LX. On Mr. Edgeworth's method of reducing observations relating to several quantities