Let $X$ be a compact complex $n$-manifold and $D$ be a smooth comdimension $1$ submanifold. Also let $U:= X\setminus D$ and $j$ be the inclusion of $U$ in $X$. Then it is well known that the following (long exact) sequence respects mixed Hodge structures:

$$ \cdots\to H^{k-2}(D)(-1)\stackrel{\gamma_k}{\to}H^k(X)\stackrel{j^*}{\to}H^k(U)\stackrel{\text{Res}}{\to}H^{k-1}(D)(-1)\to\cdots.$$

Here $\gamma_k$ is the Gysin map obtained as the Poincare dual of $H^{2n-k}(X)\to H^{2n-k}(D)$. And $\text{Res}$ is the residue map that knocks off a factor of type $\frac{dz}{z}$ from a form.

My question: Where does one needs compactness of the manifold/ variety $X$? If one were to work with non-compact $X$, would the above sequence respect the mixed Hodge structure?


Yes, it's fine also if $X$ is not compact. More generally you have for any variety $X$ and subvariety $Z$ a long exact sequence $$ \ldots \to H^k_c(X \setminus Z) \to H^k_c(X) \to H^k_c(Z) \to H^{k+1}_c(X \setminus Z) \to \ldots $$ of mixed Hodge structures, which gives your long exact sequence by Poincaré duality when $X$ is smooth and $Z$ a smooth divisor.

  • $\begingroup$ Thanks a lot. Where can I find this? $\endgroup$ May 23 '16 at 3:58
  • $\begingroup$ I know it's somewhere in Peters-Steenbrink, for instance. $\endgroup$ May 23 '16 at 5:42

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.