This has been posted on SE, but I haven't gotten a reply, so I thought I'll try my luck here.

I would like to refer you to 2.30 & 2.32 in Silverman's book Advanced Topics in the Arithmetic of Elliptic Curves.

2.30(b)[(c) in errata]: Suppose $\mathfrak{P}$ remains inert in $L'$, say $\mathfrak{P}R_{L'}=\mathfrak{P}'$. Prove that $$q_{\mathfrak{P}}^2=q_{\mathfrak{P}'},\quad a_{\mathfrak{P}}=0,\quad\psi_{E/L'}(\mathfrak{P}')=-q_{\mathfrak{P}}.$$ 2.32(a) Prove that the local $L$-series of $E$ at $\mathfrak{P}$ is given by $$ L_{\mathfrak{P}}(T,E/L) = \left\{ \begin{array}{l l} \cdots & \quad \mathfrak{P}\text{ splits}\\ 1-\psi_{E/L'}(\mathfrak{P}')T & \quad \text{$\mathfrak{P}$ inert}\\ \cdots & \quad \mathfrak{P}\text{ ramifies}\\ \end{array} \right.$$

In the beginning of the chapter, we have that if $E$ has good reduction (in this case, we have good reduction because of 2.30(c))$$L_{\mathfrak{P}}(T,E/L) = 1-a_{\mathfrak{P}}T+q_{\mathfrak{P}}T^2$$ So is the question missing a power of 2 in $T$? Should it be $1+\psi_{E/L'}(\mathfrak{P}')T^2\text{when is $\mathfrak{P}$ inert}$? Thanks for the help! (If it helps, I can print the question; say so in the comments and I'll do it.)

Edit: In Deuring's paper, he has $L_0(s,Kk_1,\mathbf{P})=1+N\mathbf{p}^{1-2s}$. So the presence of $2s$ seems to me to tally with $T^2$.


1 Answer 1


There is nothing wrong with the question. The subtlety is that $T$ is not constant with respect to the fields $L'$ and $L$. My mistake was that I took $T$ to be the same in both fields when in reality it is not.

\begin{align*} L(s,E/L)&=\prod_{\text{$\mathfrak{P}$ in $L$}}\left(1-a_{\mathfrak{P}}q_{\mathfrak{P}}^{-2}+q_{\mathfrak{P}}^{1-2s}\right)^{-1}\\ &=\prod_{\text{$\mathfrak{P}$ in $L$}}\left(1+q_{\mathfrak{P}}^{1-2s}\right)^{-1}\\ &=\prod_{\text{$\mathfrak{P}$ in $L$}}\left(1+q_{\mathfrak{P}}q_{\mathfrak{P}}^{-2s}\right)^{-1}\\ &=\prod_{\text{$\mathfrak{P}$ in $L$}}\left(1+q_{\mathfrak{P}}q_{\mathfrak{P}'}^{-s}\right)^{-1}\\ &=\prod_{\text{$\mathfrak{P}'$ in $L'$}}\left(1-\psi_{E/L'}(\mathfrak{P}')q_{\mathfrak{P}'}^{-s}\right)^{-1}\\ &=L(s,\psi_{E/L'}) \end{align*}

  • $\begingroup$ Dear @BlackAdder, even though this has happened some time ago, could you elaborate a little on what you are precisely doing in this computation? It looks like you are just considering inert primes of $L$ - how does that work to give you the $L$-series of $E/L$? $\endgroup$
    – Jupp
    Aug 26, 2019 at 14:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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