Statement of problem: Accurate delivery of torque to implant prosthetic screws is critical to generate ideal preload in the screw joint and offer protection against screw loosening. Mechanical torque limiting devices (MTLDs) are available in 2 different styles that have not been compared for accuracy. Purpose: The purpose of this study was to determine the accuracy of friction-style and spring-style MTLDs in delivering target torque values. Material and methods: Five MTLDs from each of 6 different implant manufacturers (n=30) were selected to determine their accuracy relative to their target torque values. All MTLDs were new; 3 types were of the friction style while 3 were of the spring style. To measure the output of each MTLD, a digital torque gauge with a 3-jaw chuck to hold the driver was used. Force was applied to the MTLD until either the friction style released at a precalibrated torque value or the spring style flexed to a precalibrated limit. The peak torque value registered by the digital torque gauge was recorded and the procedure was repeated 50 times for each MTLD. Statistical analysis used repeated-measures ANOVA (α=.05) to assess the accuracy of the MTLDs in delivering target torque values. Results: Both the mean absolute difference (ABSDIFF) and the mean percentage deviation (PERDEV) between measured torque values and target torque values differed significantly (P<.001) for the friction-style MTLDs (3.83 Ncm, 13.74%) and for the spring-style MTLDs (0.82 Ncm, 2.36%). Conclusions: Within the limitations of this study, MTLDs that use spring-style components are significantly more accurate than those that use friction-style components in achieving their target torque values. (J Prosthet Dent 2008;100:86-92).