Duplex ultrasonography is an accepted method to assess noninvasively arterial inflow to the penis. Optimal pharmacological agents as well as timing of the scan and stimulation during the scan continue to be debated. In an effort to achieve a more complete smooth muscle relaxation and capture what we perceived was a wide variation in interval to maximum arterial velocity, we revised our duplex protocol in January 1991. We report on 280 consecutive patients evaluated in this manner. Patients received 0.25 or 0.5 cc of a triple drug mixture containing 22.5 mg./cc papaverine, 0.83 mg./cc phentolamine and 8.33 micrograms/cc prostaglandin E1. Scans were performed at 0, 5, 15 and 30 minutes after injection in all patients. Any patient not having a full erection at 15 minutes performed private self-stimulation while in the standing position for at least 5 minutes before the 30-minute scan. If we conservatively define normal arterial inflow as a peak Doppler velocity of 25 cm. per second or greater in the best artery, only 35% of our patients achieved this velocity at 5 minutes. Of the remainder 26% and 22% did not reach normal velocity values until 15 and 30 minutes, respectively, after the injection. By delaying initial measurements of velocity until 5 minutes, could the highest inflow velocity be missed and patients diagnosed incorrectly? The group at risk would be those who had good tumescence at 5 minutes and who had presumably already decreased the inflow velocities. Of the 280 patients 74 (26%) had greater than 10% tumescence at 5 minutes. Only 6 of these 74 patients did not reach velocities of 25 cm. per second or more in the best artery at some time during their study. In conclusion, our study clearly supports delaying the initial scan until 5 minutes, since only 6 of our 280 patients (2.1%) may have been incorrectly diagnosed. The study also strongly argues for additional scans until 30 minutes and self-stimulation when necessary.