Patterns of rural settlement in Sierra Leone: methods of geographical analysis in a tropical environment

Abstract

The debate between classical empiricists and those who favour deductive and theoretical research methods is a feature of modern geography. This thesis aims to show that these approaches are complementary rather than mutually exclusive. It also hopes to demonstrate new methods and applications of location analysis in an underdeveloped area where base line information is uneven in quality. The dissertation is divided into four main parts. In the first, a model settlement pattern for Sierra Leone is devised. This model is based on a stylised subsistence village and the arguments of central place theory. The second part deals with the uses made of aerial photographs, topographical maps, pilot surveys and random sampling procedures in constructing an accurate base map of rural settlement distribution for the whole country, the first of its kind for any West African state. The third part of the work uses purely qualitative and empirical methods, A system of settlement regions is devised and described, and the settlement model is compared with the actual pattern, and with overall changes in settlement structures between 1927 and 1964. An account of rural settlement evolution using historical sources and comparative mapping is also presented. In the fourth part, a range of parametric and nonparametric tests and techniques of location analysis (e. g. set theory and nearest-neighbour analysis) is used to establish indices of settlement density and nucleation and to test the hypotheses presented in earlier sections of the work.