To assess long-term temporal trends in White Shark (Carcharadon carcharias) predation, and examine the relationship between C. carcharias and pinnipeds, we examined a 17-year (1988–2004) data set of daily shark attack surveys from Southeast Farallon Island (SEFI), California. We modeled within-season and among-year variation in the number of observed shark attacks at SEFI. Within-season, daily probability of attack was affected by hours of effort, date, tide height, and pinniped abundance. The seasonal peak in shark predation did not vary inter-annually. Daily probability of attack was positively correlated with pinniped abundance and negatively correlated with tide height. After controlling for within-season effects, 51% of annual variation in the number of observed shark attacks was explained by an inverse function of the abundance of Northern Elephant Seals (Mirounga angustirostris). The addition of shark abundance to the model did not significantly increase the model's descriptive power. This suggests a functional response where predation by C. carcharias increased with abundance of M. angustirostris, but plateaued once a critical density of M. angustirostris were available, either due to limitations of prey handling or satiation. We detected no temporal trend in annual number of observed shark attacks. Therefore, C. carcharias depends on M. angustirostris as a critical food source and will increase or decrease predation rates depending upon current seal populations.