Abstract
The structures are interpreted in terms of a regularly repeated downstream sequence of flow configurations involving (1) a hydraulic jump upstream of each rib, which acts as a local flow control, (2) a critical flow above the rib and a steep cascade into a pool immediately downstream, and (3) a gradually redeveloping supercritical flow well downstream from the pool. An analysis of the stability of the stones of a rib, combined with an empirical criterion for the existence of supercritical flow in a stream cascading over a train of regularly spaced two-dimensional obstacles, yields a model in rough quantitative agreement with field observations. Rib spacing is found to vary directly with the calibre of the constituent clasts and inversely with the bed slope, the depth of flow over the ribbed surface being of the same order as the stone size. Earlier interpretations of transverse ribs as due to antidune or modified antidune flow or to the stepwise upstream migration of a hydraulic jump seem inconsistent on both morphological and hydraulic grounds with the observed field character and relations of the bedforms.