In this paper we study the monodromy action on the first Betti and de Rham nonabelian cohomology arising from a family of smooth curves. We describe sufficient conditions for the existence of a Zariski dense monodromy orbit. In particular we show that for a Lefschetz pencil of sufficiently high degree the monodromy action is dense.(ii) In the case of M B (X o , n), considered with its natural structure of an affine algebraic variety, there exists a point x B ∈ M B (X o , n) so that the orbitis Zariski dense in M B (X o , n); or in the case of M DR (X/B, n) → B there is a leaf of the foliation defined by ∇ n DR which is Zariski dense in the algebraic Zariski topology.This theorem suggests that for families f : X → B with a "large enough" geometric monodromy one should expect Zariski dense monodromy actions on non-abelian cohomology. Geometrically families with large monodromy naturally arise from hyperplane sections and Lefschetz fibrations. In this context we prove the following Theorem B Let Z be a smooth projective surface with b 1 (Z) = 0. Let O Z (1) be an ample line bundle on Z and let n > 1 be a fixed odd integer. Then there exists a positive integer ℓ (depending only on Z and O Z (1)), such that for every k ≥ ℓ and for every Lefschetz fibration f : Z → P 1 in the linear system |O Z (k)| we have:(i) There is no meromorphic function on M B (Z o , n) an which is invariant under the action of mon n B (π 1 (P 1 \ {p 1 , . . . , p µ }, o)) (equivalently there is no meromorphic function on M DR ( Z/P 1 \ {p 1 , . . . , p µ }, n) an which is ∇ n DR -invariant);(ii) In the case of M B (Z o , n), there exist a point x B ∈ M B (Z o , n) so that the orbitis Zariski dense in M B (Z o , n); or in the case of the space M DR ( Z/P 1 \ {p 1 , . . . , p µ }, n) the foliation defined by ∇ n DR has a Zariski dense leaf.Here as usual Z is the blow-up of Z at the base points of the pencil and p 1 , . . . p µ ∈ P 1 are the points where the map Z → P 1 is not submersive.
Preliminary reductionsWe start with some general results about linear group actions on algebraic varieties, which will allow us to localize at a point the Zariski density property of an action.