Caching behavior by the American kestrel [Falco sparverius] is described. In 1972-73, 25 instances of prey caching were recorded; 10 (40.0%) involved food storage and 9 (36.0%) involved food retrieval. Six (24.0%) unsuccessful attempts to retrieve prey were observed in 1972-73. In 1973-74, 65 instances of prey caching were recorded, of which 36 (55.4%) involved storage and 23 (35.4%) involved retrieval. Six (9.2%) unsuccessful attempts to retrieve prey were observed in 1973-74. The storage of a single prey item and its subsequent retrieval occurred twice in 1972-73, and 12 times in 1973-74. Small mammals, birds, snakes, and frogs were the prey items cached. In 1973-73, 15 (88.2%) of the prey cached were small mammals, and in 1973-74, 30 (68.2%) were snakes and frogs. Prey was cached predominantly in grass clumps (87.5%), but other sites occasionally were used; these included, short broken posts (4.7%), fence posts (4.7%), and shrubs (3.1%). Kestrels stored prey throughout the day. Between 8 and 17% of all storing observations occurred during each hourly interval from 09:00 to 16:00. Most attempts at retrieval (25, 78.1%) occurred after 15:00. Kestrels continue hunting immediately after storing prey, regardless of feeding behavior beforehand. They stockpiled surplus prey items in caches, and retrieved several stored prey in succession.