IX.] THE HABITABLE WORLD. 173
by Dicaearchus. Following Pytheas as his authority, Eratosthenes
fixed the northern limit at the parallel of Thule,
.... , , . ... . , , A . Its Breadth.
which he regarded as coinciding with the Arctic
circle. In the opposite direction—since the idea which once pre-
vailed that the region between the tropics was uninhabitable had
been dispelled by the knowledge that Syene was on the tropic of
Cancer, and that Meroe and other inhabited places lay far beyond
that line—he determined the limit at the furthest spot towards the
south in which at that period men were known to exist, the land
of the Sembritae (Sennaar). In the same parallel with this he
rightly placed the Cinnamon region (Somaliland), and also, by a
happy conjecture, Taprobane (Ceylon) to the southward of India.
The interval between these two limits, and with it the breadth of
the inhabited world, he estimated, according to the rough calcula-
tions which were available in that age—following a meridian line
drawn from Meroe to the mouth of the Borysthenes, and from
thence to the parallel of Thule—at 38,000 stadia.
In computing the length of the same area from
west to east, he was able to avail himself of the calculations which
had been made, first by the companions of Alexander, and after-
wards by Megasthenes and Patrocles, of the extent of the newly
discovered portions of the continent of Asia, These however
were partly due to conjecture; and since much of the remaining
distance had to be measured by sea, it was necessary here again
to trust to the vague estimates of sailors, which owing to the
delays and uncertainties of navigation were usually in excess of
the reality. The parallel which he selected for this measurement
followed the line of the Mediterranean, and afterwards that of the
great central mountain-chain of Asia, the Taurus and the
Himalaya. Starting from a point in the Atlantic to the westward
of the Pillars of Hercules which would be on the same meridian
as the Armorican promontory in Gaul—for he regarded this as
lying further to the west than the Sacrum Promontorium in
Spain (Cape St Vincent)—he calculated the distances by sea as
far as the gulf of Issus, and thence by land to the furthest ex-
tremity of India. The total thus obtained amounted to 77,800
stadia, an estimate which exceeds the reality by about one-third1.
1 Strabo, i. 4. a, 5.