and the animal had lost 60% of total blood volume. Reinfusion of shed blood maintained normal hemodynamics and both high and low bleeding rates could be maintained without hemorrhagic shock. Bleeding rate at the arterial injury site was held at 395±78 ml/min for 8 minutes, 110±11 ml/min for 15 minutes, and 12±1 ml/min for 12 minutes. Doppler flow signatures at the site of injury were characterized by high peak and end-diastolic flow velocities at the bleeding site which varied with the rate of hemorrhage. conclusion. We have developed a hemodynamically stable model of acute arterial injury which can be used to evaluate diagnostic and treatment methods focused on control of the arterial injury site.