No label defined (Q4357790)

Let a symmetric nondegenerate \((0,2)\)-tensor field \(g_{ij}\) be given on a \(C^\infty\)-differentiable \(n\)-dimensional manifold \(M\) and let \[ \Omega^{ir}_{sj}=\textstyle{{1\over 2}}(\delta^i_s\delta^r_j- g_{sj}g^{ir}),\;\overset{*}\Omega^{ir}_{sj}=\textstyle{{1\over 2}}(\delta^i_s\delta^r_j+ g_{sj}g^{ir}). \] The map \(f_\Omega: T^1_2(M)\to T^1_2(M)\), given by \(f_\Omega(X^i_{jk})= \Omega^{ir}_{sj}X^s_{rk}\), is called an \(\Omega\)-map. Its fixed point set is given by \(\overset{*}\Omega^{ir}_{sj}A^s_{rk}= 0\). For a linear connection \(D=(\Gamma^i_{jk})\) a general transformation \(T:D\to \overline D\), given by \(\overline{\Gamma^i_{jk}}= \Gamma^i_{jk}+ Y^i_{jk}\), is called an \(\Omega\)-transformation, if \(Y^i_{jk}\) is a fixed point of \(f_\Omega\). Some necessary and sufficient conditions are deduced which characterize \(\Omega\)-transformations.