The origin of the bend in the Hawaiian-Emperor Island seamount chain is the prototypical problem at the heart of the ongoing hotspot debate: Does the bend represent a change in motion between the Pacific plate and underlying hotspot (plume?), or might it represent a change in the intraplate stress field?
Reconstructions of continental plates relative to the Pacific plate produce loci with changes in motion that correspond in age and geometry with the bend, especially the North and South American plates. This implies that the two continental plates moved very slowly relative to the Pacific hotspots, especially between 75 and 30 Ma, and that the Hawaiian-Emperor bend is the result of change of Pacific plate motion relative to the underlying hotspot reference frame (or vice versa). Additional reconstructions imply that the Pacific hotspot frame is distinct from the Atlantic and Indian Ocean hotspots over the same time period.
Passage of some hotspot traces from one plate to another suggests that the hotspots are at least partially sublithospheric phenomena and, therefore, cannot simply represent fractures propagated by intraplate stress fields. Contemporary and paleostress fields correspond to hotspot kinematic models for at least three continental plates. If the stress fields represent drag of the continental plates against the underlying asthenosphere and mesosphere, this suggests that the sub-lithospheric hotspots are embedded within the mesosphere (and do not move independent of it); both hotspot traces and stresses record the same motion.
The correspondence of North and South American and Hawaiian hotspot motions, combined with the evidence for embedding of the hotspots in the mesosphere, leads to the inference that motion of the two continental plates (along with the Eurasian) constrains the motion of the sub-Pacific mesosphere. Thus, the Pacific and Atlantic-Indian hotspot reference frames are plausibly independent and separated by the deep circum-Pacific subduction zones.