The earliest observations on bedforms were on ripples, both current ripples, and the more complex forms produced by waves. George Darwin (1883) and Heinrich Blasius (1910) made early experimental studies. Already, the connection between ripple migration and cross-lamination was discovered independently by several workers, including Sorby and Gilbert. Observation of subaqueous dunes came later, before 1914 mainly by Cornish, but he, and both Kindle and Bucher mainly described the external forms of intertidal dunes, rather than their internal stratification. Full awareness of the ubiquity of subaqueous dunes in sand-bed rivers came much later, following the introduction of echo-sounding techniques in the 1940s. The first attempt at a theory of such forms was published by Exner in 1920. Theoretical understanding was hampered by lack of a clear distinction between ripples and dunes, which was not made until experiments began in very large flumes in the 1960s. These same experiments, and observations of North Sea intertidal dunes, at last established the connection between dune migration and cross-bedding.

The 1940s-1960s were decades of intense research into sediment transport and resistance to flow in alluvial channels, carried out in hydraulic (not geological) laboratories. Selim Yalin (1965) and John Southard (1971) attempted to establish universal laws of similarity for sediment movement. Most authors were reluctant to believe that flow depth had to be considered a major determining variable, because if it is, then experiments are of limited use to study large phenomena like rivers and tidal currents. Southard's publications have convinced most workers (except for lingering Bagnoldians) that unfortunately, that is the case. The problem is acute only if the full range of bedforms must be considered: current ripples are little affected once depths exceed 10 cm; antidunes depend little on grain size (and are rare in large rivers or deep tidal currents). The main problem is the nature of dunes and the transition to plane bed, where not only are grain-size, velocity and depth all important, but lag and hysteresis effects can be significant.