The Marlborough Fault System is a zone of dextral transpression in continental crust at the southern end of the Hikurangi subduction system between the obliquely convergent Australian and Pacific plates. Detailed mapping of an area of deformed Tertiary cover sequence on the down thrown side of the Kekerengu Fault (the Kekerengu- Washdyke study area) has revealed two phases of deformation, D_e and D_l. In the study area De consists of nine kilometre scale thrust faults cutting sediments derived from extra- and intra-basinal uplift. The timing of this episode is constrained by the age of the first clastic deposits and by a previously unidentified unconformity in this area of Late Miocene age. A clear sequence of D_l events is recognized deforming all earlier structures including Pliocene aged sediments. Although elsewhere in Marlborough a regional post-Pliocene <i>ca.</i> 20° clockwise block rotation has been previously identified, in the Kekerengu-Washdyke study area one site suggested no post-Pliocene rotation and another a large <i>ca.</i> 100° clockwise rotation. This lack of D_l rotation was unexplained and the large rotation attributed to localized Early Miocene deformation. Palaeomagnetic work carried out in this study has identified six more sites which contain this large rotation (average <i>ca.</i> 118+11°). The rotation therefore appears to be a regional event, likely to be a result of the location of Marlborough in the hinge zone at the southern end of the Hikurangi Margin. One site from dykes in basement rocks does not record this large rotation, indicating that the rotation occurred in upper layers detached from unrotating rock below by an unknown structure (such as a thrust fault), or that the rotation did not occur in this area. The large rotation is believed to have been achieved by pinning of the D_e thrusts to the south of the Marlborough region. The data suggests that the D_e thrusts in Marlborough were initially NW-trending and seaward, not landward-directed as was previously supposed. Palaeomagnetic work has also added to the evidence for a lack of D_l regional rotation on the downthrown side of the Kekerengu Fault. A previously unidentified second phase of D_l folding and 'bending' within the study area appears to have accommodated the regional rotation and suggests that the Kekerengu Fault acted as the eastern boundary of the D_l rotating block.