Exercise 38, Chatper 8 of Apostol's Mathematical Analysis. Given a monotonic sequence $\{a_n\}$ of positive terms, such that $\lim_{n\to\infty}a_n=0$. Let $$s_n=\sum_{k=1}^na_k,\quad u_n-\sum_{k=1}^n(-1)^ka_k,\quad v_n=\sum_{k=1}^n(-1)^ks_k.$$

Prove that:

a) $v_n=\frac12u_n+(-1)^ns_n/2$.

I've shown this without any difficulty.

b) $\sum_{n=1}^\infty(-1)^ns_n$ is $(C,1)$ summable and has Cesaro sum $\frac12\sum_{n=1}^\infty (-1)^na_n$.

My attempt so far: By the conditions on $\{a_n\}$, $\sum_{n=1}^\infty(-1)^na_n$ converges by the alternating series test. Let $s=\sum_{n=1}^\infty(-1)^ka_k$. Note $\lim_{n\to\infty} \frac1n\sum_{k=1}^nu_k=s$, by Cesaro's theorem. Let $\sigma_n=\frac1n\sum_{k=1}^nv_k$. By part a)


So, $$\lim_{n\to\infty}\frac1n\left(\sum_{k=1}^{n-1}v_k+\frac12v_n\right)=\lim_{n\to\infty}\frac1{2n}\sum_{k=1}^nu_k=\frac12s.$$

So, I need to show that $$\lim_{n\to\infty}\frac1n\left(\sum_{k=1}^{n-1}v_k+\frac12v_n\right)=\lim_{n\to\infty}\frac1n\sum_{k=1}^\infty v_k$$.

This seems reasonable enough, but I'm not finding a way to do it.


Since $a_n \to 0$, so does $A_n:= \frac{s_n}{n} \to 0$. Since

$$v_{2n} = s_2 - s_1 + s_4 - s_3 + \cdots + s_{2n} - s_{2n-1} = a_2 + a_4 + \cdots + a_{2n},$$

we have

$$\frac{v_{2n}}{2n} < A_{2n} \to 0.$$


$$v_{2n+1} = v_{2n} - s_{2n+1} = -a_1 - a_3 - \cdots - a_{2n+1}.$$


$$\left|\frac{v_{2n+1}}{2n+1}\right| < A_{2n+1} \to 0.$$


$$\lim_{n\to \infty} \frac{v_n}{n} = 0.$$


$$\lim_{n\to \infty} \frac{1}{n}\sum_{k = 1}^{n-1} v_k = \frac{1}{2}s,$$

which implies

$$\lim_{n\to \infty} \sigma_n = \lim_{n\to \infty} \frac{1}{n+1}\sum_{k = 1}^n v_k \cdot \frac{n+1}{n} = \frac{1}{2}s\cdot \lim_{n\to \infty} \frac{n+1}{n} = \frac{1}{2}s.$$

  • $\begingroup$ Thanks, kobe. Nice answer. $\endgroup$ – Tim Raczkowski Feb 27 '15 at 18:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.