This report presents an investigation of riverine sand dunes and includes an extensive literature review on the subject. It was concluded that the principal mechanisms responsible for river bed instability, which results in the production of ripples, dunes, etc., are fairly well understood. However, several of the key constituent processes are not well formulated. Principal among these are the phase shifts that occur between the local sediment discharge, the local near-bed velocity, local bed-profile displacement, and local bed shear stress. The reasons for the current deficiencies in the theoretical models are discussed and are concluded to stem from the difficulties inherent to analysis of nonuniform, turbulent, boundary-layer-type flows. Several existing theories for prediction of dune height were verified. None of the five published predictors evaluated was found to be satisfactory. A new, "inverse" model for prediction of dune height was developed.