Suppose that $C_1, C_2$ are two curves of genus $g \geq 2$ defined over a number field $K$. Let $J_1, J_2$ respectively be their Jacobians. Suppose that $J_1, J_2$ are isogenous over $K$ and $C_1(K), C_2(K)$ are both non-empty, can $C_1(K), C_2(K)$ have different cardinalities?

For $g = 1$ and without the assumption that $C_i(K) \ne \emptyset$, the conclusion is obviously false. Take any elliptic curve $E$ over $\mathbb{Q}$ of positive rank such that the $2$-Selmer group of $E$ is non-trivial, so that there is a genus one curve $C$ which represents a non-trivial $2$-Selmer element of $E$. Then $E$ is isomorphic to the Jacobian of $C$, and the Jacobian of $E$ is itself, so that $C,E$ have isogenous Jacobians but $E(\mathbb{Q})$ is by assumption infinite but $C(\mathbb{Q}) = \emptyset$.


1 Answer 1


Yes it is possible, already for $(K,g) = ({\bf Q},2)$, and already with the first example of isogenous $C_1,C_2$ listed in the LMFDB: curve 249.a.249.1, $y^2 + (x^3+1) y = x^2 + x$, has one rational Weierstrass point, while curve 249.a.6723.1, $y^2 + (x^3+1) y = -x^5 + x^3 + x^2 + 3x + 2$, has two, so the rational-point counts don't even have the same parity. (The counts are known to be $5$ and $6$ respectively.) Further examples are isogeny class 277.a (two curves, each with a unique rational Weierstrass point, but the total counts are $1$ and $5$), and isogeny class 644.a (two curves, one with two rational points, the other not solvable over $\bf R$ --- so this violates your assumption that each $C_i(K)$ is nonempty). The first example where the counts do agree is the two curves in isogeny class 294.a, each with $4$ rational points.

By the way, for $g=1$ there are examples even with each $C_i(K)$ nonempty; for example, take $K=\bf Q$, let $C_1$ be either $X_0(11)$ or $X_1(11)$, and let $C_2$ be the third elliptic curve of conductor $11$. Then $C_1$ has five rational points and $C_2$ has only one.

  • $\begingroup$ Thanks to @Jared Weinstein for correcting the typo in the last sentence. $\endgroup$ Aug 27, 2019 at 23:42

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.