Coral reefs throughout the world have been damaged by storms, diseases, coral predators, temperature anomalies, and human activities. During the past three decades, recovery has been limited and patchy. Although a damaged coral reef cannot be restored to its original condition, interest in reef restoration is increasing. In a pilot project in the Caribbean (US Virgin Islands), storm-produced fragments of Acropora palmata, A. cervicornis, and Porites porites were collected from donor reefs and transplanted to nearby degraded reefs. Sixty coral fragments were attached to dead-coral substrate (usually A. palmata skeletons), at similar depths from which they had been collected (1-3.5 m), using nylon cable ties. Seventy-five intact colonies were designated as controls. Study colonies were assessed at 6-month intervals for 2 years (1999-2001) and annually thereafter (through 2004). One-fourth of the 135 colonies and fragments monitored were alive at the conclusion of the 5-year study. Survival of control and transplanted A. cervicornis and P. porites was very low (median survival 2.4 and 1.8 years, respectively), with no significant differences between transplant and control colonies. Site and depth did not contribute significantly to A. palmata colony survival, but colony size and transplant/control status did. Probability of survival increased with colony size. Median survival for A. palmata was 1.3 years for transplant and 4.3 years for natural colonies when not controlled for size. A. palmata was the only viable candidate for reef rehabilitation. Storm swells were the primary cause of mortality.