We map behavioral characters related to mating system onto a phylogeny of the New World blackbirds (family Icteridae) in order to test hypotheses on the evolution of polygyny in Red-winged Blackbirds (Agelaius phoeniceus). The two hypotheses we test are in the sense that, unlike most polygyny hypotheses, they allow the an- cestral mating system to differ from the present one in characters other than female pref- erences for mated versus unmated males. In one model, polygyny evolves from the typical territorial monogamy system of most terrestrial passerines; in the second model, polygyny evolves from a system resembling that of monogamous species, with marsh breed- ing and without male territoriality. Both hypotheses assume that female-biased parental care coevolves with polygyny. Our reconstruction suggests that the closest non-polygynous an- cestor of Red-winged Blackbirds was characterized by monogamy, male territoriality, equal sharing of parental care between the sexes, and terrestrial breeding. Further, polygyny and female-biased care are suggested to have evolved on the same branch as marsh nesting. These results refute our second hypothesis in which polygyny evolves from Agelaius mo- nogamy, while providing provisional support for the first model in which polygyny evolves from territorial monogamy. Received 27 March 1998, accepted 17 June 1998. THE EVOLUTION OF POLYGYNY in birds has been a particular focus of mating-systems re- search. Hypotheses emerging from work on avian polygyny have been extended both to other taxa and to other mating systems. Polyg- yny is defined in the ornithological literature as a mating system in which one male forms long- lasting breeding associations with more than one female at a time (Wittenberger 1981). Po- lygyny hypotheses attempt to explain why a fe- male would mate with a male that is already mated. Existing polygyny models thus concen- trate to a large extent on the evolution of a sin- gle characteristic, female preference for mated versus unmated males, and assume that other characteristics form a fixed background against which this one trait evolves. We have argued in favor of extending such hypotheses to take a longer view of the evolution of polygyny, be- ginning with an ancestral mating system that may differ from the present one in many attri- butes in addition to female preferences (Searcy and Yasukawa 1989, 1995). We have proposed two such long-term models to explain the evo- lution of polygyny in Red-winged Blackbirds (Agelaius phoeniceus; Searcy and Yasukawa