Fulton 2.4

Posted on November 15, 2017

2.17. Let V=V(Y^2-X^2(X+1))\subset\mathbb{A}^2, and \overline{X},\overline{Y} the residues of X,Y in \Gamma(V); let z=\overline{X}/\overline{Y}\in k(V). Find the pole sets of z and z^2.

The pole set of z is \{(0,0)\}.

First observe that z=\overline{Y}/\overline{X}=\overline{Y}/^2\overline{XY}=\overline{X(X+1)}/\overline{Y}. Therefore z is clearly defined at all points (x,y) where x\ne 0 or y\ne 0. It remains to show that z is not defined at (0,0).

Suppose \overline{F}/\overline{G}=\overline{Y}/\overline{X} for some F,G\in k[X,Y]. Then \overline{YG}=\overline{XF} and YG-XF\in (Y^2-X^2(X+1)). So YG-XF=H(Y^2-X^2(X+1)) for some H\in k[X,Y]. Hence Y(G-HY)=X(F-HX(X+1)). So G-HY\in (X). Therefore 0=G(0,0)-(HY)(0,0)=G(0,0).

The pole set of z^2 is empty.

Observe that z=\overline{Y}^2/\overline{X}^2=\overline{X^2(X+1)}/\overline{X}^2=\overline{X+1}.

2.18. Let \mathcal{O}_{P}(V) be the local ring of a variety V at a point P. Show that there is a natural one-to-one correspondence between the prie ideals in \mathcal{O}_{P}(V) and the subvarieties of V that pass through P.

It suffices to show that there is a one-to-one correspondence between prime ideals in \mathcal{O}_{P}(V) and prime ideals in \Gamma(V) containing in I(P).

Note from Proposition 3 that an ideal I\subset\mathcal{O}_{P}(V) is generated by I\cap\Gamma(V). Thus if I is prime, then J=I(I\cap\Gamma(V)) is a prime ideal in \Gamma(V).

Suppose J\subset\Gamma(V) is a prime ideal, I(P)\subset J. Note that I(J), the ideal generated by J in \mathcal{O}_{P}(V), is a proper ideal since J does not contain any units. Suppose a/b,c/d\in\mathcal{O}_{P}(V) such that (ac)/(bd)\in I(J). Since 1/b,1/d,1/(bd) are units, they are not in I(J). Thus ac\in I(J) and therefore ac\in J. So a or c must be in J\subset I(J). Thus I(J) is prime.

2.19. Let f be a rational function on a variety V. Let U=\{P\in V \ | \ f \text{ is defined at } P\}. Then f defines a function U to k. Show that this function determines f uniquely.

Suppose f=a/b=c/d, a,b,c,d\in\Gamma(V). Observe that for all P\in U, f(P), f(P)b(P)d(P)=a(P)d(P)=c(P)b(P). Moreover, for all P\in V\setminus U, 0=a(P)d(P)=c(P)b(P). Therefore ad=cb. So a/b=c/d.

2.20. Let V and f be as in the example given in this section.

Show that the pole set of f is exactly \{(x,y,z,w) \ | \ y=0 \text{ and } w=0\}.

Clearly f is defined everywhere y\ne 0 or w\ne 0. It remains to show that f is not defined when y=w=0.

Suppose f=\overline{F}/\overline{G}. Then \overline{XG}=\overline{YF} and XG-YF\in(XW-YZ). Thus XG-YF=H(XW-YZ) and X(G-HW)=Y(F-HZ), H\in k[X,Y,Z,W]. So G-HW\in (Y). Let P=(x,0,z,0), x,z\in k. Observe that G(P)-(HW)(P)=0. Since (HW)(P)=H(P)(0)=0, G(P)=0. Thus f is not defined at P for any choice of representative.

Show that it is impossible to write f=a/b where a,b\in\Gamma(V), and b(P)\ne 0 for every P where f is defined.

Suppose f=\overline{F}/\overline{G} such that \overline{F}/\overline{G} is defined at all points f is defiend. By the same argument above, G-HW\in (Y). So if Y=0, G=HW. If W=0, then G=AY, A\in k[X,Y,Z,W]. So G=AY+HW.

Since \overline{F}/\overline{G} is defined at all points f is defined, V(G)\cap V is the pole set of f. Thus I(V(G)\cap V)=(Y,W), since (Y,W) is the ideal of the pole set. However, for any choice of A,H, observe that I(V(G)\cap V)=(AY+HW, XW-YZ)\ne (Y,W), a contradiction.

2.21.* Let \varphi:V\to W be a polynomial map of affine varieties, \widetilde{\varphi}:\Gamma(W)\to\Gamma(V) the induced map on coordinate rings. Suppose P\in V, \varphi(P)=Q. Show that \widetilde{\varphi} extends uniquely to a ring homomorphism (also writter \widetilde{\varphi}) from \mathcal{O}_{Q}(W) to \mathcal{O}_{P}(V). (Note that \widetilde{\varphi} may not extend to all of k(W).) Show that \widetilde{\varphi}(\mathfrak{m}_{Q}(W))\subset\mathfrak{m}_{P}(V).

Suppose \psi:\mathcal{O}_{Q}(W)\to\mathcal{O}_{P}(V) is a homomorphism such that \psi(a)=\widetilde{\varphi}(a) for any a\in\Gamma(W). If a is a unit in \mathcal{O}_{Q}(W) then \psi(a^{-1})=\psi(a)^{-1}=\widetilde{\varphi}(a)^{-1}. So for any a/b\in\mathcal{O}_{Q}(W), \psi(a/b)=\widetilde{\varphi}(a)/\widetilde{\varphi}(b). Therefore \widetilde{\varphi} extends uniquely to a homomorphism \mathcal{O}_{Q}(W)\to\mathcal{O}_{P}(V).

Suppose f\in\mathfrak{m}_{Q}(W). Then \widetilde{\varphi}(f)(P)=f(Q)=0. So \widetilde{\varphi}(\mathfrak{m}_{Q}(W))\subset\mathfrak{m}_{P}(V).

2.22.* Let T:\mathbb{A}^n\to\mathbb{A}^n be an affine change of coordinates, T(P)=Q. Show that \widetilde{T}:\mathcal{O}_{Q}(\mathbb{A}^n)\to\mathcal{O}_{P}(\mathbb{A}^n) is an isomorphism. Show that \widetilde{T} induces an isomorphism \mathcal{O}_{Q}(V)\to \mathcal{O}_{P}(V) if P\in V^T, for V a subvariety of \mathbb{A}^n.

Since \widetilde{T}:\Gamma(\mathbb{A}^n)\to\Gamma(\mathbb{A}^n) is an isomorphism, it extends uniquely to an isomorphism \widetilde{T}:\mathcal{O}_{Q}(\mathbb{A}^n)\to\mathcal{O}_{P}(\mathbb{A}^n), by Problem 2.21.

Let \pi_V:\mathcal{O}_{Q}(\mathbb{A}^n)\to\mathcal{O}_{Q}(V) and \pi_{V^T}:\mathcal{O}_{P}(\mathbb{A}^n)\to \mathcal{O}_{P}(V^T) be the natural homomorphisms (see Lemma 2.13). Observe that \ker(\pi_{V^T}\circ\widetilde{T})\subset\mathcal{O}_{Q}(\mathbb{A}^n) is equal to the ideal generated by \ker(\pi_{V^T}\circ\widetilde{T})\cap\Gamma(\mathbb{A}^n), similarly for \ker\pi_V. Thus since \ker\pi_V\cap\Gamma(\mathbb{A}^n)=\ker(\pi_{V^T}\circ \widetilde{T})\cap\Gamma(\mathbb{A}^n), \ker\pi_V=\ker(\pi_{V^T}\circ\widetilde{T}). So \pi_{V^T}\circ\widetilde{T} descends, by Lemma 1.1, to an isomorphism \mathcal{O}_{Q}(V)\to\mathcal{O}_{P}(V^T).