Most often kites that were catching their prey by skimming close to the ground did not return to a perch but ate while they were flying. Associations of groups on posts at edges of fields, in trees or in flight were ephemeral as each bird seemed driven by a restless urge to be in motion. The kites generally gave the impression of catching their prey effortlessly and casually in the course of their flights. However, on July 20, 1961, one flying over a pond was seen to swoop three times in rapid succession at a dragonfly without catching it. The kite then flew higher, circled, and swooped three times more at the dragonfly, catching it on the last attempt. Most of the insects preyed upon are slower and less elusive than dragonflies, which are largely immune to the attacks of flying predators because of their great prowess in flight. Only on rare occasions could the kind of prey captured be observed in the field. Food habits were studied by collecting pellets of the kites at the Park, and analyzing them. The pellets were usually disgorged early in the morning while the kites were still on their night roosts in large cottonwoods. Often several kites roosted in the same tree. The pellets were of characteristic appearance, elliptical, approximately 15 millimeters in diameter, 30 millimeters long, pinkish or purplish, composed of insects' exoskeletons compacted, and comminuted to about the consistency they would have after passing through a meat grinder. A total of 205 pellets was collected--37 on August 20, 1960; 56 on July 18, 1961; 60 on August 4 and 5, 1961, and 52 on August 21 to 23, 1961. A total of 453 separate items was tentatively identified. Obviously the material was far from ideal for the identification of prey, which had to be reconstructed from minute fragments. The kites are dainty feeders and discard the larger and less digestible parts such as wings, legs, and heads. Often it was uncertain how many individuals or how many kinds of insects were represented in a pellet. Probably most pellets contained many individuals of the same species, but these were not separable. Hence, only 2.2 items per pellet were found, whereas Sutton found an average of 22.2 items in each of the 16 stomachs that he examined. Best information concerning kinds of prey utilized was obtained soon after the fledglings had left the nest; on various occasions these still clumsy young dropped nearly intact insects that were delivered to them by the adults. These insects, recovered from beneath the perches, were the basis for all specific and generic determinations; other material was determinable only to order or to family.