Diepkloof Rock Shelter offers an exceptional opportunity to study the onset and evolution of both Still Bay (SB) and Howiesons Poort (HP) techno-complexes. However, previous age estimates based on luminescence dating of burnt quartzites (. Tribolo etal., 2009) and of sediments (. Jacobs etal., 2008) were not in agreement. Here, we present new luminescence ages for 17 rock samples (equivalent dose estimated with a SAR-ITL protocol instead of classical MAAD-TL) as well as for 5 sediment samples (equivalent dose estimated with SAR-single grain OSL protocol) and an update of the 22 previous age estimates for burnt lithics (modified calibration and beta dose estimates). While a good agreement between the rock and sediment ages is obtained, these estimates are still significantly older than those reported by Jacobs etal. (2008). After our own analyses of the sediment from Diepkloof, it is suspected that these authors did not correctly chose the parameters for the equivalent dose determination, leading to an underestimate of the equivalent doses, and thus of the ages.From bottom to top, the mean ages are 100±10ka for stratigraphic unit (SU) Noël and 107±11ka for SU Mark (uncharacterized Lower MSA), 100±10ka for SU Lynn-Leo (Pre-SB type Lynn), 109±10ka for SUs Kim-Larry (SB), 105±10ka for SUs Kerry-Kate and 109±10ka for SU Jess (Early HP), 89±8ka for SU Jude (MSA type Jack), 77±8ka for SU John, 85±9ka for SU Fox, 83±8ka for SU Fred and 65±8ka for SU OB5 (Intermediate HP), 52±5ka for SUs OB2-4 (Late HP).This chronology, together with the technological analyses, greatly modifies the current chrono-cultural model regarding the SB and the HP and has important archaeological implications. Indeed, SB and HP no longer appear as short-lived techno-complexes with synchronous appearances for each and restricted to Oxygen Isotopic Stage (OIS) 4 across South Africa, as suggested by Jacobs etal. (2008, 2012). Rather, the sequence of Diepkloof supports a long chronology model with an early appearance of both SB and HP in the first half of OIS 5 and a long duration of the HP into OIS 3. These new dates imply that different technological traditions coexisted during OIS 5 and 4 in southern Africa and that SB and HP can no longer be considered as horizon markers. © 2012 Elsevier Ltd.
