The problem of an elliptical rigid inclusion in an elastic solid, with homogeneous stress at a large distance, is solved. The greatest stress differences are found to be at or near the ends of the flaw and to be determined mainly by the greatest principal stress. The result differs from that of G. I. Taylor for flaws filled with compressible but weak material, where the greatest stress-differences, for similar but differently oriented flaws, would occur for flaws at 45 degrees to the greatest principal stress. The result would suggest that in dynamic metamorphism of rocks weak crystals would tend to form at 45 degrees and strong ones would tend to form needles parallel to the greatest principal stress. Doubts are expressed about this interpretation, and an alternative is suggested, namely, that stress has little to do with the formation of minerals, but that they are sheared and rotated into parallel directions by flow in the surrounding rock after they are formed.