MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

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

Is it true that the dual of a smooth hypersurface $X$ of $\mathbb{P}^n$ of degree $d\ge 2$ is a hypersurface? If yes, could you give me a simple proof ? Or a reference?

Note that in this case, the dual is birational to $X$.

share|cite|improve this question
up vote 7 down vote accepted

Yes, the dual is a hypersurface. Here is a simple proof when $n\geq 4$. First of all, by the Lefschetz hyperplane theorem, the restriction homomorphism, $$ r : \text{Pic}(\mathbb{P}^n) \to \text{Pic}(X),$$ is an isomorphism. In particular, every invertible sheaf on $X$ is either ample, trivial, or anti-ample.

Denote the Gauss map on $X$ by, $$ G : X \to (\mathbb{P}^n)^\vee, \ \ x \mapsto \mathbb{T}_xX, $$ where $\mathbb{T}_x X$ denotes the tangent hyperplane to $X$ at $x$. The invertible sheaf $G^*\mathcal{O}(1)$ is either ample, trivial or anti-ample. If $G$ is nonconstant, then $G^*\mathcal{O}(1)$ has positive degree Hilbert polynomial, hence $G^*\mathcal{O}(1)$ is neither trivial nor anti-ample. Therefore $G^*\mathcal{O}(1)$ is ample. In particular, $G$ can have no positive dimensional fiber, since this would give a subvariety of $X$ on which $G^*\mathcal{O}(1)$ is not ample. Therefore $G$ is finite, and hence $\text{dim}(G(X))$ equals $\text{dim}(X)$.

Finally, if $G$ were constant with image $[H]$, for $H$ a fixed hyperplane in $\mathbb{P}^n$, then for every $x\in X$, the tangent hyperplane to $\mathbb{T}_x X$ equals $H$. In particular, for every $x\in X$, $x$ is contained in $H$. Thus $X$ is contained in $H$, contradicting that $X$ is a smooth hypersurface of degree $d>1$.

Edit. I just want to add, I heard this argument long ago, but I do not know who first proved this result. Fyodor Zak has done important work on the Gauss map, so you might consult his work to try and find a reference.

Second edit. The OP asks about $n=3$. One could probably adapt the "fast" argument above using Noether-Lefschetz, but it is probably more honest just to compute $G^*\mathcal{O}(1)$. Indeed, the morphism $G$ arises from the short exact sequence sequence, $$ 0 \to \mathcal{O}_{\mathbb{P}^n}(-d)|_X \xrightarrow{u} \Omega_{\mathbb{P}^n/k}|_X \to \Omega_{X/k} \to 0,$$ and the Euler sequence (restricted to $X$), $$ 0 \to \Omega_{\mathbb{P}^n/k}|_X \xrightarrow{v} \mathcal{O}_{\mathbb{P}^n}(-1)^{\oplus(n+1)}|_X \to \mathcal{O}_{\mathbb{P}^n}|_X \to 0. $$ The adjoint of $v\circ u$ is an invertible quotient, $$ \mathcal{O}_{\mathbb{P}^n}(1)^{\oplus(n+1)}|_X \to \mathcal{O}_{\mathbb{P}^n}(d)|_X, $$ which is equivalent to an invertible quotient (after twisting), $$\mathcal{O}_{\mathbb{P}^n}^{\oplus(n+1)}|_X \to \mathcal{O}_{\mathbb{P}^n}(d-1)|_X. $$ This invertible quotient defines the morphism $G:X\to (\mathbb{P}^n)^\vee$. Thus, $G^*\mathcal{O}(1)$ is isomorphic to $\mathcal{O}_{\mathbb{P}^n}(d-1)|_X$. Thus, if $d>1$, then $G^*\mathcal{O}(1)$ is ample. Hence $G$ is finite.

share|cite|improve this answer
Thanks for the answer. This is exactly what I needed. But how does it work if $n=3$ ? – Jérémy Blanc Feb 21 '14 at 13:00
Thanks for the edit too. – Jérémy Blanc Feb 22 '14 at 5:57

A reference for this statement is Dolgachev's beautiful book Classical Algebraic Geometry, Chapter 1.

share|cite|improve this answer
Thanks for the reference, but where exactly? Actually, I had already read this chapter partially just before asking the question, but probably not the right result. – Jérémy Blanc Feb 21 '14 at 13:01
In the file of the book I have, it is explained in Section 1.2.2, soon after the Reflexivity Theorem. – Francesco Polizzi Feb 21 '14 at 13:06
What I had read is "Let us apply this to the case when X is a nonsingular hypersurface. The polar map is a finite map, hence the dual of a nonsingular hypersurface is a hypersurface." and it is true that this is quite easy, but I wanted the argument (given by Jason Starr above). Thanks to both of you for your help. – Jérémy Blanc Feb 22 '14 at 5:55

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.