Abstract
Evidence of direct mortality to fishes exposed to oil is very limited. The 1989 Exxon Valdez oil spill had major impacts on marine mammals and seabirds, but was not implicated in the 1993 collapse of the Prince William Sound herring population because of the four year gap. However, we use several independent evidences, including changes in the predation behavior of Steller sea lions, to show that the collapse actually began in 1989. We show the failure to detect the actual timing of the collapse was due to deficiencies in the fishery model used to assess herring population abundance and lack of understanding about the vulnerability of herring. Finally, we show that the oil spill actually had greater impacts than originally believed as a result of catastrophic impacts on the ecosystem from the herring collapse