Faulting exerts an important control upon drainage development in active extensional basins and thus helps determine the architecture of the sedimentary infill to a synrift basin. Examples of the interaction between faulting and drainage from the western United States and central Greece may be grouped into a relatively small number of classes based upon the structural position of a drainage catchment: footwall, hangingwall, fault offset and axial. Our examples illustrate the diversity of erosional effects that might arise because of variations in the spacing, orientation and segmentation of faults and their interactions. Where basement lithology is similar, footwall catchments are generally smaller, shorter and steeper than those of the hangingwall. Footwall-sourced alluvial fans and fan deltas are: generally smaller in area than those sourced from similar lithologies in the hangingwall. Wide fault offsets often give rise to large drainage catchments in the footwall. The development of axial drainage depends upon the breaching of transverse bedrock ridges by headward stream erosion or by lake overflow. Once breaching has occurred the direction of axial stream flow is controlled by the potential developed between basins of contrasting widths. Fault migration and propagation leads to the uplift, erosion and resedimentation of the sedimentary infill to formerly active basins, leading to the cutting of footwall unconformities. The outward sediment flux from structurally controlled catchments is modulated in an important way by lithology and runoff. The greatest contrasts in basement lithology arise when fault migration and propagation have occurred, such that the sedimentary fill to previously active basins is uplifted, incised and eroded by the establishment of large new drainage systems in the footwalls of younger faults. Drainage patterns in areas where faults interact can shed light on the relative timing of activity and therefore the occurrence of fault migration and propagation. Facies and palaeocurrent trends in ancient grabens may only be correctly interpreted when observations are made on a length scale of 10–20 km, comparable to that of the largest fault segments.