EVER since geography was re-established as an organised discipline, the essence of which is the study of terrestrial distributions and their inter-relations, geographers have been sifting and collating data of extremely varied character. The facts which have thus been incorporated in the body of geographical literature have them selves usually been established by workers in other fields, while geographers have drawn deductions from them, in many cases without having the opportunity to test their validity on the ground. As a result, generalisation and causation in regard to very large sections of the continents must necessarily rest on a rather insecure foundation. The question therefore arises—how can this be remedied?