Abstract Using the results of I it is shown that there cannot exist a paramagnetic solution of the (1/z)1 equations for the dipolar and quadrupolar moments in the case of the Blume—Emery—Griffiths model except of purely quadrupolar interaction for S = 3 2 . This is also in contrast to the Heisenberg—biquadratic interactions and MFT. The singular points of the dipolar susceptibility for the Heisenberg—biquadratic interactions are calculated numerically for the sc lattice in the paramagnetic limit. The dipolar susceptibility for the Blume—Emery—Griffiths model is derived in the quadrupolar limit for an arbitrary spin. The zero-temperature case of the dipolar susceptibility in the quadrupolar limit for the Blume—Emery—Griffiths and Heisenberg—biquadratic interactions ( S = 3 2 ) is solved numerically for sc, bcc and fcc lattices. The poles of the quadrupolar susceptibility for the purely quadrupolar Heisenberg—biquadratic interactions and purely quadrupolar Blume—Emery—Griffiths model ( S = 3 2 ) are calculated also for the three cubic lattices. It is stated that there is a small difference of the (1/z)1 theory and the series analysis approach in the location of the critical points for the Ising and Heisenberg models.