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 just beginning to learn algebraic geometry. An exercise in Reid, p. 24 is to prove that if $Q(x,y,z)$ is a quadratic form over a field $k$ with at least 4 elements, and $Q$ vanishes on the zero set of the conic $\{xz=y^2\}\subset \mathbb{P}^2_k$, then $Q=\lambda (xz-y^2)$. I have solved the problem, I assume in the way the author intended, but in trying to integrate the solution with my previous knowledge, I feel that I am missing something. Can you help me get a sense of "the real story"?

The solution I have: $\{xz=y^2\}\subset \mathbb{P}^2_k$ is parametrized by $(u:v)\in \mathbb{P}^1_k \mapsto (u^2:uv:v^2)\in \mathbb{P}^2_k$. Thus $Q(u^2,uv,v^2)$ vanishes for all choices of $u,v$. This is a homogeneous polynomial in two variables, so I can interpret its zero set as a set of elements in $\mathbb{P}^1_k$. It is degree 4 so unless it is identically zero as a polynomial, it cannot have more than $4$ zeros in $\mathbb{P}^1_k$. But $\mathbb{P}^1_k$ has at least 5 points, (at least) 4 for the elements of $k$ and also $\infty$. Since $Q(u^2,uv,v^2)$ vanishes on all of them, it must be zero as a polynomial. Now an elementary calculation with generic coefficients for $Q$ shows it has the desired form. (Auxiliary question: is this correct?)

This kind of reasoning in terms of projective space is new to me so I sought an alternative solution in the hopes that it would help me understand this solution better.

ATTEMPTED alternative solution: Since $Q$ and $xz-y^2$ are low degree and I want to show $(xz-y^2)|Q$, divide $Q$ by $xz-y^2$ in the ring $k(z,y)[x]$ to obtain a remainder $R\in k(z,y)$ since $xz-y^2$ is degree 1 in $x$. Clear the denominator in $R$ to obtain an element of $k[z,y]$. Essentially this amounts to replacing $zx$ with $y^2$ throughout $z^2Q$. Thus $R$ is a homogeneous degree at most 4 polynomial in $y$ and $z$, and by construction it vanishes whenever $xz=y^2$.

I want to conclude that $R$ is zero as a polynomial, but I can't do what I did last time because I do not know that $R$ vanishes for every $(y:z)\in \mathbb{P}^1_k$. In particular, I do not know if it vanishes on $(1:0)$, although it does vanish on $(u:1)$, all $u\in k$ (i.e. on a copy of $\mathbb{A}^1_k\subset \mathbb{P}^1_k$), because let $x=u^2$. Thus, e.g. if the field is $\mathbb{F}_4$, I only know that the associated inhomogeneous polynomial to $R$ has four zeros, and this means I do not know it is the zero polynomial.

It seems to me that morally, what is going on is that I'm attempting to parametrize $xz=y^2$ by taking $(y:z)\in\mathbb{P}^1$, solving $xz=y^2$ for $x$, and then mapping $(y:z)\in\mathbb{P}^1_k \mapsto (x:y:z)\in\mathbb{P}^2_k$, but this isn't working because if $(y:z)=(1:0)$ then there is no $x$ solving $xz=y^2$. But I remain confused. In particular:


1) Is there a way to do this without referring to projective space? We are proving a statement about low-degree polynomials in 3 variables. I feel like I should be able to do this with very simple commutative algebra. (If $k$ is algebraically closed, then the result is immediate from Hilbert's Nullstellensatz, but this feels like way too much power and doesn't work if $k$ is, e.g., $\mathbb{F}_4$.) Can you see such a solution?

2) Can you help me understand better "why" the second (attempted) solution above didn't work? (Apologies for the vague question. I am going for "moral / conceptual" satisfaction, a slippery thing. I will appreciate any attempt to help.) Is there a way to complete that solution using projective or non-projective ideas?

Thanks. Apologies in advance for the imprecision of the questions.

share|cite|improve this question
up vote 1 down vote accepted

Your first solution is correct. The second one also, but you have to go further. You have $$z^2Q(x,y,z)=(xz-y^2)F(y,z)+R(y,z).$$ Over $\mathbb F_4$, your hypothesis implies that $R(y,z)=c(y^4-yz^3)$ for some constant $c$. Then $F(y,0)=cy^2$. But the term $cxzy^2$ doesn't show up in $R(y,z)$ neither in $z^2Q(x,y,z)$, so $c=0$ and you are done.

share|cite|improve this answer
Yes, got it, this works. So if I want to avoid talking about projective space, the proof splits into two cases based on whether $k=\mathbb{F}_4$ or not? Any thoughts on why this happens? (I.e. why did using projective ideas make the proof more economical?) – Ben Blum-Smith Nov 15 '11 at 15:45
I think this is because with the projective space $P^2$, you don't privilege any point, while with second method, you need a special treatement for the point $x=1,y=z=0$. – user18119 Nov 15 '11 at 23:40

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.