In this paper the general characteristics are defined of the recent land-use—transport interaction models developed ultimately from the work of the Centre for Land Use and Built Form Studies and the Martin Centre. In honour of the Centre's anniversary, we may perhaps refer to the general form as the 'Martin Centre Model', whilst acknowledging that most of its refinement and application has been undertaken by private consultancies building upon the Centres tradition. The Martin Centre Model is characterised by a generalised approach integrat ing an input - output framework, a spatial distribution model, and a microeconomic approach to land prices and utilities of location. Numerous models corresponding to this description have been used since the mid-1970s, with many variations both of the spatial-economic models and of the associated transport models. The strengths of this approach include the scope for transport to influence the location of activities; the interactions between land-using activities themselves, and between them and the supply of land and/or buildings; and the ability to evaluate a wide range of alternative or complementary policies, in the land-use and transport fields. The weak nesses can be divided into two broad classes: first, the problems of calibrating a complex model which generates many variables for which observed data may be available; second, when the model is successfully calibrated, whether its assumptions and behaviour are reasonable, especially from the point of view of land-use and transport planners using it in practice. Problems of both kinds are particularly marked when one comes to consider the calibration of the model and its behaviour over time. A number of these problems are considered, and suggestions are made about some future areas of 'research into practice' that may overcome or avoid them.