Attribution: You must attribute the work in the manner specified by the author or licensor ( but no in any way that suggests that they endorse you or your use of the work).Noncommercial ‒ you may not use this work for commercial purpose.No Derivative works ‒ You may not alter, transform, or build upon this work.Sharing ‒ Individual scientists are hereby granted permission, without fees or further requests to GSA, to use a single figure, a single table, and/or a brief paragraph of text in other subsequent works and to make unlimited photo copies of items in this journal for noncommercial use in classrooms to further education and science.

We thank Spencer for his comment and welcome the opportunity to debate the results of our study and its implications. We are glad that he supports our principal finding: subhorizontal corrugated surfaces adjacent to the Mid-Atlantic Ridge at 15°45′N are exhumed oceanic detachment faults (MacLeod et al., 2002). That the (low-temperature) fault rocks we recovered from the surfaces imply deformation mechanisms very different from those predicted in previous models (e.g., Spencer, 1999; Tucholke et al., 2001) is indeed intriguing but demonstrates the importance of testing models by direct observation.

Spencer argues that the spreading-parallel corrugations that occur on the surface of the 15°45′N structure (and other core complexes) could not form by the model we propose because it implies the hanging wall is weaker than the footwall, and he asserts that grooves could not form under such circumstances. Instead he proposes the reverse: the hanging wall must be strong and the footwall weak in order for the former to cast grooves into the latter. However, because high-temperature plastic deformation is absent in proximity to the detachment, we can rule out models in which a strong brittle hanging wall casts a weak ductile footwall (Spencer, 1999), at least for the 15°45′N structure (but see below). Spencer makes the additional suggestion that strong gabbro or diabase hanging-wall rocks could cast grooves into a weak serpentinized footwall. We believe this model is also unlikely. Oceanic crust at slow-spreading ridges is now known to be far more heterogeneous than previously recognized: a regular magmatic crustal layer is not always developed, and serpentinized peridotite may form a significant proportion of the geophysically defined “crust” (Cannat, 1993). The Mid-Atlantic Ridge between 14°30′ and ~16°N is the best-known example of magma-starved ocean floor: extensive areas of serpentinized peridotite are capped by thin, discontinuous lava flows (Cannat et al., 1997), indicating efficient serpentinization of the upper lithosphere at the ridge axis. Gabbro and diabase are volumetrically minor. It is therefore highly unlikely that the hanging wall of the 15°45′N detachment was a continuous, strong layer with gabbro/diabase or unserpentinized peridotite rheology; the mechanical behavior of both footwall and hanging wall was probably controlled by the rheology of peridotite and its alteration products. Indeed, where we did map gabbro it was in the form of an isolated pluton intruding serpentinized peridotite in the footwall of the detachment, not the hanging wall. The margins of this strong intrusion do not coincide directly with the highs on the corrugated surface, suggesting that strength contrasts between hanging wall and footwall are not the direct, or at least not the only, cause of groove formation. For grooves to be cast in a weak footwall by a strong hanging wall, as Spencer proposes, serpentinization can only have occurred in the former and not the latter. This is implausible. That our peridotite samples from the footwall are indeed partially or completely serpentinized has no relevance to the question of whether the detachment nucleated at a serpentinization front or not, because serpentinization of the footwall would have started as soon as movement and uplift on the fault commenced. We emphasize that the fault zone material (talc, chlorite ± tremolite schists; Escartín et al., 2003) appears to have been much weaker than either the footwall or hanging wall; this is necessary for strain to have remained localized on a single structure for ~1 m.y. and is demonstrated by undeformed mesh-textured serpentinized peridotites in the footwall very close to the corrugated fault surface (e.g., MacLeod et al., 2002, their Fig. 3C).

From the limited observations we have to date it is clear that oceanic core complexes do not all have to form in the same way. We have also carried out a rock drill survey of the Atlantis Bank core complex (SW Indian Ridge), and find that the surface of this massif is formed predominantly of gabbros deformed under high-temperature plastic- or melt-present conditions and with subhorizontal fabrics (MacLeod et al., 1999). Ocean Drilling Program drilling has shown that deformation diminishes in intensity at deeper levels (e.g., Dick et al., 1991), suggesting not only that a detachment was active in the presence of magmatism, but that it must have rooted into a ductile magmatic lower crust. Such hot detachments must form in response to fundamentally different mechanisms of strain accommodation and localization from cold detachments such as the 15°45′N structure.

In conclusion, although Spencer's continuous-casting model may provide an explanation for the formation of grooves on continental and maybe some hot oceanic core complexes, it is not the only mechanism, and it does not explain how they formed at the cold 15°45′N detachment. Linkage of precursory faults, as we suggested, could generate corrugations, as could differing depths to the serpentinization front resulting from spatial variations in the permeability structure of the upper crust. There are clearly many ways to form irregularities on a fault surface, but only those parallel to the transport direction are stable and have the potential to amplify and be preserved.