A remark on Raghunathan's vanishing theorem (Q1068961)

Let G be a simple Lie group and \(V_ 0\) an irreducible real representation of G. Following a result of \textit{M. S. Raghunathan} [Am. J. Math. 87, 103-139 (1965; Zbl 0132.021)] the first cohomology group \(H^ 1(\Gamma,V_ 0)\) of a uniform discrete subgroup \(\Gamma\) of G vanishes unless G is locally isomorphic to SO(n,1) and \(V_ 0=H^ jV\), where V is the standard representation of SO(n,1) and \(H^ jV\) denotes the harmonic polynomials of V of degree j, or G is locally isomorphic to SU(n,1) and \(V_ 0=S^ jV\), the symmetric power of the standard representation or its dual \((j=0,1,...).\) The paper under review deals with a non-vanishing result in the case \(G=SO(n,1)\) and non-trivial coefficients. Given a uniform, discrete, torsion free subgroup \(\Gamma\) of G, denote the associated locally symmetric space by \(M=\Gamma \setminus H^ n\) \((H^ n= hyperbolic\) n- space) and recall that \(H^*(\Gamma,V_ 0)=H^*(M,\tilde V_ 0)\) where \(\tilde V_ 0\) is the corresponding local system on M. Then the author proves: Suppose M is a compact hyperbolic n-manifold admitting a two-sided, non-singular, totally geodesic hypersurface (where two-sided means with trivial normal bundle) then the first cohomology group \(H^ 1(M,H^ jV)\) does not vanish for \(j=1,2,... \). This result complements previous ones of the author [Ann. Math., II.Ser. 104, 235-247 (1976; Zbl 0364.53020)]. Its proof rests on methods used in \textit{D. Johnson} and the author: ''Deformation spaces associated to compact hyperbolic manifolds'', to appear.