Gray whales Eschrichtius robustus exist as 2 geographically and genetically distinct populations in the eastern and western North Pacific. Subjected to intensive commercial whaling during the 19th and 20th centuries, the western population presently numbers approximately 100 individuals and is regarded as one of the most endangered baleen whale populations in the world. Since 1997, ongoing studies of western gray whales have resulted in a photographic dataset that can be used for mark-recapture survival estimation. Pollock's robust design was applied to 129 individual whale encounter histories spanning 25 monthly capture occasions from 1997 to 2003. Using Akaike's Information Criterion (AICc) model selection, models incorporating individual heterogeneity in residency patterns and higher temporary emigration probabilities for younger whales provided better fits to the data. Non-calf and calf (1st year post-weaning) survival were estimated as 0.951 (SE = 0.0135, 95% CI = 0.917 to 0.972) and 0.701 (SE = 0.0944, 95% CI = 0.492 to 0.850), respectively, averaging across the best models (n = 13) in order to account for model uncertainty. The non-calf survival point estimate is similar to mark-recapture estimates for Gulf of Maine humpback whales, but lower than an indirect estimate for the eastern gray whale population. Although no statistically robust direct estimates of baleen whale calf survival exist for comparison to the current study, the calf survival estimate is markedly lower than a value suggested for Gulf of Maine humpback whales. Estimation of survival is necessary for assessing the status of western gray whales, which can contribute to increased protection, conservation, and management planning for this critically endangered population.