holds true for all n ≥ 2 and s > 1 (cf. [2]). This is a standard argument that is used
in several text books on analytic number theory, for instance, in connection with the
Riemann zeta-function (cf. [1]). The authors of the present note, who are concerned
with lectures on number theory for undergraduates, discussed the problem of how to
derive an upper bound for
∞
k=n k−s better than the one given in (1) by using only
elementary arguments. The first result of these discussions is the following