Fabrics present in the serpentinized peridotite breccia are identical on both mesoscopic and microscopic scales (Pl. 2, Pl. 3, Pl. 4). The breccia is cohesive, disorganized, and highly angular, with a porphyroclastic texture. Except for few samples, no preferred orientation of clasts is observed, and a randomly oriented fabric prevails. Fragment and particle size ranges from decimetric to microscopic. The fragments can be fitted together, indicating no rotation or shear between fragments. Frequently, fragments of serpentinite cannot be clearly distinguished from the matrix or groundmass, which suggests that the matrix is generated by comminution and alteration of the serpentinized peridotite porphyroclasts. Porphyroclast sections show sharp angular edges and are frequently grouped as obviously derived from a single porphyroclast (Fig. 4). In some samples porphyroclasts are formed from deformed serpentinites or former serpentine breccia. Polyphase cracks are pervasive and can be clearly observed throughout the breccia intervals.
On the basis of the aforementioned fabric characteristics and of the specific microscopic textures and structures described below, we consider the serpentinite peridotite breccia to be a cataclasite (Twiss and Moores, 1992; Marshak and Mitra, 1988).
Serpentinized peridotite breccias and cataclasites are extensively veined on meso- and microscopic scale (Pl. 2). Vein infilling is made up of carbonates (mainly low-Mg calcite) and fibrous serpentine. This serpentine veining, similar to that encountered in the calcite-free serpentinized peridotite breccia at Site 897 (from Cores 149-978C-66R to 70R, and from Cores 149-897D-16R to 20R; Shipboard Scientific Party, 1994a), is abundant inside serpentinized peridotite boulders, clasts and fragments. It may also coexist with calcite in thicker veins that bound bigger serpentinite boulders within the breccia (Pl. 2, Fig. 6). Porphyroclasts that include inherited serpentine veins are common (Pl. 2, Figs. 1–3).
Calcite veining is higher in the upper part of the breccia sequence (Upper Breccia in Unit IV, Shipboard Scientific Party, 1994b) and increases dramatically in breccia clasts included among the sediments at the bottom of the recovered sequence (from Core 149-899B-26R downhole).
We observed three main types of vein geometry:
1. Roughly orthogonal vein sets cross-cutting porphyroclasts and matrix (Pl. 2, Figs. 1, 3, 4).
2. Planar crack-like thick veins. These veins usually develop anti-axial fillings that show several screens of wall rock from multiple crack-seal events, and large lateral branches (Pl. 2, Figs. 4, 6).
3. A complex network of veins bounding angular porphyroclasts or breccia fragments (Pl. 2, Figs. 5, 6). This network is similar to the jigsaw puzzle pattern considered diagnostic of hydraulic implosion brecciation (Sibson, 1986).
Calcite vein fillings are granular and/or spatular and possess subhedral to anhedral crystals. When crystal size increases toward the vein center, euhedral forms may occur. Crystal orientation developed perpendicular to the vein walls, thereby indicating that veins formed as tensile fractures (Pl. 2, Figs. 4, 6). Neocrystallized calcite in the breccias is largely confined to veins, and very few porphyroclasts are exclusively composed of neocrystallized calcite. This suggests that episodes of formation of thick calcite veins postdate the main processes of brecciation. Samples containing up to 40% of vein filling exhibit significant alteration and calcite replacement of the host rock, suggesting fluid-rock chemical reactions.
These three types of veins are mutually cross cutting (Pl. 2), and are frequently offset by microfaults. Vein geometry is not affected by the fabric of the wall rock, indicating that the breccia, matrix, and porphyroclasts were cohesive at times of vein formation. The structure of both the serpentine- and calcite-filling provides evidence that the opening of tensional fractures was rapid relative to filling and that fluid-assisted processes produced brecciation.
Microscope studies on selected samples from Site 899 point to a two-stage, polyphase deformational history (Pl. 3, Pl. 4) in the serpentinized peridotite breccias. We recognized a first stage of deformation (stage D) in the porphyroclasts, and, in the whole breccia, identified a second stage, F, that includes three phases (F1, F2, and F3) and yields the cataclastic fabrics. Deformation phases were ordered on the basis of the overprinting of structures and relationships of mineral growth, with the structures of the various phases of deformation providing insight into deformation mechanisms.
The first stage of deformation (D) resulted in serpentinization, structures of pervasive shear foliation, and the coeval development of serpentine veins. Mylonite and cataclasite textures in serpentinite clasts (Pl. 3, Figs. 1, 4), and serpentine veins also restricted to clasts (Pl. 2, Figs. 1–3), are representative of stage D deformation.
Basement samples from Holes 897C and 897D have CS structures and rotational fabrics interpreted as stage D deformation. The penetrative deformation, including the fracture cleavage, in serpentinized peridotites from Hole 897C is also interpreted to have formed during stage D (e.g., Samples 149-897C-71R-2, 22-23 cm, and 71R-2, 8–17 cm; Shipboard Scientific Party, 1994a). The discontinuous or diffuse veins filled with dark serpentinite or white chrysotile, as well as the cataclasites observed in the lower sections of Holes 897C and 897D (below 677 mbsf in Hole 897C and below 780 mbsf in Hole 897D), are also considered to be stage D deformation (Fig. 5A).
Shear foliation structures indicate a low-temperature ductile-brittle deformation of the serpentinite during stage D. Assuming a hypothetical serpentinite composed of antigorite-chrysotile-lizardite, this ductile-brittle field is thought to have developed in P-T conditions of 0.2–0.4 GPa and below 350°–500 °C (Wicks, 1984).
The second deformation stage (F) reveals at least three main phases of brittle deformation (F1, F2, and F3). The F1 and F2 phases of deformation resulted in structures of cataclastic brecciation (Pl. 3 and Pl. 4) and veining (Pl. 2). The F3 deformation phase resulted in pervasive brecciation and calcite vein filling.
The F1 deformation phase resulted in serpentinization and penetrative cataclastic fracturing, with the matrix formed exclusively by comminuted serpentine grains (Pl. 4; Pl. 3, Figs. 2–6; Pl. 4). Associated with the F1 deformation phase, a second generation of scarce serpentine veins developed, cross-cutting the porphyroclasts and matrix. The F1 phase cataclastic structures comprise cataclastic foliation (Pl. 3, Fig. 5) (Chester et al., 1985), cataclastic flows (see flow structures, Pl. 4, Fig. 4), and cataclastic lineation (Pl. 4, Fig. 6). To better define the cataclastic lineation (Tanaka, 1992), strain markers from oriented sections of samples from Leg 149 would have been necessary. Within the porphyroclasts, however, we recognized all the phases of cataclastic evolution (Pl. 3, Figs. 5, 6; Pl. 4, Figs. 3–5). Matrix intrusion into open fractures is common for the cataclastic lineations in the Site 899 samples, thus indicating fluid-assisted processes. Fluids favor cataclastic deformation by reducing the effective confining pressure, which leads to strain softening and hosted cataclastic flow. Cataclastic flows develop through successive cycles of softening and hardening, followed by veining and, most likely, stress corrosion and pressure solution (Babaie et al., 1991; Blenkinsop and Sibson, 1992).
The F2 deformation phase resulted in renewed cataclastic fracturing with a matrix formed of serpentine and calcite, associated microfaults of small displacement, and cataclastic flows. No cataclastic fo-
liation or lineation was observed for this deformation phase (Pl. 3, Figs. 2, 3).
The F3 deformation phase shows no cataclastic features; no cataclastic lineation or flow structures have been observed. It involves pervasive fracture and veins superimposed on preexisting F1 and F2 phase fabrics, resulting in restocked brecciation and fractures (veins cross-cutting matrix and clasts) filled by calcite and subordinate serpentine (Pl. 2, Figs. 3–6; Pl. 3, Fig. 3).
One of the main conclusions from our fabric studies is that two major cataclastic events are recorded in the serpentinized peridotite breccias:
1. Cataclasis from stage D deformation. Fabrics from this event exist only inside clasts, boulders and intervals of serpentinized peridotite. The resulting cataclasites are associated with mylonite textures, shear foliation, and calcite-free serpentine veining.
2. Cataclasis from stage F deformation. This later, multiphasic, event resulted in serpentinization and penetrative cataclastic brecciation associated with extensive calcite and scarce serpentine veining.
Cataclasites from crystalline rocks, with fabrics similar to those of the serpentinized peridotite breccias recovered at Sites 897 and 899, are known to be formed by brittle faulting and fracturing in different tectonic settings: (1) wrench tectonics (Blenkinsop and Sibson, 1992; Tanaka, 1992; Chester et al. 1993), (2) thrusting in accretionary prisms or nappe-stacking (House and Gray, 1982), and, (3) extensional faulting (Malavieille, 1993). However, the most clear-cut examples of cataclasites have been reported from highly seismic "weak" transform fault zones (e.g., the San Andreas Fault System in southern California). In these transform zones, great amounts of overpressured fluids are found along the fault zones, and cataclasites are associated with large fracture-damage zones (Chester and Logan, 1987; Blanpied et al., 1992; Chester et al, 1993).
Well-defined examples of cataclastic fabrics in serpentinite rocks are scarce in the current literature (e.g., Hoogerduijn Strating and Vissers, 1994). Nevertheless, as the P-T conditions for brittle behavior of serpentinite rocks are below 0.15 GPa and less than 350°C (Raleigh and Paterson, 1965; Wicks, 1984), faulting below these conditions may result in serpentinite cataclasites.
In order to characterize the fault zone where the serpentinized peridotite breccias that were recovered during Leg 149 would have originated, we include examples of tectonic cataclasites from the subcontinental peridotites of the Ronda Massif (Betic Cordillera, Spain). The hand-specimen samples shown in Fig. 5B and Pl. 4, Fig. 2, belong to tabular decametric bodies of serpentinized peridotite breccias cropping out in a brittle fault zone to the south of the Ronda Massif (Sánchez-Gómez et al., 1995). This serpentinized peridotite breccia has cataclastic fabrics identical to those of the serpentinite breccias from Sites 897 and 899, and comprises a similar wide range of clast sizes; it has also recorded several phases of cataclastic deformation (Pl. 4, Fig. 2). At the outcrop, the cataclasite zone includes elements of undamaged peridotite up to 1 m in diameter that make up more than 30% of the entire rock. High-displacement narrow shear zones bound cataclasite elements, less damaged serpentinite blocks, and peridotite gouges. This assemblage of fault rocks can be observed in a zone about 200 m wide.