An Empirical Study of the Areal Distribution of Houses in Puerto Rico

Abstract

A FOCUS for the geography of settlement was recently identified by K. H. Stone (I965): 'The suggested core of settlement geography is the buildings-where are they and why are they there?' (p. 347). 'The focus on buildings is to provide for this part of geography a clearly definable core which is an expression of man-land relations presently inadequately studied and known' (p. 355). This study contributes to answering the 'where' question for the map distributions of houses in twenty-one rural areas of Puerto Rico. An important finding is that pertinent properties of these map distributions are adequately summarized by the negative binomial frequency function. This regularity indicates that the distributions are characterized by clusters or groups of houses and suggests the presence of a pervasive influence underlying the locations of houses. Though several models are known to generate areal distributions described by the negative binomial function, it is shown that these models are not appropriate for the rural settlement distributions in the twenty-one sample areas. Because an alternative model is not formulated, this study does not answer the 'why' question. However, empirical regularities are identified that may contribute to the formulation of an appropriate model.