Pages 155-161, Language: EnglishLibman / NichollsTwenty five extracted human central incisors were divided into five groups nd prepared for complete cast crowns. Test teeth had cast dowel cores fabricated with the ferrule height varying from .5 to 2.0 mm in .5-mm increments. The five control teeth did not have cast dowel cores. A 4.0-kg load was applied to each of the restored teeth at an angle of 135 degrees to the long axis of each tooth. This load was applied cyclically at a rate of 72 cycles per minute. The load application point was predetermined by a waxing jig that was used to wax all crowns. The primary variable was the ferrule length. The independent variable was the number of load cycles required to create preliminary failure. Preliminary failure was defined here as the loss of the sealing cement layer between crown and tooth. An electrical resistance strain gauge was used to provide evidence of preliminary failure. The results of this study showed that the 0.5 mm and 1.0 mm ferrule lengths failed at a significantly lower number of cycles than the 1.5 mm and 2.0 mm ferrule lengths and control teeth.