Experimental evidence that the reaction of kyanite to form sillimanite is at least bivariant

Mixtures of the sillimanite and kyanite starting materials used in two published investigations were placed side by side at 750 degrees C and 10 + or - 0.5 kb P water (Althaus, 1967; Richardson and others, 1968). After five days, kyanite in one mixture had nearly disappeared while in the other sillimanite had nearly disappeared. The differences must be attributed to the starting materials. The two kyanites were essentially identical, but the sillimanites differ appreciably in Fe 2 O 3 -content, and this is thought to be the most likely cause of the divergent results for the two investigations.