Nikora and Goring’s paper provides a thorough account of the turbulence properties of open-channel flow over fixed and weakly mobile gravel beds. One of the most striking results in the paper is the apparent large reduction in the value of the von Karman constant ~hereafter referred to as k or the von Karman parameter! during conditions of weak bed-load transport. Confirmation of this result would be significant as k is embedded in many bedload transport models and it would provide additional insight into the physics of sediment transport. The paper advances two possible explanations for the observed decrease of the von Karman parameter. First, the authors note that ‘‘Potentially, different @flow # mechanisms may dominate at different particle Stokes numbers.’’ A useful account of this effect may be found in Elgobashi ~1991!. Second, they suggest that the reduction in the von Karman parameter is associated with behavior that is peculiar to near-threshold conditions, and is related to other critical phenomena known to occur at phase transitions. The purpose of this discussion is to augment the second of these explanations by providing a more mechanistic account of the physical processes which may produce the observed reduction in k. As pointed out by Nikora and Goring, drag reduction is widely considered to result from an alteration of turbulence structure in a flow. Specifically, a reduction in the spacing or intensity of largescale events, often termed bursts or sweeps, has been shown to reduce drag. Consequently, a substantial literature has formed on methods of inducing drag reduction in boundary layers ~Lumley and Blossey 1999!. One such method employs wall-mounted sensors to detect the early development of large-scale flow structures. A body force is then applied to the near-wall flow in order to disrupt the further growth of the detected structure, thereby preventing it from contributing substantially to drag. Such a method has been found to produce a 20‐25% reduction in drag ~Blossey and Lumley 2000!. In this discussion it is suggested that near threshold a loose sediment bed behaves in a similar manner. Consider the case of a mixed-grain-size sediment bed over which weak bed-load transport is observed ~i.e., similar conditions to those reported in the paper!. The sediment flux is sporadic, suggesting that grain motion occurs when the near-bed velocity is high ~Nelson et al. 1995!, but it is not sustained during periods of low velocity. Therefore, weak bed-load transport takes place in a regime in which the general ambient condition of the surface is static but, during the passage of large-scale turbulent structures, patches of grain motion occur. A qualitative similarity with the drag reduction method described in Blossey and Lumley ~2000! is therefore apparent as particles, entrained into the flow by a large turbulent event, provide a near-bed body force retarding the flow close to the point ~a few wall units from the boundary! and time ~near the leading edge of the flow structure ! that Blossey and Lumley advocate for drag reduction in boundary layers. This provides a mechanistic explanation for the behavior observed by Nikora and Goring, supporting their assertion that weak bed-load transport belongs to a special class of drag reducing flows. Note that Blossey and Lumley’s ~2000! method has been applied to hydraulically smooth flow in which the mechanisms of turbulence production near the wall are relatively well known. In hydraulically rough flows, such as that described in the paper, the mechanism of near-wall turbulence production is less well understood. It is supposed here that the precise mechanisms of turbulence production are not important and that the crucial similarity between loose sediment boundaries and Blossey and Lumley’s method lies in the application of a body force to the near-wall fluid disrupting the growth of large-scale turbulent