Ocean island chains are trackways of moving lithospheric plates relative to the underlying mantle. Mantle hotspots act in a similar manner to a candle that would burn a line in a sheet of paper were one to be passed over it. The largest, most coherent and best studied ocean island chain is that of the Hawaiian Islands and the Emperor Seamounts  in the NW Pacific. The volcanoes that built the chain range in age continuously from Late Cretaceous (81 Ma) at the northern tip of the Emperor Seamounts where they touch the Kamchatka Peninsula to the present in the Big Island of Hawai’i itself. So far, so good for the hotspot-track hypothesis. But the chain is bent into a WNW segment (Hawaii) and one that trends NNW (Emperor). That might seem to be superb evidence that the direction of West Pacific sea-floor spreading underwent a sudden, 60º change around 47 Ma (the age of the Diakakuji seamount at the apex of the bend). However, measurements in 2001 of palaeomagnetic latitude in sea-floor cores along the chain revealed clear palaeomagnetic evidence that the Hawaiian hot spot has not always been fixed relative to moving lithospheric plates.  From Late Cretaceous to Late Eocene times the hotspot seems to have been was shifting southwards relative to the north magnetic pole at a rate comparable with that of sea-floor spreading, and then became stationary to explain the 60° bend in the chain (See American Geophysical Union 2001 Fall Meeting in EPN for January 2002).

Further work has been done since 2001, and a review of the huge oddity that bucks John Tuzo Wilson’s 1963 theory of hotspots fixed in space and time is timely (Tarduno, J. et al. 2009. The bent Hawaiian-Emperor hotspot track: inheriting the mantle wind. Science, v. 324, p. 50-53). Data have moved on to suggest that the hotspot is indeed the head of narrow mantle plume originating deep down, perhaps even near the core – mantle boundary (CMB). But could such a massive structure change it’s behaviour so that its head would move? Some have suggested the development of a propagating crack in the Pacific lithosphere and then its closure, but no evidence points unerringly that way. After considering a range of possible mechanisms, the authors suggest that the great bend records past changes in mantle flow beneath the West Pacific, so that the plume would itself have bent in the vertical dimension. Seismic tomography has revealed apparently low-angled zones of hot, low-velocity mantle, such as one that may (or may not) connect with the Afar plume beneath the triple junction of the East African Rift, the Red Sea and the Gulf of Aden after rising from the CMB south of Cape Town. They are tantalising results, because the resolution is simply not good enough to be sure. It needs an order of magnitude better tomographic resolution of mantle features to truly make more headway.