We conducted a prospective, double-blind clinical trial to evaluate the differences in operating time, intraoperative blood loss, postoperative pain at 3 hours and 1 week, and delayed (>24 hr) bleeding associated with ultrasonic harmonic scalpel tonsillectomy and conventional tonsillectomy. The study was carried out on 28 patients with recurrent tonsillitis and/or adenotonsillar hypertrophy who underwent harmonic scalpel tonsillectomy on one side and cold dissection tonsillectomy with suction electrocautery hemostasis on the other. The harmonic scalpel was associated with significantly less intraoperative blood loss (mean: 6.2 vs. 58.8 ml; p < 0.0001) and less early (3 hr) postoperative pain as determined by scores on a 10-point visual analog scale (mean: 3.5 vs. 4.4; p = 0.0042); although the difference in early pain scores is statistically significant, it is probably not clinically significant. Pain scores at 1 week were nearly identical (mean: 2.7 vs. 2.6; p = 0.9246). The length of operating time was similar (mean: 10.9 vs. 7.7 min; p = 0.0022). An unanticipated finding was the fact that delayed bleeding, which occurred in 3 patients (10.7%), occurred only on the harmonic scalpel side. We conclude that the only clearly demonstrable advantage that the harmonic scalpel had over cold dissection was that it caused less intraoperative blood loss.