Within the past few years thirteen U.S. cities have considered construction of rapid transit systems while three cities, Atlanta, Los Angeles, and Seattle, have conducted referenda to obtain voter approval for the construction of a rapid transit system) In Atlanta the referendum was rejected in 1968 but a somewhat different plan was approved by a slim majority in 1971. This paper analyzes, within the framework of the rational voter model, the outcomes of the two public transit referenda held in Atlanta. The objectives of the paper are to provide a test of the rational voter hypothesis and to provide insights into voter preferences for public mass transit.Previous tests of the rational voter model have analyzed results of referenda on public works projects, Dildine (1972) and Frey and Kohn (1972); votes on tax or finance issues, Smith (1975); and voter turnout, Silberman and Durden (1975), Settle andAbrams (1976), and Barzel and Silberberg (1973).Although several referenda on mass transit issues have been held, there has been little economic analysis of the voting patterns in these referenda.