We prove a sharp decoupling for nondegenerate surfaces in R-4. This puts the progress in [1] on the Lindelof hypothesis into a more general perspective.