“…The mild behavior of the mean anomaly is due to the additional term in (40) for the secular part of mean motion n ˚when compared with the usual secular rate in (42). We checked that if the later is used instead of the former, the error in the mean anomaly grows by about two orders of magnitude at the end of the one-day interval in the current example, reaching an amplitude close to one degree, which translates into about two hundred km along-track as opposed to the km level obtained when using n For a second example we take a low-eccentricity orbit with the initial conditions a " 7707.27 km, e " 0.01, I " 63.4 ˝, Ω " 180 ˝, ω " 270 ˝, M " 0, (58) corresponding to the configuration of the popular Topex orbit [44], and the propagation is likewise carried out up to one day, which now amounts to about 13 orbits. We checked that the agreement between the mean elements dynamics and the average dynamics of the true orbit preserves in each case an analogous number of digits to the previous example.…”