Abstract.A new dual-Petrov-Galerkin method is proposed, analyzed, and implemented for third and higher odd-order equations using a spectral discretization. The key idea is to use trial functions satisfying the underlying boundary conditions of the differential equations and test functions satisfying the "dual" boundary conditions. The method leads to linear systems which are sparse for problems with constant coefficients and well conditioned for problems with variable coefficients. Our theoretical analysis and numerical results indicate that the proposed method is extremely accurate and efficient and most suitable for the study of complex dynamics of higher odd-order equations. 1. Introduction. Over the last thirty years, spectral methods have been playing an increasingly important role in scientific and engineering computations. Most work on spectral methods is concerned with elliptic and parabolic-type equations; there has also been active research on spectral methods for hyperbolic problems (see, for instance, [11,7,14] and the references therein). However, there is only a limited body of literature on spectral methods for dispersive, namely, third and higher odd-order, equations. In particular, relatively few studies are devoted to third and higher oddorder equations in finite intervals. This is partly due to the fact that direct collocation methods for higher odd-order boundary problems lead to very much higher condition numbers-more precisely, of order N 2k , where N is the number of modes and k is the order of the equation-and often exhibit unstable modes if the collocation points are not properly chosen (see, for instance, [17,21]).In a sequence of papers [22,23,25,26], the author constructed efficient spectralGalerkin algorithms for elliptic equations in various situations. In this paper, we extend the main idea for constructing efficient spectral-Galerkin algorithms-using compact combinations of orthogonal polynomials, which satisfy essentially all the underlying homogeneous boundary conditions, as basis functions-to third and higher odd-order equations. Since the main differential operators in these equations are not symmetric, it is quite natural to employ a Petrov-Galerkin method.The key idea of the new spectral dual-Petrov-Galerkin method is the innovative choice of the test and trial functional spaces. More precisely, we choose the trial functions to satisfy the underlying boundary conditions of the differential equations, and we choose the test functions to satisfy the "dual" boundary conditions.Recently, Ma and Sun [19,20] studied an interesting Legendre-Petrov-Galerkin method for third-order equations. The main difference between this paper and [19,20]