MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I'm starting to study hyperKähler manifolds and I have a question for you. I hope it's not too stupid...

I know that given a Kähler metric $g$ on an hyperKähler manifold $X=(M,I)$ ($M$ is the differential structure and $I$ the complex structure) there are $J$, $K$ complex structures on $M$ with quaternionic relations such that for every $(a,b,c)\in S^2$ the linear combination $aI+bJ+cK$ is a complex structure which makes $g$ Kähler.

The form $\sigma_I := g(J-,-)+ig(K-,-)$ should be a $(2,0)$-form for $X$ which gives the isomorphism $K_X\simeq \mathcal{O}_X$.

But I have trouble proving it: let's take $u$ and $v$ eigenvectors for $I$ (seen as an almost complex structure) with eigenvalue $i$. Then $$\sigma_I(u,v)=g(Ju,v)+ig(Ku,v)=2g(Ju,v)$$ But $\sigma_I$ should be a nowhere zero $(2,0)$-form, so it should be $2g(Ju,v)\neq 0$. Why is that? I just know that $g(J-,-)$ is a $(1,1)$-form for $(M,J)$.

share|cite|improve this question
A short notes here, hope it helps!… – Jay Nov 6 '13 at 11:08
up vote 2 down vote accepted

Not sure which of the following is giving you trouble, so here is both:

  • $\sigma_I$ is nondegenerate (i.e. $\sigma_I(u,\cdot) = 0$ implies $u=0$) for the simple reason that its real and imaginary parts are nondegenerate, since $g$ is.

  • $\sigma_I$ is a (2,0)-form relative to the complex structure $I$ because, letting $z^a$ be complex coordinates for that structure and $(\partial/\partial z^a,\partial/\partial\bar z^a)$ the resulting basis of $TX\otimes\mathbf C$, we have $I\partial/\partial\bar z^a = -i\partial/\partial\bar z^a$ and hence \begin{align} \sigma_I(\partial/\partial\bar z^a,\cdot) &=g(J\partial/\partial\bar z^a,\cdot)+ig(K\partial/\partial\bar z^a,\cdot)\\ &=g(KI\partial/\partial\bar z^a,\cdot)+ig(K\partial/\partial\bar z^a,\cdot)\\ &=g(-iK\partial/\partial\bar z^a,\cdot)+ig(K\partial/\partial\bar z^a,\cdot)\\ &=0. \end{align} Thus we get $\sigma_I=\sigma_{ab}dz^a\wedge dz^b$ where $\sigma_{ab}=\sigma_I(\partial/\partial z^a,\partial/\partial z^b)$; i.e. no $d\bar z^a$ terms, as claimed.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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