Pb, Nd, and Sr isotopic results for lavas of the Cretaceous Ontong Java and Manihiki oceanic plateaus fall well within the modern-day oceanic island or hot pot field. The data provide no evidence of old continental basements but indicate a major involvement of 'Kerguelen-type' or 'EM-I'-like mantle in the sources of both plateaus, which appear to have probably been formed, at least in part, by hotspots. However, the presently active hotspots that Pacific plate reconstructions suggest might have been possible plateau sources lack Kerguelen-type isotopic compositions. Either these hotspots did not participate in the formation of the two plateaus, or if they did, Kerguelen-type material must have been volumetrically much more important early in their existence. Two hypotheses for the origins of these plateaus which involve hotspot sources are consistent with the sparse available geochemical, geochronological and geophysical data. The first holds that the plateaus formed cataclysmically in association with surfacing plume heads; the second posits a relatively steady but robust hotspot at or near a ridge crest and requires a much longer period of formation. A near-ridge origin appears to be indicated by evidence that most of the Pacific plateaus were built largely on relatively young ocean crust. However, we suggest that a near-ridge origin is also compatible with the plume head concept in that plume heads appear very likely to become associated with spreading axes through their influence on rift propagation, which should be substantially greater than for ordinary hotspots. In either case, the lack of hotspot tracks (seamount chains) attached to the two plateaus would be a consequence of ridge migration or rift propagation in a near-ridge setting.
