The Romanche transform offsets the Mid-Atlantic Ridge (MAR) axis by about 950 km in the equatorial Atlantic. Multibeam and high-resolution mulitchannel seismic reflection surveys as well as rock sampling were carried out on the eastern part of the transform with the R/V Akademik Strakhov as part of the Russian-Italian Mid-Atlantic Ridge Project (PRIMAR). Morphobathymetric data show the existence on the northern side of the transform of a major 800-km-long aseismic valley oriented 10¿ to 15¿ from the active valley; it disappears about 150 km from the western MAR segment. The aseismic valley marks probably the former location of the Romanche transform (''PaleoRomanche'') that was active up to roughly 8--10 Ma, when the transform boundary migrated to its present position. A temporary microplate developed during the migration and reorientation of the transform. This microplate changed its sense of motion as it was transferred from the South American to the African plate. A prominent transverse ridge extends for several hundred kilometers parallel to the transform on its northern side, reaching its shallowest part (shallower by over 4 km than the predicted thermal contraction depth) in a zone opposite the eastern MAR axis/transform intersection (RTI). Flat-top peaks on the summit of the transverse ridge are capped by acoustically transparent, weakly stratified, shallow water platform/lagunal/reef limestones. This limestone unit is a few hundred meters thick and overlies igneous basement.

Evaluation of the seismic reflection data as well as study of samples of carbonates, ventifact basaltic pebbles and gabbroic, peridotitic, and basaltic rocks recovered at different sites on the transverse ridge, suggest that (1) the summit of the transverse ridge was above sea level at and before about 5 Ma; (2) the transverse ridge subsided since then at an average rate 1 order of magnitude faster than the predicted thermal contraction rate; its summit was flattened by erosion at sea level during subsidence; (3) the transverse ridge is an uplifted sliver of lithosphere and not a volcanic constructional feature; and (4) transtensional and transpressional tectonics have affected the transverse ridge. Hypotheses on the origin of the Romanche transverse ridge include (1) lateral heat conduction across the RTI; (2) shear heating; (3) lithospheric flexure due to thermal stresses in the cooling lithosphere; (4) viscoelastic deformation of the lithosphere; (5) hydration/dehydration of mantle peridotites; and (6) longitudinal flow of melt and igneous activity across the RTI. These processes cannot by themselves explain the transverse ridge, although some of them could contribute to its formation to a small extent.

Vertical tectonics due to transpressional and transtensional events related to a nonstraight transform boundary and to regional changes in ridge/transform geometry is probably the primary process that gave rise to the uplift of the transverse ridge and to its recent subsidence. Uplift may have been caused primarily by thrust faulting induced by transpression related to the oblique impact of the lithospheric plate against the former (PaleoRomanche) and the younger transform boundaries, before and during the transition to the present boundary. After migration of the transform boundary to its present position, transpression was replaced by transtension and by subsidence of the transverse ridge. An aseismic axial rift valley impacting against the transform valley about 80 km west of the present RTI suggests eastward ridge jumping that probably followed transform migration. Localized transtension or transpression due to bends in the orientation of the transform may have caused intense although localized vertical movements, such as those that formed an ultradeep (>7800 m) pull-apart basin along the transform valley. ¿ American Geophysical Union 1994