Tendon defects remain a major concern in plastic surgery because of the limited availability of tendon autografts. Whereas immune rejection prohibits the use of tendon allografts, most prosthetic replacements also fail to achieve a satisfactory long-term result of tendon repair. The tissue engineering technique, however, can generate different tissues using autologous cells and thus may provide an optimal approach to address this concern. The purpose of this study was to test the feasibility of engineering tendon tissues with autologous tenocytes to bridge a tendon defect in either a tendon sheath open model or a partial open model in the hen. In a total of 40 Leghorn hens, flexor tendons were harvested from the left feet and were digested with 0.25% type II collagenase. The isolated tenocytes were expanded in vitro and mixed with unwoven polyglycolic acid fibers to form a cell-scaffold construct in the shape of a tendon. The constructs were wrapped with intestinal submucosa and then cultured in Dulbecco's Modified Eagle Medium plus 10% fetal bovine serum for 1 week before in vivo transplantation. On the feet, a defect of 3 to 4 cm was created at the second flexor digitorum profundus tendon by resecting a tendon fragment. The defects were bridged either with a cell-scaffold construct in the experimental group ( n= 20) or with scaffold material alone in the control group ( n= 20). Specimens were harvested at 8, 12, and 14 weeks postrepair for gross and histologic examination and for biomechanical analysis. In the experimental group, a cordlike tissue bridging the tendon defect was formed at 8 weeks postrepair. At 14 weeks, the engineered tendons resembled the natural tendons grossly in both color and texture. Histologic examination at 8 weeks showed that the neo-tendon contained abundant tenocytes and collagen; most collagen bundles were randomly arranged. The undegraded polyglycolic acid fibers surrounded by inflammatory cells were also observed. At 12 weeks, tenocytes and collagen fibers became longitudinally aligned, with good interface healing to normal tendon. At 14 weeks, the engineered tendons displayed a typical tendon structure hardly distinguishable from that of normal tendons. Biomechanical analysis demonstrated increased breaking strength of the engineered tendons with time, which reached 83 percent of normal tendon strength at 14 weeks. In the control group, polyglycolic acid constructs were mostly degraded at 8 weeks and disappeared at 14 weeks. However, the breaking strength of the scaffold materials accounted for only 9 percent of normal tendon strength. The results of this study indicated that tendon tissue could be engineered in vivo to bridge a tendon defect. The engineered tendons resembled natural tendons not only in gross appearance and histologic structure but also in biomechanical properties.