PURPOSE: To investigate the optimal delay and acquisition time for discrimination of the arterial from thevenous phase in ultrafast 3D contrast-enhanced MR angiography of the carotid artery. MATERIALS AND METHODS: Wereviewed the MR angiographic findings of 233 patients in whom carotid stenosis and cerebrovascular disease weresuspected. On the basis of delay and acquisition time they were divided into four groups. In the first three,contrast material was injected manually, and after the optimal time, mechanical injection was used for the lastgroup. On the basis of signal intensity of the carotid artery, image quality was graded in four steps.Discrimination of the arterial from the venous phase was graded in three steps, based on the degree of venousenhancement. RESULTS: The best grade of image quality was 70% in the first group, 85% in the second, and 95% inthe third. In discrimination of the arterial-venous phase, the most definite grade was 50% in the first group, 62%in the second, and 75% in the third. Between manual and mechanical injection groups, there was no significantdifference in image quality and discrimination of the arterial-venous phase. CONCLUSION: These results suggestthat for ultrafast 3D contrast-enhanced MR angiography of the carotid artery, with manual injection of contrastmaterial, 8-second delay time and 7-second acqusistion time are optimal.