Birds of several species have been reported to kill immature birds of other species in contexts other than predation. Contexts include: (1) takeover of specialized nesting sites such as cavities (e.g., Kendeigh 1941, Pinkowski 1977), (2) brood parasitism (reviewed by Payne 1977), and (3) elimination of potential competitors for food (suggested by Picman 1977, 1980 for Long-billed Marsh Wrens [Cistothorus palustris]; Belles-Isles and Picman 1986 for temperate House Wrens [Troglodytes aedon]). The latter context is problematical because it cannot be inferred without evidence of actual or potential food limitation. In this note I document interspecific infanticide of tropical House Wrens by Rufousand-White Wrens (Thryothorus rufalbus). The infanticide appears to be based on competition for food since nestlings were killed only during an 8-week food shortage within a 4-year period. The infanticide was detected during a study of House Wrens conducted between 1982 and 1985 near the rural village of La Laguna, Republic of Panama (approximately 25 km northeast of Panama City). Six attacks occurred between 23 May and 14 July 1983; three were observed and three have been inferred (Table 1). On 23 and 24 May, a Rufous-and-White Wren was perched at the entrance of a nest box containing three House Wren nestlings (3 days old on 23 May). The box was mounted on a 1.25 cm diameter pole at a height of 1.6 m. The intruder thrust its head and beak rapidly and repeatedly into the nest and threw out one nestling on the first day. This nestling, as well as one of the remaining nestlings, had lacerations on the head and abdomen and puncture wounds on the abdomen. Thereafter one nestling was missing each day for two days. Parental House Wrens returning to the nest were driven away by the intruder during both days of observation. In a second case, on 9 July, an observer frightened a Rufous-and-White Wren from its perched position at the entrance of the nest box shortly after it commenced thrusting its head repeatedly inside the box. This nest contained six nestlings 10 days old. No mortality or injury resulted from this interrupted attack, and there was no evidence of repeated visits (i.e., no nestling loss or injury) for 3 days, after which a snake ate the nestlings. This second nest was located 3 km from the first, suggesting that two different Rufous-and-White Wrens were responsible for these incidents. In a third case, on 19 June, a Rufous-andWhite Wren repeatedly pecked a House Wren fledgling for 15 sec in a family group of recent fledglings and parents. The fledgling escaped and was observed 5 days later. This incident took place approximately 100 m from the first incident. The House Wren breeding season extended through October of 1983 (approximate length 24 weeks), but no further direct interactions between the two wrens were observed. Three additional nest losses (one partial) between late May and mid-July of 1983 may have been caused by Rufous-and-White Wrens because of similar patterns of lacerations and puncture wounds on the bodies of dead or dying nestlings (Table 1). One nestling had a severely broken tibiotarsus as well. Rufous-and-White Wrens were seen or heard within 25 m of these nests on the day of detection of infanticide. The nests were located at least 1 km apart and between 200 to 1,300 m from the nests mentioned above, suggesting that the Rufous-and-White Wrens seen in the vicinity of these nests were different individuals. Some alternative causes of the infanticide can be eliminated. Predators were probably not responsible for the killings because the dead or dying nestlings were not eaten. Infanticide by other House Wrens, associated with takeovers of mates or territories during each