I am looking at page 930-931 Of Riley, Hobson, Bence.
They go from:
e'j,= Sij ei
X'i = (S^{-1})ij Xj
Define L as inverse of matrix S
X'i = Lij Xj, since rotations of coordinate axes are rigid, transformation matrix L is orthogonal, thus the inverse transformation is
Xi = Lji X'j and
Lik...