Seismic reflection profiles from the sediment rich Alaska subduction zone image short, frontally accreted, imbricate thrust slices and repeated sequences of long, underthrust sheets. Rapid landward increases in wedge thickness, backthrusting, and uplift of the forearc are observed, suggesting underthrusting beneath the wedge. These features and a widely varying frontal wedge morphology are interpreted to be caused by different modes of accretion active concurrently along the trench at different locations. Episodic wedge growth is observed in high basal friction experiments using sand as an analog material. Two phases of an accretionary cycle can be distinguished: frontal accretion of short imbricate thrust slices, alternating with underthrusting of long, undeformed sheets. The phase is shown experimentally to depend upon the surface slope of the wedge. Mechanical analysis of the forces at work predicts these two modes of deformation due to the varying frictional forces and yield strengths for a temporally varying wedge geometry. Maximum length of thrust slices is calculated for experimental conditions and confirmed by the observations. For a steep frontal slope (at the upper limit of the Mohr‐Coulomb taper stability field) the overburden is too great to permit underthrusting, and failure occurs repeatedly at the wedge front producing short imbricate slices. The wedge grows forward, lowering the surface angle to the minimum critical taper. For a shallow frontal slope the reduced overburden along an active roof thrust permits sustained underthrusting, causing frontal erosion and backthrusting, steepening the wedge and thus completing the cycle.
The occurrence and preservation of aragonite in eclogite-facies rocks of north-eastern Corsica is linked to an uncommon microtexture. Aragonite exclusively occurs as oriented fibres in garnet crystals of a graphitic, more or less siliceous marble that immediately overlies a serpentinite body of the meta-ophiolitic unit. The arrangement of the fibres is grossly radial, but more clearly sectoral in subhedral garnet, the fibres growing perpendicular to the garnet/matrix interface. Raman mapping reveals that the carbonate is calcite in the matrix and in poikiloblastic garnet cores, and that the fibres in the garnet mantle are aragonite alone in the case of a carbonate matrix, and both aragonite and quartz (in distinct fibres) in a quartz–carbonate matrix. These features are interpreted as prograde intergrowths, the result of garnet nucleation and growth in a calcite and then aragonite matrix (±quartz). Upon further heating and/or decompression, the aragonite matrix transformed back to calcite while the carbonates included in garnet retained their original structure, in spite of the relatively high temperature attained (ca. 500 °C). These aragonite relics are one more example of the preservation of a high-pressure polymorph through mechanical shielding of inclusions in a rigid host. The aragonite–garnet intergrowths are similar to quartz–garnet intergrowths described in amphibolite-facies graphitic schists. They are evidence that oriented inclusions in garnet are not necessarily precipitates (‘exsolutions'). Unlike precipitates, their orientation is controlled more by the shape of the garnet growth front than by symmetry constraints.
