The kinetic and thermodynamic CH3/CD3 isotope effects (k(H)/k(D) and K(H)/K(D)) in the forward and reverse directions of ten nonidentity methyl transfer reactions, X- + CH3Y --> XCH3 + Y-, and the kinetic isotope effects of five identity methyl transfer reactions have been computed ab initio using a 6-31++G(d,p) basis set. The kinetic isotope effect is smaller in the exoergic direction in rive of the nonidentity reactions and larger in the exoergic direction in the other five. Over the entire data set there is no correlation between (k(H)/k(D))XY and either DELTAG(double dagger) or DELTAG. Instead, as in the Marcus treatment of rate-equilibrium relationships, (k(H)/k(D))XY is determined by the kinetic isotope effects of the identity (X,X) and (Y,Y) reactions, by the thermodynamic isotope effect, and by the thermodynamic driving force DELTAG/4DELTAG(double dagger)0, where DELTAG(double dagger)0 is 0.5(DELTAG(double dagger)(XX) + DELTAG(double dagger)(YY)). In a variant of the Marcus treatment, (k(H)/k(D))XY has also been related to the looseness of the transition structure: with all other effects constant, an inverse (k(H)/k(D))XY becomes more inverse as the transition structure loosens. The same trend is seen with identity reactions: a decrease in (k(H)/k(D))XX is associated with loosening of the transition structure, and an increase in the barrier. Taken together, the observations on identity and nonidentity reactions suggest that when an S(N)2 transition structure is altered by variation of the nature of the entering and leaving groups, a decrease in k(H)/k(D) reflects a looser geometry, which leads to inhibition, not catalysis. This raises the possibility that the decrease in k(H)/k(D) which accompanies enzymic catalysis of methyl transfer may be caused by factors other than changes in transition state geometry. The arguments in support of the-idea that k(H)/k(D) decreases in such reactions because of specific compression of the transition state have, therefore, been reexamined. Some problems have been identified.