“…For example, at > 3∕2, we do not expect to observe any pitch angle diffusion for equatorial electrons [see Delcourt et al, 1995;Young et al, 2008], while at < 3∕2 this diffusion is important (see section 3 and Shibahara et al [2010]). Additionally, in the vicinity of the loss cone (at small pitch angles), the nonadiabatic scattering is nondiffusive (i.e., the averaged jump is not equal to zero and is always positive, see Delcourt et al [1994Delcourt et al [ , 1995), while analytical estimates give only Δ ∼ sin 0 with zero average value for a random uniform distribution of 0 [Howard, 1971;Cohen et al, 1978;Birmingham, 1984;Chirikov, 1987;Varma, 2003]. Moreover, as we discuss in Appendix B, most analytical expressions for Δ should be tested by numerical calculations because these expressions were derived using not well justified methods.…”