the degree. With this distance, and with his own observed arc, he made such a computation of the earth's circumference as gave him 500 stadia to the degree. But the absurdity of thus employing the 700-stadium degree as an element in the computation, by which he obtained the 500-stadium degree, did not prevent even Ptolemy from adopting the latter estimate. A degree of 500 stadia at the equator gave him one of 400 stadia on the parallel of Rhodes. On this basis, reckoning from the Fortunate Isles (Ferro 18° 9' 45" west), he obtained the following longitude: Carthage 32° 20', Rhodes 56°, Issus 66° 30', Indus 122° 30', Ganges 146°, Thinæ (conjectured to be Tenasserim, in farther India) 177° 30'—errors respectively of about four, ten, twelve, thirty, forty, and sixty-one degrees. These large errors he had no means of recognizing, but when he came to his latitudes he did have a corrective. Accepting, as he was disposed to do, Eratosthenes's distances, all his own latitudes became too high. Pytheas's Thule, instead of being at the arctic circle, would have been beyond the north pole. He therefore did what is so often done, allowed one error to force him into another, viz., the use of a degree of one length (500 stadia) for his longitudes, and of another length (700 stadia) for his latitudes. Like those of Eratosthenes, therefore, his latitudes are tolerably correct. A few features of this map should be noticed in comparison with former maps. Thule, which was unnoticed by Strabo, reappears, though it is not the Thule of Pytheas, but an island much nearer to Britain. The Sicilian Straits are no longer on a parallel with the Columns of Hercules, nor the straits and Rome on the same meridian. Alexandria and Rhodes are on different meridians, as also the Hellespont and Byzantium. The Caspian becomes again an inland sea. In the East, the great peninsula of the Deccan disappears, the island of Taprobane occupying its place. The Indian Ocean is an inland sea, Africa being connected by unknown lands with the lands of the far East.
In our map we have followed Gosselin's opinion that the Golden Chersonese was the region about the mouths of the Irrawaddy River, not the Malay Peninsula. But the most important of the minor features of the map, with which we must close our sketch, is its representation of the sources of the Nile.
Nili caput quærere was a work proposed to itself by the ancient as seriously as by the modern world. In the days of Herodotus the source of the river was considered a question of antiquity. The ancient dynasties, the Persian conquerors, and later the Greeks and the Romans, all made more or less of effort to solve the problem. We have seen how Herodotus answered the question. In the days of Eratosthenes, opinions were far more correct. Speaking of the Astaboras (Atbara), and the Astapus or the Astasobas (Blue Nile), he says, "Certain authors pretend that this last name applies to another river, which flows from lakes situated to the south, and forming the principal affluent of the Nile." This is the first definite reference to the south-
