No-take marine reserves can be effective biodiversity conservation and fisheries management tools, but as yet there is little indication of whether rotational management, i.e. alternately closing and opening an area to fishing, might also confer such benefits on coral reef fish stocks. Using data taken from the state of Hawaii's long-term reef monitoring program, we have assessed the effects of more than 2 decades of rotational management on fish stocks at the Waikiki-Diamond Head Fishery Management Area (FMA) on Oahu, Hawaii. Fish biomass tended to increase during the 1 to 2 yr closure periods, but the scale of these increases was insufficient to compensate for declines during open periods. The net effect was that, between 1978 and 2002, total biomass declined by around two-thirds. Coincident with this decline was the virtual disappearance of large fishes (> 40 cm) of fishery-target groups: acanthurids, scarids and mullids. Such fishes, although initially common, were only rarely recorded in surveys after 1990. In 1988, a portion of the FMA was converted into the permanently closed Waikiki Marine Life Conservation District (MLCD). Assessment of the relative effectiveness of permanent and rotational closure is complicated by declines in habitat quality, particularly within the MLCD, caused by overgrowth of much of the reef by the alien algae Gracilaria salicornia, which began in the early 1990s. However, the initial effect of full closure was a reversal of the previous downward trend in fish biomass, and, even in the post habitat-decline period, biomass of target species within the MLCD has been nearly twice as high as in the FMA. Additionally, there have been no declines or even downward trends in maximum size of target families in the MLCD. Overall, rotational management, as implemented at the Waikiki FMA, has not been an effective means of conserving fish stocks or revitalizing public fishing grounds.