West Antarctica is generally thought to have behaved as a single terranesince mid-Cretaceous time. Subsequently some limited internal deformation,below the resolution of paleomagnetic methods, may have taken place, andthere has been some extension between East and West Antarctica. However,the most dynamic part of the Antarctic plate during Late Cretaceous andTertiary times was the Pacific margin of West Antarctica and adjacentoceanic areas. Unfortunately this is one of the least studied sectors ofthe Antarctic margin, due to its remoteness and inaccessibility.We present new reconstructions of the Pacific margin of Antarctica based onconstraints from marine magnetic data and regional free-air gravity fields.Recent results from interpretation of seismic reflection and gravityprofiles collected in the Bellingshausen Sea are also incorporated. Thereconstructions show regional constraints on the complex tectonic evolutionof the Amundsen and Bellingshausen Seas during and after the breakupbetween New Zealand and West Antarctica. The Late Cretaceous to earlyTertiary development of the region involved interactions between thePacific, West Antarctic, Phoenix (Aluk), Charcot and Bellingshausen plates.The West Antarctic margin consists of two main sectors. The western sectorcomprises the parts of the margin that Chatham Rise and Campbell Plateauseparated from at about 90 Ma and 80 Ma, respectively. Part of this sector,probably including much of the Amundsen Sea continental shelf, wassubsequently affected by independent motion of the Bellingshausen plate. Inthe eastern sector, Phoenix plate subduction continued through the LateCretaceous and early Tertiary, then terminated progressively from southwestto northeast as segments of the Antarctic-Phoenix ridge migrated to theAntarctic Peninsula margin. Between these two sectors is an area withcomplex gravity anomalies near Peter I Island. We suggest that a fragmentof the former Charcot plate, containing the oldest ocean floor adjacent toWest Antarctica, is preserved in this area. Some implications of our newreconstructions can be tested by onshore and offshore deployments ofseismometers along the Pacific margin. Ideally such deployments should beco-ordinated with controlled-source seismic investigations.