The currently dominant neoclassical models of urban ground rents and housing distribution as well as recent alternatives along neoricardian or Ricardian-Marxian lines display major weaknesses which derive from the problems inherent in the concept of differential rents and its pure competition presuppositions. The paper argues that an effective alternative can be developed on the basis of Chamberlin's theory of monopolistic competition. The formal structure of such a model simplifies greatly the matter of the determination of land prices and incorporates with ease difficult aspects such as the role of public controls. In addition, it offers a much needed theoretically valid integration of land price theory with operationally powerful Lowry-type models. The most interesting implication, however, of a monopolistic competition model is that the venerated 'law' of differential rents must be thoroughly rejected on both theoretical and practical grounds.