1 Schemes 2.

1
Below is a fairly detailed outline of the standard proof that for a ring A and
f ∈ A, the locally ringed space


D(f ), OX D(f ) on X = Spec(A)

is isomorphic to Spec(Af ).
Here D(f ) ⊆ Spec A is the **standard open set** of all prime ideals
p ⊂ A such that f ∈ / p. Also Af denotes the **localization** of A obtained
by inverting powers of f .
—
1. The Underlying Topological Spaces Are Homeomorphic
1.1 Bijective Correspondence of Points
Recall that

Spec(Af ) = q q is a prime ideal in Af .

Each prime q ⊆ Af corresponds to a prime ideal p ⊆ A (its “contraction”)

such that p ∩ S = ∅, where S = {1, f, f 2 , . . . } is the multiplicative set we
are localizing at.
- **Equivalently**, saying p ∩ {1, f, f 2 , . . . } = ∅ means f ∈
/ p. - Hence
such p ∈ Spec(A) precisely belongs to D(f ).
This yields a natural **bijection** of sets

Spec(Af ) ←→ D(f ) ⊆ Spec(A),

by sending q 7→ q ∩ A. (And conversely, each prime p ∈ D(f ) extends

uniquely to a prime q ⊆ Af .)
1.2 Topology Compatibility
Under this bijection, the Zariski topology on Spec(Af ) is the same as
the subspace topology on D(f ) ⊆ Spec(A).
- In Spec(Af ), the basic open sets have the form

D(g) = {q ∈ Spec(Af ) | g ∈
/ q},

for g ∈ Af . - If g̃ ∈ A is a representative of g ∈ Af , then under the bijection,

D(g) in Spec(Af ) corresponds to D(g̃) ∩ D(f ) inside Spec(A).
Hence the map is a homeomorphism of topological spaces:

Spec(Af ) −→ D(f ) ⊆ Spec(A).

1
 2. The Sheaf of Rings Matches: Isomorphism of Locally Ringed Spaces
We must also check that this homeomorphism respects the **structure
sheaf** OSpec(A) restricted to D(f ) versus the structure sheaf OSpec(Af ) .
2.1 Definition of the Structure Sheaf on Spec(Af )
By construction (or by the usual universal property of localization),
∼
 −1
OSpec(Af ) ( D(g) ) = Af g . Concretely:

OSpec(Af ) ( D(g) ) = (Af )g = Af g (since localizing twice is the same as one combined localizatio

2.2 Restriction of OSpec(A) to D(f )

On the other side, recall that OSpec(A) (D(h)) ∼
= Ah . When we restrict
OSpec(A) to the open subset D(f ), its basic opens are of the form D(h) ∩
D(f ) = D(f h). Hence




OSpec(A) D(f ) D(f h) = OSpec(A) D(f h) = Af h .

2.3 Checking the Isomorphism on Basic Opens

Under our homeomorphism Spec(Af ) ∼ = D(f ), the basic open D(g) ⊆
Spec(Af ) corresponds to D(g̃) ∩ D(f ) = D(f g̃) ⊆ Spec(A). Comparing the
rings of sections:
- **Left side**:


OSpec(Af ) D(g) = (Af )g = Af g .

- **Right side**:



OSpec(A) D(f )
D(f g̃) = Af g̃ .

But under “lifting” g 7→ g̃, we see Af g̃ ∼

= Af g .
Hence these two rings are **canonically isomorphic** for every basic
open of the form D(g). By uniqueness of sheafification and the fact that the
basic opens form a basis for the Zariski topology, these local identifications
glue to give an **isomorphism of sheaves of rings** over the homeomorphism

Spec(Af ) ∼
= D(f ).

2.4 Local Rings and Locally Ringed Space Isomorphism

To check it is an isomorphism **of locally ringed spaces**, one finally
checks that the induced map on stalks (local rings) is an isomorphism. At
a point q ∈ Spec(Af ) corresponding to p ∈ D(f ) ⊆ Spec(A), one has:
- The stalk OSpec(Af ),q ∼
= (Af )q . - The stalk OSpec(A),p localized to D(f )
is Ap , and this is isomorphic to (Af )q because q sits over p.

2
 Everything matches up so that the maps on maximal ideals (or prime
ideals in the stalks) are consistent. Hence it is indeed an **isomorphism of
locally ringed spaces**.
Thus we conclude:

(D(f ), OX |D(f ) ) ∼
= Spec(Af )

as locally ringed spaces.

3
1 Schemes 2.2
**Solution Explanation**
We want to show that if (X, OX ) is a
scheme (over some base), and
U ⊆ X is an open subset, then U, OX U is itself a scheme, often called
an **open subscheme** of X. Here, OX U denotes the restriction of the
structure sheaf OX to U . The scheme structure on U is then inherited from
X.
—
1. Recall the Definition of a Scheme
A **scheme** (X, OX ) is a topological space X together with a sheaf
of rings OX (called the structure sheaf), such that every point of X has
an open neighborhood V ⊆ X for which (V, OX |V ) is **isomorphic** (as a
locally ringed space) to Spec(A) for some ring A.
Concretely, for every x ∈ X, there is an open affine neighborhood V ⊆ X
and an isomorphism of locally ringed spaces

(V, OX |V ) ∼


= Spec(A), OSpec(A) .

—
2. Restricting to an Open Subset U ⊂ X
We are given an open U ⊆ X. We consider the restricted sheaf OX |U .
We need to check:
1. The topological space U equipped with the restriction of OX is a
**locally ringed space** (i.e., its stalks are local rings). 2. It is covered by
open affines of the form Spec(A).
2.1. Locally Ringed Space Condition
- For each u ∈ U , let OX |U denote the restriction of the presheaf OX to
U . Then the stalk at u, (OX |U )u , is **by definition** the same as OX,u ,
because restricting to a smaller open set does not change the germ of a
section at u. - Since OX,u is a local ring (as X is a locally ringed space),
the restriction (OX |
U )u remains a local ring.
Hence U, OX |U is a **locally ringed space**.
2.2. Cover by Affine Opens
Given u ∈ U , because X is a scheme, there exists an open affine neigh-
borhood V ⊂ X of u, say

(V, OX |V ) ∼
= Spec(A).

But U ∩ V is an open subset of V (since U and V are both open in X).

Therefore, the induced structure on U ∩V is the restriction of OX |V to U ∩V .

1
Because Spec(A) is itself a topological space, an open subset of Spec(A) is
(by definition) an **affine open** if it can be written as Spec(A)f = { p ∈
Spec(A) : f ∈/ p} for some f ∈ A.
Hence we get an isomorphism of locally ringed spaces

(U ∩ V, OX |U ∩V ) ∼


= Spec(A)f , OSpec(A) Spec(A) , f

showing (U ∩ V ) is **affine** in the restricted sense.

Thus every point u ∈ U has an open neighborhood

U ∩V ⊆ U isomorphic
to Spec(A)f , which is affine. Hence U, OX |U is **locally affine**, and so
it is a scheme.
—
3. Conclusion
By restricting the structure sheaf OX to the open subset U , we obtain a
locally ringed space (U, OX |U ). Locally, it is isomorphic to open subschemes
of affine schemes (because X is covered by affines, and we further restrict
to open subsets). Therefore, (U, OX |U ) **is itself a scheme**, called the
**open subscheme** of X determined by U .

Hence, any open subset U ⊆ X with the restricted sheaf OX |U is a scheme.

2
1 Schemes 2.4
Below is a restatement in more conventional notation and typography, show-
ing how one obtains the natural map


α : HomSch (X, Spec A) −→ HomRings A, Γ(X, OX ) .
—
1. Set-up and Notation
1. **A** is a (commutative) ring. 2. **(X, OX )** is a scheme. 3. We
are given a morphism of schemes

f : X −→ Spec A.

1.1 The Associated Map f # on Structure Sheaves

By the definition of a morphism of schemes, we have

f # : OSpec A −→ f∗ OX .

Concretely, this is a morphism of sheaves of rings on the topological space

Spec A (with its Zariski topology).
1.2 Taking Global Sections
Taking global sections of both sides on Spec A gives us an induced ring
homomorphism



A = Γ Spec A, OSpec A −→ Γ Spec A, f∗ OX .

But


Γ Spec A, f∗ OX ≃ Γ X, OX
by the definition of the direct image sheaf f∗ OX . Hence we get a ring
homomorphism
A −→ Γ(X, OX ).
—
2. The Natural Map α
Given any morphism f : X → Spec A, the above process yields a
**ring** homomorphism

Φf : A −→ Γ(X, OX ).

Thus we get a **map of sets** (in fact, a map of hom-sets):

 
α : HomSch X, Spec A −→ HomRings A, Γ(X, OX ) ,

1
where each f is sent to the induced homomorphism Φf .
—
3. Why It Is “Natural”
1. **Functorial in A**. One checks that if A → A′ is a ring homomor-
phism, then there is a induced map Spec A′ → Spec A, etc. The construction
of α is compatible with such morphisms in the usual sense of functoriality
(one formalizes this by looking at the contravariant functor Spec (·) from
rings to schemes).
2. **Functorial in X**. Likewise, if X ′ → X is a morphism of schemes,
then a morphism X → Spec A induces one X ′ → Spec A, and everything
commutes.
Hence α reflects the usual duality between the category of rings (con-
travariantly) and the category of affine schemes, extended to the fact that
morphisms into an affine scheme Spec A are equivalent to ring homomor-
phisms out of A on global sections.
—
4. Historical or Conceptual Note
- In classical algebraic geometry, this map α is intimately connected with
the **Yoneda lemma** viewpoint:


HomSch (X, Spec A) = HomSch Spec Γ(X, OX ), Spec A
if X were affine; more precisely, for affine X, we have X ≃ Spec Γ(X, OX ).
- In the nonaffine case, one still obtains the map α by restricting attention
to the induced map on global sections, although now X is not necessarily
Spec Γ(X, OX ).
Either way, the construction of α is straightforward from the definitions
and is often used as a step in describing how morphisms of schemes into
affine schemes are encoded by ring homomorphisms on global sections.
—
5. Final Summary
1. From a scheme morphism f : X → Spec A, we automatically get a
**structure sheaf map** f # : OSpec A → f∗ OX . 2. Taking global sections
recovers a **ring** homomorphism A → Γ(X, OX ). 3. This process is
**functorial**, giving a map

 
HomSch X, Spec A −→ HomRings A, Γ(X, OX ) .
4. This is precisely the map α in the statement.

Hence the natural map α arises by sending f 7→ A → Γ(X, OX ) .

2
1 Schemes 2.5
Description of Spec(Z)
The **spectrum of the ring of integers**, Spec(Z), is the set of prime
ideals of Z, equipped with the Zariski topology. The prime ideals of Z are:
1. The zero ideal (0), which corresponds to the **generic point** of
Spec(Z). 2. The prime ideals (p) for each prime number p, which correspond
to the **closed points** of Spec(Z).
Thus, the underlying topological space consists of the points:

Spec(Z) = {(0)} ∪ {(p) | p prime}.

The Zariski topology on Spec(Z) is defined by the basic closed sets:

V (I) = {p ∈ Spec(Z) | I ⊆ p}
for ideals I of Z. Specifically:
- V (0) = Spec(Z), since every prime ideal contains (0). - V (p) = {(p)},
which means each closed point is a singleton. - V (n) = {(p) | p divides n}
for n > 0, meaning that the closed set defined by an integer n consists of
the prime ideals corresponding to the prime factors of n.
The structure sheaf OSpec(Z) is the sheaf associated with the **affine
scheme** Spec(Z), meaning its sections over D(f ) (the standard distin-
guished open set associated with f ∈ Z) are given by Z[ f1 ].
Spec(Z) as a Final Object in the Category of Schemes
In the **category of schemes**, a final object (or terminal object) is an
object T such that for every scheme X, there exists a **unique morphism**
X → T.
We claim that Spec(Z) is the final object in the category of schemes.
That is, for every scheme X, there exists a **unique** morphism of schemes:

X → Spec(Z).
Existence of a Morphism X → Spec(Z)
A morphism of schemes f : X → Spec(Z) corresponds to a **ring ho-
momorphism** on affine open sets. Specifically, for any affine open U =
Spec(A) ⊂ X, a morphism X → Spec(Z) induces a ring homomorphism:

Z → A.
But since Z is the **initial object in the category of rings**, there is
a unique ring homomorphism Z → A, namely the one sending 1 to the
multiplicative identity in A.

1
 Since every scheme is covered by affine open sets Spec(A), and the
morphisms on these opens glue uniquely due to the uniqueness of the in-
duced ring homomorphisms, we obtain a **unique** scheme morphism X →
Spec(Z).
Uniqueness of the Morphism
Suppose f, g : X → Spec(Z) are two morphisms. Then on each affine
open U = Spec(A) ⊂ X, these morphisms induce ring homomorphisms
Z → A. But as mentioned, this homomorphism is uniquely determined.
Hence, f and g must agree on all affine open sets, meaning f = g globally.
Conclusion
Since every scheme X admits a unique morphism to Spec(Z), this shows
that Spec(Z) is the **final object** in the category of schemes. This re-
sult reflects the idea that Spec(Z) is a universal base space for arithmetic
geometry, encapsulating the **global arithmetic** of all schemes.

2
1 Schemes 2.6
**Spectrum of the Zero Ring** The **zero ring** 0 is the unique ring in
which 0 = 1, meaning that every element is equal to 0. Explicitly, it consists
of only one element {0}, and the only possible multiplication and addition
satisfy 0 + 0 = 0 and 0 · 0 = 0.
The **prime spectrum** of a ring A, denoted Spec(A), consists of all
prime ideals of A, equipped with the Zariski topology.
- In the **zero ring**, the only ideal is {0}, which is also the entire ring
itself. - However, an ideal I in a ring is **prime** if whenever ab ∈ I, then
either a ∈ I or b ∈ I. - Since the only ideal {0} satisfies this property vac-
uously (because there are no nonzero elements), it **is not a prime ideal**
in the usual sense. - In fact, in commutative algebra, the **zero ring has no
prime ideals** because the definition of a prime ideal typically assumes the
presence of multiplicative elements. - Thus, we conclude that:

Spec(0) = ∅.

That is, the spectrum of the zero ring is the empty topological space.
—
**The Zero Ring as an Initial Object in the Category of Schemes** A
category has an **initial object** if there exists an object X such that for
every other object Y , there is a unique morphism X → Y .
1. **The Category of Rings**: - In the category of commutative rings
with unity, the zero ring is an **initial object**. - This follows because for
any ring R, there is a unique ring homomorphism 0 → R, which sends the
unique element 0 to 0 in R. - However, there is no ring homomorphism
R → 0 unless R itself satisfies 0 = 1, which only happens if R is also the
zero ring.
2. **The Category of Schemes**: - The **category of schemes** has
objects given by schemes (locally ringed spaces satisfying certain conditions)
and morphisms given by morphisms of locally ringed spaces. - The empty
space ∅ with the **trivial structure sheaf** (i.e., the sheaf of rings being
the zero ring at every open set) corresponds to Spec(0), which is the empty
scheme. - For any scheme X, there is a **unique morphism** ∅ → X (since
maps from the empty set to any space are unique). - Thus, Spec(0) serves
as the **initial object** in the category of schemes.
—
**Conclusion** - The **spectrum of the zero ring is empty**: Spec(0) =
∅. - The **zero ring is an initial object** in the category of rings. -

1
The **empty scheme** Spec(0) is the **initial object** in the category
of schemes.
This highlights an important feature of algebraic geometry: the algebraic
structure of the zero ring translates naturally into the **initial topology and
categorical properties** of its associated geometric space.

2
1 Schemes 2.7
Below is a standard argument in algebraic geometry showing that giving a
morphism
Spec K −→ X


is equivalent to giving a pair x, k(x) ,→ K consisting of a point x ∈ X and
an inclusion of fields k(x) → K. Here k(x) denotes the **residue field** of
X at x.
—
1. From a Morphism Spec K → X to a Point x ∈ X and k(x) → K
1. **A morphism Spec K → X determines a point x ∈ X.**
- Since Spec K has a unique (closed) point corresponding to the zero
prime ideal in K, its image in X is a single point x. - Concretely, if

ϕ : Spec K −→ X


is a morphism of schemes, then x := ϕ unique point of Spec K is a point
in X.
2. **We get an induced local ring map OX,x → K.**
- By definition of a morphism of locally ringed spaces, we get a commu-
tative diagram on the stalk:


OX,x −→ Γ Spec K, OSpec K = K,

respecting local structures (it sends the maximal ideal of OX,x into the
unique maximal ideal 0 ⊂ K). - Thus we have a **local homomorphism**
OX,x → K that kills the maximal ideal mx ⊂ OX,x .
3. **Hence we get a map on residue fields k(x) → K.**
- Because mx is in the kernel of OX,x → K, this induces a ring homo-
morphism
k(x) = OX,x /mx −→ K.
- Since OX,x /mx is a field, this map is in fact an **injection** (or **inclu-
sion** of fields), because the only prime ideal in K is the zero ideal. Hence
we get
k(x) ,→ K.
Putting these together, any morphism Spec K → X gives rise to:
1. A **point** x ∈ X. 2. An **inclusion** of residue fields k(x) ,→ K.
—
2. From a Point x ∈ X and k(x) ,→ K to a Morphism Spec K → X

1
 Conversely, start with:
- A point x ∈ X. - An **injective** ring map k(x) → K.
We want to construct a morphism of schemes Spec(K) → X with image
x.
1. **Lift k(x) → K to a local map OX,x → K.**
- Because OX,x ↠ k(x) is surjective (the quotient by mx ), any injection
k(x) ,→ K extends uniquely to a local ring homomorphism OX,x → K that
kills mx .
2. **Use sheaf/lifting properties to get a map on an open neighborhood
of x.**
- By the usual property of stalks, such a local map OX,x → K is given
by a (compatible) family of maps from sections over some neighborhood of
x. This (in the language of local morphisms of locally ringed spaces) yields
a **scheme morphism** from Spec K into X, sending the point of Spec K
to x.
3. **Uniqueness**: Two different extensions would differ on the maxi-
mal ideal mx and so would give different local maps.
Hence from the data (x, k(x) ,→ K), there is a **unique** morphism

Spec(K) −→ X

mapping the unique point of Spec K to x and inducing the given injection
k(x) → K on residue fields.
—
3. Summary of the Equivalence
Putting both directions together, we get a natural **1-to-1 correspon-
dence**:
n o n o
morphisms Spec K → X ←→ (x, k(x) ,→ K) x ∈ X .

Hence giving a morphism from Spec K to X is equivalent to giving

1. A **point** x ∈ X, and 2. A **field inclusion** k(x) ,→ K.
That completes the proof.

2
1 Schemes 2.8
**Solution Explanation**
Let X be a scheme over a field k. We want to show that giving a k-
morphism
Spec k[ε]/(ε2 ) −→ X

is equivalent to specifying:
1. A point x ∈ X whose residue field is k (i.e. a k-rational point, meaning
k(x) = k), 2. An element of the Zariski tangent space Tx X.
Here the **Zariski tangent space** Tx X at the point x is defined as

∨
Tx X = mx /m2x ,

the dual of the vector space mx /m2x over k(x). Since k(x) = k in our situa-
tion, this is a vector space over k.
—
1. Setup: Morphisms from Dual Numbers
1. **Ring of dual numbers**: We consider the local Artin ring R :=
k[ε]/(ε2 ). It has a unique maximal ideal (ε), satisfying ε2 = 0.
2. **Residue field**: The residue field of R at (ε) is isomorphic to k.
Concretely,
R/(ε) ∼= (k[ε]/(ε2 ))/(ε) ∼
= k.
Hence any morphism Spec(R) → X selects some point x ∈ X whose residue
field must contain k. If we want a morphism of k-schemes (i.e. the structure
map X → Spec k is respected), then that point x must actually have residue
field equal to k.
—
2. From a Morphism to a Point in X plus a Tangent Vector
**Step 1:** Let
f : Spec(k[ε]/(ε2 )) −→ X
be a morphism of k-schemes. Because Spec(R) has a single (closed) point
corresponding to the ideal (ε), the image of that point under f is some
point x ∈ X. Since R/(ε) ∼= k, the residue field of x must be k. Thus x is
k-rational.
**Step 2:** Now look at the induced local homomorphism on stalks at
x. Denote by OX,x the local ring of X at x with maximal ideal mx . We get
a local ring homomorphism

OX,x −→ k[ε]/(ε2 )

1
sending mx into (ε). Since ε2 = 0, the image of m2x must be zero. This
homomorphism therefore factors as

OX,x −→ OX,x /m2x −→ k[ε]/(ε2 ).

**Step 3:** Restrict further to the quotient mx /m2x . Because m2x goes to
zero, we get an induced linear map of k-vector spaces

mx /m2x −→ (ε)/(ε2 ).
But (ε)/(ε2 ) ∼
= k · ε as a k-vector space (one-dimensional). Thus any
such local homomorphism amounts to choosing an element

[α] ∈ Homk mx /m2x , k · ε .

Dualizing (or simply identifying the space of k-linear maps mx /m2x → k

with (mx /m2x )∨ ), we see that specifying this local homomorphism is equiva-
lent to specifying an element of

(mx /m2x )∨ = Tx X.
Hence:
- The point x is determined by where (ε) maps, - The tangent vector is
determined by how the ring homomorphism kills m2x but may have a non-zero
image on mx .
—
3. From a Point plus a Tangent Vector to a Morphism
Conversely, given:
1. A k-rational point x ∈ X (so mx ⊂ OX,x and k(x) = k), 2. A tangent
vector θ ∈ Tx X (i.e. an element of Homk (mx /m2x , k)),
we can construct the corresponding morphism Spec(k[ε]/(ε2 )) → X by
defining a local ring homomorphism

OX,x −→ k[ε]/(ε2 ),

as follows:
- On the residue field level, OX,x /mx → k[ε]/(ε2 )/(ε) ∼
= k, we take the
isomorphism corresponding to the chosen rational point x. - On mx /m2x →
(ε)/(ε2 ), we use the linear map induced by θ. Concretely, if θ(f ) ∈ k is the
image of f ∈ mx modulo m2x , then we send it to θ(f ) · ε ∈ k · ε.

2
 Extending multiplicatively (and respecting ε2 = 0), one obtains a unique
local homomorphism of rings OX,x → k[ε]/(ε2 ). This patching is enough to
define a morphism of schemes

Spec(k[ε]/(ε2 )) −→ X

that sends the closed point of Spec(k[ε]/(ε2 )) to x and encodes the chosen
tangent vector θ.
—
4. Conclusion
Summarizing both directions:
- **Given** a k-morphism Spec(k[ε]/(ε2 )) → X, we get a unique k-
rational point x ∈ X (residue field k) and a vector in Tx X. - **Conversely,**
any choice of x ∈ X with k(x) = k and a vector in Tx X uniquely determines
such a morphism.
Hence to give a morphism

Spec(k[ε]/(ε2 )) −→ X

over k is **equivalent** to giving a k-rational point x ∈ X plus an element

of the Zariski tangent space Tx X.

3
1 Schemes 2.9
Below is a standard argument showing that every nonempty irreducible
closed subset of a scheme has a **unique** generic point. The proof hinges
on two key facts:
1. Every **irreducible** closed subset Z ⊆ X in a scheme X has at least
one generic point (in fact, any irreducible topological space has a generic
point in the “spectral” sense).
2. The underlying topological space of a scheme is **T0 ** (Kolmogorov).
In a T0 -space, an irreducible set can have **at most** one generic point.
Putting (1) and (2) together yields the **existence** and **uniqueness**
of the generic point of Z.
—
1. Terminology and Setup
- A **generic point** ζ of a closed set Z ⊆ X means that

{ζ} = Z.

Equivalently, ζ is contained in every dense open subset of Z.

- A **scheme** X has an underlying topological space (the spectrum-
type topology). This underlying space is always **T0 **, meaning: for any
two distinct points x, y ∈ X, there is an open set that contains one of them
but not the other.
- A subset Z ⊆ X is **irreducible** if it cannot be written as a union
Z = Z1 ∪ Z2 of two proper closed subsets.
—
2. Existence of a Generic Point
**General fact** (true for the underlying “spectral” topology of any
ringed space like Spec R, or any irreducible topological space that arises in
the Zariski-type setting):
¿ If Z is a nonempty irreducible closed subset in a scheme X, then there
is at least one point ζ ∈ Z whose closure equals Z.
In more algebraic language, when Z = V (I) in Spec R, irreducibility of
Z translates to I being prime (or at least an intersection of primes in some
contexts), and a **minimal** prime in that intersection corresponds to a
point whose closure is exactly V (I). Concretely, one shows that the set
\
U
U ⊆Z
nonempty open in Z

1
is nonempty because Z is irreducible (no proper closed subset covers it).
Any point ζ in this intersection must lie in every nonempty open of Z,
hence {ζ} = Z.
Thus a generic point **exists**.
—
3. Uniqueness of the Generic Point in a T0 -Space
Now we show **any two distinct points** in an irreducible space Z
**cannot both** serve as generic points (i.e., there is at most one generic
point).
1. Suppose for contradiction that Z has **two** distinct points ζ1 , ζ2 ∈
Z such that
{ζ1 } = Z and {ζ2 } = Z.
In a T0 -space, there is an open set U ⊂ X containing ζ1 but **not** con-
taining ζ2 .
2. Intersect U with Z. Then U ∩ Z is an open subset of Z containing ζ1
but excluding ζ2 .
3. But if ζ2 were also a generic point, ζ2 would lie in **every** nonempty
open subset of Z. That contradicts the fact that ζ2 ∈ / U ∩ Z.
Hence no two distinct points can both be generic. This shows **at most
one** generic point.
—
4. Putting It All Together
- By irreducibility, we have **at least one** generic point. - By T0 , we
have **at most one** generic point.
Therefore, a nonempty irreducible closed subset Z in a scheme X has
**exactly one** generic point ζ.

Hence every nonempty irreducible closed subset of a scheme has a unique generic point.

2
1 Schemes 2.10
The spectrum of a ring, denoted Spec(R[x]), is the set of all prime ideals
of R[x], equipped with the Zariski topology. Let’s analyze its structure and
compare it to R and C.
Step 1: Prime Ideals in R[x]
In R[x] , prime ideals fall into two main categories: 1. Maximal Ideals
The maximal ideals of R[x] correspond to the points in C , but in a spe-
cific way: • If α ∈ R, then(mα ) = (x − α)[Link] α + iβ ∈ C \
R, then(x−(α+iβ))(x−(α−iβ)) = x2 −2αx+(α2 +β 2 )def inesarealirreduciblepolynomial, andtheide
2αx+(α2 +β 2 ))isamaximalideal.T hus, themaximalidealsinR[x]correspondto :
Realpointsα ∈ R, as(x−α).P airsof complexconjugatepoints(α±iβ), appearingasquadraticpolynom
2αx+(α2 +β 2 )).2.T heM inimalP rimeIdealT heonlynon−maximalprimeidealisthezeroideal(0), wh
Thus, we conclude:

Spec(R[x]) = {(0)}∪{(x−α) | α ∈ R}∪{(x2 −2αx+(α2 +β 2 )) | α, β ∈ R, β ̸= 0}.

Step 2: Zariski Topology on Spec(R[x]) • The basic closed sets in Spec(R[x])

are given by sets of the form:

V (I) = {p ∈ Spec(R[x]) | I ⊆ p},

where I is an ideal of R[x] . • The closure of a point depends on the corre-

sponding ideal: • The point (0) (the generic point) is dense in Spec(R[x])
because every nonzero polynomial vanishes somewhere in C. • A point (x -
α)(correspondingtoα ∈ R)isclosed, meaningitf ormsasingle−pointclosedsetintheZariskitopology.
2αx + (α2 + β 2 ))(correspondingtoaconjugatecomplexpair)isalsoclosed.
Thus, the Zariski topology is very different from the usual topology: •
The only dense point is (0) . • All other points (corresponding to maximal
ideals) are closed. • The space is not Hausdorff in the usual sense.
Step 3: Comparison with R and C
Comparison with R • The set R naturally corresponds to the maximal
ideals (x - α)f orα ∈ [Link], Spec(R[x]) contains additional points cor-
responding to pairs of complex conjugates. • The topology on R (usual
topology) is completely different from the Zariski topology on Spec(R[x]),
where every proper closed set is finite.
Comparison with C • The set C is much larger than R , but in Spec(R[x]),
we only see complex numbers appearing in pairs of conjugates. • Spec(C[x])
is just C ∪{(0)}, whichisbiggerthan Spec(R[x]), since complex points appear
individually in C , but only in conjugate pairs in Spec(R[x]).

1
 Summary • Spec(R[x]) consists of real numbers (corresponding to max-
imal ideals (x - α))andcomplexconjugatepairs(correspondingtoideals(x2 −
2αx+(α2 +β 2 ))).T heZariskitopologyisverydif f erentf romtheusualtopology :
thezeroideal(0)isdense, [Link], Spec(R[x])
is larger because it includes points representing complex numbers in conju-
gate pairs. • Compared to C , it is smaller because individual complex
points do not appear—only their conjugate pairs do.
Thus, Spec(R[x]) is a topological space that encodes both real and com-
plex algebraic information but in a way that differs significantly from the
usual number systems.

2
1 Schemes 2.11
We analyze the spectrum of the polynomial ring k[x] over the finite field
k = Fp .
**Step 1: Structure of Spec(Fp [x])**
The prime spectrum Spec(Fp [x]) consists of prime ideals in Fp [x], which
fall into two categories:
1. **The zero ideal** (0): - This corresponds to the **generic point**,
which represents the function field Fp (x). - Its residue field is Fp (x), the
field of rational functions over Fp .
2. **Maximal ideals** (f (x)) where f (x) is an irreducible polynomial
over Fp : - Each maximal ideal corresponds to an irreducible polynomial
f (x), and the residue field is Fp [x]/(f (x)), which is a finite field extension
of Fp . - If f (x) has degree d, the residue field is Fpd , the finite field with pd
elements.
Thus, the spectrum Spec(Fp [x]) consists of: - A generic point correspond-
ing to (0) with residue field Fp (x). - One closed point for each irreducible
polynomial f (x), with residue field Fpd where d = deg(f ).
**Step 2: Counting Points with a Given Residue Field**
To count the number of points in Spec(Fp [x]) with a given residue field
Fpd , we count the number of monic irreducible polynomials of degree d over
Fp . This number is given by:
1X
Nd = µ(k)pd/k
d
k|d

where µ(k) is the Möbius function. This formula counts the number of
irreducible polynomials of degree d over Fp .
Thus, for each d, there are Nd closed points in Spec(Fp [x]) with residue
field Fpd .
**Conclusion** - Spec(Fp [x]) consists of a generic point corresponding
to (0) and closed points corresponding to irreducible polynomials. - The
residue fields of closed points are the finite fields Fpd where d is the degree
of an irreducible polynomial. - The number of closed points with residue
field Fpd is given by the number of monic irreducible polynomials of degree
d over Fp , computed using the formula involving the Möbius function.