On some properties of a set of probability measures (Q1100800)

Let (S,\(\rho)\) be a metric space and \({\mathcal P}(S)\) denote the set of probability measures defined on (S,\(\zeta)\) (\(\zeta\)-stands for the \(\sigma\)-field generated by open sets of S). We prove, that if \(\tau\),\(\nu\in {\mathcal P}(S)\), and \(L(\nu,\tau)=c\), (L denotes the Lévy-Prokhorov distance) then for every \(0<t<1\) exists a probability measure \(\eta\in {\mathcal P}(S)\) such that \(L(\tau,\eta)=tc\) and \(L(\eta,\nu)=(1-t)c\), with the assumption that the metric space (S,\(\rho)\) satisfies the following condition: For every \(0<t<1\) and every x,y\(\in S\) such that \(\rho (x,y)=c<1\), there exist \(z\in S\) such that \(\rho (x,z)=tc\) and \(\rho (z,y)=(1-t)c\).