Rotation of principal components: A reply