The Latin editions of Ptolemy are the earliest, and are greatly to be preferred to the Greek. They were printed from a translation into Latin by Jacob Angelus, and consist of an edition at Bologna bearing the date 1462, but which is believed not to be the true date; one at Vincenza in 1475, which is really the earliest edition; that of Rome, 1478, the first with maps; Ulm, 1482 and 1486; Rome, 1490 and 1507; Venice, 1511; Strasburg, 1513, 1520, and 1522, in which the text is compared with an old Greek MS.; an edition in 1525, which bears to be a correction of Jacob Angelus’s translation by I. de Regiomonte, and in which the principal changes introduced into the later Greek editions first make their appearance; and the two editions by Servetus in 1535 and 1541. The principal Greek editions are those of Erasmus in 1533 and 1546, Montanus in 1605, and Bertius in 1619. A recent edition has appeared, by Dr. F. G. Wilberg, in 1838, from a collation of nine MSS. with the editions of 1482, 1513, 1533, and 1535, and with a MS. at Milan, another at Vienna, and two Latin MSS. collated by Mannert. The author has himself collated for this work the Latin editions of 1482, 1486, 1520, 1522, 1525, 1535, with the Greek editions of 1605 and 1619, and with Wilberg’s edition, and he agrees with Mannert in giving the preference among the early editions to the Ulm edition of 1482, and the Strasburg editions of 1520 and 1522. In the so-called corrected edition of 1525 he has no confidence. The variations occur both in the names and in the latitudes and longitudes. In cases where all the editions agree, there can be no doubt as to the genuine text used. When they differ, he has laid down the positions according to the variant readings, and selected the one which best corresponded with the appearance of the country. The agreement is mainly in the position of the towns, and the variations in the features of the coast, and are, therefore, more easily corrected. Footnote 56: The Vedra might more naturally be supposed to be the Tyne, but an altar found at Chester-le-Street, on the Wear, on which the name Vadri occurs, indicates the Wear as the river, to which indeed the name bears a greater resemblance. There is no variation in the position of these two places. Footnote 57: The early Latin editions have, instead of Ienae aestuarium, Fines aestus. It is possible that this may be the correct reading, and that Wigtown Bay may have marked the utmost limit to which the Roman troops penetrated in Agricola’s second campaign. Footnote 58: The early Latin editions all read Leva. The edition of 1525 first altered it to Deva, and is followed by the late editions, and also by Wilberg; but the distance both from the Firth of Tay in the south and from Kinnaird’s Head corresponds more exactly with the mouth of the North Esk than with that of the river Dee. Footnote 59: