Seeking to understand why the cubic centre in MgO:Cr3+ has the same 10Dq value as emerald, ab initio cluster and periodic supercell calculations have been performed. It is found that the equilibrium Cr3+-O2- distance, R, in MgO:Cr3+ is equal to 2.03 Å and thus 0.06 Å higher than that measured for the emerald. Calculations carried out on the isolated CrO69- complex at R = 2.03 Å give 10Dq = 14 510 cm-1, which is 10% smaller than the experimental figure for MgO:Cr3+. Nevertheless, when the internal electric field, ER(r), due to the rest of the lattice ions is also taken into account, the calculated 10Dq = 16 210 cm-1 coincides with the experimental value. Accordingly, the colour shift for different oxides doped with Cr3+ can be well understood on the basis of this extrinsic contribution to 10Dq usually ignored in a ligand field description. The calculated electrostatic potential, VR(r), related to ER(r), is found to be attractive when the electronic density is lying along 〈110〉 directions and |r| > 1 Å driven by the first shell of twelve Mg2+ ions. The action of VR(r) upon the CrO69- complex slightly decreases the energy of t2g(xy, xz, yz) orbitals with respect to that for eg(3z2-r2, x2-y2) orbitals, thus enhancing the 10 D q value by 0.2 eV. However, the addition of VR(r) induces very small changes in the electronic density, a relevant fact that is related to the 2E(t2g3)→4A2(t2g3) emission energy being nearly independent of the host lattice along the series of Cr3+-doped oxides.