Determination of the growth pattern of a harvested species is an important step toward a proper assessment and suitable management of the stock. Although coconut crab Birgus latro populations in most regions have been severely depleted, few reliable biological data are available. We assessed the growth of B. latro using a mark-recapture method based on the use of passive integrated transponder (PIT) tags on Hatoma Island, located southwest of Okinawa, Japan. Of 781 crabs tagged (337 males, 444 females; 14.89 to 58.03 mm thoracic length (ThL)), 17 males and 20 females were recaptured. The application of PIT tags provided field data for growth of individuals for up to 2326 d and confirmed that coconut crabs molt during the winter dry season on Hatoma Island. The curves relating ThL with age, estimated from data of 29 crabs that had spent one or multiple winter dry seasons at liberty, showed that the coconut crab is an extremely slow growing species, with males showing much higher mean asymptotic ThL (69.87 mm) than females (42.79 mm). The Brody growth coefficient (K)was 0.061 yr �1 for males and 0.091 yr �1 for females. This study is the first to provide growth estimates of wild coconut crabs based on mark- recapture over multiple years. Potentially long-lived, slow growing species such as coconut crabs may be susceptible to overharvesting, and a cautious approach is recommended to the determina- tion of resource management regimes for B. latro.