Characteristic length and timescales for a turbid surge are used to estimate bedload transport by the surge, deriving estimates for the conditions under which deposited material will be mobilized as bedload, and of the relative importance of bedload in determining the overall deposit geometry. A critique is provided of the common modelling assumptions which underlie these estimates and of how their consistency can be checked. For large turbidity currents, such as those which emplaced the Marnoso Arenacea turbidites in northern Italy, model predictions of overall geometry are not easy to reconcile with the field data: some possible reasons for this are discussed. The estimates obtained from these characteristic scales are consistent with the widespread presence in turbidites of sedimentary structures which indicate bedload transport; some of these structures from the Marnoso Arenacea Formation are reviewed briefly. However, the estimates suggest that bedload transport is not a major factor responsible for the geometry of the large turbidites in this formation, which exhibit a broad thickness maximum in their proximal region that contrasts with the downstream-thinning geometry of smaller beds. This effect suggests that the explanation for this theoretically unexpected geometry should be sought in other physical mechanisms.