Comments on Aksnes' intermediary

The Hamiltonian offered by Aksnes as an intermediary to the main problem in the theory of an artificial satellite is proved to be separable in spherical coordinates. But it is not of Staeckel type, and its latitude equation cannot be integrated by a quadrature. To circumvent the difficulty, a new concept of intermediary is proposed. From this standpoint, Aksnes' theory is revised (i) to simplify a possible automation of Aksnes' literal developments and (ii) to bring forth the connection with the radial intermediary based on a preparatory elimination of the parallax from the main problem.