1

1

Lets $X$ be a simply connected projective toric variety of dimension $n$.

Lets $\tau_1,\cdots,\tau_k$ be the set of $(n-1)$-dimensional cones of corresponding fan which is in one-to-one correspondence with invariant curves on $X$.

Both $H_2(X,\mathbb{Q})$ and $N_1(X,\mathbb{Q})$ are quotients of the vector space $\sum \mathbb{Q} \cdot \tau_i $ with respect to different equivalence relations. ( Please correct me if I am wrong) and we have a surjective map $H_2 \rightarrow N_1$.

My question is whether this is an isomorphism?

Dual to this picture we have $H^2(X) \cong Pic(X)$ and therefore a surjective map $H^2(X,\mathbb{Q}) \rightarrow N^1(X,\mathbb{Q})$; which I am asking again whether this is an isomorphism.

For those who don't know: $N^1(X)=$ formal linear sum of Cartier-divisors modulo numerical equivalence and $N_1(X)=$ formal linear sum of curves modulo numerical equivalence; Where numerical equivalence is given by intersection theory between Cartier divisors and curves.

flag
Note that $N^1=Pic/Pic^0$, so the questions is: Is there a non-trivial line bundle in $Pic^0(X)$ when $X$ is toric. – Mohammad F.Tehrani Apr 15 2011 at 15:07

2 Answers

1

The answer is yes and can be found in section 6.3 of the book : Toric varieties,(David Cox, John Little, Hal Schenck) There they prove:

Numerical equivalence = algebraic equivalence

which gives the desired result.

link|flag
where have you got the book? :) – Dmitri Apr 17 2011 at 21:32
@Dmitri A draft of the book is freely available: cs.amherst.edu/~dac/toric.html – AlB Apr 17 2011 at 22:20

Your Answer

Get an OpenID
or

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