XVI.] PTOLEMY. 341

But the cause which contributed more than any other to produce
this result was the completeness of his system, which' communi-
cated to his statements an appearance of finality that did not
really belong to them. The evidence on which Ptolemy relied in
determining the position .of places on the face of the glohe was
largely derived from the same source which had been used by his
predecessors — the computations of distances made by travellers
and navigators, whose estimates were from the nature of the case
inaccurate. But when the evidence was withdrawn from view, as
it was in Ptolemy's work, and the results were embodied in maps
and tables of distances symmetrical in form, the definite character
which they thus acquired caused them in the course of time to be
regarded in the light of exact statements, as if they were based on
scientific observations. It should be remembered, however, that
Ptolemy himself is not responsible for this result of his mode of
treating his subject, for in his Introduction he makes no secret of
the imperfection of his materials l : nor ought the mistakes in his
work which are revealed by modern discovery and modern science
to blind us to its preeminent merits,

With regard to the measurement of the circumference of the
earth Ptolemy followed Marinus in accepting Posi-
donius' erroneous estimate of 180,000 stadia9, which about the*

fell short of the reality by one-sixth. It resulted
from this that, as he adopted from Hipparchus the
division of the equator and other great circles into 360 degrees,
he made every degree only 500 stadia (50 geographical miles)
instead of 600 stadia (60 geographical miles), which is the true
computation. - This mistake at once affected his calculation
of distances on his map, for in consequence of it he over-
estimated them: thus, if he discovered from his authorities —
itineraries or otherwise — that the interval between two places was
500 stadia, he would express this on his map by a degree, which
in reality is 600 stadia ; and when the estimate was made on a
large scale, the error in excess became very great. This was
especially felt when he came to deal with the

second important question of general scientific of the Habit-
, , i- , . i 1*1 * ^' -LI able World.
geography, that of th& length of the habitable
1 i. a. 4. a v. supra, p. 19*.