| The grouper species roi (Cephalopholis argus) was introduced to Hawaii in 1956 to create a new fishery. It has become the dominant predator on coral reefs in the Main Hawaiian Islands, but the envisioned fishery failed due to concerns about ciguatera fish poisoning from roi consumption. This situation raises resource management and ecological questions, including (1) How much prey does the alien species roi consume, and what are the ecological effects on native species? and (2) Is roi as ciguatoxic as its reputation suggests, and could better understanding of factors associated with toxicity lead to a safe fishery for roi?;My dissertation addresses these questions, based on gastric evacuation rate experiments with roi, stomach content analysis, assays of ciguatoxin concentrations in roi tissue, and on the analysis of benthic habitat maps and underwater visual censuses of roi and prey fish densities.;Regarding (1), the estimated roi population size in 7.8 km2 of reef habitat along the West coast of Hawaii Island was 56,290 individuals (biomass 34 metric tons (t)), with an annual prey consumption of 93.7 t of fish (equivalent to 8.2 million individuals) and 5.5 t of crustaceans. Consumption equaled 11.2% of the reef fish community standing stock biomass, and was focused on early life history stages. Compared to the aquarium fish industry, the most important fishery targeting small reef fish in Hawaii, roi consumption was focused on a wider variety of fishes and on smaller individuals (mean weight 11.4 g versus 25 g). It removed fewer individuals (73% of industry take) of yellow tang (Zebrasoma flavescens), the main species collected by the industry, but removed 24 times more fish overall. The large scale of consumption suggested that roi may play a substantial role in shaping native reef fish communities.;Regarding (2), ciguatoxin concentrations in 18.2% of roi exceeded acceptable levels for human consumption. Toxicity was highly variable on all analyzed geographic scales, very weakly correlated with roi size, and not correlated with fitness and length-at-age of individuals. I concluded that a safe fishery for roi was not feasible in Hawaii due to poor predictability of toxicity. |
