In general, the mathematical and astronomical treatises were earlier known to the West than the purely metaphysical works: this was the case in the eleventh and twelfth centuries; in the thirteenth the schoolmen were familiar with the whole body of Aristotle’s works. Thus the influence of Aristotle on natural science was early important, either through Arabian commentators or paraphrasers, or through translations made from the Arabic, or directly from the Greek.[355] Jourdain affirms that it was the influence of Aristotle and his interpreters that kept alive in the Middle Ages the doctrine that India and Spain were not far apart. He also maintains that the doctrine of the sphericity of the earth was familiar throughout the Middle Age, and, if anything, more of a favorite than the other view. The field of the later ecclesiastical and scholastic writers, who kept up the contentions over the form of the earth and kindred subjects, is too large to be here minutely surveyed. Such of them as were well known to the geographical students of the centuries next preceding Columbus have been briefly indicated in another place;[356] and if not completely, yet with helpful outlining, the whole subject of the mediæval cosmology has been studied by not a few of the geographical and cartographical students of later days.[357] So far as these studies pertain to the theory of a Lost Atlantis and the fabulous islands of the Atlantic Ocean, they will be particularly illustrated in the notes which follow this Essay. [Illustration: Wm. H. Tillinghast] NOTES.