We are developing an age model for calibration of the Antarctic radiocarbon time scale, based on the ultra-high resolution Holocene record from two sites drilled in the Palmer Deep (PD), Antarctic Peninsula, during ODP Leg 178. The PD is an enclosed, bathymetric depression, >1400 m deep, composed of three basins (I, II, III) separated by bathymetric sills. The basins efficiently trap the products of coastal and near-shore productivity in an oceanic system that is strongly seasonal in its productivity cycle.Site 1098 (Basin I) recovered a 46 m core, consisting of bioturbated and laminated muddy, diatom ooze with minor turbidites, and a basal diamict (till unit). Site 1099 (Basin III) recovered a 107 m condensed section of diatom ooze, similar to 1098, but expanded by biogenic turbidites. Diamicton between 33-35 m coincides with the seismic mid basin reflector (MBR), underlain by black, silty diatom muds. The entire section may be Holocene in age, with the MBR representing a mid-Holocene event. Alternatively the MBR may represent the last glacial maximum (LGM). AMS radiocarbon dates are used to test these hypotheses. Multiple analytical methods are also used to investigate the millennial-scale natural variability of paleoproductivity recorded in PD. These methods include analyses of laminae by X-radiographs, thin sections, diatom assemblages, grain-size distribution, and organic geochemistry. They are used to address such questions from the PD as: - what are the duration and nature of 200 year, high productivity cycles? - what is the nature of triple oscillations (T-Events) driven by variation in solar irradiance? - what is the origin of the MBR in Site 1099 and is it correlatable to the diamict in Site 1098?