Ozone produced by discharge or photolysis of oxygen has unusually heavy isotopic composition ((18)O/(16)O and (17)O/(16)O ratio) which does not follow normal mass fractionation rule: δ(17)O ∼ 0.52(*)δ(18)O, expressed as an anomaly Δ(17)O = δ(17)O - 0.52(*)δ(18)O. Ozone molecule being an open isosceles triangle can have the heavy isotope located either in its apex or symmetric (s) position or the base or asymmetric (as) position. Correspondingly, one can define positional isotopic enrichment, written as δ(18)O (s) or δ(18)O (as) (and similarly for δ(17)O) as well as position dependent isotope anomaly Δ(17)O (s) and Δ(17)O (as). Marcus and co-workers have proposed a semi-empirical model based in principle on the RRKM model of uni-molecular dissociation but with slight modification (departure from statistical randomness assumption for symmetrical molecules) which explains many features of ozone isotopic enrichment. This model predicts that the bulk isotope anomaly is contained wholly in the asymmetric position and the Δ(17)O (s) is zero. Consequently, Δ(17)O (as) = 1.5 (*) Δ(17)O (bulk) (named here simply as the "1.5 rule") which has been experimentally confirmed over a range of isotopic enrichment. We now show that a critical re-analysis of the earlier experimental data demonstrates a small but significant departure from this 1.5 rule at the highest and lowest levels of enrichments. This departure provides the first experimental proof that the dynamics of ozone formation differs from a statistical model constrained only by restriction of symmetry. We speculate over some possible causes for the departure.