A contradiction permeates all studies of settlement forms. On the one hand is the theoretical impossibility of creating a viable classification of an apparently infinite variety of forms when there is evidently an absence of any direct connection, in either time or space, between most of them. Identical or similar forms can be generated in wholly different ways and at diverse times. On the other hand, it is also evident from the most superficial observations that settlement forms have close and complex relationships with human culture, reflecting lifestyles and aspirations. While there is no doubt that settlement forms can and should be studied as part of the societies generating themand in this it is interesting to note how much important work concerning them derives from the investigations of anthropologists-there are important questions about their role as cultural indicators when their characteristics are plotted on distribution maps. While the interpretations which identified distinctive settlement forms with racial groups have now been largely abandoned, we are, nevertheless, left with a powerful dilemma. The observable spatial variations in the characteristics of settlement plans continue to beg many questions. These can be seen both in landscapes and on maps, at continental scales (Figures 7.4, 7.5 and 7.6) as well as quite local scales. To use these rich data, classification is an essential step, because in order to talk about rural settlement forms at all there must be a framework for description and analysis, comparison and generalisation. Nevertheless, classification must, above all, not be an end in itself but a tool to aid enquiry and rigorous investigation. In this, the morphological approach can be used as a powerful research tool, complementing, amplifying and extending the limited written record. The volume of evidence is very great. As with the architecture of individual buildings, careful structural analysis of particular cases coupled with careful comparative studies leads to broader hypotheses and questions.