The oil shortages and price rises of the 1970s led to a broad spectrum of policies designed to save fuel in transport. One such group of policies concentrated on utilising the longstanding observation that congestion-free driving improves vehicular fuel economy. In particular, improving the average speed of urban traffic was believed to reduce urban fuel use. Strategies to create free-flowing traffic networks through traffic engineering and management techniques became an integral part of most transport energy conservation strategies. At the same time urban systems studies were appearing which suggested that urban land use factors are fundamental in determining transport energy demand. Some cities with denser, more centralised developnent patterns and congested traffic systems appeared to have lower per capita transport energy use than cities with low density, dispersed land use and free-flowing traffic networks. Conflicts thus began to be recognised between these land use studies and the traffic studies but they were never systematically examined. This thesis focuses on explaining and attempting to resolve these apparent conflicts by drawing out the relative importance of freeflowing, fuel efficient traffic and broader land use/transport factors in determining urban system fuel efficiency. It does this by showing how a holistic understanding of transport energy use can be achieved by recognising the different levels of know1edge involved and by integrating research a t all these levels. Adopting this approach, a detailed examination is made of transport energy use in Perth, Western Australia. Spatial and temporal patterns of urban driving and vehicular fuel consumption are derived using a computer-instrumented vehicle within a framework of urban ecology. These patterns are described in detail and are found to be understandable in terms of a simple model of intersection and vehicle-based traffic events. Using this simple model, urban driving and fuel consumption patterns are then systematically linked back to their urban ecological framework. Centrality, as expressed by urban activity intensity emerges as the major underlying factor in both the urban ecological framework and driving patterns. Driving cycles for central to outer areas as well as the whole city are produced and a detailed examination is made of driving cycle development procedures, including the issue of data resolution. Two new methods of driving cycle synthesis are formulated. Driving cycles together with travel survey and land use/transport model data are then used to estimate spatial differences in transport energy use per capita within the urban ecological framework. This case study is used to test two conceptual models of transport energy conservation implied by the traffic and land use studies: a linear assumptions model which suggests reductions in urban system fuel use through freer-flowing traffic and a feedback assumptions model which introduces some possible broader system effects of free-flowing and congested traffic. The results suggest very strong support for the feedback model and a reversal of the linear model. Urban structural factors such as distance from the CBD and density are found to be the major explanatory factors of Perth's transport energy use. The much higher vehicular fuel consumption in Perth's inner areas due to restricted traffic flows are completely offset by energy conserving land use and transport features which result in lower per capita energy use in these areas; the lower vehicular fuel consumption in uncongested outer areas is swamped by land use factors which result in longer travel distances and less use of energy conserving modes, resulting in markedly higher per capita energy use, i .e. a trade-off is found between a fuel efficient traffic system and a fuel efficient urban system. It is concluded that policies which propose fuel savings from blanket improvements in urban trraffic flow, such as major road building programmes will be counter-productive, while urban planning policies which encourage higher density, more compact cities will tend to lower energy use. It is suggested that a key area of research is now how to combine selective traffic system improvements with urban land use policies so that both may work in concert to produce a more fuel efficient urban system.