On the equireciprocal point problem (Q1056385)

Let \(C\) be a plane curve. A point \(P\) in the plane is called an equireciprocal point (ER point) of \(C\), if \(C\) is star-shaped at \(P\) and if every chord \(XY\) of \(C\) through \(P\) satisfies \((1/\vert XP\vert)+ (1/\vert YP\vert) = \alpha\) \((\alpha = \) ER constant of \(C\) at \(P)\). Extending work of \textit{A. Hallstrom} [``Equichordal and equireciprocal points.'' Bogazici Univ. J. Sci. 2, 83--88 (1974)] the author shows that there exist non-elliptical converse curves with a given pair \(P\), \(Q\) of ER points and with ER constant \(\alpha\) satisfying \(\alpha < 4/\vert PQ\vert\). These curves cannot be twice differentiable, because this condition implies that they be ellipses, which is also proved in the paper under review. Furthermore in most cases the ER constant does not depend on the ER point: If \(C\) has two ER points \(P\), \(Q\) and the chord through \(P\) and \(Q\) meets \(C\) at \(X_0\) and \(Y_0\), then \(C\) has either infinite curvature at these points or the ER constants at \(P\) and \(Q\) are equal. The methods of proof use the theory of invariant manifolds of dynamical systems.