In P art A, a brief outline is given of the 2 principal ways of determining force constants, the method of force constant refinement and the method of force constant display, with particular reference to the assessment of error in the frequency data employed. If D ennison's R ule is assumed to be valid, small isotopic frequency shifts may be almost unaffected by ignorance of the anharmonicity, and in cases where such shifts arise from substitution at an atom in a bridge position, such shifts may be as useful or even more useful than data on Coriolis coefficients in defining the force field. This directly contradicts an earlier conclusion by D uncan and M ills[1]. However isotopic shifts resulting from substitution at a terminal atom, as in deuteration, are likely to be so influenced by uncertainty in anharmonicity as to be virtually useless for the purpose. In the utilization of small isotopic shifts both the force constant refinement and force constant display methods require to be modified so as to calculate the shifts Δν on each mode, these being then compared with the experimental values of the shifts. In P art B, the method of small isotope shifts is applied to CF 4, for which purpose data on ν 3 and ν 4 of 12CF 4 and 13CF 4 are reported. Due to a slight Fermi resonance between ν 3 and 2ν 4 the shift Δν 3 carries an uncertainty of about ±1 cm −1 and this renders it virtually useless for defining the force field. However a shift of 2·5 ± 0·1 cm −1 on ν 4, is far more effective in fixing the force constants than either of the measured Coriolis coefficients. The values obtained after correcting for Fermi resonance and anharmonicity are as follows: ω 3 1311·9 cm −1, ω 4 631·3 cm −1, F 33 6·489, F 34 −0·827, Δω 3 39·4 cm −1, Δω 4 2·5 cm −1, F 44 1·010, mdyn/Å With this force field the intensity data of S churin [18] yield the following values of the bond polar properties: (1) (2) ∂μ/∂ r ±6·38 ±4·62 μ ±0·25 ±1·77