Although sea floor spreading of some sort has created marginal basins, it is not known whether this spreading is similar to that at midocean ridges. Back arc spreading in the Mariana Trough occurs along an axial topographic high, but there is rapid variation in style along strike and less well defined symmetry than there is at midocean ridges. The topography is very rough, despite an apparently fast spreading rate. This lack of correlation between spreading rate and topographic relief also occurs in other marginal basins of the western Pacific. In addition, the apparently different curvatures between the frontal and remnant arcs of some marginal basins, such as the Mariana Trough, might result from nonrigid deformation of the frontal arc. There is no single finite pole of rotation about which the Mariana trough can be closed without seriously violating observed fracture zone directions and the geology at both ends of the arc. South of 14¿N the frontal arc is truncated by the east-west fractures so that marginal basin crust is juxtaposed against the Pacific plate near the possible triple junction between the Philippine and the Pacific plates and the Mariana platelet. The present data suggest that the Mariana Trough has opened as if the frontal arc were a horizontal beam pinned at its northern end and deforming laterally so that fracture zones parallel to the spreading direction have themselves undergone minor extension. That is, the Mariana Trough has pulled loose from the Philippine plate at its southern end and has been bent and fractured so that extension has occurred not only parallel but also perpendicular to the direction of principal rotation.