Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

I am reading Hulek' Elementary Algebraic Geometry, p.103

Let $V$ be an irreducible affine hypersurface, say $V=V(f)\subset\Bbb{A}^n$. Then the coordinate ring is by definition $k[V]=k[x_1,\ldots,x_n]/(f)$. Suppose $f$ contains the variable $x_1$. Then $$k(V)=k(x_2,\ldots,x_n)[x_1]/(f)$$ can somebody very kindly explain me this identification?

share|cite|improve this question
Where exactly is that result stated ? – Georges Elencwajg Jan 21 '14 at 10:46
Oops sorry I'm wrong (I wrote that down on some notes and…) it is actually in Hulek' Elementary Algebraic Geometry p.103 last line. Let me edit the post. – Heitor Fontana Jan 21 '14 at 11:11
Thanks a lot for answering so quickly, dear Heitor. I can stop browsing Shafarevich now :-) And +1 for this nice question. – Georges Elencwajg Jan 21 '14 at 12:16
up vote 6 down vote accepted

By assumption $k[x_1,\dotsc,x_n]/(f) \cong k[x_2,\dotsc,x_n][x_1]/(f)$ is an integral domain. It follows that $f$ is irreducible over $k[x_2,\dotsc,x_n]$, as a polynomial in $x_1$. By assumption it is not constant. Hence, Gauss' Lemma tells us that it stays irreducible over $k(x_2,\dotsc,x_n)$. This means that $k(x_2,\dotsc,x_n)[x_1]/(f)$ is a field. Since it contains $k[x_1,\dotsc,x_n]/(f)$ as a subring and is generated by it, it must be its field of fractions.

share|cite|improve this answer
I hadn't read your answer while I composed mine (and had lunch). In a way I'm glad to realize there isn't a clever shortcut. Anyway +1 for your answer. – Georges Elencwajg Jan 21 '14 at 12:09

The crucial point is that $f$ , which is irreducible in $k[x_2,\ldots,x_n][x_1]$, is still irreducible in $k(x_2,\ldots,x_n)[x_1]$: this is not trivial but follows from a result of Gauss on UFD's.
Hence $L:=k(x_2,\ldots,x_n)[x_1]/(f)$ is a field.

The obvious $k$-morphism $k[V]=k[x_1,\ldots,x_n]/(f)\to L=k(x_2,\ldots,x_n)[x_1]/(f)$ extends to a morphism $$k(V)=\text {Frac} (k[x_1,\ldots,x_n]/(f))\to L=k(x_2,\ldots,x_n)[x_1]/(f)$$ Again this is not obvious but results from Gauss: if a polynomial $g\in k[x_1,\ldots,x_n]$ is not a multiple of $f$ in $k[x_2,\ldots,x_n][x_1]$, then $g$ will not be a multiple of $f$ in $k(x_2,\ldots,x_n)[x_1]$ either.

Once we have this morphism of $k$-extensions $k(V)\to L$ the result follows easily: like all morphisms with source a field it is injective and surjectivity is clear once you realize that a polynomial in $h\in k[x_2,...,x_n]$ is invertible in $k(V)$, because $h$ cannot be a multiple of $f$ since $f$ contains $x_1$ and $h$ does not.

A great reference on Gauss's results on UFDs is Artin's Algebra , Theorem (3.9) of Chapter 11, page 401.
Beware that the results in this theorem are elementary but quite subtle and that it is easy to make mistakes in their statements.

share|cite|improve this answer
"Again this is not obvious but results from Gauss:" Isn't this just the fact that $A \to A[S^{-1}]$ is injective when $A$ is an integral domain and $0 \notin S$? – Martin Brandenburg Jan 21 '14 at 13:57
@Martin: the point is that a morphism $u:k[V]=k[x_1,\ldots,x_n]/(f)\to L=k(x_2,\ldots,x_n)[x_1]/(f)$ extends to a morphism $\hat {u}: k(V)\to L$ if and only if $u$ is injective. And injectivity amounts to the implication that a non-multiple of $f$ in $k[x_1,...,x_n]$ is sent to a non-multiple of $f$ in $k(x_2,\ldots,x_n)[x_1]$. – Georges Elencwajg Jan 21 '14 at 14:54
I know that. I wanted to indicate that injectivity is easy when you use localizations ... – Martin Brandenburg Jan 21 '14 at 15:42

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.