1
$\begingroup$

Clearly, my question can be asked more generally, but suppose for simplicity that $X$ is a smooth surface and that $0 \to E \to F \to K \to 0 $ is exact with $E ,F$ rank two bundles on $X$ and $K$ a line bundle on a (smooth?) divisor $D \subset X$. Can one give information on the cokernel of the injection $Sym^n(E) \to Sym^n(F) $ ? For example is this cokernel some obvious sum of linear algebra constructions (various symmetric powers) of $E$ , $F$, and $K$ ? I can only say that I have thought about this question for a while and no answer has occurred to me.

Adam

$\endgroup$
1
$\begingroup$

Hi Adam,

Say $n=2$. Let $EF\subset Sym^2(F)$ be the subsheaf generated locally by products of sections of $E$ and $F$. Then a diagram chase should give an extension $$0\to (EF)/Sym^2(E)\to Sym^2(F)/Sym^2(E)\to Sym^2(F)/EF\to 0$$ of sheaves supported on $D$. To get a better sense of this, say $E=O_X\oplus O_X$, $F=O_X\oplus O_X(D)$ and $K=O_D(D)$ (which is locally what's going on). Then $$Sym^2(E) = O_X^3,\ EF=O_X\oplus O_X(D)^2,\ Sym^2(F) = O_X\oplus O_X(D)\oplus O_X(2D)$$ so that the above sequence is an extension of $O_D(D)$ by $O_D(D)$. So perhaps in general, it is an extension of $K$ by $K$, but I admit I haven't checked this. In which case, the pattern for $n>2$ shouldn't be hard.

$\endgroup$
1
  • $\begingroup$ Hi Donu, Thanks I think your logic should work. I will think about it some more and thanks for the help. Adam $\endgroup$
    – meh
    Apr 15 '12 at 4:56

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.