The procedure of fitting an ellipsoid to vitrinite reflectance anisotropy is described. Several authors on this subject use incorrect fitting formulae. The correct formulae are given and a least- squares procedure is developed to give confidence regions for the principal reflectances and their orientations. A FORTRAN program is offered that computes the principal reflectances using the correct formulae. A test is constructed for whether the principal reflectance orientations are the same at two locations. Because of the ellipsoid shape, the reflectance anisotropy can be compared to the strain ellipsoid. As an example, the methods are used to compare two coal blocks from an openpit mine in Alberta, Canada. The test shows that the principal reflectances of the two blocks have different orientations, indicating that the orientation of the principal strain axes is different at the two sites.