ABSTRACT
For the agricultural geographer working in tropical Africa no one scale is intrinsically more appropriate than another and both large- and small-scale investigations present difficulties. Ideally, each investigation in agricultural geography should involve four stages: the identification of a problem; the formulation of a hypothesis or hypotheses; the collection of the relevant data; and the testing and modification of the hypothesis to provide an adequate explanation. In many ways the problems of the statistician in tropical Africa resemble those of the agricultural geographer in that he must attempt to describe, with quite inadequate resources, the characteristics of large areas about which comparatively little is known. Nevertheless, these impressions will have to be checked by sample surveys, for which tropical Africa is well suited, with its relatively small variation in size of farm and traditional agricultural practices over large areas.
