A consecutive series of eighty-seven saddle-bridges with free-end saddles, executed by the author during the years 1958-72, was investigated in 1982. The aim was to assess the long-term prognosis. Five years after application 98% of the bridges were still functioning. The corresponding percentage after 10, 15 and 20 years was 82, 69 and 49%, respectively. The risk that saddle-bridges cease to function is very low during the first 5-year period but significantly higher thereafter. Soldering fractures constitute an important cause of the removal of appliances during the first 8 years following application. After this time the most frequent causes are caries and loosening of the teeth. Factors such as sex, time of application, patient's age, length of saddle and whether the bridges were fitted in the mandible or maxilla, have not had a significant influence on the prognosis. Furthermore, no statically significant difference has been found between the long-term prognosis for full and partial saddle-bridges. On the other hand, patients with a removable full denture in the opposing jaw show a significantly worse prognosis (P less than 0.05) than patients without a denture in this jaw. The results of the present study indicate that the short-term prognosis for saddle-bridges is very favourable and that the long-term prognosis is so good that saddle-bridges can constitute an alternative therapy even in jaws with as few as three remaining teeth.