Apolar locus of a polynomial (Q2844644)

We shall begin with several notations. The author recalls the definition of apolar monic polynomials \( p, q \) and the classical theorem of Grace. The smallest disc containing all zeroes of the algebraic polynomial \( p(z) \) of degree \( n \geq 2 \) is denoted by \( D(p) \). Then \( D(p) \) contains at least one zero of \( q \) and \( D(q) \) contains at least one zero of \( p \) (Grace theorem). For every \( p(z) \) the minimal closed and simply connected domain \( \Omega(p) \subset D(p) \), called the apolar locus of \( p \), is defined in such a way that \( \Omega(p) \) contains at least one zero of \( q(z) \) and \( \Omega(q) \) contains at least one zero of \( p(z) \). A polynomial \( q(z) \), apolar to \( p(z) \), is extreme for \( \Omega(p) \) if all zeroes of \( q(z) \) are located on the contour of \( \Omega(p) \). This is the main result of the paper under consideration: If \( q \) is extreme for \( \Omega(p) \), then all zeroes of \( q \) are located on a circle. In Statement 3 a link between the apolar locus, the Grace theorem and Rolle's theorem for complex polynomials is established.