Purpose: This study determined the effect of bonding a composite core on the number of load cycles to cement failure for teeth restored with complete crowns. Materials and Methods: Ten extracted human maxillary central incisors were divided into two groups of five each. The tooth roots were at least 11 mm long, allowing an 8-mm post. Under copious irrigation, each tooth was cut to a flat plane 1 mm coronal to the buccal cementoenamel junction, perpendicular to the long axis of the tooth. The canal space was prepared to a final diameter of 1.25 mm, 8 mm deep. Each tooth was rebuilt with a 050 Parapost and Corestore. Final preparation height was 7 mm, including a 1-mm ferrule. For group 1, the core was bonded to the dentin using autocured Scotchbond Multipurpose Plus. For group 2, the core was not bonded. Following crown cementation, all teeth were subjected to a fatigue load of 4 kg at a rate of 280 cycles per minute. Results: One tooth with a bonded core failed at 19,880 cycles, and the other four in this group did not fail after 100,000 cycles. All of the teeth without a bonded core failed at a cycle count of less than 100. The Mann-Whitney nonparametric test showed a significant difference between the two groups (P .01). Conclusion: Bonding of a composite core to dentin prior to crown cementation provided a significantly stronger crown retention under fatigue loading.