Several studies have found incomplete calculus removal during periodontal treatment with traditional hand curets, sonic, and ultrasonic instruments. This study evaluated the speed and effectiveness of subgingival calculus removal with new diamond-coated ultrasonic tips on single-rooted teeth. Single session subgingival scaling and root planing was performed on 80 teeth with 5 to 12 mm probing depths in 15 patients. Each patient provided groups of 4 teeth that were randomly treated with either hand curets (HAND); standard smooth ultrasonic tip (US); or fine grit (FINDIAM) or medium grit (MEDDIAM) diamond-coated ultrasonic tips. The time taken to reach the therapeutic endpoint of a clean, smooth root surface in a defined region on each tooth with each instrument by the 3 therapists with differing experience levels was recorded. The teeth were then atraumatically extracted, stored in a surfactant, photographed at 10X, and the percent of calculus present in the area of the pocket or on a comparable control surface calculated by histometric point counting. ANOVA and paired t tests showed that mean percent remaining calculus on treated versus control surfaces was HAND 4.6 +/- 5.3 versus 57.5 +/- 28.2, US 4.7 +/- 6.4 versus 54.4 +/- 25.9, FINDIAM 4.3 +/- 5.2 versus 37.5 +/- 22.1, and MEDDIAM 3.4 +/- 4.2 versus 50.7 +/- 20.1, respectively (all P < 0.01). The mean time in seconds to reach the clinical endpoint ranged from HAND 289 +/- 193, US 194 +/- 67, FINDIAM 167 +/- 71, to MEDDIAM 147 +/- 92. All powered instruments were significantly faster than HAND (P < 0.05), but did not differ from each other. On a 0 = "smooth" to 3 = "rough" scale, most often HAND resulted in "smooth" surfaces (10/20), the powered tips of all types "slight" surface roughness (10/20 each), and US the most "moderate" roughness (7/20). There were no differences in percent calculus remaining, surface roughness, or time spent among the 3 treating clinicians despite their varying experience levels. The results of this study showed that percent calculus remaining was <5% with all the instruments given time ad libitum on a given root surface. Root roughness was generally slightly greater with all 3 powered tips. All of the powered instruments took significantly less time than the HAND. Both DIAM tips took less time than US. Diamond-coated ultrasonic tips appeared to be much more efficient than HAND or US in removing calculus in moderate-deep probing depths on single-rooted teeth in vivo.