Using the 160-m-long flume at Tsukuba University we undertook an experiment to provide a first estimate of the virtual velocity of sand in the size range 0.5–2.0 mm. For the flow velocity used in our experiment this sediment-size range would conventionally be regarded as suspended sediment. The virtual velocity was found to be 37–41% of the flow velocity. Paradoxically, virtual velocity decreases as particle size decreases. Such a lower virtual velocity of finer sediment is not inconceivable. First, trapping of the sediment appears to be a function of bed roughness, and there is a probable relationship between bed roughness and trapping efficiency for particles of different sizes. Second, finer particles are more likely to find sheltered positions on a rough bed and thus experience lower mobility, relative to the more exposed coarser grains, as observed for bedload transport. Third, the virtual velocity of particles undergoing bedload transport has been found, in some instances, to be lower for finer clasts. We combine our data with previous studies of virtual velocity of bedload to develop, for the first time, a hypothesis for a holistic analysis of sediment movement in rivers. Copyright © 2017 John Wiley &amp
Sons, Ltd.