I am reading Silverman's The Arithmetic of Elliptic Curves. I am wondering if with the definition of morphism he gives, we can conclude that if $\phi:V_1\rightarrow V_2$ is a morphism of projective varieties then $V_1\cong \phi(V_1)$. Throughout my mathematical career, I have seen cases where a morphism induces an isomorphism to the image, and cases where it doesn't. It usually depends on how restrictive you want the morphism condition to be. I don't really know what happens in this case, and would appreciate some insight. Here are the definitions Silverman uses for (iso)morphism of projective varieties:

Let $V_1$ and $V_2 \subset \mathbb{P}^n$ be projective varieties. A rational map from $V_1$ to $V_2$ is a map of the form $$ \phi: V_1 \longrightarrow V_2, \quad \phi=\left[f_0, \ldots, f_n\right], $$ where the functions $f_0, \ldots, f_n \in \bar{K}\left(V_1\right)$ have the property that for every point $P \in V_1$ at which $f_0, \ldots, f_n$ are all defined, $$ \phi(P)=\left[f_0(P), \ldots, f_n(P)\right] \in V_2 . $$

A rational map $$ \phi=\left[f_0, \ldots, f_n\right]: V_1 \longrightarrow V_2 $$ is regular (or defined) at $P \in V_1$ if there is a function $g \in \bar{K}\left(V_1\right)$ such that (i) each $g f_i$ is regular at $P$; (ii) there is some $i$ for which $\left(g f_i\right)(P) \neq 0$.

A rational map that is regular everywhere is called a morphism. Finally:

Let $V_1$ and $V_2$ be varieties. We say that $V_1$ and $V_2$ are isomorphic, and write $V_1 \cong V_2$, if there are morphisms $\phi: V_1 \rightarrow V_2$ and $\psi: V_2 \rightarrow V_1$ such that $\psi \circ \phi$ and $\phi \circ \psi$ are the identity maps on $V_1$ and $V_2$, respectively.

  • $\begingroup$ No, consider $P^n\times P^m \to P^m, (x, y) \to y$ for example, with $P^n\times P^m$ projective as it can be embedded into some $P^N$. $\endgroup$
    – C. Brendel
    Commented Mar 9 at 13:50

1 Answer 1


Let $E\subset\Bbb P^2$ be an elliptic curve in Weierstrass form. Then the projection to the $x$-axis gives a 2-to-1 surjective map from $E\to\Bbb P^1$, and $E\not\cong\Bbb P^1$ since they have different genera.

Even if you require bijectivity, this won't hold: consider the normalization map $\Bbb P^1\to V(y^2z-x^3)\subset\Bbb P^2$, which is bijective but not an isomorphism, since the source is smooth but the target is singular.

One positive result in this direction is Zariski's main theorem - a quasi-finite separated birational map of varieties with normal target is an open immersion. Applying this to the case of a morphism between projective varieties, we have that a quasi-finite birational morphism of projective varieties with normal target is an isomorphism. So you need some assumptions.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .