Throughout its range, the peregrine falcon (Falco peregrinus) preys almost exclusively on birds, but also captures bats and terrestrial mammals, and occasionally items such as reptiles (Ratcliffe 1980; Cade 1982) and fish (White and Roseneau 1970). Beebe (1960) described an instance of carrionfeeding and suggested this behavior was not rare. However, with the exception of brief comments by White and Roseneau (1970) and Ratcliffe (1980), and a recent account by Holland (1989), I can find no other published records of this type of feeding. Here I describe two accounts of carrionfeeding and discuss possible causal factors. At 1024 hours (PST) on 17 January 1989, I encountered a peregrine falcon (later identified from photographs as an immature female of the subspecies pealei) hunting a flock of dunlins (Calidris alpina) at Grayland Beach on the outer coast of Washington state. I watched the falcon continuously. By 1053 hours it had hunted four times and made 19 attempts to capture prey. The flights each covered distances of 400-1000 m along the beach. After the fourth flight the falcon landed on the beach, and at 1058 hours stepped onto the intact carcass of a common murre (Uria aalge) that lay amidst the tidal debris. I had noticed this carcass over two hr earlier. The falcon began plucking feathers within five sec and was pulling flesh from the carcass at 1104 hours. It ate steadily until 1149 hours. It immediately wiped its bill and then preened until 1154 hours when it flew 90 m south along the beach. Before I left the area at 1208 hours, I checked the carcass and found that the pectoralis muscles had been stripped. The murre appeared to have washed onto the beach the day before. The second observation of carrion-feeding, on 13 January 1990, also was made at Grayland Beach. At 1333 hours (PST) I saw an immature female F. p. anatum flying low above the beach. It landed on the beach and stepped onto the intact carcass of a white-winged scoter (Melanitta fusca). It began plucking feathers at 1337 hours and by 1348 hours had stripped the pectoralis muscles from the left side. It then paused briefly before plucking feathers from the other side of the carcass. The falcon flew from the carcass when a vehicle passed nearby at 1356 hours. I used the opportunity to inspect the carcass and found no lacerations or hemorrhaging on the back that might be associated with predation by a falcon. The falcon returned to the carcass and resumed feeding after I departed. Scavenging is thought to occur when hunting conditions are poor (Ratcliffe 1980). On the day of the first observation a very high tide (2.8 m at 0758 hours) had flooded the mudflats and saltmarsh of Grays Harbor, forcing thousands of dunlins to move to the outer beach where they roosted and foraged. Peregrine falcons and merlins (F. columbarius) follow these movements to outer beaches (Buchanan, unpubl. data). The movement of dunlins to the outer beach on this date probably occurred > three hr before high tide. Dunlins had not yet begun movement back to the estuary when I left the beach at 1208 hours (i.e., dunlins roosted at the outer beach for > six hr). Success rates of peregrine falcons hunting dunlins are significantly lower at beaches than at adjacent estuaries in Washington (Buchanan, unpubl. data). Consequently, falcons that follow the flight of dunlins from the estuary to outer beach roost areas have options to 1) wait (several hours) for the dunlins to return to the estuary, 2) hunt in an area where they are less efficient predators, or 3) seek other prey. Factors contributing to the decision to scavenge were likely cumulative over a period beginning long before the hunting flights I observed. Immature falcons are generally less efficient hunters than adults because of their inexperience (Cade 1982). The falcons observed scavenging by Beebe (1960) and Holland (1989) were in immature plumage, as were both falcons discussed here. Although I did not observe the second falcon hunting, conditions that day were similar to those of the first (2.8 m high tide, extended shorebird roosting period on the outer beaches). I suspect the scavenging events resulted from sub-optimal conditions and ineffective hunting of a very small prey item.