EAGLE Mission: The first-order structure of oceanic rifts is controlled by the supply of magma (Phipps-Morgan & Chen 1993), whereas that of continental rifts is controlled by faulting (e.g. Hayward & Ebinger 1996; Jackson & Blenkinsop 1997). Thus, asthenospheric processes controlling magma supply must come to dominate over lithospheric processes as rifting proceeds to seafloor spreading, but there is no consensus in when or how this transition occurs. Nor has the geochemical evidence for this transition been explored. Reconstructions of some conjugate margins show a lateral offset in the high strain zones within the upper and lower crust which has been interpreted as evidence for lithospheric or crustal-scale detachment faults that accommodate extensional strain until break-up (e.g. Driscoll & Karner 1998). Taylor et al. (1999) and Dunbar and Sawyer (1989) suggest that the geometry of break-up is controlled by pre-existing lithospheric weaknesses. Alternatively, Hebert et al. (2000) suggest that oceanic rifts propagate through the emplacement of discrete mantle upwellings. Models of continental break-up and the initiation of oceanic rift segmentation predict profound differences in 3-D geometry of the crust and upper mantle during break-up, but existing data are inadequate to distinguish between these models. This critical stage of break-up is largely based on interpretations of ancient passive margins where the thermal response of the process has long since decayed away, and the structural and stratigraphic records of two widely-separated continents are deeply buried by thick post-rift sequences. There is a clear need to study this transitional stage in an active rift; such a study is the focus of this proposal. Models of continental break-up show that it is achieved through the combined processes of mechanical weakening of the lithosphere by stretching and intrusive heating, as well as dynamic processes within the asthenosphere (Dunbar & Sawyer 1989; White & McKenzie 1989). What remains unclear is the relative importance of magmatic and dynamic processes in the localization of strain from a >50-km-wide rift to a ~10-km-wide seafloor spreading center marking a new plate boundary. Much of our knowledge comes from successfully rifted margins, which show a sharp rise in Moho topography near the ocean-continent boundary with or without a zone of transitional crust lying between ‘typical’ oceanic and thinned continental crust (e.g. Pickup et al. 1996). Magmatic margins, the focus of the present study, have thicker igneous crust against the rifted margin, and show characteristic seaward-dipping reflectors that mark the location of extrusive magmatic centers (e.g. White & McKenzie 1989; Menzies et al. 2002). Studies of passive margins worldwide have highlighted the ubiquity of a high velocity (7.4 km/s) lower crustal layer beneath the stretched continental crust, interpreted as magmatic underplate (Holbrook & Kelemen 1993). Boutilier and Keen (1999) showed that small-scale convection leads to underplating after break-up, but the volume, distribution and mechanics of magmatic intrusion/underplating prior to break-up remain enigmatic owing to a lack of data from transitional rifts. Differences between models have profound implications for the timing and amount of heat transfer along passive margins. The temporal evolution of a magmatic margin is illustrated in a suite of cross sections of the Afro-Arabian rift system in different stages of development (Figure2). The Asal rift, Djibouti, illustrates styles of lithospheric stretching and magmatic processes immediately after the onset of seafloor spreading (2A). Both geophysical and geological data from the ‘near break-up’ transitional rift (2B) (predicted structure, in part after Ebinger and Casey, 2001) show a very narrow zone of strain and magma injection that is similar to 2A. Young continental rifts, on the other hand, commonly show asymmetric rift basins bounded by steep border faults that accommodate most of the strain across the rift (2C). These snapshots of magmatic margin evolution indicate that strain and magmatism migrate from the border faults to a narrow zone within the rift, suggesting that break-up does not occur along the abandoned border fault (2B). Diking processes accelerate and localize strain, and produce different rift morphologies, depending on the relative importance of faulting and diking. Existing data are inadequate to determine whether detachment fault models of breakup proposed for weakly magmatic margins (e.g. Driscoll & Karner 1998) or dike/magma supply models proposed for oceanic rifts (e.g. Rubin 1992; Poliakov & Buck 1998) are appropriate in magmatic rifts prior to break-up. In the oceanic setting, along-axis mantle flow patterns preserved in olivine crystal alignments predicted by the periodic upwelling models are matched by observations of mantle anisotropy (Blackman & Kendall 1997), but there are few data from continental rifts adequate to assess 3-D upper mantle flow. Our project aims to image 3D variations in crustal thickness and upper-mantle structure, to characterize the distribution of strain and magmatism across a typical transitional rift sector, and to map upper-mantle anisotropy, thereby providing a snapshot of the lithosphere immediately prior to separation (e.g. Figure 2B) so that we may distinguish between proposed models for continental break-up. By integrating the geophysical results with new geochemical evidence on primitive basalt lavas, we will document the critical link between structural and magmatic evolution of the rift system as rifting becomes completed. Our work will improve understanding of continental fragmentation above ancient plumes (e.g. India-Seychelles-Madagascar), and the thermal history of magmatic passive margins (e.g. south Atlantic). In addition to the U.S. experiment team, this study includes three UK universities with complementary expertise: Leicester (controlled-source experiments in Africa); Leeds (seismic networks in Africa and the antipodes) and Royal Holloway University of London (RHUL) (integrated studies of margins and rifts).