Arthrotomies of middle carpal joints were done on 13 horses, and a 1-cm partial thickness, round defect was made on the radial facet of both third carpal bones. In one joint, 1-mm diameter 1-cm deep holes were drilled within the defect, and one joint was used as a control. Horses were assigned to 2 groups--group 1 (n = 6 horses), 5 drill holes; group 2 (n = 7 horses), 11 drill holes. At 1 and 3 weeks after surgery, differences between joints in synovial fluid total protein values, WBC counts, or results of mucin precipitate tests were not significant (P = 0.005). Physically and radiographically, horses were the same during the 12 initial weeks they were housed in stalls and the 9 weeks they were kept in paddocks. Twenty-one weeks after surgery, horses were euthanatized. Joints with drill holes had a significantly greater area (P less than 0.05) of healthy fibrocartilage new tissue: group 1--33 to 68% new tissue, compared with 0 to 23% new tissue in controls; and group 2--22 to 64% new tissue, compared with 0 to 37% new tissue in controls. Differences between healing of defects with drill holes in groups 1 and 2 were not significant. Thickness of new tissue over drill holes was 33 to 61% of thickness of cartilage adjacent to the defect, and thickness of tissue between drill holes was 11 to 43% (group 1) and 8 to 79% (group 2) of the thickness of cartilage adjacent to the defect. In all defects with drill holes, new tissue in the form of fibrocartilage was detected deep in drill holes, whereas fibrous tissue was observed superficially and adjacent to drill holes.