-phase
transformation in (Mg,Fe)
SiO
. These authors
suggested that the instability associated with phase
transformations in subducting slabs may cause deep
earthquakes. This hypothesis implies that the phase
transformations from olivine to 
-phase or olivine to
-phase in the subducting slabs occur at much higher
pressures than the equilibrium situation due to the low
temperatures. However, the prediction of the depth extent to
which metastable olivine might be present is difficult due to
the large uncertainties in the temperatures in subducting
lithosphere and the kinetic parameters (Rubie and Ross,
1994). In addition, the effects of strain rate (time-scale)
on shear instability associated with phase transformations
have not been studied. This leads to a large uncertainty in
the application of laboratory results to the earth.
Furthermore, the focal mechanism of recent Bolivia earthquake
indicates a very large fault plane (Kikuchi and Kanamori,
1994) which is difficult to be compatible with a simple
metastable olivine model for deep earthquakes which predicts
a thin metastable tongue at this depth (e.g., Kirby et al.,
1991). An alternative model for deep (and intermediate)
earthquakes has also been proposed by Ogawa (1987) and Hobbs
et al. (1988). They proposed that a positive feedback due to
heat generation during deformation might cause melting that
could lead to faulting under high pressures. Such a process
provides a possible mechanism for intermediate earthquakes
that occur within a lithosphere (not at the interface between
lithosphere and surrounding mantle), for which mechanisms
involving dehydration or phase transformations do not
apply.