Urban consolidation is a major issue on the agenda of Australian cities. The significance of this is explored in the three papers in this collection. Richard Cardew reviews the papers by Troy and Bunker, introducing some additional considerations with particular reference to Sydney. He sees urban consolidation as an ongoing process which will continue with metropolitan growth and rising land values. However, he argues, the flats boom and the massive increase in household formation - shown in trends in headship ratios - which occurred in the 1960s and 1970s are unlikely to reoccur. The era of most rapid consolidation is past. Pat Troy offers the most critical view of present urban consolidation policies. He questions many of the benefits claimed for consolidation, in particular the assumption that it will lower requirements for public sector investment in infrastructure through more efficient use of area services such as schools and hospitals, and network services such as water, sewerage, power, transportation and communication. He argues that these and other assumed benefits are based on demographic trends unlikely to be realised, and on infrastructure savings which are illusory. Troy is especially critical of the claim that higher urban densities will lower the cost of housing, pointing out that multi-unit housing tends to be at the higher end of the market. Troy concludes his paper with a programmatic call for an increase in the supply of dwellings and a set of recommendations for achieving this. Ray Bunker's paper reviews the history of urban consolidation as part of metropolitan planning over the last ten years. Like Troy, he questions many of the assumptions invested in urban consolidation policies, and argues that while a degree of consolidation is occurring, it is but one means invoked to serve a number of ends, and the pursuit of those ends themselves involves other instruments, some of which may be more effective. Further, consolidation needs to be gradual, locally differentiated and responsive, and these local dimensions need to be expressed more poweifully.