The Mariana forearc region is thought to be non-accretionary (Hussong, Uyeda, et al., 1981; Fryer, Pearce, Stokking, et al., 1992). The differences between the serpentine muds and the serpentinized peridotite clasts discussed earlier indicate additions of "exotic" materials to the serpentinite muds. From the similarities between the serpentinite muds and the serpentinized peridotites (Fig. F3), it is evident that the serpentinized ultramafic materials control the composition of the serpentinite muds. However, the major and trace element geochemistry of the serpentinite muds suggest the admixture of a slab (MORB-like) component. In Figure F15 we have calculated binary mixing arrays between averages of the serpentinized peridotite and metamorphic schist compositions compiled in Table T8. Interestingly, independent of element and ratios chosen, the serpentinite muds always plot on (or very close to) the mixing lines between average serpentinized peridotite and metamorphic schist compositions (Fig. F15). Moreover, their positions on Figure F15 always correspond to additions of 3%–5% metamorphic clast materials to the serpentinized peridotites.
The metamorphic schists incorporated in the serpentinite muds originated from a basaltic protolith, either from a trapped Philippine Sea plate or, more possibly, from the subducting Pacific slab. If the samples have Pacific plate origin, then they will be representative of the metamorphic pressure and temperature conditions existing at the slab/mantle interface and will thus be appropriate for testing the current models for mantle wedge pressure-temperature conditions postulated by Peacock (1996) and Peacock and Wang (1999). By contrast, the presence of trapped Philippine sea crust in the Mariana forearc (see De Bari et al., 1999) metamorphosed to HP/LT conditions implies that this crust was actively involved in the proto-Mariana intra-arc rifting and then tectonically emplaced to considerable depths (high pressures) in the Mariana forearc mantle. Philippine plate origin scenarios are hard to explain with the existing velocity structure of the forearc region. Hf and Nd isotopes, which are insensitive to alteration and low-temperature metamorphism, may be used in the future to distinguish between these two potential sources, as was successfully done for western Pacific samples by Pearce et al. (1999).