facie in Orbe Luna_, cap. vii.]

[Footnote 454: See Augustine, _De civitate Dei_, xvi. 9;
Lactantius, _Inst. Div._, iii. 23; Jerome, _Comm. in Ezechiel_,
i. 6; Whewell's _History of the Inductive Sciences_, vol i. p.
196.]

[Footnote 455: See above, p. 266.]

[Footnote 456: For an account of the cosmography of the Divine
Comedy, illustrated with interesting diagrams, see Artaud de
Montor, _Histoire de Dante Alighieri_, Paris, 1841.]

[Footnote 457: It was first printed without indication of place
or date, but probably the place was Paris and the date
somewhere from 1483 to 1490. Manuscript copies were very
common, and Columbus probably knew the book long before that
time. There is a good account of it in Humboldt's _Examen
critique_, tom. i. pp. 61-76, 96-108. Humboldt thinks that such
knowledge as Columbus had of the opinions of ancient writers
was chiefly if not wholly obtained from Alliacus. It is
doubtful if Columbus had any direct acquaintance with the works
of Roger Bacon, but he knew the _Liber Cosmographicus_ of
Albertus Magnus and the _Speculum Naturale_ of Vincent de
Beauvais (both about 1250), and drew encouragement from them.
He also knew the book of Mandeville, first printed in French at
Lyons in 1480, and a Latin translation of Marco Polo, published
in 1485, a copy of which, with marginal MS. notes, is now in
the Colombina.]

[Illustration: Annotations by Columbus.]

[Sidenote: Ancient estimates of the size of the globe and the length of
the Oecumene.]

The principal consideration, of course, was the distance to be
traversed; and here Columbus was helped by an error which he shared with
many geographers of his day. He somewhat underestimated the size of the
earth, and at the same time greatly overestimated the length of Asia.
The first astronomer to calculate, by scientific methods, the
circumference of our planet at the equator was Eratosthenes (B. C.
276-196), and he came--all things considered--fairly near the truth; he
made it 25,200 geographical miles (of ten stadia), or about one seventh
too great. The true figure is 21,600 geographical miles, equivalent to
24,899 English statute miles.[458] Curiously enough, Posidonius, in
revising this calculation a century later, redu