PLATE TECTONICS 347 Figure Map of plate boundaries showing the relative velocities across them from the data of DeMets C, Gordon RG, Argus DF, and Steins (1990) Reproduced from Lowrie W (1997) Fundamentals of Geophysics, Fig 1.11 Cambridge: Cambridge University Press circles about Euler poles A great circle at right angles to a small circle passes through the pole Thus, if the azimuths of several transforms along a plate boundary are determined, great circles can be constructed normal to them, and should intersect at the Euler pole (Figure 8) Transform faults are readily recognized by their morphology (for example, a narrow linear valley for ridge–ridge transforms along mid-ocean ridges) In practice, both spreading rates and transform azimuths are used together to solve for pole positions, sometimes supplemented by other data, such as earthquake focal mechanisms, to estimate relative motion directions Euler poles may be calculated for individual plate pairs, for groups of plates, or for the global plate system (Figure 9) The most recent determination of global plate motions was performed by C DeMets and others from Northwestern University, Illinois, USA, and provides a remarkably precise and self-consistent description of these motions (see ‘Further Reading’) As stated above, an important attribute of plate kinematics is its ability to predict plate motions An interesting example of this is the possibility of determining convergence rates across subduction zones Even at ocean–ocean subduction zones, one plate is destroyed, together with the record of magnetic lineations carried on it Thus, there was no direct way of measuring such motion until the recent development of sufficiently precise geodetic methods However, the relative motions of the plate pair can be determined by global fits as described above, and then the motion at any point on the common plate boundary, including subduction zones, can be calculated from the Euler pole data In recent years, geodetic methods have been developed to the level at which they can begin to measure plate motions directly Where plate boundaries exist on land (such as the Mid-Atlantic Ridge in Iceland or the San Andreas Fault in California, USA), standard geodetic methods, such as electronic distance measurement, can be used at a local scale (over ranges of a few kilometres) On a slightly larger scale of tens to hundreds of kilometres, precise relative position determinations (to precisions of a few millimetres) can be made by careful use of the Global Positioning System satellite network Relative positions between widely separated continents can be determined by Very Long Baseline Interferometry, in which the variation in phase of radio signals from distant quasars is used Repeat measurements by these methods over times of a few years can now resolve plate motions, and give results that, in general, agree well with the more traditional determinations The rotation rates of the major plates about their Euler poles range from about 2.1 per million years for the Cocos–Pacific pair, to about 0.1 per million years between Africa and Europe or Africa and Antarctica, and only 0.03 per million years for India–Arabia Many of the minor plates (so-called ‘microplates’) rotate much faster than this, at tens of degrees per million years In terms of linear rates, the fastest plate divergence rate at present is on the East Pacific Rise between the Pacific and Nazca plates, at