We developed an estimator for the probability of sighting sea otters (Enhydra lutris) by shorebased counters, based on simultaneous double-surveys. We then estimated probability of sighting sea otters in California and evaluated the estimator's principal assumptions. The overall probability of sighting sea otteis on 5 replicated double-surveys at each of 6 study areas was 0.945. Estimated probability of sighting did not vary (P > 0.05) among study areas or over time and was not correlated (P > 0.05) with distance of otters from observers. Probability of sighting was affected (P 3. Observer teams may have varied slightly in sighting ability, but this variation did not substantially affect population estimates. J. WILDL. MANAGE. 52(1):70-76 The sea otter population in California ranges between Santa Cruz (36056'N) and Shell Beach (35009'N) (U.S. Fish and Wildl. Serv. 1984). The population probably contained <100 individuals in the early 1900's (Ralls et al. 1983), but due to legal protection, it gradually increased in size and range (U.S. Fish and Wildl. Serv. 1984). Few survey data are available from before the mid-1960's, and various methods were used through the 1970's, making trends in the population difficult to assess. In 1982, the California Department of Fish and Game and the U.S. Fish and Wildlife Service began a collaborative program to use a single method to survey otters, thereby providing a better means of assessing future population trends. Currently, sea otters are counted from shore throughout approximately 70% of their range in California, with the remaining area counted from aircraft. Estimates of otter population size are important to evaluate past survey data, interpret consequences of incidental mortality from net entanglement, and determine how many animals might be removed to establish another colony in California. We first developed an estimator for the probability of sighting sea otters by shorebased observers. We then estimated this probability in a field study, and evaluated its consistency over space and time. Finally, we ex mined the estimator's principal assumption, that all animals in the population are equally sightable, by analyzing the effects of activity, group size, and distance from observer on the probability of sighting. There are few reports This content downloaded from 157.55.39.27 on Sat, 26 Nov 2016 04:16:33 UTC All use subject to http://about.jstor.org/terms J. Wildl. Manage. 52(1):1988 PROBABILITY OF SIGHTING * Estes and Jameson 71 of double-survey methods in the literature on wildlife population assessment. We are aware only of Caughley's (1974) use of double-surveys to estimate bias in aerial counts of African elephants (Loxodonta africana). Although our paper deals specifically with sea otters in California, the problems, methods, and analyses have broader application in wildlife ecology. We thank J. A. Ames, J. L. Bodkin, E. R. Farout, R. A. Hardy, B. B. Hatfield, M. C. Kenner, M. L. Riedman, J. E. Vandevere, and F. E. Wendell for assisting with the field counts. The manuscript benefitted from comments by R. G. Anthony, D. P. Demaster, J. R. Gilbert, D. B. Siniff, and 3 anonymous referees. This work was supported by the California Department of Fish and Game and the U.S. Fish and Wildlife Ser-