||Following the Rogers Report of 1999 and PPG3 of 2000, the density of new development in English cities was substantially increased. The overt motive for his was that fuel use was to be reduced. The main evidence for the view that density increase would reduce fuel use per head was a graph published by Newman and Kenworthy which showed that fuel use per head appeared to be higher in lower density cities. A review of the literature published over the last twelve years casts considerable doubt on this conclusion. Firstly, the evidence on which Newman & Kenworthy based their graph is severely criticised. It is argued that the price of fuel is far more important than residential density in determining fuel use, that even substantial increases in density would reduce fuel use very little, and that , most importantly, density can be changed only very slowly, whilst many would argue that reductions in fuel use and CO2 emissions are urgent. Seceondly, there is evidence to show that at very high densities fuel use is higher rather than lower as people seek to travel out of the city in order to get away to less dense environments. This is compatible with the large body of evidence that people prefer lower densities to higher densities. Thirdly, there is a conflict with a policy of containment such as exists in Britain. If high density housing is built separate from other destinations for work or pleasure then people will travel across the intervening space to get to those destinations however high the residential density may be at the origin or the destination. Given British green belt policy with its high level of commuting across the greenbelts, mainly by car, it is evident that the policy of containment increases fuel use and CO2 emissions. Any policy which took seriously the need for reductions in fuel use would allow for urban expansion on the edge of existing cities. It can be argued that the fact that such expansion is not promoted demonstrates that the need to reduce fuel use is simply being used as a justification for existing policies, not to design new ones.