Submarine gravity currents, especially long run-out flows that reach the deep ocean, are exceptionally difficult to monitor in action, hence there is a need to reconstruct how these flows behave from their deposits. This study mapped five individual flow deposits (beds) across the Agadir Basin, offshore north-west Africa. This is the only data set where bed shape, internal distribution of lithofacies, changes in grain size and sea floor gradient, bed volumes, flow thickness and depth of erosion into underlying hemipelagic mud are known for individual beds. Some flows were 30 to 120 m thick. However, flows with the highest fraction of sand were less than 5 to 14 m thick. Sand was most likely to be carried in the lower 5 to 7 m of these flows. Despite being relatively thin, one flow was capable of transporting very large volumes of sediment (ca 200 km3) for large distances across very flat sea floor. These observations show that these relatively thin flows could travel quickly enough on very low gradients (0·02° to 0·05°) to suspend sand several metres to tens of metres above the sea floor, and maintain those speeds for up to 250 km across the basin. Near uniform hemipelagic mud interval thickness between beds, and coccolith assemblages in the mud caps of beds, suggest that the flows did not erode significantly into the underlying sea floor mud. Simple calculations imply that some flows, especially in the proximal part of the basin, were powerful enough to have eroded hemipelagic mud if it was exposed to the flow. This suggests that the flows were depositional from the moment they arrived at a basin plain location, and that deposition shielded the underlying hemipelagic mud from erosion. Reproducing the field observations outlined in this exceptionally detailed field data set is a challenge for future experimental and numerical models.