The crystal‐fields (CFs) causing electron states splittings of the same second moment σ2 can produce different total splitting ΔE magnitudes. Based on the numerical data on CF splittings for the representative sets of CF Hamiltonians ${\cal H}_{\rm CF} = \sum\nolimits_k \sum\nolimits_q B_{kq} C_q^{(k)} $ with fixed indexes either k or q, the potentials leading to the extreme ΔE have been identified. For all CFs the admissible ranges (${\rm \Delta }E_{{\rm min}} $, ${\rm \Delta }E_{{\rm max}} $) have been found numerically for $1 \leq J \leq 8$. The extreme splittings are reached in the CFs for which are the definite superpositions of the $C_q^{(k)} $ components with different rank k = 2, 4 and 6 and the same index q. Apart from few exceptions, the lower limits occur in the axial fields of ${\cal H}_{\rm CF} (q = 0) = \ B_{20} C_0^{(2)} + B_{40} C_0^{(4)} + B_{60} C_0^{(6)} $, whereas the upper limits in the low symmetry fields of ${\cal H}_{\rm CF} (q = 1) = B_{21} \left( {C_1^{(2)} + C_{ - 1}^{(2)} } \right) + B_{41} \left( {C_1^{(4)} + C_{ - 1}^{(4)} } \right) + B_{61} \left( {C_1^{(6)} + C_{ - 1}^{(6)} } \right)$ with real CF parameters. Mixing the components with different q yields a secondary effect and does not determine the extreme splittings. The admissible changes with J from to $2.40\sigma $, whereas the from to $4.10\sigma $. The maximal gap has been found for the states $|J = 4\rangle $. Not all the nominally allowed total splittings, preserving condition, are physically available, and in consequence not all virtual splitting diagrams can be observed in real CFs.