Reply to Comment on: Diffusion theory of multidimensional activated rate processes: The role of anisotropy

The problem of two dimensional overdamped anisotropic diffusion is governed by two small parameters, (i) the thermal energy e=kBT/AV, where AV is a reference activation energy (e.g., the height of the saddle point above the bottom of the reactant well), and (ii) the anisotropy parameter S = qx /q,, where qx and vY are the two damping coefficients (assuming the friction tensor is diagonal). Therefore the two limits, (a) first e-0, then S-+0, and (b) first S -0, then e+O, must be considered separately, because it is not a priori clear that they are interchangeable. Indeed, there are cases when they are not, as correctly pointed out in Ref. 2. It should be pointed out however that the analysis presented in Ref. 1 is concerned with the limit (a). The limit (b) is considered there only for the case A > 0, for which the limits (a) and (b) are indeed interchangeable (here A = V, at the saddle point). The results for the case A&O in Ref. 1 are valid only in the limit (a), so that the comment “the cases A<0 can be handled in a similar manner” is misleading, as correctly pointed out in Ref. 2. Unfortunately, some of the statements, as well as the result