We are able to use calculus to compute the radius of convergence of a power series. Now I want to look at a analytical way to that procedure. Say, we have a power series $\sum a_nx^n$. What if we should know that $\frac{a_{n+1}}{a_n} \to z$. What does that mean for the radius of convergence of $\sum a_nx^n$. Of course, the radius of convergence should be $\frac{1}{z}$. But how can I proof that when $\frac{a_{n+1}}{a_n} \to z$, the radius of convergence has to be $\frac{1}{z}$.
We could use the inequality $ \limsup a_n^{1/n} \leq \limsup \frac{a_{n+1}}{a_n}$ to find $\limsup a_n^{1/n}$ and compute the radius of convergence. But how can I show that $\frac{a_{n+1}}{a_n} \to z \Rightarrow \limsup a_n^{1/n} = z $?
This answer is for an earlier version of the question.
$\lim \sup a_n^{1/n}=z$ does not imply that $\frac {a_{n+1}} {a_n} \to 1$. For example, if $a_n=2n$ for $n$ even and $n$ for $n$ odd then $\lim \sup a_n^{1/n}=1$ but $\frac {a_{n+1}} {a_n}$ does not have a limit. Note that ratio test only gives a sufficient condition for convergence and this condition is not necessary.