The System MgSiO3--CaMgSi2O6

Subsolidus and liquidus phase relations along the join MgSiO 3 —CaMgSi 2 O 6 have been determined from the results of dry and hydrothermal runs. Two-pyroxene mixtures which crystallize within the solvus in dry runs form cryptoperthites, and X-ray methods must be used to determine their compositions and locate the boundaries of the solvus. Mg-rich and Ca-rich pyroxenes can, however, be distinguished in well-crystallized hydrothermal runs by optical techniques. The results obtained indicate that solid solution along this join is more restricted than was found by Atlas (1952). The solvus intersects the solidus over a composition interval of 42 wt per cent. Additional data on the rhombic enstatite ⇌ protoenstatite inversion are consistent with Atlas's value of 985°±10°C. This inversion has proved to be very sensitive to pressure, and the dT/dP slope of the transition is 84°±10°/kb. The crystallization of natural pyroxenes from basaltic magmas is reviewed in the light of the experimental data. The solid solution shown by Fe-poor pyroxene pairs from layered intrusions such as the Skaergaard is remarkably restricted. The temperature of crystallization of these pyroxenes as deduced from the solvus on the join MgSiO 3 —CaMgSi 2 O 6 is about 1000° C. In making this estimate it is assumed that any subsolidus exsolution has not proceeded beyond the stage of lamellae formation. This temperature is below the dry solidus of basalt, and the result indicates that either the solid solution shown by these natural pyroxenes is influenced by impurities such as Al or Fe‴, or the melting interval of the magmas which formed the intrusions was lowered by a substantial vapor pressure of H 2 O. It is suggested that pyroxenes which have crystallized from extrusive basalts and which have compositions that plot in the central portion of the pyroxene quadrilateral are metastable solid solutions formed by quick cooling. Available information is not sufficient to clarify the relationship between pigeonite and the various polymorphs of MgSiO 3 . It is probable, however, that there is a first-order inversion between pigeonite and protoenstatite, and possible phase relations between these forms in the system MgSiO 3 —FeSiO 3 are discussed.