Regioselective Halo- and Carbodesilylation of (Trimethylsily1)-1-methylpyrazoles'

an increase in the extent of bridging of the Hg rather than to an increase in the extent of solvation. Analyses of plots of 6AS* vs. u and u+ reveal that in addition to the linear relationships found for solvent A', fairly linear ones also exist for solvent C ( r = 0.956 and 0.947, respectively). The values of p (-1.38) and p+ (-1.21) obtained from the correlation lines are 58% smaller than their corresponding standard values for this solvent. In this solvent, bAH* is not a linear function of bAS* nor u (a+), thus these deviations probably indicate that the linearity of these plots are not significant. In conclusion, these results show that the rates and the activation parameters for this reaction are functions of both the substituent and the solvent. The fact that AG* for each styrene increases as the methanol concentration decreases reflects the destabilizing effect that a reduction in the solvent's polarity has on the polar transition state formed in step 2. However, the fact that AG* for each styrene increases by the same amount in going from one solvent to another reflects the destabilizing effect that a reduction in the solvent's polarity has on the ion pair formed in step 1. Thus solvent effects on both steps are important. Unfortunately, it is not clear which, if either, step is the more sensitive to these effects. A part of the ambiguity arises from the fact that p-dioxane plays dual and competing roles-complexing agent vs. solvating agent-in this reaction. It is our hope that studies in progress, using solvents of the same dielectric constant as p-dioxane which cannot form a complex with the reactive Hg+2 species present in solution, will resolve this matter.