1. Algebraic Varieties in Pn.
Let (ζ0, · · · , ζn) denote the standard coordinate system on Cn+1and (ζ0 : · · · : ζn) be the corresponding homogeneous coordinate system on Pn.
Definition 1.1. A polynomial f (Z0, · · · , Zn) in S = C[Z0, · · · , Zn] is called homogeneous of degree d ≥ 0 if f (λZ0, · · · , Zn) = λdP (Z0, · · · , Zn) for all λ ∈ C. The set of all homogeneous polynomials of degree d together with 0 is denoted by Sd.
Then we have a direct sum decomposition of S :
S =M
d≥0
Sd
such that Si· Sj ⊂ Si+j for any i, j.
Given a homogeneous polynomial f in S, we say that a point P (a0 : · · · : an) ∈ Pn is a zero of f if f (a0, · · · , an) = 0 for any representative (a0, · · · , an) of P (a0 : · · · : an). It is easy to check that the notion of zero of a homogeneous polynomial in Pn is well-defined.
When P is a zero of f, we denote f (P ) = 0. The set of zeros of f, is denoted by V (f ). In general, let T be a family of homogeneous polynomials over C. We denote V (T ) the set of all common zeros of polynomials in T, i.e. V (T ) is the set of all P ∈ Pn so that f (P ) = 0 for all f ∈ T.
Definition 1.2. We say that Y ⊂ Pn is an algebraic set if there exists a family T of homogeneous polynomials so that Y = V (Y ). We say that a subset U of Pn is a Zariski open set of Pn if Pn\ U is an algebraic set.
Theorem 1.1. The Zariski open sets in Pn forms a topology. We call this topology the Zariski topology on Pn.
Proof. To do this, it is equivalent to show that (1) ∅ and Pn are algebraic.
(2) Any intersection of algebraic sets is also algebraic.
(3) Any finite union of algebraic sets is algebraic.
Since V (1) = ∅ and V (0) = Pn, (1) is obvious. To prove (2), we only need to show V (S
iTi) =T
iV (Ti) for any family {Ti} of sets of homogeneous polynomials. Again, this is obvious.
Let T1, · · · , Tmbe sets of homogeneous polynomials. Let T be the set {f1· · · fm: fi∈ Ti}.
Then elements of T are homogeneous polynomials. If P ∈ S
iV (Ti), then P ∈ V (Tj) for some j. Hence fj(p) = 0 for all fj ∈ Tj. Then P is a zero of all f in S, i.e. P ∈ V (T ).
Conversely, if P ∈ V (T ), then f (P ) = 0 implies that P ∈ V (Ti) for some i. Hence P ∈ S
iV (Ti). We conclude thatS
iV (Ti) = V (T ). We completes the proof. An ideal a in S is said to be a homogeneous ideal if
a=M
d
(a ∩ Sd).
If a is a homogeneous ideal in S, we define
V (a) = V (T ), where T is the set of homogeneous polynomials in a.
Definition 1.3. We say that X ⊂ Pn is a projective variety if X is irreducible algebraic set. A quasi projective variety is an open subset of a projective variety.
1
2
Let Y be any subset of Pn. The homogeneous ideal I(Y ) of Y is the ideal generated by f ∈ S such that f a homogeneous polynomial and f (P ) = 0. The homogeneous coordinate ring of Y is defined to be the quotient ring S(Y ) = S/I(Y ).
Let Ui = {(ζ0 : · · · : ζn) ∈ Pn: ζi 6= 0}. Then Ui is the complement of the zero set V (ζi) of ζi for 0 ≤ i ≤ n and hence Ui is Zariski-open. We find that {Ui : 0 ≤ i ≤ n} forms an open covering of Pn. We also consider the map
ϕi : Ui → Cn
defined by ϕi(ζ0 : · · · : ζn) = (ζ0/ζi, · · · , ζn/ζ0). Then ϕi is a bijection for all 0 ≤ i ≤ n. Let us claim that ϕi are homeomorphism in the Zariski-topology on Pn and Cn.
Let Y be a Zariski closed subset of Cn. We want to show that ϕ−1i (Y ) is Zariski closed in Pn.
Let T be a set of polynomials in C[X1, · · · , Xn]. Given a polynomial f (X1, · · · , Xn) of degree d, we define a homogeneous polynomial ϕ∗f of degree d in variables Z0, · · · , Zn by
(ϕ∗if )(Z0, · · · , Zn) = Zidf (Z0/Zi, · · · , Zn/Zi).
We set ϕ∗T = {ϕ∗f : f ∈ T }.
Lemma 1.1. Ui∩ V (ϕ∗iT ) = ϕ−1i V (T ) for any set of polynomials T in C[X1, · · · , Xn].
Proof. Let P ∈ ϕ−1i (Y ) and f ∈ T. We have ϕ∗if (P ) = 0. Hence P ∈ V (ϕ∗iT ). (By definition P ∈ Ui.) Conversely, if P ∈ V (ϕ∗iT ) ∩ Ui, then for all f ∈ T, f (ϕi(P )) = 0. Hence P ∈
ϕ−1i V (T ). We prove the statement.
If Y is a Zariski closed set in Cn, Y = V (T ) for some T ⊂ C[X1, · · · , Xn]. Hence ϕ−1i Y = Ui∩ V (ϕ∗iT ) is closed in Ui. We see that ϕi is continuous in the Zariski topology.
Let f (Z0, · · · , Zn) be a homogeneous polynomial of degree d. We set fϕi(X1, · · · , Xn) = f (X1, · · · , Xi−1, 1, Xi, · · · , Xn).
For any family T of homogeneous polynomials in C[Z0, · · · , Zn], we set Tϕi = {fϕi : f ∈ T }.
Lemma 1.2. ϕi(Ui ∩ V (T )) = V (Tϕi) for any set T of homogeneous polynomials in C[Z0, · · · , Zn].
Proof. Let Q ∈ ϕi(Ui ∩ V (T )). Then we can find P ∈ V (T ) ∩ Ui so that ϕi(P ) = Q. (If Q = (a1, · · · , an) ∈ Cn, P = [a1, · · · , ai−1, 1, ai, · · · , an].) Since P ∈ V (T ), f (P ) = 0 for all f ∈ T. We see that fϕi(Q) = f (P ) = 0. Hence Q ∈ V (Tϕi).
Let Q ∈ V (Tϕi). Then fϕi(Q) = 0 for all f ∈ T. Hence f (ϕ−1i (Q)) = fϕi(Q) = 0. We see that ϕ−1i (Q) ∈ Ui∩ V (T ) and hence Q ∈ ϕi(Ui∩ V (T )). We complete the proof. Hence if Y is a Zariski closed set in Ui, then there exists a family T of homogeneous polynomials in S so that Y = Ui ∩ V (T ). Then ϕ(Y ) = V (Tϕ) in Cn and hence ϕi is a closed mapping in the Zariski topology.
Proposition 1.1. ϕi : Ui → Cn are homeomorphisms in the Zariski topology.
Corollary 1.1. If Y is a projective (resp, quasi-projective) variety, then Y is covered by open sets Y ∩Uifor 0 ≤ i ≤ n which are homeomorphic to affine (reps. quasi-affine) varieties via ϕi defined above.
Proof. Since {Ui} is an open covering of Pn, {Ui∩ Y : 0 ≤ i ≤ n} forms an open covering for Y. If Y is a projective variety, it is Zariski closed. Hence for 0 ≤ i ≤ n, Ui∩ Y is Zariski closed in Ui. Since ϕi is continuous in the Zariski-topology, ϕi(Ui∩ Y ) is Zariski closed in
3
Cn. The irreducibility of Ui∩ Y implies the irreducibility of ϕi(Ui∩ Y ). Hence ϕi(Ui∩ Y ) is an affine variety in Cn for each 0 ≤ i ≤ n. Since ϕi is a homeomorphism, Ui ∩ Y is homeomorphic to the affine variety ϕi(Ui∩ Y ) in Cn.