This paper describes a (non-empirical) Hartree-Fock-Slater (Xα) LCAO study of the Mössbauer parameters in a series of thirteen Sb compounds: SbX 3(X = F, Cl, Br, I), Sb(CH 3) N Cl 3- N ( N = 1,2,3), SbX 5(X = F, Cl) and Sb(CH 3) 3X 2(X = F, Cl, Br, I), in relation with the chemical bonding in these compounds. The calculated isomer shifts agree very well with the experimental data: correlation coefficient 0.99, Δ R/ R = −1.08×10 −3. The quadrupole splitting for the Sb(III) compounds is consistently (correlation coefficient 0.97) too small by a factor of 2.7, for the Sb(V) compounds the results are more scattered and even smaller. The proportionality factor could be partly due to the uncertainty in the 121Sb nuclear quadrupole moment and partly to the neglect of core polarization effects in the (frozen core) HFS LCAO calculations. Test calculations on HCl, which are also reported in this paper, show that the HFS LCAO method can yield accurate core polarization corrections to the quadrupole splitting (Sternheimer factors), but only at the cost of an expensive numerical integration scheme. The results suggest further that the structures of some of the Sb compounds might be different from what has previously been proposed.