Let $E/F$ be a cyclic extension of degree $p^n$, where p is prime. Let $L$ be an intermediate field such that $[E:L] = p$. If $E = L(\alpha)$, prove that $E = F(\alpha)$.
I've tried to work it out, but my line of reasoning hasn't gotten me anywhere yet. We have $$ p = [E:L] = [E:L(\alpha)]\cdot[L(\alpha):L] $$
By assumption $[E:L(\alpha)] = 1$, so the order of $\alpha \in E$ must be $p$.
I'm not sure why adding an element of order $p$ to $F$ will give me $E$, but it must make use of the fact that $E/F$ is a cyclic extension (which I take to mean its Galois group is in fact cyclic).
Since E/F is cyclic of degree $p^n$, for any integer m less or equal to n, there is a unique subextension $F_m$/F of degree $p^m$ , and E is the union of the ascending chain of the $F_m$ ' s . Your hypothesis is that E = $F_{n-1}$(a) . If F(a) were not equal to E, a would belong to a certain $F_m$ with m < n , hence a would be in $F_{n-1}$ : contradiction .