We studied the demography of an insular California Spotted Owl (Strix occidentalis occidentalis) population in southern California for 12 years. We used model selection based on information theory to examine the relationship between weather and reproduction and survival. Mean annual fecundity was 0.139 (SE = 0.050) for subadult females and 0.345 (SE = 0.028) for adult females. Adult females had higher fecundity than subadult females during all years, and fecundity in both age classes was higher when a wet year preceded a dry spring (i.e. breeding season). A model incorporating these factors explained 100% of the estimated temporal process variation in fecundity. Mean apparent survival was 0.796 (SE = 0.012), 0.880 (SE = 0.041), 0.692 (SE = 0.062), and 0.368 (SE = 0.038) for adult, second-year subadult, first-year subadult, and juvenile (first-year) owls, respectively. We found no temporal process variation in survival. Using a Leslie projection matrix, we estimated the finite rate of population change to be 0.906 (SE = 0.018) over the entire period of study (1987-1998), which indicated that the population declined ∼9% per year during the study. That rate of decline was higher than a rate (λ1991-1998 = 0.921, SE = 0.020) we estimated for a shorter period (1991-1998) that matched the time interval used in a recent meta-analysis of Spotted Owl population dynamics. We believe that both the present estimates and those of the meta-analysis are valid, given their respective goals. The study population was characterized by relatively high, constant survival of territorial birds, low and variable annual reproduction, and relatively low juvenile survival. Because weather was strongly correlated with reproduction, fecundity rates for the species may decline during short-term droughts and when storms occur during the breeding season. Weather extremes may not, however, be sufficient to affect temporal variation in survival of Spotted Owls in this part of their range.