The proliferation and apparent intransience of urban squatter settlements in Third World countries have produced numerous explanations of their cause, nature and solution. In Fiji, it is estimated that nearly one-fifth of the population of the Suva Urban Area were squatters in 1976 and that squatter numbers, in recent years, have increased by over 10 percent annually. These increases have occurred despite development plan objectives which aim to reduce regional and class disparities and a public housing programme which has housed about one-fifth of the urban area's population since the late 1950s. In the early 1960s many scholars supported the view that squatters were 'marginal' to the life of the city and this view is still held by many administrators, in Fiji and elsewhere. Studies commencing in the mid 1960s, however, have led to a revision of opinions on squatter marginality and the emergence of two major schools of thought among scholars. The Dependency School considers squatting a product of the type of urbanization experienced by Third World countries and sees no solution short of major changes to the 'system'; the Progressive Development (or Progressive Housing) School, on the other hand, sees squatting as a housing problem caused by the rate of Third World urbanization and inappropriate public housing programmes. The latter School advocates the encouragement of self-help activities among the poor to close the gap between housing supply and demand. The public housing programme in Fiji has been influenced by such arguments and during recent years there has been some official support for site and service, squatter upgrading and core housing schemes. The present work proceeds from a statement of major theoretical questions derived from the literature, and a description of urbanization and the housing situation in Fiji, to test the assumptions of the Marginality, Dependency and Progressive Development Schools. Data were obtained on urbanization and housing in Fiji, and sample surveys were conducted among Fijian and Indian squatters and other low-income households in Suva City and Urban Area in 1976. The data were used to consider six propositions which were tested via 36 hypotheses; comparisons were made between ethnic, squatter and other low-income, and city and urban area households. The propositions were found to be substantially correct. Squatter behaviour was shown to be as modern and as urban as the behaviour of other low-income people, and households with 'modern' attributes tended to be more 'successful' in the city. Modern and traditional behaviour, however, were not found to be antithetical and, at least among Fijians, the households which were the most traditional tended to be those which were most 'successful.' Squatters were also shown to make a positive contribution to the economy of the city and to be no more an obstacle to rational urban land use than many other land users. The views of the Marginality School were therefore found wanting. Squatters were shown to participate in self-help activities which led to increasing adequacy in housing. The squatter residential environment was found to permit greater all round flexibility than official public housing and was therefore considered more appropriate for poor households which relied heavily on the use of the residential environment for supplementary informal economic activities and kinship networks. In particular, squatting was found to allow Fijians and Indians to adopt strategies of survival and improvement based on their respective cultural mores. Some support was therefore provided for the views of the Progressive Development School. The levels of improvement, however, were found to be generally insufficient and to involve too few households to accept these views without reservation. Improvements occurred among more modern households, among owner-occupiers and those with some security uf tenure as claimed by the Progressive Development School, but income was found to be the most critical variable. Income increased among Fijians by accretions of kin and among Indians by accretions of time, but few households had sufficient income to permit them to alter significantly their position in society or to ensure minimally adequate housing. The views of the Dependency School that squatting is a residential manifestation of poverty, and not a housing problem as such, were therefore accepted. In accepting this explanation of the causes and nature of squatting, however, the writer maintained that practical measures advanced by the Progressive Development School could lead, if not to the solution of the problem, at least to its improvement. At the macro level, the solution to the urban squatter question in Fiji was seen to lie in the implementation of the egalitarian proposals of the VIth Development Plan, a change in Fiji's relations with other countries, and a reassessment of the position of the urban poor. At the micro level, much improvement can be achieved by the formulation of a housing policy which gives close attention to the needs and aspirations of squatters indicated in the present study. The distinction between macro and micro level 'solutions' is seen to offer a compromise position between the viewpoints of the two most prominent schools of thought on the urban squatter question.