Ballast water transport of nonindigenous species (NIS) is recognized as a significant contributor to biological invasions and a threat to coastal ecosystems. Recently, the use of ozone as an oxidant to eliminate NIS from ballast while ships are in transit has been considered. We determined the toxicity of ozone in artificial seawater (ASW) for five species of marine organisms in short-term (< or = 5 h) batch exposures. Larval topsmelt (Atherinops affinis) and juvenile sheepshead minnows (Cyprinodon variegatus) were the most sensitive to oxidant exposure, and the mysid shrimp (Americamysis bahia) was the most sensitive invertebrate. Conversely, benthic amphipods (Leptocheirus plumulosus and Rhepoxinius abronius) were the least sensitive of all species tested. Mortality from ozone exposure occurred quickly, with median lethal times ranging from 1 to 3 h for the most sensitive species, although additional mortality was observed 1 to 2 d following ozone exposure. Because ozone does not persist in seawater, toxicity likely resulted from bromide ion oxidation to bromine species (HOBr and OBr-), which persist as residual toxicants after at least 2 d of storage. Total residual oxidant (TRO; as Br2) formation resulting from ozone treatment was measured in ASW and four site-specific natural seawaters. The rate of TRO formation correlated with salinity, but dissolved organic carbon and total dissolved nitrogen did not affect TRO concentrations. Acute toxicity tests with each water over 48 h using mysid shrimp, topsmelt, and sheepshead minnows yielded results similar to those of batch exposure. Addition of sodium thiosulfate (Na2S2O3) to ozonated waters resulted in TRO elimination and survival of all organisms. Our results provide necessary information for the optimization of an efficacious ozone ballast water treatment system.