200 THE ELEMENTS DOWN TO PLATO’S TIME
We also have
r sin 6 = \ AB — r'sin 6' (1)
In order that the lune may be squareable, we must have, in
the first place, r 2$ _ T ,C ±Q\
Suppose that 6 = m 6', and it follows that
r' — V m. r.
Accordingly the area becomes
-|r 2 (m sin 26' — sin 2 m 6') ;
and it remains only to solve the equation (1) above, which
becomes sin m Q' — v / m < s i n
This reduces to a quadratic equation onty when m has one
of the values o q 3 e 5
'3*
The solutions of Hippocrates correspond to the first three
values of m. But the lune is squareable by ‘ plane ’ methods
in the other two cases also, Clausen (1840) gave the last four
cases of the problem as new 1 (it was not then known that
Hippocrates had solved more than the first) ; but, according
to M. Simon 2 , all five cases were given much earlier in
a dissertation by Martin Johan Wallenius of Abo (Abveae,
1766). As early as 1687 Tschirnhausen noted the existence
of an infinite number of squareable portions of the first of
Hippocrates’s lunes. Vieta 3 discussed the case in which m = 4,
which of course leads to a cubic equation.
(¡3) Reduction of the problem of doubling the cube to
thè finding of two mean proportionals.
We have already alluded to Hippocrates’s discovery of the
reduction of the problem of duplicating the cube to that of
finding two mean proportionals in continued proportion. That
is, he discovered that, if
a:x = x:y = y :b,
then a ?J \x ?J — a:b. This shows that he could work with
compound ratios, although for him the theory of proportion
must still have been the incomplete, numerical, theory
developed by the Pythagoreans. It has been suggested that
1 Creile, xxi, 1840, pp. 375-6.
2 GeschicJite der Math, im Altertum, p. 174.
Yieta, Variorum de rebus mathematicis responsorum lib. viii, 1598.