Let $k$ be a field, let $f \colon X \to Y$ be a morphism of $k$-varieties, and assume $X$ and $Y$ are smooth and projective. Let $H(\_)$ be a classical Weil cohomology theory (i.e. one of $\ell$-adic étale, Betti, algebraic de Rham, crystalline). Of course it depends on $k$ which of the theories are applicable, but I do not want to spell all the cases out here.

What conditions can one impose on $f$ to assert that $f^{*} \colon H^{i}(Y) \to H^{i}(X)$ is injective/surjective?

Currently I can only think of the rather trivial: if $f$ admits a section (resp. retraction), then $f^{*}$ is injective (resp. surjective).

[Edit] For example, is it true that $f^{*}$ is injective if $f$ is dominant? [/Edit]

  • $\begingroup$ Someone may hit this with the CW-hammer if you think a “one property/condition” per answer style would be best. $\endgroup$
    – jmc
    Jun 24, 2014 at 8:53
  • 2
    $\begingroup$ Do dominant non-surjective morphisms of smooth projective varieties really exist? I would have thought some valuative criterion would get violated. $\endgroup$
    – S. Carnahan
    Jun 24, 2014 at 9:21
  • 3
    $\begingroup$ Well, over $\mathbb{C}$ I would say that they do not exist. Using the analytic topology, the image of the morphism must be compact (since $X$ is compact and $f \colon X \to Y$ is continuous), hence closed (since $Y$ is Hausdorff). On the other hand, the image contains a dense open set (since the map is dominant), so the image is actually the whole of $Y$ (by a dimension argument). $\endgroup$ Jun 24, 2014 at 10:02
  • 7
    $\begingroup$ @FrancescoPolizzi: over any field, the image of a projective variety is closed, and that's all you use, right? (Or did I misunderstand something?) $\endgroup$
    – user5117
    Jun 24, 2014 at 11:34
  • 1
    $\begingroup$ @ArtiePrendergast-Smith – Thanks for your comment. In the end I should have realised this myself. Anyway, I didn't know Kleiman's result, so Timo's answer + your comment really answers my question! $\endgroup$
    – jmc
    Jun 24, 2014 at 19:36

3 Answers 3


Kleiman, Algebraic Cycles and the Weil Conjectures, Proposition 1.2.4: Let $f: X \to Y$ be surjective. Then $f^*: H^*(Y) \to H^*(X)$ is injective.

Let me recall the proof.

Let $x$ be a closed point of the generic fibre of $f$ and set $Z := \overline{\{z\}} \subseteq X$ and let $z = cl(Z)$ be the cycle class of $Z$. Since the cycle class map commutes with $f_*$, one has $f_*(z) \neq 0$.

Now assume $a \in H^*(Y)$ is in the kernel of $f^*$. Then one has $f^*(ab)z = f^*(a)f^*(b)z = 0$ for every $b \in H^*(Y)$, so by the projection formula $0 = f_*(f^*(ab)z) = abf_*(z)$, so by Poincaré duality $a = 0$.

  • $\begingroup$ Ok, cool. Do you know if dominant is enough? $\endgroup$
    – jmc
    Jun 24, 2014 at 8:55
  • 12
    $\begingroup$ If $X$ and $Y$ are projective, then dominant implies surjective. $\endgroup$ Jun 24, 2014 at 11:45
  • $\begingroup$ @DanPetersen – Of course, I should have realised this myself. Thanks! $\endgroup$
    – jmc
    Jun 24, 2014 at 19:32

For posterity, and as it was a bit tricky to find a reference, note also that the statement holds for possibly non-projective, quasi-projective smooth varieties, or, more generally for non-compact smooth Kähler manifolds:

Theorem (Wells): Let $f:X\to Y$ be a proper surjective morphism of smooth complex manifolds and assume that $X$ is Kähler. Then the following induced homomorphisms are injective

  1. $f^*:H^q(Y,\Omega^p_Y)\to H^q(X,\Omega^p_X)$
  2. $f^*:H^k(Y,\mathbb{C})\to H^k(X,\mathbb{C})$
  3. $f^*:H^k(Y,\Omega^p_Y(E))\to H^k(X,\Omega^p_X(f^*E))$ for a holomorphic vector bundle $E$.

The reference is Wells Comparison of de Rham and Dolbeault cohomology for proper surjective mappings.


I do not know any effective, general criterion. But the Lefschetz hyperplane theorem for complex algebraic varieties tells you that if you have a hyperplane section $X\subset Y$ where $Y$ is a complex projective algebraic smooth variety of dimension $n$ such that $Y-X$ is smooth then the induced map $$H^*(Y,\mathbb{Z})\rightarrow H^*(X,\mathbb{Z})$$ is an isomorphism for $*<n-1$ and an injection for $*=n-1$.


Edit: Let $Y$ be a smooth complex projective variety and let $D$ be a normal crossing divisor ($dim_{\mathbb{C}}(Y)=n$). If the contraction $Y/D$ is algebraic projective and if the map $Y\rightarrow Y/D$ is also algebraic then $$H^i(Y,\mathbb{C})\rightarrow H^i(D,\mathbb{C})$$ is surjective for $i\geq n$.

  • $\begingroup$ Thanks for your answer! This is indeed a criterion, but not exactly the flavour that I was looking for (no offense!). $\endgroup$
    – jmc
    Jun 24, 2014 at 8:50
  • $\begingroup$ Sorry, but I don't understand the last paragraph. Are there some missing hypotheses? $\endgroup$ Jun 24, 2014 at 16:10
  • $\begingroup$ OK, I see. I misunderstood your notation. In algeometer-speak, you're saying: suppose $D$ can be contracted to a point, then... And you're getting this from the decomposition theorem. $\endgroup$ Jun 24, 2014 at 18:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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