Let M be a smooth manifold and G a compact connected Lie group acting on M by isometries. In this paper, we study the equivariant cohomology of X = T * M , and relate it to the cohomology of the Marsden-Weinstein reduced space via certain residue formulae. In case that X is a compact symplectic manifold with a Hamiltonian G-action, similar residue formulae were derived by Jeffrey, Kirwan et al. [26,25].D̺(X) = d(̺(X)) − ι X (̺(X)), X ∈ g, ̺ ∈ (S(g * ) ⊗ Λ(X)) G ,where d and ι denote the usual exterior differentiation and contraction, the equivariant cohomology of the G-action on X is given by the quotient H * G (X) = Ker D/ Im D, which is canonically isomorphic to the topological equivariant cohomology introduced in [2] in case that G is compact, an assumption that we will make from now on. The main difference between ordinary and equivariant cohomology is that the latter has a larger coefficient ring, namely S(g * ), and that it depends on the orbit structure of the underlying G-action. Let us now assume that X admits a symplectic structure ω which is left invariant by G. By Cartan's homotopy formula,where L denotes the Lie derivative with respect to a vector field, implying that ι X ω is closed for each X ∈ g. G is said to act on X in a Hamiltonian fashion, if this form is even exact, meaning that there exists a linear function J : g → C ∞ (X) such that for each X ∈ g, the fundamental vector field X is equal to the Hamiltonian vector field of J(X), so that d(J(X)) + ι X ω = 0.An immediate consequence of this is that for any equivariantly closed form ̺ the form given by e i(J(X)−ω) ̺(X) is equivariantly closed, too. Following Souriau and Kostant, one defines the momentum map of a Hamiltonian action as the equivariant mapAssume next that 0 ∈ g * is a regular value of J, which is equivalent to the assumption that the stabilizer of each point of J −1 (0) is finite. In this case, J −1 (0) is a smooth manifold, and the corresponding Marsden-Weinstein reduced space, or symplectic quotientis an orbifold with a unique symplectic form ω red determined by the identity ι * ω = π * ω red , where π : J −1 (0) → X red and ι : J −1 (0) ֒→ X denote the canoncial projection and inclusion, respectively. Furthermore, π * induces an isomorphism between H * (X red ) and H * G (J −1 (0)). Consider now the mapand assume that X is compact and oriented. In this case, Kirwan [28] showed that K defines a surjective homomorphism, so that the cohomology of X red should be computable from the equivariant cohomology of X. This is the content of the residue formula of Jeffrey and Kirwan [26], which for any ̺ ∈ H * G (X) expresses the integral(1)following residue formula. Let ̺ ∈ H * G (T * M ) be of the form ̺(X) = α + Dβ(X), where α is a closed, basic differential form on T * M of compact support, and β is an equivariant differential form of compact support. Fix a maximal torus T ⊂ G, and denote the corresponding root system by ∆(g, t). Assume that the dimension κ of a principal G-orbit is equal to d = dim g, and denote the product of the pos...