visible near the shore to absorb the water of the river (στόμα τὸ καταπῖνον τὸ ῥεῦμα τοῦ ποταμοῦ), so it is plain that the water cannot maintain its separation and its sweetness, whereas the spring Arethusa is perfectly good to drink.” I have translated here the sense rather than the words of Strabo; but the phenomena of “rivers falling into chasms and being drunk up,” for a time, is exactly what happens in Greece. It did not appear to Strabo impossible that the Alpheius might traverse this great distance underground; nor do we wonder at this, when we learn that a more able geographer than he (Eratosthenês) supposed that the marshes of Rhinokolura, between the Mediterranean and the Red sea, were formed by the Euphrates and Tigris, which flowed underground for the length of 6000 stadia or furlongs (Strabo, xvi. p. 741: Seidel; Fragm. Eratosth. p. 194): compare the story about the Euphrates passing underground, and reappearing in Ethiopia as the river Nile (Pausan. ii. 5, 3). This disappearance and reappearance of rivers connected itself, in the minds of ancient physical philosophers, with the supposition of vast reservoirs of water in the interior of the earth, which were protruded upwards to the surface by some gaseous force (see Seneca, Nat. Quæst. vi. 8). Pomponius Mela mentions an idea of some writers, that the source of the Nile was to be found, not in our (οἰκουμένη) habitable section of the globe, but in the Antichthon, or southern continent, and that it flowed under the ocean to rise up in Ethiopia (Mela, i. 9, 55). These views of the ancients, evidently based upon the analogy of Grecian rivers, are well set forth by M. Letronne, in a paper on the situation of the Terrestrial Paradise, as represented by the Fathers of the Church; cited in A. von Humboldt, Examen Critique de l’Histoire de la Géographie, etc., vol. iii. pp. 118-130.