Recent surveys of recovery plans indicate that criteria, such as population sizes, for delisting species from the U.S. Endangered Species Act (ESA) are often unrealistically low by scientific standards. We describe the delisting criterion for the threatened southern sea otter (Enhydra lutris nereis) developed by the Southern Sea Otter Recovery Team. A major oil spill is the most serious threat to this sea otter population. After extensive modeling of oil spills, the recovery team concluded that it was not scientifically defensible to develop a delisting criterion in terms of a single probability of extinction over a specified time period. Instead, the team decided to define a size at which it would consider the population endangered and to consider the population threatened as long as a major oil spill might reduce it to that size. The effective population size (N_e) for endangered status was set at 500, estimated to be about 1850 otters. Using a spill the size of the Exxon Valdez spill (250,000 bbl), the oil spill model was iterated to generate a frequency distribution of the number of sea otters contacted by oil, from which the team estimated that less than 800 otters would be killed by 90% of the simulated spills. Thus, the delisting criterion was set at 1850 + 800 = 2650 individuals. There have been several proposals to improve the Endangered Species Act by providing quantitative guidance, in the form of specific probabilities of extinction within some time frame or specific criteria like those used by the World Conservation Union as to the levels of extinction risk represented by the terms “threatened” and “endangered.” Experiences of the Sea Otter Recovery Team indicate that guidelines should not be overly rigid and should allow flexibility for dealing with specific situations. The most important consideration is to appoint a recovery team that is both technically well qualified and unconstrained by pressures from management agencies.