Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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 trying to figure out why there are 27 lines on a smooth cubic hypersurface in $\mathbb{P}^3$ using Chern classes, without looking up the proof in 3264 and All That. One thing confusing me is the following:

We want to find the top Chern class of the bundle $E$ over $\mathbb{G}(1,3)$ whose fiber over a line $L$ is the space of cubic forms on $\mathbb{P}^3$, restricted to $L$. After all, a smooth cubic hypersurface gives a section of this vector bundle, and the top Chern class will hopefully tell us that this section needs to vanish 27 times.

The first way I thought of to study $E$ is to note that it is a quotient of the trivial bundle $\mathbb{G}(1,3)\times\mathbb{C}^{20}\to\mathbb{G}(1,3),$ where $\mathbb{C}^{20}$ parametrizes cubic forms on $\mathbb{P}^3$ and the map $$\mathbb{G}(1,3)\times\mathbb{C}^{20}\to E$$ of bundles maps $(L,f)\mapsto(L,f|_L)\in E.$ Then these fit in the short exact sequence $$0\to F\to\mathbb{G}(1,3)\times\mathbb{C}^{20}\to E\to0,$$ where $F$ is the bundle whose fiber over $L$ is the cubic forms of $\mathbb{P}^3$ that vanish on $L$.

Now the part I'm getting confused about. Properties of Chern classes tell us that the total Chern classes satisfy $$c(\mathbb{G}(1,3)\times\mathbb{C}^{20})=c(E)c(F).$$ Since $\mathbb{G}(1,3)\times\mathbb{C}^{20}$ is a trivial bundle, I'm pretty sure it has trivial total Chern class $c(\mathbb{G}(1,3)\times\mathbb{C}^{20})=1.$ But this would mean that $c(E)$ and $c(F)$ are units, i.e. trivial themselves, right? Which is very hopefully not the case. Help? (I could be way off base on much of this, since I came up with it right as I was falling asleep last night.)

(Edited to fix minor errors.)

share|cite|improve this question
up vote 2 down vote accepted

The cohomology ring $H^*(G(1,3),\mathbb Z)$ has many units and to say that $c(E)\in H^*(G(1,3),\mathbb Z)$ is one of them is far from implying that $c(E)=1$ and even further from implying that $E$ is trivial.

share|cite|improve this answer
Thanks - I am now back on the right track and it was kind of a silly issue! – Rob Silversmith Nov 16 '12 at 20:18

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.