A 1960—71 study of populations of color—banded ant—following antbirds of three species on a tropical—forested lowland reserve, Barro Colorado Island, showed that the small species (Spotted Antbird, Hylophylax naevioides) remained stable at about 20 pairs/km2. A medium—sized species, the Bicolored Antbird (Gymnopithys bicolor), decreased from about 3 pairs to 1.5 pairs/km2. A large species, the Ocellated Antbird (Phaenostictus mcleannani), declined from 1.5 pairs/km2 to near extinction–only one female remained in early 1971. Two of three other species that regularly follow army ants showed relatively stable populations, but a third large species (Barred Woodcreeper, Dendrocolaptes certhia) declined from two pairs to local extinction. Prior to 1960 a very large ground—cuckoo that follows ants had already become extinct there. Thus, the three largest of the seven original species that regularly followed ants were gone or nearly gone by 1970. The decrease in numbers of regular ant—following birds was not made up by increases in occasional followers. Detailed studies of antbirds showed no clear reasons for declines, except that annual mortalities of adults were high in Ocellated Antbirds (about 30%) compared to Spotted Antbirds (15%—17%) and nest losses perhaps higher in the former (96% compared to 91%). Nest mortalities were slightly lower (88%) and adult mortalities intermediate (about 25%) in Bicolored Antbirds. Female Ocellated Antbirds had higher mortalities than males. The antbirds renest repeatedly during long nesting seasons, up to 14 times per year for Ocellated Antbirds. However, to replace females of this species under Barro Colorado conditions 19 nestings per year would be needed. Concurrent listing of all birds of the island showed that 45 species of breeding birds, 22% of the avifauna present when the island was made a reserve, had disappeared by 1970. No new species replaced them. Of the lost species 13 are forest birds, in danger if forests are cut elsewhere. The other species, second—growth and forest—edge birds, have been crowded out by growth of the forest. Loss of species from this tropical reserve, especially the part apparently caused by the small size and isolation of the reserve, poses problems for conservation and ecological studies of tropical biotas. It is suggested that large future reserves have corridor zones to each other, that is, that intensive human use not preempt too much area nor interrupt immigration of animals or plants from one refuge to another.