In the framework of the provision of one pure public good, we characterize the class of strategy-proof voting schemes on single-plateaued preferences over a convex and closed subset of the real line (the set of feasible levels of the public good).