Tectonic activity continually re-paves the oceanic part of the Earth, though not in the manner of the awesome night-time machines seen frequently by owlish drivers as they negotiate the contraflows and cones on highways, large and small.  Slab-pull helps ease plates apart, forcing asthenospheric mantle to rise and partially melt as pressure falls off.  Or, at least that is widely believed, for active mid-ocean processes can only be observed at second-hand through samples scraped from the exposed ridge surface for analysis.  What once lay at the guts of spreading centres emerges only when slabs of ocean lithosphere slide nicely over continental margins because of compressive forces related to plate subduction.  Gravity demands that such obduction is a rare and special process, since oceanic lithosphere is denser than that of continents.  Indeed, as ocean floor ages and cools it become increasingly likely to founder into the deep mantle.  Ophiolites represent oddly buoyant parts of the ocean floor, almost certainly because they were once thermally anomalous or quite young at the time of their emplacement.  There is no guarantee that they represent run-of-the-mill oceanic lithosphere.  However, structures in them, especially a subsurface layer made of innumerable basaltic dykes and little else, show concretely that magmatism was dominated by continual extension; exactly as expected for a former spreading centre.  The most studied ophiolite is that of the Semail Mountains in Oman, which exhibits every definitive layer of lithosphere that point to magmatism in an extensional oceanic environment.  The crustal part is not the best guide to the ophiolite’s genesis, because melt chemistry varies so much with pernickety vagaries of melting and fractionation.  It is the mantle sequence that reveals what went on (Le Mée, L. et al. 2004.  Mantle segmentation along the Oman ophiolite fossil mid-ocean ridge.  Nature, v. 432, p. 167-172).  Laurent Le Mée and colleagues from the University of Nantes focus on chemistry and mineralogy of the well-preserved ultramafic rocks in the Oman ophiolite’s mantle layers.  Their results show how a whole number of petrogenetically important chemical features vary systematically parallel to the original axis of spreading, to define three distinct axial segments.  Within each are other regular fluctuations that define segments of lesser magnitude.  This along-axis chemical variability can be modelled in terms of large variations in the degree of mantle melting (between 10-30%), with the lowest degree coinciding with the major segment boundaries.  Those discontinuities also tally with increased numbers of mantle-cutting dykes (not the crustal sheeted dykes).  Major segments probably formed from regional upwellings of asthenosphere, whereas those with shorter wavelengths reflect individual diapirs.  Along active spreading centres, segmentation of chemical affinities in basalt lavas seems to link with various magnitudes of transform faulting, and it is this local tectonics that shows up so nicely in the Oman mantle sample.