As silicate perovskite is widely believed to be the dominant phase of
the Earth's lower mantle, the perovskite melting curve represents an
upper bound to the temperature of this region, and places important
constraints on the thermodynamic and rheological properties of the lower
mantle. Again, experimental results reported over the last four years
disagree. Sweeney and Heinz [1993] reported spatially averaged
melt temperatures for 
perovskite between
30 and 94 GPa in a Nd:YAG laser-heated diamond cell.
Their results support earlier studies which concluded that the melting
temperature of perovskite is around 3000 K at 60 GPa, and the extrapolated
melting temperature at the CMB pressure is 4500 
500 K.
In contrast, Zerr and Boehler [1993] measured a melting temperature
of 5000 
200 K in their 
laser heating experiments at 62.5
GPa. Extrapolation to the CMB pressure yields a melting temperature between
7000 and 8500 K. A number of discussions of the relative merits of these
experiments have been reported [ Heinz et al., 1991; Heinz
et al., 1994; Boehler and Zerr, 1994]. The potential
implications of a high melting temperature for perovskite on mantle
dynamics have just begun to be addressed [ van Keken et al., 1994].
Extrapolation of recently measured melting curves (to 32 GPa) of MgO and
(Mg,Fe)O indicates these materials melt at 3500-5000 K at lower mantle
pressures [ Zerr and Boehler, 1994], but molecular dynamics
calculations of MgO melting temperatures are
higher by several thousand Kelvin [ Cohen and Gong, 1994].
The effect of minor components, including volatiles, on the melting
behavior of lower mantle materials has not yet been studied.