W A TRITING in the political science profession's most prestigious journal, Jeffrey Pressman has observed that effective political leadership in a fragmented system is proclaimed a necessity in American cities. And political scientists have proclaimed a model of the brokerentrepreneur who can fill that need.' Indeed, the political science profession's thinking on the subject of mayoral leadership has been greatly influenced by the classic writings of Edward Banfield2 and Robert Dahl.3 Both Banfield's and Dahl's can be viewed as variants of a larger model emphasizing the virtues of near-consensual leadership.4 The good mayor, according to the Banfield/Dahl model, protects his precarious position in a system of dispersed power by acting either to build or to maintain a broad-based, overarching coalition within the city; he will take little, if any, action unacceptable to any of the major elements in this coalition. The Banfield/Dahl norm of consensus-oriented leadership is reflected in the writings of other mayoral observers such as Robert Salisbury,5 Raymond Wolfinger,6 James Cunningham,7 Alexander George,8 Anne Greer,9 and Andrew McFarland.10 Only recently has the normative bias of this model been challenged. Though both the broker and the entrepreneur are consensus-oriented mayors, only the latter possesses strong goal orientations. The broker, for the most part, is a reactive mayor. He protects his fragile power position by acting to implement only those policies which have first gained a seal of approval12 from all the community groups directly concerned with an issue. He will not advance programs of his own for to do so would only lead him to intervene in power costly situations: Instead he waits for the community to agree upon a project. When agreement is reached, or when the process of