EARLY ERRORS. 
381 
wholly without some guiding principle. In ge- 
neral, every geographical feature was extended 
in the direction which it followed, when it first 
merged into the unknown space. In some cases, 
indeed, a certain flexure is necessary, in order to 
complete the figure of a continent, an ocean, or 
other grand geographical feature. Such was the 
line by which, in the system of Ptolemy, the 
African continent was carried round till it met 
the farthest extremity of Asia. The estimate of 
distance was also a point on which the early geo- 
graphers were naturally liable to much error. 
The most usual was that of exaggeration. As- 
tronomical observations were yet rare, and very 
imperfect. The materials were derived almost 
entirely from travellers employed on mercantile 
or military objects. The difficulties and dangers 
of an unknown tract, the windings of the road, 
and the disposition to magnify their own achieve- 
ments, combined in inducing such persons to form 
a high idea of the space which they had travers- 
ed. Sometimes, on the other hand, rumour con- 
veys the knowledge of a grand remote feature, 
while the intermediate space is but imperfectly 
known. That space then appears less, and the 
distant object nearer, than it really is. The pil- 
lars of Hercules, made known to the early Greeks 
by the exploits of that celebrated adventurer, 
were placed by them very little beyond Sicily. 
