We investigated the efficiency of basic fibroblast growth factor (b-FGF) released from a gelatin sponge in the regeneration of tracheal cartilage. A 1-cm gap was made in the midventral portion of each of 10 consecutive cervical tracheal cartilages (rings 4 to 13) in 15 experimental dogs. In the control group (n = 5), the resulting gap was left blank. In the gelatin group (n = 5), a gelatin sponge alone was implanted in the gap. In the b-FGF group (n = 5), a gelatin sponge containing 100 mug b-FGF solution was implanted in the gap. We euthanatized one of the five dogs in each group at 1 month after implantation and one at 3 months and examined the implant sites macroscopically and microscopically. In the control and gelatin groups, no regenerated cartilage was observed in the tracheal cartilage gap at 1 or 3 months. The distances between the cartilage stumps had shrunk. In the b-FGF group, fibrous cartilage had started to regenerate from both host cartilage stumps at 1 month. At 3 months, regenerated fibrous cartilage filled the gap and had connected each of the stumps. The regenerated cartilage was covered with regenerated perichondrium originating from the host perichondrium. Shrinkage of the distance between the host cartilage stumps was not observed in the b-FGF group. We succeeded in inducing cartilage regeneration in the gaps in canine tracheal cartilage rings by using the slow release of b-FGF from a gelatin sponge. The regenerated cartilage induced by b-FGF was fibrous cartilage.