Chrysotile fibres, and also single crystals of lizardite, were dehydrated hydrothermally, and the resulting pseudomorphs studied by X-ray rotation photographs. The initial products are normally forsterite and disordered material; talc forms more slowly. At water pressures up to 600 kg/cm2, both forsterite and talc show strong preferred orientation. At higher pressures the talc is oriented but the forsterite tends to be unoriented. The orientation of forsterite formed at 550° C and 500 kg/cm2 differs from that found when chrysotile is dehydrated by heating in air. The orientation relationships suggest that, in the dry reaction, the forsterite orientation is controlled mainly by the octahedral layers of the chrysotile, while in the hydrothermal reaction it is controlled mainly by the tetrahedral layers. This does not support the generally accepted mechanism for the dry reaction. New mechanisms are suggested, for both dry and hydrothermal reactions, similar to that already suggested by us for the dehydration of brucite. The formation of talc in the hydrothermal reaction, and of enstatite in the dry reaction above 1000°, are also discussed. The dehydration of talc, to give enstatite, has also been studied using single crystals, and the mechanism of this process is briefly discussed.