1631, Gool had called Descartes's attention to this problem from Pappus. It was considered a difficult problem by contemporary mathematicians. The object was to find the locus of a point P from which line segments AP, BP, CP, DP may be drawn in any prescribed manner whatsoever, relative to four straight lines a, b, c, d, so that AP • BP = CP • DP. Descartes chose a fixed origin O on the line a, set OA = x, the obliquely placed "applicate" AP = y. Then, by means of a relationship based on similarity, he expressed BP, CP, DP as linear functions of x and y. He eliminated the constant term of the resulting equa- tion by a suitable choice of O and obtained . He constructed and discussed the resulting curve, established the type of the conic section from the coeffi- cients and also recognized the pair of straight lines occurring in the case of an exact square root. This was followed by an investi- gation of the complicated "locus to several lines" and also of the curves generated by mechanical geometric movement in which the reference line segments were taken in special positions of the greatest possible suitability. Occasional use was made of negative applicates; none, of negative abscissas. Avoiding infinitesimal (hence, "approximate") analysis, Des- cartes solved the problems of normals to an algebraic curve as follows (purely algebraic method). He set the applicates per- pendicular to the axis, took the point P 0 (x 0, y 0 ), on the curve, the point M (t, 0) on the axis and let the circle whose center was M and which passed through P 0 intersect the curve again at P(x, y). Then he substituted the value in the equation of the curve and obtained the equation f(x, t) = 0 with x as the unknown, t as the parameter, having x - x 0 as a linear factor. Now, Descartes required that the factor x - x 0 be split off again. Thus he found t and from -7- |