EGYPTIAN GEOMETRY
123
difficulties ’, as it calls itself, is rough mensuration. The
following are the cases dealt with which concern us here.
(1) There is the rectangle, the area of which is of course
obtained by multiplying together the numbers representing
the sides. (2) The measure of a triangle is given as the pro
duct of half the base into the side. And here there is a differ
ence of opinion as to the kind of triangle measured, Eisenlohr
and Cantor, taking the diagram to represent an isosceles tri
angle rather inaccurately drawn, have to assume error on
the part of the writer in making the area \ah instead of
\aV{b 2 — %a 2 ) where a is the base and b the ‘side’, an error
which of course becomes less serious as a becomes smaller
relatively to h (in the case taken a = 4, h = 10, and the area
as given according to the rule, i.e. 20, is not greatly different
from the true value 19-5959). But other authorities take the
triangle to be right-angled and h to be the side perpendicular
to the base, their argument being that the triangle as drawn
is not a worse representation of a right-angled triangle than
other triangles purporting to be right-angled which are found
in other manuscripts, and indeed is a better representation of
a right-angled triangle than it is of an isosceles triangle, while
the number representing the side is shown in the figure along
side one only of the sides, namely that adjacent to the angle
which the more nearly represents a right angle. The advan
tage of this interpretation is that the rule is then correct
instead of being more inaccurate than one would expect from
a people who had expert land surveyors to measure land for
the purpose of assessing it to tax. The same doubt arises
with reference to (3) the formula for the area of a trapezium,
namely ■|(a + c) x b, where a, c are the base and the opposite
parallel side respectively, while b is the ‘ side ’, i.e. one of the
non-parallel sides. In this case the figure seems to have been
intended to be isosceles, whereas the formula is only accurate
if b, one of the non-parallel sides, is at right angles to the base,
in which case of course the side opposite to b is not, at right
angles to the base. As the parallel sides (6, 4) in the case
taken are short relatively to the £ side ’ (20), the angles at the
base are not far short of being right angles, and it is possible
that one of them, adjacent to the particular side which is
marked 20, was intended to be right. The hypothesis that