Since 1981, I treated 81 scaphoid fractures: 14 wrists (group A) with acute unstable fracture, 37 (B) with ununited fracture accompanied by dorsiflexed intercalated segment instability (DISI), and 30 (C) with ununited fracture without DISI. The methods and results were evaluated retrospectively. Bone union occurred in 100% of group A wrists, 92% of group B, and 90% of group C. By a modification of Cooney's clinical scoring chart, mean scores were 91 in group A, 84 in group B, and 87 in group C. In 43% of group A, a scapholunate gap remained after surgery. Of 21 wrists in group B treated by bone graft and screw fixation, union was not achieved in one; in another 13, the R-L angle was corrected to within 10 degrees of the unaffected side. In group C, DISI did not develop if bone union occurred. There was correlation between shortness of the scaphoid and DISI remaining. The clinical scores of wrists with and without DISI were not different, except for patients who did hard manual labor, in half of whom the score was poor when DISI remained. Cineradiography of ununited fractures was done before and after surgery. For three wrists in group B, plain X-ray films showed that DISI improved, but cineradiography showed that DISI was present; even without static DISI, there may be dynamic DISI. Treatment of acute unstable fractures of the scaphoid by reduction and internal fixation generally gave satisfactory results. Perhaps when ligament injury may have caused instability, ligament repair or reconstruction is needed. For ununited fractures with DISI, correction of the DISI was needed as well as bone union. If union occurred with the scaphoid shorter than normal, DISI was present, so restoration of the length of the scaphoid during surgery is mandatory.