Megaflutes are large, concave-up, erosional scours within turbidite successions and have been described from modern and ancient settings; they are interpreted to represent the passage of erosive, largely bypassing, turbidity currents. Here, two categories of megaflute are distinguished: (1) allogenic, corresponding to the case above, where erosion is associated with sediment bypass leading to deposition farther into the basin, e.g., the “classical” megaflutes of the Ross Formation; (2) autogenic, where the erosion occurs during aggradation of the host bed. Both types are illustrated with examples from the Carboniferous succession of northern England. The distinction is important because the formative processes are different, leading to distinct facies associations for each group. For example, allogenic flutes are interpreted to form in systems with large variations in flow magnitude and imply significant down-dip sediment accumulation, whereas autogenic megaflutes are thought to be formed by unsteady flows and imply less significant bypass. Autogenic megaflutes occur above inversely graded, poorly sorted sandstones, which improve in sorting and decrease in mud content upwards to the megaflute surface, which is overlain by a well-sorted mantle of the coarsest sediment found within the bed. Detailed grain-size profiles suggest that autogenic megaflutes develop in response to waxing of flows from a depositional to an erosional state, cannibalizing the already deposited sand bed. As the flow wanes, reverting to a depositional state, the coarsest-grained fraction of its remaining load is deposited first, forming the basal part of a variably well-developed, normally graded and overall finer-grained interval. Thus, it is interpreted that flow unsteadiness promotes the formation of these features, which has implications for the reconstruction of paleoenvironmental conditions, specifically the generation mechanisms, return periods, flow routing into the basin, and the resultant flow characteristics of turbidity currents.