Kessler, Strickland, or modified Becker repairs, all augmented with a running circumferential epitenon suture, were performed for simulated zone II flexor tendon lacerations in the index, long, and ring fingers of 12 fresh-frozen cadaveric specimens. Each hand was tested with a tensiometer built for curvilinear testing of human flexor tendons in an intact hand. Each tendon was cycled 100 times, then examined for gapping before testing to failure. Maximum load to failure, including tendon load and pinch force, was recorded for each tendon. We propose that combining the advantages of cyclical testing and a curvilinear model is the most effective way of testing flexor tendon repairs capable of undergoing an early active motion protocol. None of the repaired tendons failed during the cyclic portion of testing. The average gapping after cycling for the 3 suture techniques was 0.12 +/- 0.35 mm for the Kessler technique, 0. 00 +/- 0.00 mm for the Strickland technique, and 0.19 +/- 0.26 mm for the modified Becker technique. The maximum tendon loads to failure were 33.8 +/- 6.8 N for the Kessler technique, 30.4 +/- 5.64 N for the Strickland technique, and 76.3 +/- 9.02 N for the modified Becker technique. There was a statistically significant difference between the modified Becker repair and the other 2 repairs for maximum tendon load and pinch force to failure. The results of this study show that all 3 tendon repair techniques can withstand forces reported with passive motion, but only the modified Becker repair allows sufficient strength above those forces that are estimated for active motion during tendon healing.