You are on page 1of 304

Solutions to Hartshorne's Algebraic Geometry

Andrew Egbert
October 3, 2013

Note: Starred and Formal Schemes questions have been skipped since for the most part we skipped those
in class.

Contents
1

I. Varieties

18

1.1

I.1 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

18

1.1.1

Ex, I.1.1 g x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

18

1.1.2

b. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

18

1.1.3

Ex, I.1.2 x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

18

1.1.4

Ex, I.1.3 x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

18

1.1.5

Ex, I.1.4 x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

19

1.1.6

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

19

1.1.7

Ex, I.1.6 x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Ex, I.1.5 x

19

1.1.8

Ex, I.1.7 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

20

1.1.9

b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

20

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

20

1.1.10 (c) x.

1.2

1.1.11 d x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

20

1.1.12 Ex, I.1.10 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

21

1.1.13 (b). x g and I.2.7.a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

21

1.1.14 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

21

1.1.15 d. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

21

1.1.16 e. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

21

I.2 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

22

1.2.1

I.2.1 x homogenous nullstellensatz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

22

1.2.2

I.2.2 projective containments x

1.2.3

I.2.3.a. containments. x

1.2.4
1.2.5

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

22

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

22

b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

22

c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

22

1.2.6

d. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

23

1.2.7

e. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.2.8

b. x g and below

1.2.9

c. x g

23

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

23

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

23

1.2.10 I.2.5 (a) x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

24

1.2.11 (b). x

24

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.2.12 I.2.7 a. x g
1.2.13 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

24

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

24

1.2.14 I.2.8 x g
1.2.15 I.2.9 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

24

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

25

1.2.16 I.2.10 x
1.2.17 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

25

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

26

1.2.18 c. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.2.19 I.2.11 x g (use p2)
1.2.20 b. x g
1.2.21 x.

26

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

26

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

26

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

27

1.2.22 c. x

1.2.23 I.2.12 x g and below

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

27

1.2.24 .b. x g and above and below (part d) . . . . . . . . . . . . . . . . . . . . . . . . . . . .

27

1.2.25 c. x

27

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.2.26 d. x g and above

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

27

1.2.27 I.2.13 x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

28

1.2.28 I.2.14 x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

28

1.2.29 I.2.15 x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

28

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

29

1.2.30 b. x
1.2.31 c. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.2.32 I.2.16 x g
1.2.33 b. x g

29

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

29

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

29

1.2.34 I.2.17 x g complete intersections and below . . . . . . . . . . . . . . . . . . . . . . . . .

29

1.2.35 b. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

30

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

30

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

30

1.2.36 starred.
1.3

26

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I.3 x
1.3.1

I.3.1 x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

30

1.3.2
1.3.3

b. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

30

c. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

30

1.3.4

d. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

30

1.3.5

e. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31

1.3.6

I.3.2 g bijective, bicontinuous but not isomorphism. x . . . . . . . . . . . . . . . . . . .

31

1.3.7

b. frobenius not isomorphism. x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31

1.3.8

I.3.3a. x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31

1.3.9

b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31

1.3.10 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

32

1.3.11 I.3.4 x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

32

1.3.12 I.3.5 x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

32

1.3.13 I.3.6 x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

32

1.3.14 I.3.7 a x. g
1.3.15 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

32
32

1.3.16 I.3.8 x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

33

1.3.17 I.3.9 x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

33

1.3.18 I.3.10 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

33

1.3.19 I.3.11 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

33

1.3.20 I.3.12 x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

34

1.3.21 I.3.13 x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

34

1.3.22 I.3.14 x g (and below) projection from point
1.3.23 b. x g

. . . . . . . . . . . . . . . . . . . . . . . .

34

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

34

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

35

1.3.24 I.3.15 x
1.3.25 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

35

1.3.26 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

35

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

35

1.3.27 d. x g

1.3.28 I.3.16 x g
1.3.29 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

36

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

36

1.3.30 I.3.17 x. g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

36

1.3.31 b. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

36

1.3.32 c. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

36

1.3.33 d. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

36

1.3.34 e. x.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

37

1.3.35 I.3.18.a x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

37

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

37

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

37

1.3.36 b. x g
1.3.37 c. x

1.3.38 I.3.19 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.3.39 I.3.20 x g and below
1.3.40 b. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.3.41 I.3.21a. x

38
38

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

38

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

38

1.3.43 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39

1.3.44 d. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39

1.3.45 e. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39

1.3.42 b. x

1.3.46 I.4.1 x g
1.3.47 I.4.2 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39

1.3.48 I.4.3 x g and below . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39

1.3.49 b x. g

40

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.3.50 I.4.4 x g all parts

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

40

1.3.51 b. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

40

1.3.52 c. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

40

1.3.53 I.4.5 x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

40

1.3.54 I.4.6 x g and below . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

41

1.3.55 b. x g
1.3.56 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

41

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

41

1.3.57 I.4.7 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.3.58 I.4.8 x g
1.3.59 b. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

41
42

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

42

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

42

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

43

1.3.61 I.4.10 x g
I.5 x

41

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.3.60 I.4.9 x g
1.4

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

38

1.4.1

I.5.1.a x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

43

1.4.2

I.5.1.b. x g what kind of singularities . . . . . . . . . . . . . . . . . . . . . . . . . . . .

43

1.4.3

c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

43

1.4.4

d. x g what kind of singularity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

43

1.4.5

I.5.2a. x pinch point x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

44

1.4.6

b. x conical double point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

44

1.4.7

c. x

44

1.4.8

I.5.3 x g

1.4.9

b. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

44

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

45

1.4.10 I.5.4 x g
1.4.11 b. x g
1.4.12 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

45

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

45

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

45

1.4.13 I.5.5 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.4.14 I.5.6 x g

45

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

46

1.4.15 b. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

46

1.4.16 c. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

46

1.4.17 d. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

46

1.4.18 I.5.7 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

47

1.4.19 b. (important) x g
1.4.20 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

47

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

47

1.4.21 I.5.8 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

47

1.4.22 I.5.9 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

47

1.4.23 I.5.10 x g:a,b,c

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

48

1.4.25 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

48

1.4.26 I.5.11 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

48

1.4.27 I.5.12 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

49

1.4.28 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

49

1.4.29 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

49

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

49

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

50

1.4.31 I.5.13 x
1.4.32 I.5.14 x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

50

1.4.33 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

50

1.4.34 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

51

1.4.35 d (starred) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

51

1.4.36 I.5.15 x g:a,b

52

1.4.37 b. x
I.6 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

52

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

52

1.5.1

I.6.1 x g:a,b,c

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

52

1.5.2

b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

52

1.5.3

c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

52

1.5.4

I.6.2 x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

53

1.5.5

b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

53

1.5.6

c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

53

1.5.7

d. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

53

1.5.8

e. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

54

1.5.9

I.6.3 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.5.10 I.6.4 x g
1.5.11 I.6.5 x

54

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

54

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

54

1.5.12 I.6.6 x g:a,c

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

54

1.5.13 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

55

1.5.14 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

55

1.5.15 I.6.7 x g
1.6

48

1.4.24 b. x

1.4.30 d. x.

1.5

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I.7 x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

55

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

55

1.6.1

I.7.1 x g

1.6.2

b. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

55

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

55

1.6.3

I.7.2 x g arithmetic genus of projective space.

1.6.4

b. x g

1.6.5

c. x g important

1.6.6

d. g x

1.6.7

e. x

1.6.8

I.7.3 x g

1.6.9

I.7.4 x

. . . . . . . . . . . . . . . . . . . . . . .

56

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

56

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

56

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

56

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

56

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

57

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

57

1.6.10 I.7.5 x g:a,b upper bound on multiplicity x . . . . . . . . . . . . . . . . . . . . . . . . .

57

1.6.11 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

57

1.6.12 I.7.6 x Linear Varieties x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

58

1.6.13 I.7.7 x

58

1.6.14 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.6.15 I.7.8 x contained in linear subspace. x

. . . . . . . . . . . . . . . . . . . . . . . . . . .

58
58

2

II Schemes

59

2.1

59

II.1 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.1

x II.1.1 Constant presheaf

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

59

2.1.2

II.1.2 x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

59

2.1.3

b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

59

2.1.4

c. x.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

59

2.1.5

II.1.3x surjective condition. x

2.1.6

2.1.3.b. x g Surjective not on stalks x . . . . . . . . . . . . . . . . . . . . . . . . . . . .

60

2.1.7

II.1.4 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

60

2.1.8

b. x

60

2.1.9

x II.1.5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.1.10 II.1.6.a x map to quotient is surjective

60

. . . . . . . . . . . . . . . . . . . . . . . . . . .

61

2.1.11 II.1.6.b x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

61

2.1.12 II.1.7 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

61

2.1.13 II.1.7.b x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

61

2.1.14 II.1.8 x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

61

2.1.15 II.1.9 x Direct sum of sheaves x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

62

2.1.16 II.1.10 x Direct Limits x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

62

2.1.17 II.1.11 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

62

2.1.18 II.1.12 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

62

2.1.19 II.1.13 Espace Etale x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

63

2.1.20 II.1.14 x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

63

2.1.21 II.1.15 Sheaf Hom x: g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

64

2.1.22 II.1.16 g Flasque Sheaves x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

64

2.1.23 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

64

2.1.24 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

64

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

65

2.1.25 d. x g

2.1.26 e. x sheaf of discontinuous sections

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.1.27 II.1.17 x g skyscraper sheaves (important)
2.1.28 II.1.18 x g Adjoint property of f^(-1)
2.1.30 b. g x extending by zero
2.1.31 c. x

65

. . . . . . . . . . . . . . . . . . . . . . . . .

65

. . . . . . . . . . . . . . . . . . . . . . . . . . . .

65

2.1.29 II.1.19 x g Extending a Sheaf by Zero (important?)

. . . . . . . . . . . . . . . . . . . .

66

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

66

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

66

2.1.32 II.1.20 x Subsheaf with Supports

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

67

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

67

2.1.34 II.1.21 g sheaf of ideals x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

67

2.1.35 b. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

68

2.1.36 c. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

68

2.1.33 b. x

2.1.37 d. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

68

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

69

2.1.39 II.1.22 x g Glueing sheaves (important) . . . . . . . . . . . . . . . . . . . . . . . . . . .

69

II.2

2.1.38 e. x
2.2

59

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

69

2.2.1

II.2.1 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

69

2.2.2

II.2.2 x g induced scheme structure.

. . . . . . . . . . . . . . . . . . . . . . . . . . . .

69

2.2.3

II.2.3 a. x reduced is stalk local x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

70

2.2.4

b. reduced scheme x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

70

2.2.5

c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

70

2.2.6

II.2.4 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

71

2.2.7

II.2.5 x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

71

2.2.8

II.2.6 x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

71

2.2.9

II.2.7 x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

71

2.2.10 II.2.8 x g Dual numbers + Zariski Tangent Space
2.2.11 II.2.9 x g Unique Generic Point (Important)

72
72

2.2.12 II.2.10 x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

72

2.2.13 II.2.11 x g (Spec Fp important)

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

73

2.2.14 II.2.12 x g Glueing Lemma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

73

2.2.15 II.2.13 x quasicompact vs noetherian. a . . . . . . . . . . . . . . . . . . . . . . . . . . .

73

2.2.16 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

74

2.2.17 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

74

2.2.18 d. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

74

2.2.19 II.2.14 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

74

2.2.20 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

74

2.2.21 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

75

2.2.22 d. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

75

2.2.23 II.2.15x g (important)

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

75

2.2.24 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

75

2.2.25 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

76

2.2.26 II.2.16 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

76

2.2.27 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

76

2.2.28 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

76

2.2.29 d. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

77

2.2.30 II.2.17 Criterion for aneness x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

77

2.2.31 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

77

2.2.32 II.2.18 g x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

78

2.2.33 b. g x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.2.34 c. x g (important)
2.2.35 d. g x
2.3

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

78
78

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

78

2.2.36 II.2.19 x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

79

II.3 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

79

2.3.1

II.3.1 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

79

2.3.2

II.3.2 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

79

2.3.3

II.3.3 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

80

2.3.4

b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

80

2.3.5

c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

80

2.3.6

II.3.4 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

80

2.3.7

II.3.5 x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

81

2.3.8

b. x g

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

81

2.3.9

c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

81

2.3.10 II.3.6 x g Function Field

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

81

2.3.11 II.3.7 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

82

2.3.12 II.3.8 x Normalization

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

82

2.3.13 II.3.9 x g Topological Space of a Product . . . . . . . . . . . . . . . . . . . . . . . . . .

83

2.3.14 b. g x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

83

2.3.15 II.3.10 x g Fibres of a morphism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

84

2.3.16 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.3.17 II.3.11 x g Closed subschemes

84

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

84

2.3.18 b. (Starred) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

85

2.3.19 c. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

85

2.3.20 d. x g scheme-theoretic image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

85

2.3.21 II.3.12 x Closed subschemes of Proj S . . . . . . . . . . . . . . . . . . . . . . . . . . . .

85

2.3.22 b. x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.3.23 II.3.13 x g Properties of Morphisms of Finite Type

. . . . . . . . . . . . . . . . . . . .

86
86

. . . . . . 87 2. . . . . . .4. . . . . . . . . . .7 x R-scheme .22* (Starred) . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. . . . . .44 d. . 98 2. . . . . . . . . . 93 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4. . . 98 2. . . . . . . . . . . . . . . . . . . .56 II. . . . . . . . . .3. . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . .4. . .37 c. . . . . . . 92 2. . . . . . 102 . . . . . . . . . . . . . . 89 2. . . . . . . . . . .5 x g center is unique by valuative criterion . . . . . . . . . . . . . . . . . 87 2. . . . . . . 88 2. . . . . . . 96 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .36 b. .42 b. x g . . . . . . . . .4. . . . . . . . . . . . . . . .4. . . . . 96 2. . . 89 . . . . . . . . . . . . . . .28 f. x .3. . . . .12 c. . . . . . . . . . . . . . . . . . . . . . . . . . 88 2.3. . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . . . .46 b (starred) . . . . . . . . . 87 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. .3. . . . . . . . . . 93 2.4 x . . . . . . . .8 d. . . . . . . 93 . . . . . . . . . . . . . . . . . . . . . . . . .45 II. . . . . . . . . . . . .27 e. . x g . . . . . . . x . . . . . . . . . . . . .4. . . . 93 2. . 88 2. .3. . . . . . . . . . . . .51 c. . .33 c. . . . . . . . . . . .3. . . . . . . . . . . . . . .3. . . . . . . .3. . . . . . . . . . . . . . 101 2. . . . 96 2. . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . x . .4.6 x g . . 95 2. . . . . . . . . . . x g . . . . . . . . . . . . . . . .53 e. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29 g. . . . . . . . . . . . . . . . . . . . . . 90 2. . . . . . . .40 f. . . .3.4. . .3. . . 91 2. . . . . . . . . . . . . . . . . . . . . . . x . . . . . . . . . . . . . . . . . . . . . . .16 e. . . . . . . . x . . .3 II. . . . . . . . . . .3. .47 c. . . . . . . . . . . . . . . . . x . . . . . . . 100 2. . . . . . . . . . . . . . . . . . . . . .3. . . .1 II. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 2. . . . . . . . . . . . . . . .3. . . .3. .3. . .4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .39 e. . . . . . . .4. . . . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . 91 2. . .15 II. . x . . . . . . . . . . . . . .4.3. . . . . . . .3. 89 2. . . .4. . . .49 II. . . . . . . . 101 2. . . .5 II. . .3. . . 97 2. . . . . x . . . .15 x . . . . . . . 94 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x . . . . . . . . . . . . . . . . . . . . .4 2. . . . . . . . x . . . . . . . . .3. . . . 94 . . . . . . . . . .3. . . . . . . . . . .3. . . . . . . . .3. . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19 x . . . . . . . . . . . . . . . . . . . . . . . . . 99 2. . .3. . . . . . . . . . . . . . . .30 II. . . . . . . . . . . . . . . . .4. x g . . . . . . . . . . 94 2. . . . . . . . . .34 II. . . . x . . . . . . . . . . . . .9 II. . . . . . . . . . . . . . . . . . . . . . 86 2. . . . . .3. . . . . . . . . . . . . . . . . . . 95 II. . x . 86 2. . . . .2 II. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6 b. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4. . . . . . . . . .3. . .4. x g . . . . . . . . .3. . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 2. . . . . . . . . . .11 b. . . . . . x 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32 b. 88 2. . 101 2. . . . x . . . . . . . . . . x g . . . . . x . .21 x . . . . . . . . .7 (starred) .4. . . .16 x g Noetherian Induction .3. . . . . . . . x . . . . . . . . . . . . . . . . . . . . . . x g specialization . . . . . . .54 f. .1 x g Nice example valuative crit . . . . . 100 2. .3. . . . . . . 97 2. . . . .2 x . . . . . . . . x . . x . . . . . . . . . 90 2. . . . . . . . . . . . . . . . . . . . x g . . . . . . .4 II. . .3. . . . . . . . . . . . .10 II. . . . 99 2.57 II. 92 2. . . . . . . . . . . . . . .4 x stopped g'ing here . . . . . . . . . . . . . .4. . . . . . . .3. . . . .8 x . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 2. . 94 2. . . . . . . . . . . . . . . . . . . . . . . . .35 II. .41 II. . 98 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x . . . . .23 x . . .52 d. .4. . . . . . . .25 c. .3. .2. . . . . . . . . . . .14 x g . . . . 92 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55 II. . . . . . . . .38 d. . . . . . . . . . . . . . . . . . x g 2. . . . .18x Constructible Sets . . . . . . . . . . . . . . . . . . . . . . .50 b. . . . . . .20 x g Dimension . . . . . . . .3. .3. . . . . . . . . . . .3 x g . . . x . . . . .26 d. . . . . . . . . . . . . . . . . .14 e. . . 94 2. . . . . .13 d. . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. .17 x Zariski Spaces 2. . .24 b. . . .3. . . . . . . . . . . . . .43 c. . . . . . . 89 2. . . . . . . . . . . . . . . . . . . . . . . . . . .48 d. . . . . . . . . . .4. . . . . . .31 II.3. . . . . . . . . . .3. . . . . . . 90 2. . .

. .5. .5. . . .5. .25 II. . 108 2. . . . . . . . . . . . . . . .15 x Extension of Coherent Sheaves . . . . . . . . . . . . . . . . . .37 (d) x . . . . . . . . . . 114 2. . . . . . .5. . . . . . . . . . . . .6 (a) x g 2. . . . . . . . . . . . . . . . . . . . . . . . .5 II. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . .28 (c) x . . . . . . . . . . . . . . . . . . . 105 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 2. . . .21 (b) x . . . . . . . . . . . . . . . . . . . . . . .5. . . . . . . . .17 II. . . . 106 2. . . .5. . .26 II. . . . 113 2. . . . . . . . . 112 2. . . . . .22 (c) x . . . . . . . . . . . . . . . . .5. . . . . . .11 x g . . . . . . . . . . . . . . . . 104 2. . . . . . . . . . . . . . . . . . . . . .2. . . . .5. 103 2. . . .5.7 x 2. . . . . . . . .24 b. . . . . . . . . . . . . . . . . . .5. . 106 2.11 (c) x g .33 II. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10 Chow's Lemma (starred) . . . x . . . . . . . . . .34 II. . .12 x Examples of Valuation Rings 2. . . . . .4. . . . . . . . . . . . . . . . . . . . . . . .5. . 113 2. . . . . . . . . . . . . .5. . . .29 (d) x . . . . . . . . . . . . . . . . . .15 (d) x . . . . . . . .4 (d) x g Projection Formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 2. . . . . . .9 x . . . . .28 (3) x . .4. . . . . . . . . .5. . . . . . . . . . . . . . . . . . .18 b. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 2. . . . . . . . . .14 (c) x g . . . . 114 2. . . . . . . . . . . .5. . . . .5. . . . . . . . . . . .4. . . 108 2. . . . . . . . . . .5. . . . 106 2. .5. . . . . . . . .7 II. . . . . . . . . . . . . . . . . . . .5. . . . . . . . . . . . . . . . . . . . . 116 2.4. . . . . . 103 2. . . . . . . . . . .5. . . .12 x g .23 II. . . . . . . . . . . . . . . . . . . . . 116 . . . . . . . . . . . . . . . .20 II. . . . . . . . . . . . .26 b. . . . . . . . . . . . . .36 (c) x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 2. . . . . 102 2. . . . . . . .4.4. . . . . . . . . 114 2. . . . . . . . . . .5. . . . . 112 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13 x . . . . . . . . . . . .5. . . . . . . .6 (b) x . .5. . . . . . . . . . . . . . . . . 106 2. . . . . . . . . . .5. . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . 105 2. .4. . . . . . . . . . . .5. . . . . . . . . . . . . .27 (b) x . . . . . . . . . . . x . . . . . . . . . . . . . 110 2. . . . . . . . 110 . . . . . . . . . . . . . . . .5. . . . . . . . . .4. . . . . . . . .38 II. . . . . . .5. . . . . .5. . . . 116 2. . . . . . . . . . . . x g closed immersion is nite. . . . . . . . . . . . . . . . .2 (a) x . . . . . . . . . . . 109 2.5. . . . . . . . 109 2. . . . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x . . . . . . . . . . .9 II. . . . . . . . . . . . . . . . . . . . . . . . . . . . .5. . . . . . . . . . . . . . . . . . . . .19 (c) x . . 107 2. . . .1 II. . . 104 . . .18 II. . . 108 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 2. . . . . . . . . . . 107 2. . . . . . .5. . .5. . . . . . .5. . . . . . . . . . . . . . . . . . . . . . . . part of starred . . . .5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17 f. . .5. . . . . . . . . . . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 . . . .8 II. . . . . . . . .5. . . . . . . .5. . . . . . . . . . . .24 (b) x . . . (1) x . . . . . . . . .5 x . . . . . . . .5. . .32 (b) x g . . . . .5. . . . . . . . . . . . . . . . . . . . . . . . . . .25 (c) x . . . .8 x . . . . . . .5. . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . 106 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4. . . . .4. . . . .5. . . . . .20 b part of starred . . . . . . . . . . . . . . . . . . . . . 106 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 2. . 105 2. . . . . . . . . .27 (2) x . . . . . . . . . . .11 x .5. . . . .4 x . . . . . . . . . . . . . . . . . . . . . 107 2. . . 112 2. . . . . . . . . . . . . . . .1 g x 2. . . . . . 111 2. .5. . . . . . .13 (b) x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31 II. . . . . . . . 114 2. . . .used stein factorization . . . . . .5 II. . . . . . . . . . . . . .5. . . . . . . . . . . .3 (c) x g . . . .10 x . . . . . . . . . . . . . . . . .2 (b) g x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9 x g important . . . . . . 103 2. . . . . . . . . . . . . . .5. . . . . . .5. . . . . . . . . . . . . . .5. . . . . . . . . . . .35 (b) x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12 II. . . . . . .22 d. . . . . 102 2. . . .19 II. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21 c. . . .23 II. . . . . . . . . . . . . . . . . . . . . . . . . 111 2. . . . . . . . . . . . . . . . 113 2. . . . . .14 x . . . . . part of starred . . . . . . . . . . . . . . . . . . . .5. . . . . . . . . . . . 104 2. . . . . . . . .10 b. . . . . . . . . .3 x g . . . . . . . . . . .5. . . . . . . .5. . . . . . . . .16 (e) x . .5. . . . . . . . . . . . . . . . .5. . . . . . . . . .5 (a) x g . . . . . . . . . 103 2. . . .30 II. .4.5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 2. . . . . . .5. . . . . . . . . . . . . .

. . . . . .2 x . . . . . .6. .53 II. . 134 . . . . . . .5. . . . . . . . . . . .4 x . . . . . . .46 (d) x . . . . . . . . . . . . . . . . . . . . .4 b. . . . . . . .5. . . . . . . . . . . . . . . . . . .48 II. . . . . . . . . . . . . . . . . . . 134 2. . . . . . . . . .5. . . . . .7 II. . . . . . . . . .7. . . . . . . . . . . .6. . . . . . . . . . . . . . .2 (starred) Varieties in Projective Space 2. .13 b. . . . . .5. . . . . . . . 137 2. . . .10 (b) x .5. . .6. . . . . . . . .7 (starred) 2. . .2 II. . . . .16 xc a. . . . . . . .8 (a) x g (use easier method) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 2. . . 127 2. . . . . . . .6. . . . . . . .7. 117 2. . . . . . . . . . x 2. . . .11 x g ample large multiple is very ample.5 (b) x g 2. . . . . .7. . .12 x g:1st paragraph .3 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49 (b) x g . . . . . .7. . . . . . . . . . . . . . . .7 x Projective Morphisms . . . . 136 2. . . . . . . .51 d. . . . . . . . . . . . . .9 II. . . . . . . . . . . . . 132 2. . . .7. . . . 128 2. . . . . . . . . . . . . . . . . . . . .6. .6. . . . . . . . . . . . . . . . . . . . . . . . .56 d. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 2. .9 c. . . . . . . . . . . . . . . . . . . . . . . .43 II. . . . . . . 136 2. . . . . . . . . . . . . . . . . x . . . . . .5. 118 . . . . . . .3 II. . . . . . . . . . . . . . . . . 118 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55 c. . . . . . . . . . . . . . . . . . . . . . . . . . . .5. . . . . . g Tensor Operations on Sheaves . . . . . . . . . . . . . . . . . .44 (b) xc g 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 . .47 (e) xc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x 2. .6. . . . . . . . .6. .10 d. . . . .2. . . . . . . . . . . . . . . .5. 124 2. . 121 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40 (c) x . . 134 2. . . . . . . . . . . . . . . . . . . . . . . . .7. . . . . . . . . . . .12 II. . . .6. . . . .5. . . . . . . . . . . . . . . . . . . . . 134 2. . . .5. . . . . . . . . . . . . . .6 x Divisors . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7. . . . . . . . . . . . . . . . . . . .5 II. . . .42 e. . . . . . . . . . . 123 2. . . . . .7. . . . . . . . . . . . . . . . . . . 136 2. . . . .6. . . . . . . . . . . . . . . . . . . . .3 II. . . . . . . . . . . . . . 123 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6. 119 2. . . . . . . . . . . .6. . 137 . . . . . . . . . . . 135 . . . . . . . . . . . . . . . . . .1 II. .5. . . . . .6. . . . . . . . . . . . . . . . . . . . . . .7 II. . . . x . . . . . . . . . x . . . . . . . .6 x g . . . . . . . . . . . . . . . . . . . .1 II. .7 x g Some Rational Surfaces 2. .7. . . x .6. . .5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6. . . . . . . . . . . . . . . . . . . . . . . . .5 x g 2. .12 II. . . . . . . . . . . . . . .6. . . . . . . . . . . . . . . . . . . . . . . .7. . . . . x . . . x . . . . . . . . . . . . . . . . . 117 2.5. .5. . . . .6. . . . . . . .6. . .16 II. . . .2 II. . . .5. . . . . . . . . . . . . . . . . . . . . .7.5. . .7. . . . . . . .7. . . . . .6. . . . . . . . . . . . . . . . . . . . . . . .7. . . . . . . . .7. x g . . . . 117 2. . . . . .10 x g The Grothendieck Group K(X) . . . . . . . . . . . . . . . 138 . . . . . . 126 2. . . . . . . . . . . . . . . . .7. . . . . . . . . . . . . . . . . . . . . . . . . . . .7. . . . . . . . . . . . . . . . . . . . . . . . .6. . . . . . . . . . . . 131 2. . . . . . . . . . . . . . . . . . . .15 (c) x . . . . . . . . . . . . . .6 (c) x g . . . . . .6. . . . . . . . 128 . . . . . . . . . . . . . . . .6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9 (starred) Singular Curves (starred) . . . .45 (c) cx . . . . . . . . . . . . . . . . . . .1 x g . 123 II. .5. . . . . . . . . . . . . . . . . . 128 2. . x . . . . . . . . . . . . . 136 2. . . . . . . . . . . . . . . . . . . . . .6 b. .6. . . . . . . . . . . . . . . . . . . . .17 x Ane Morphisms . . . . 116 2. 135 .1 x g 2. . . . . 121 2. . . . . . . . . . . . .6. . . . . . . . . .8 II. . . . . . . .6. . .5. . . . x 2. . . . . . . . . . . . . . . . . . . . . . . . x . . . . x g . . . . x g . . 120 2. . . . . . . .7 (d). . . . . . . . . . 133 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7. . . . . . . . . . . . . . . . . . . . . . . .41 (d) x . . . . . . . . . . . 124 . . . . . . . . . . . . . . . . . . . . . . . . .5. . . . . . . . .4 II. . . . . . . . .6 . . . . . . . . . .4 a. . . . . .39 (b) x . . . . . .7. . . . .6 x g The Riemann Roch Problem 2. .7. . . . . . . . . .14 (b) x . . . . . . . 131 2.11 (c) x .7. . . . . 119 2.18 x Vector Bundles 2. . . 122 2. . . . . . . . . . . . . . . . . . . .8 b. . . . 135 2. . . . . . . . . . . . . . . . . . . . . .54 b. . . .5. . . . . . . . . .7. . . . . . . . . . . . x . . . . . . . . . . . . . . . x g . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 2. . . . . . . . 124 2. . . . 127 . . .50 (c) x g (used for stein factorization) . . . . . 136 . . . . . . . . . . . . . . . . . . . . . . . . . . . .13 II. . . .15 b. . . . 137 . . . . . . . . . . . . . . . . . . . . . . . . . .5. . . . . . . . . . . . . . . . 119 2. . . . . .14 II. . . .52 e. . . . . 130 2. . .

. . . . . . .2 III. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 2. . . . . . . . . . . . . . . . . . . . . . . 144 2. x g 2. . . . . . .8 . . . . . . . . . . . .8. . . . . . . . 141 2. . . . . .7. . . . . .11 x .4 x Complete Intersections in Pn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 . . . . . . . . . . . . .18 II. .19 b. . . . . . . . . . . . . . . . . . . . .22 II. . . . . . . . . . . . . . . . . x . . . . . . . . 145 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7. . . .8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19 b.8. .1. . .6 x Innitesimal Lifting Property . . .8. . . . . . . . . . . . . . . . . .8. . . . . . . . . . . .2. . . . . . . .29 b. x g . .16 c. . . . . . . . . 143 2. . . 144 2. . . . . . . . 141 . . . . . . .2. . . . . . . . . . . . . 153 . . . . . . . . . . . . . . . . . . . . . . . . . . . .7. . . .9 II. 146 2. . . . . . x . . . . . . . . . . . . . . .7. . . x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2. . . . . . .8 c. . . . . . . . . . . . .20 c. .1 II. .8 excision x 3. . . . . .5 III. . . . . . . . 138 2. . . . . . . . . . . . . . . . . . . . .21 II. . . .11 c. . . .7. x g . . . . . .5 II. . . . . . . . . . . . . . .8. . . . . . . . . . . .3 x Cohomology with Supports . . . . . . . . . . . . . .1 3. . . . . . . . . . . . . 148 2. . . . . .8. . . . . . . . . . . . . 141 2. . . . .2. . . . . . . . .3 c. . . . . . 150 III Cohomology 3. . . . . . . .7. . . . . . . . . . . . . 145 . . . . . . 153 3. . . . . . . . . .17 b. .8. . . . . . . . .10 b. . . . . . . . . . . . . . . . . . .7. 154 . . .10 x P^n Bundles over a Scheme 2. . . . . . . 150 2. . . . . . . . . . . . . . . . . . . . 145 2. 153 3. . . . . . . . . . . . . . . . x g . . . . . . . . .3 III. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 x . . .1 III. . 155 . .7. . . . . . . . . . . . . . . . . . .8. . x g . . . . . . . . . .8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9 x g 2. . . . .1 III. . x g . . . . . .8. . . .26 II. . . .9 3 II. . . . . . . . . 142 II. . . . . . . . . . . . . . . . . . . . . . . . . . .8 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 2. . . . . . . . . . . . .7. . .8.13 A Complete Nonprojective Variety * 2. . . . . . . . . . . . .12 d. . . . . . . . . . .7. . . . . . .7. . . . . . . . . . . . . . . . . . . . .2. . . . . . . . . . . . . . . . . .15 g. . . . . . . . 139 2. . . . . . . . . . . . . . . x Flasque Global sections are exact 3. . .8. . . . . . . . . . . . . . . . . . . . . .9 Formal Schemes . . . . . .16 II. .7. . . .2. . . 140 2. . . . .2. . . . . . 147 2. . . .7. . . . . . . . 142 2. . . .2. . . . . . . . . . . . . . . .14 x g 2. . . . . . . . .8. . .2 x asque resolution g . . . . . . . . . . . . . .20 II. . . . . . . . . . . . . . . . . . . .14 f. . . .8. 149 2.4 b. . . . . . . . . . . . . . . . . . . . . . .8. . . . . . . . . . . x g . 147 2. . . . . . . . . . . . . . . .8. . . . . .8. . . . . . . . . . . . . . . . . . . . .7. . . . . . . . . . . . . . 140 2. . . . . . . . . . . . . . . .8 x sections vs quotient invertible sheaves 2. . . . . . . . . . . . . . . .8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 .2. 144 2.8. . . . . . . . . . . . . . .skip . . . . . . . . . . .2 x 153 . . . . . . . . . . . . . . . . .8. . . . . .3 x g Product Schemes . . . . . .24 b. . . . . . . . . . . . . .22 d. . . .a x g 3. . . . . .23 II. . . . . . . . . . . . . . . . . . . . . . . . 140 2. . . . .17 II. . . . . . . . . . . . . . . . . . . . . .7. . . .12 x g . .7 b. . . . . . . . . . . . . x g . . . . . . . . . . . . . .8. . . . . . . . . . . . . . . . . . . .8. . 142 2. . . . . . 139 . . . . . . . . .5 x g Relative Canonicals Important! . . . . . . . . . . .28 II. . .2. . . . . x g . . .7. . . . . . . . . x . . . . .8. . . . . . . x . . . . . . . x . . . . . . . . . . .8. .2 x . 149 2. . x g . . . . . . . . . . . . . . . .4 x Mayer-Vietoris . . . . . . . . . . . 153 . . . . 155 .2. . .7. . . . . . . . . . . . . . . . .7. . 154 . . . x . . . . . 143 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 3. . . . . . . . . . . . . . . 153 3. . . . . . . . . . . .2. . . . . . . . . . . . . . . . . . . . . . . . . . 143 2. x g 2. . . . . . . . . .2. . . . . . . . x 2. . . . . . . .2 b. . . . . . . . . . . . . . . . . . . . . . . . . . .7. . . . . x . . . . . . . . . . . . . . . . . .7. . . . . . . .18 II. . . . . . . . . . . . 139 . x . . . . . . . . . . .6 II. . . . .2 III. . . . . . . . . . . . . . . . . . .8. . . . . . . . . . . . . . . . . . . . . . . . . . . . .8. . .8. . . . . . . . . . . . . 146 2. . . . . . . . .27 II. . .7. . . . . .8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 . . . . . . . 150 2. .21 b. . . . . . .2. . . .25 c. . . . .13 e. . . . . .c x 3. .6 x . . . . . . . .4 d. . . . . . . . . . . . .8. . . . . . . . . . . .8 x Dierentials .7 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 x . . .2. . . . .9 II. . . . . . . . . . . . . . . . . . .2. . . 147 2. . . . . . . . 154 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . .5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 x .5. . . . .5. . . . x . . . . . .3. . . .4 c. .5 x . .4. . . . . . . . . .9 x . . . . . . . . . . . . . . . . . . . 171 . 3. . . . . . . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11 III. . 163 . . . . . x is left exact. . . . . . . . .4. . . . . .18 III. . . . . . . . . . . . . . . . . . . . 168 3.13 b. . . . . . . . . . .1 x g 3. .12 III. . . . . . . . . . . . . .13 b. . . . . . . . . . . 166 3. . . . .4. . . . . . . . .4 x 3. . . . . . . . . .6 III. . . . . . . .5. . . . . 170 3. . . . . . . . . 156 3. . . . . . . . . . . . . . . . . . . . x . 157 3. . . . . . . .2 x . . . . . . . . .4. . . . 161 3. . . . . . . 166 3. . . . . . . . . .4. . . . . . . . . . . .4. .11 III. . . .3 g nice x 3.4. . . .5. 174 . .1 x g pushforward cohomology ane morphism . . . . . . . .4 b. . . . . . . . . . . . . . . . . .4 . . x . 162 .7 x 3. . . . . . . . . . . . . . . . . . .10 c. . . x . . . . . . . . . .5. . .1 III.3. x 3. . . x . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 3. . . . .12 b. . . . .7 b.15 e. . . 169 . . . . . . . . . . . .4 x Cech Cohomology .5. . . . . . . .10 III. . .6 c.6 x . . . . . . .5 x . . . . . . . . . . . 164 3. . .8 x cohomological dimension . . . . 157 3. . . . . . . . . . . . . . . . . . . . . . 158 3. . .4. . . . . . .3 b. . . . . . . . . . . . . . . 165 3. . . . . . . . .4 III. . . . .3. . . . . . .2 III. . . . . . . . . . . . . . . .5. . . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x 3. . . . . . . . . . . . . . . . . . . . . . 159 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 . . . . .6 III. . . . . . . .6 x .6 x curves on a nonsingular quadric 3. . . . . . . . . . . . . . . . . . 158 . . . . . . . . . . . 169 . . . . . . . . . .5. .11 d. . . . . . . . .4.3. . . . .16 III. . . .2 III. 162 3. . . . . . . . . . . . . . . . . . . .2. . . . . . . . . .14 c x. . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 3. . . . . . . . . . . . . . . . . . x g . . . . .b. . . . . . . . . . . . . . . 161 III. . . 160 3. . . . . . . . . . . . . . . . . . . . . . . . . x 3. . . .4. . . . . . 161 3. . . . . . . 172 3. . . . . . . . . .8 c. . . . . . . . .10 III. . . . . . . . . . . . . . . . . . . . . . . . . . . .3 III. . . . . . . . .7 x . . . . . . . . . . . . . .4. .4. . .3. . . . . . . . . . . . . .2. . . . . . . . . . . . . .4. . . . 3. . . . . . . . . . . . . . . . .5 x g 3. . . . . . . . . . . . . . . . . . . .3. . . . . . . . .4. . . . . . .7 b. . .10 part c. . . . . . . x . . . . . . . . .4.5. . . .3. . . .1 III. .4. . . . . . . 171 . . . . . . . . . . . . . . . x . . .2 x .3. . . . . . . . 157 III. . . . .6 x . . . . . . .3 b. .5 . . . . .5 x Cohomology_Of_Projective_Space .8 x Localization not injective non noetherian. . . .4. . . . . . . . . . . . . . . . . . . . . . . . . 158 3. . 157 Γa 3. . . . . . . .2 x 3. . . . . . . . . . . 156 . . . . 167 . . . . . . . . . . . . . . . x .1 III. . . .5. . .4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5. . . . . . x .3. . . . . . .8 III. . . . . . . . . . . . . .4.4. . x g Important genus is birational invariant for curves!! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 3. . . . . . . . . . . . . .3. .4. . . . . 172 3. . . . . . 168 III. . x . . .3 . 170 . .5. . . . . . . . . . . . . . . .3 3. . . . . . . . . . . . . . . . . x g complete intersection cohomology. . . . . . . . . . . . . . . .4. . . . . .7 b. . . . . . . . . . . . . . .2. . . . . . .13 III. . . . . . . . 160 . . . . . . . . . . . 159 3.1 x . . . . . . . .3.4. . .3. . . . . .7a g Cohomology of circle 3. . . . . .3. . . . . . . . . . . . . . . . . . . . .4. . . . . .3a. . . . . . . . . . . . . .2. . . . .3. .12 III. . . . . . . . . . . . . 173 . . . . . . . . . . . . . . . . . x 3. . . . . . . . . . . . . . . . . . . . . . . . .5 III. . . . . . . . . .2.8 III. . . . .4. . . . . . . . . . . . . . . . . . x g 3.5. . . .5. 167 3. . . . . . . . 164 3. . . . . . . . . . . . . . 161 3. . . . . . . . . . . . . .5. . . . . . . . . . . . . . . . . . . . 156 3. . . .3. . . . . . . . . .2 III. . . . . . . . . . . . . . . . . . . 166 3. . . . . . . . . . . . . . . .2. .3. . . . . . . . . 162 3. . . . . . . . . . . . . . . . . . . .11 III. . . . . . .5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. .5 III. . . . . . . . . . . . . . . . 165 . .9 III. . . . . . . . x Arithmetic genus 3.5. . . . . . . . . . . 160 . . . . . . . . . . . .9 b. . . . . . . .4.10 (starred) 3. . . . . . . . . . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2. . . . . . . .12 x III.3 x Cohomology of a Noetherian Ane Scheme . . . . . .13 b. . . . . . . . . . . . . . . . . . . .4. . . . . . .3. . . . . . . . . . . . . . . . . . . . . . . 163 3. . . . . . .4 x Cohomological Interpretation of Depth 3. .5 x . . . . . . . . . . . . . . . . . . . .17 III. . . . .11 x 3. . . . . .5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 3. . . . . . . . . . . . . . . . .3. . .4. .9 III. . . . .3. . . . . . . . . x g . . . . . . . 170 . . . . . . . . . . .

. . . . . .7 III. . . . . . . . . .7 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x . . .9. . . . . . . . . 191 . . . . . . . . . . . . . . . . . . . . . . . . . .3 b. . . x . . . . . . .5. . . 181 3. . . . . . . . . . . . . . . . . . . . . . . .9 x Flat Morphisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8. . . . . . . .16 b. . . . . . . .4 III. . . . . . .10 x Duality for Finite Flat Morphism . . . . . . . . . . . . x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6. . 177 . . . . .9 x 3. . . . . .1 III. . . . . . . . . .24 III. . . . . . . . . 189 . . . . . . . x . . . . . . . . . . . . . .7. . . .6. . . . . . . 184 3. . . . . 183 3. . . . . . . . . . 186 3. . .5. . . . . 176 3. . . x 3. . .1 III. . . . . 178 3. . . . . . . . .8 3. . . . .4 III. . . . . . . .9. .7 3. . .5. . . . . . . . . . . . . . . .5 x . . . . . .6. . . . . . . . . . . . . . . . . . . . . . . 182 . . . . . . .6. . . .20 b. . . . . . . . . . . . . . . . . . . . . . . . . . x g nite pullbacks ampleness 3. . . . . . . . . . . . . . . . . . .6. . . . .5. . . . . . . . . . . . . . . . . . . .a x g 3. . . . . . . . . . . . . . . .8. . . .6. . .6. .5. . . . .12 III. . . . . . . . . . . . . . . .8. . . . . . . . .3 III. . . . x . . . . . . . . . . . . . . . . . . . . .10 III. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6. . . . . . . . . .8.2 III.17 c.8 e. . . . . . . . . . . . . . .8. . . . . . . . . . . . . . . . . . . . . . . . . . .18 d. . . . .6. .4 x . . . . . .1 x . . . . . . . . . . . . . . . . . . . . . .20 d. . . . . . . . . 180 3. . . . . . . . . .6. . .2 x 3. . . . . . . . . . . . . . . . . . . . . . . . . . . .17 III. . . . . . . . . . . . . . . . . . . . . . . 185 3. . . x . . . . . .6. .5. . . . . . . . . . . . . . . . . . . . . . . . .6. . . . . . . . . . . . . .2 x twisted cubic 3. . . . . . . . . . . .6.15 III. . .2 III. 191 3. . . . . . . . . 189 III. . . . . . . . . . . x . . . . . .5. x 3. . . . . . . . . . . . 175 3.1 x g Special Case Kodaira Vanishing . . . . . . . . . . . . . . . . . 175 . . . . . . . . . . . . . . . . .23 III. x g 3. . . . . . 191 . 176 . . . . . . . . . . . . . . . . . . . . x . . . . . . . . .2. . . 186 III. . . . . . . . .7 x g . . .15 III. . . . . . . . . . . . . . . . . . . . . . . . . 182 3.9.5. . . . x . . . . . . . . . . . . . . . . . . . . .8. x . . . . .7. . . . . . . . 190 3. . . . . . . . . . . . . .18 b. . . . . . . . . . .6. . x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x g . . . . . .6. . . . . . .5 b. . . . . . . . . . .8 x . . . . . . . . . . . . .16 x g ample i red is ample 3.6. .6 . . . . . . . . . . . . . . . . . .a x 3. . . . . . . . . . . . . x . . . . . . . . . . 188 3.4 b. . . . . . .8. . . . . . 180 3. 187 . . . . .5. . . 175 . . . . . . . . . . .6. . . . . . . . . . . . . . . . . . . . . . . . . . .4 (starred) .6 x Ext Groups and Sheaves . .19 III. . . . . . . . .5. . . . . .3 III. . . . . . . . . . .5. . . . . . . . . . . . . . . .3. . . . . . . . . . . . . . . . . . . . 190 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7. . . . 182 3. . . . . . . . . . .5. . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 . . . . . . . . .8. . . . . . x . . . 179 . .6 III. . . . . . . . . 189 3. . . . 184 . . . . . . . . . . .10 x g 3. . . . . . . . . . . . . .6. . . . .6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8. . . . . . .7 d. . . . .9. .8 x Higher Direct Images of Sheaves .4 III. . . 186 3. . .6.9. . . . . . . . .7. . . . . . . . . . 177 III. 185 3. . . . . . . . . . . . . . . . . .7.7. . . . . . . . . .1x . . . . . . . 191 . . . . . . . . .8. . . .5. . . 178 . . . . . . . . . . . . . . . . .8 b. . .8. . . . . . . . . . . . . . . . . . . . . . . . . . .6. . . . .3 x g 3. . . . . . . . . . . . . . . . . . . . . . . . .8. . . . . 190 3. . . . . . .9 x g Nonprojective Scheme 3. . . .7. . 188 . . . . . . . . . . . . . . . . . . . . .5 b. x 3. . x . . . . . . x . . .1 x g Leray Degenerate Case . . . . . . .2 x g 3.3 x Cohomology of dierentials on Pn . . . . . . 180 3. . . . . . . . . . . . . . 184 3. . . . . . . . . . .8. . . . . . . . . . . . . . . . . . . . . . x 3. . . . . . . . . . . . . . . . . . . . . . . .a. . . . . . . .9. . . . . . . . . . . . 181 3. . . . . . . . . . . . . . . . . .a. . . . .7 x Serre Duality Theorem . . . . . . . . . . . . . . . . . . . . . . . . 186 III. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6. . . . . .7. .6. . . . . . . . . . . .14 c. . . . . .1 III. . . . . .9 3. . . . . . . . . .6. .6. .6. . . . . . . . . . . . . 183 3. . . .13 III. . . . . . . . . . . . . . . . . . . . . . .14 b. . . . . . . . . . . 178 3. . . . . . . . . . . . . . . . . . . . . . . . 174 3. . . . 188 3. . . . . . . . . . .5. . . . . . . .6. . . . . . . . . . . . .9. 191 3. . . . . . . .21 c. . . . . . . . . . . x . . . . . . . .2 III. . . 176 . . . . .3 III. . . . . . . . . . .6. . . . . . . . . .4 x 3. . . . . . . . . . . . . . 187 3. . . . . . . . . . . . . . . . . . . . .19 c. . . . . . . . . .11 b. .2 III. . . . . . . . . . . . . . .1 III. . . . . . . . . . . . . .5. . .6. . . . . . . . . . .6. . . .3 x g Projection Formula derived 3. . . . . . . . . . . . . . . . . . . . . . . 176 . . . . . . . . 174 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6. . . .22 d. .9 c. . . .6 c. . . . . . .

. . . . . . 193 3. .13 III. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1.2 x Hurwitz Theorem . . . . . . . . . . . .9. . . . . . . . . . .1 x g upper semi-continuous tangent dimension . . . . . . .10 g x . . . . . . . . . . . . .6 x g nite morphism to P . . . . . .1x g projective space simply connected . . . . . . . . . . . . . . .9. . .9. . . . . . . 197 3. . . . . . . . . . . . . . . . . . . . . . . .9. . . . . .9. . . . . . . . . . 206 . . . . . . . . . . . . . . . . . . . . . . . .14 III.1 III. . . .9. .1. . . x . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 x etale neighborhood x . . .5 4. . . . . x g 4. . . 209 . . . . . . . . . . . . . . .1. . . . . . . . .10. . . . . . . . . . . . . . . .6 x g Etale Cover of degree 2. . . . . . . . . . . . . . . . . .12. . . . . . . . . . . . . . 197 3. . . . x 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9. . . . . . . . . . . . . . . . 197 3. . . . . .2 III.7 x g 4. . . . . . . . . . . .10. . . . . . . . . . . . . . . . . . . . . . . .11. . . . . . . . . . 207 . . . x . . . . . . 193 3. . . . .10. . . . . . . .1 x IV. . . .10. . . . . . . 204 3. . . . . . . . .6 III. . . . . . . . . . . . . . . . .1. . . . . .9. . . .4x . 200 3. . . . . . . . .1. . . .4 x 4. . . . . . . . . . . . . . . . . .8 b. . . . . . . . . 193 3. . . . . . . . . . . . . . . . . . . . . . . . . . .14 IV. . . . .1 III. . . . . . . . . . . .9. . . . .12. .12 III. . . . . . . . . . . . . 203 3. . . . . . .8 b. . . .12. . . . . . . . . . . 198 3.11 x Theorem On Formal Functions . . . . . .4 x . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10. . . 204 3. 192 3. . . 194 3. . . . . . . . .15 b. . . . . . . .1. . . . . . . . . . . . .7 III. . .11. . . . . . .8 x g arithmetic genus of a singular curve .11 x g dierence of very amples. . . . . .10. . . . .1. . . . . . . . . . . . . . .11 III. . . . . . . . . . . . . 206 . . . . . . . .1. . . . . . . . . . . . . .10 x g 3. .5 x Very Flat Families 3. . . . . . . . . . .12. . . .3 x . . . . . . .12 x Semicontinuity . 208 . .10. 193 . . . . . . . . . . 207 . . . . . . . . . . 196 3. . . . .5 c. . . . . .1. . . . . . . . . . .12. . . . . . . . . . . . . . .1. . .3 III. . . . . . 209 L (D) 4. . . . .3 g Nonproper Curve is ane 4. . . . . .1. . . . .2. . . .1 x regular != smooth always . . . . . . . . . . x . . . . . . . . . . . . . . . . .13 x. . . . . . .1. . . . . .3 III. . . . . . . . . . . . . .1 x Riemann_Roch_Theorem . . . .10 d. . . 194 3. . . . . . . . .12 III. . . .9.1. . . . . . Alternative riemann-roch 4. . . . . . . . . . . .4 x Principle of Connectedness . . . . . . . . . . .1. . . 194 . . . . . . . . . . .11. . .1. . . . . . . . 208 4. . . . .1.1 III. . . .9. . . . . 204 3. . . . . 207 4. . . . . . . . . . 4. . . . .12.4 III. . . . . . 196 . . . . . . . .10. . . . . . . . . . 209 IV. . . . . . . . . . . . . . . . . . . . . . . .9 c. 4. . . . . . . . 205 4 IV Curves 4.10 III. . . . .5 III.9. . . . . . . . .3. . . . . . . 205 4. . .9 x rigid example 3. 209 . . . . . . . . . .5 III. . . . . . . 200 3. . . . . . . . . . . x . .12. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208 . . . . . . . . . . . . 196 . .10 (b) x g Genus 0 is nonsingular. . . . . . . . . . . . . . .7 x Serre's linear system with moving singularities 3. . . . .12. . . . . .2 III. . . . . . . . .10. . . . . . . . . . . . . . . .7 x . 206 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x . . . . . . . . . . . .2. . 198 3. . . . . . . . .12. . . . . . .1 205 IV. . . . . .3 x IV. . . . . . . . . . . . . . . .1. . . . . . . . . .6 III. . . . . . . .2 III. . . . . . . . .2 g Regular Except pole at Points 4. . .11.10. . . .9 IV. . . . . .1. . . 203 3. . . . . . . . . . .2 x g . . .1. . . . . . . . . . . . . . . .12. .1. . . . . . . . .12 x g Invertible sheaves are 4. . . . . . . . . . . . . . . . . . . .7 III. . . . . . . . . .1. . . . . . . . . . . . . . . . . . . . . 210 . 201 . . . . . . . . . . . . .1. . .10. . . . . . . . . x . . . . . . . . . . . . . . . . .4 x open nature of atness . 207 1 IV. . . . . . . . . . . . . . . . . . . .2 x IV.4 III. . . . . . .2 x .1 x g higher derived cohomology of plane minus origin. . . . . . . . . . . .9. . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 4. . .11. . . . . . . . . . . .4 IV. . . . . . . . .16 III. . . . . . . . . .10. . . . . .3 III. . . . . . . 200 3. . . . . . . . . . .9. .10. . . . .11. . . . . . .1 g Regular except at a point 4. . . . . . . . . . .2 x g 3. . . . . . . . . . .4 III. . . . . . . .2 .1 IV. . .3 x . .9. . . .6 IV. . . . . .5 x g Dimension less than degree . . 201 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 . . . . . . .11. . . .3 x Rational Normal Quartic .11 III. . . . .9. . .9. . . .10. . . . . . . . . . . . . 199 . . . . . . . . . . 197 3.10. . . . . . . . . .6 x .5 x Picard Group of projective bundle . . . . . . . . . . . . . . . . . .9. . . . . .10 x Smooth Morphisms . . . . 199 3. . . . .9. . . . . . . . . . .7 IV. . .11. .1. . . . . . . . . 201 3. . . . .8 b. . . . . . . . . .11 x interesting g 3. . . . . . . . . . . . . . . .

. .8 x . . . . . . . . . . . . 222 4. . . . . . . 211 . . . x . . . .2. . . . . . . . . . . . . . . .1 x g 4. . . . . .5 IV. . . . . . . . . . . . . . . . . . . . . . . . .3 x g 4. . . . . . . . . .3. .1 IV. . . . . 226 . . . . 218 IV. . . . . . . . . .2.2. .6 conclusion. . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 . . . . 210 4. . . . . . . . . . . . . . . . . . . . . . . . . . .18 IV. . .3. .2 x g :a. . . . . . . . . . . . . . . . . . x branch divisor . . . . . . . . . . . x g No strange curves in char 0!!!! 4. . .2. . . . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . .15 IV. . . . . . . . . . .3. . . . . . .2. .2.2.2. .14 b. . . . . . . . . . . 215 . . . . . . . . . . . . . . . . . . .16 IV. . . x g . . . . . . . . . . . . . . .4 c.2. . .x 4. . . . . 220 . . . .4 c. . x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2. . .4 x g Funny curve in characteristic p . . . .19 b. . . . . . . . . 222 . . x . . .3. . . . . . . . . . . . . . . . . . . . . . . . . .6 x g Curves of Degree 4 4. . . . . . . .3. . .3 b. .3 4. .17 b. . . . . . . . . . . . . . . .3. . .4 x . . . 218 . . . . . . . . . . . . . . . . . . . . x g 4. . . . . . . .7 b. . . . . . . . . . . . . 212 c. . . . . . . . . . . . . . . . .3. . . . . . . . 218 4. . . . 218 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 4. . . . . . . . . . . . . . . . . . . . . . . . . . 4. . . 220 . . x g . . .4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2. . . . . . . . . . . . . . . . . . . . . . . 216 4. . . . . . . . . . . . x . . .3 x g . . . . . 224 IV. . . . . . . . . . . . . . . . . . . . . . .2. . . . . . . x g . . . . . . . . . . . . . . . . . . . . x . . x g 4. .3. . 220 . . . . . . . . . . . . . . . . . . x multiple tangents. . . 217 4. . . . . . . . . . . . . . . .3. . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . x g . .4 Elliptic Curves .4. . .21 d. . x . . . . . .23 b. . . . . . . . .2. . . . . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . .2. . . . . . . .2. . .1 x g 4. . . . 220 4. . . . . . . . . . . . .2. . . . . . . .10 x g . . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . . . . . x g . . . . . . . . . . . . . . . . . .12 c. . . . . . . . . . . . . .7 IV.3. . 212 4. . . . . . . . . . .3. . . . .2. . . .24 c. . . . . . . . . . . 211 . .2. . . . . . . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . 226 4. . . . . . x . . . . . 217 4. . . . . .13 g. . 222 4. . .2 x 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . .3. . . .2. . . . 219 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214 4. . 216 . . . . . . . . . 219 4. . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . . . . . . 216 4. . . . . . . . . . . . . . . . .11 x g . . . . . . . . . . . . . . . .22 IV. . . . . . . x g . . . 214 4. . . . . . . x g . . . . . . . . . . . . 221 4. . .10 d. . . . . x g 4. .11 e. . . . . .2. . . 227 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6 x g pushforward of divisors .3. . . . . . . .12 f. . . . . .20 IV. . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . . . .4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220 . . . . . . . . . . . . . . . . . . . . .3. . . . . . . .4 x g . . . . . .2. .2. . . . . . . . . . . . .2. . .18 IV. . . . . . . . . . . . . . . . . . . . . .19 IV. . . . . . . . . . . . . . . . . . . . . . . .10 IV.3.4. . . . . . . . . x g 4. . . 217 4. . . . . . . . . . . . . . .3. . . . .4. . . . . . .4. 228 . . . . . . .2 x g classication of genus 2 curves . .9 x g . . . . . . . . . . . .2 IV. . . .20 c. . . . . . . 223 4.3. . . . 211 . . .17 b. . . . . . . . . . . . . . . . . 219 4. . . . .3. . . .3. . . . . . . . . . . . . . .2. . . . . . .22 IV. . . . . . . .9 .4 IV. .4.21 b. . . . . . . . .2. . . . .3 Embeddings In Projective Space . . . . . . . . . . . .3. . . . . . . . . . .2. . . . . . . . . . .3. . . . . . . .4. . 214 4.5 x Automorphisms f a curve in genus >= 2 4. . .8 b. . . . . . . . . . . . . . . . . . .3.3.3. . . . . . . 222 4. . . . . . . . . . . . . . . . . . . x g . . . . . . . . . . . . . . . . . . . . . . 223 4. . . . x 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 4. . . . . . . . . . 217 4. . . . . .2. . . . . . . . . . . 211 . . . . . . . . . . . . . . . . . . . . . . . . . .8 c. . . . . . . . . . . . . . . . . . .7 x . 228 . 215 4. .b. . . .5 . . . . . .6 IV. . x 4. . . . . . . . . . . . . . . . . . . . . . . . . .2.3 b. . .3 IV. . . . . .13 IV. . . .2 IV. . . . . . .12 x g (just explain the advanced method) 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 IV. . . . . . . 221 4. . . . . . . . . . . . . . . . . . . . . . . . .14 h. . . . . . . . . .11 b. . . . . . . . . . . .3 x inection points gauss map . . . . . 214 4.4. . . . . . . . . . . . . . . . . . .2 IV. . . . . . . . . . . . . .7 x Etale Covers degree 2 . . . .16 IV. . . . . . . . . . . . . . . . x g . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . .15 IV. . . . . .c . . . . 223 4. . . . . . .2. . . . . . .3. . .5 x g 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x 4. . . . . . 219 4. . . . . . . . . . . . . . . . . . . .9 d. . . . .

. .22 b. . . . . . . .4. .4 x g 4. . . . . . . . . . . . 243 4. . . . . . 243 4. . .30 b. . . x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4. . . . . x . . . . . . . . . . . . . . . . . . . . . . . . . .not algebraic geometry . . . . .6. . . . .4. . . . . . 244 4. . . 242 4. . . . . . . . .4. . . . . . .7 c. . . . . . . . . . . . . . . . . . x . . . . . . . . . . . . . . . . . . . . . . .5. . .14 x Fermat Curve and Dirichlet's Theorem . . . 236 . . . . . . . . .3 IV. . . .17 a. . . . . . . . .11 c.4. . . . . . . . . . . . . 236 4. .4. . . . . . . .5 Canonical Embedding . .5 x g Curves of Genus 5 4. . .5 IV. . . . . . . . .6. . . . . . .32 d. . . . . . . . . . . . . . .a x g . . . . .38 IV.4. . . .34 IV. . . . . . . . . . . . . . . . . . . 249 4. . .4. . .3 x g Moduli of Curves of Genus 4 . .2 IV. . . . . . . . . . . . . . . . . .16 x . . . . . . . . . 242 4. . . . . . . . . . . .4. . . . . . . . . . . . . . . 230 4. . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x g 4. . x . .4. . . . . . . . . . . .4. 239 4. . . . . . . . . .4. 245 4. . x g . . . . . . . . . .4. . .5. . . . . . . . . . 241 4.25 b.36 IV.4. . 240 . . . . . . . . . . . . . 247 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29 IV. .2 IV. .6 b. . . . . . . . . . . . . . . . . . . . . .13 b. . . . .10 b. . . . . .39 Slight issue? . . . . . 240 4. . . . . . . . .4. . . . . . x . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . .1 IV. . . . . . . . . . . . .4. .20 IV. . . . x osculating hyperplanes 4. . x . . . x Hasse's Riemann Hypothesis for Elliptic Curves 4.4. . . . . . . . .28 IV. . . . . . . . . . . . . x . . . . . . . . . . . . . . . . . . 244 4. . . . . . . . . . . . . .7. .2 x g . . . .31 c. . . . . . . . . . . . . .4. . . . . . . . . . . . .4. . . . . . . . . . . . . 234 .x g kernel of frobenius 4. . . x . . . . . .27 IV. . . . . . . . . . . . . . . . . . .4. . . . . . . . . . . .15 x . . . . . . . . . . .33 e. . . . . . . . . . . . . . . . . . . . .6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248 . . . . . . . . . 246 . . 246 . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . 233 4. . . . . . 237 4. . . . . . . . . . . . . . . 4. . . . . . . . . . . . . . . . . . . . . . . . 239 4. . . . . . . 230 . . . . . . .9 IV. . . . . . . . . . . . . . . . . . . . . .4. . . .1 IV. . . . . . . . . .21 x skip . . . . . . . . . .9 x g isogeny is equivalence relation. . 249 . .18 x . . . . . . . . . . . . . .5. . . . . 242 4.4. . . . . . . . . . . . . . . . .19 b. . . . . . . . . .6 IV. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .42 IV. . . .4. . . . . . . .15 e. . . . . . . . . .5. . . . .4. . . . . . . 238 4. . x . . . . . . . . .7 b. . .4. . . .4. . . . . . . . . . . . . . . . .4 IV. . . . . . . . . . . .4. . . . . . . . . . . . . . . . . x . . . . . . . . . . . . . . . . . . . . . . .4. . .5. . . . .24 IV.5. . .11 x g . . . 234 4. . . . . . . .4. . . . . . . . . .5. .a x 4. . .12. . . . . . . . . . . . . . . . . . . . . . . . . . . .26 IV. . . . . . . . 247 . . . . . .4. . . . . . . . . . . . . . . .5. . . . . . . . .5.17 IV.5 x g 4. . . . . . . . . . . x . . . . . . . . . . . . . . 234 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4. .4. . . . . . . . . . . . . . . . . . 235 . . . . . . . . . . . . . . . . . . . . 237 4. . . . . . . . . 244 . . . . . . . . .18 IV. . . . .6. . . . . . . . 244 4. . . . . . . . . . . . . . . .7 x g Dual of a morphism . . . . . . . . . . . . . 4.5 x 4. . . . . . . . . . . . . . . . . . . . .5. . . . . . . . . . . . . . . . . . . . . . .8 x Algebraic Fundamental Group . . . . . . . . . . . . . .10 x picard of product on genus 1 4. . . . . . . . . 239 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4. .4. . . . . . . . . . . . . . . . . . . . . . . . . .6. . . . . . . . . . . . . . . . .4. . . . . .41 d. . 236 4. . . . . .40 c. . . . . . . . .4. . . . . . . 236 . . . . . . . . .4. . .9 IV. . . . . . . .12 IV. . . . . . x g . . .13 x . .4. . . 241 4. .4. . . . . . . . . . 237 4. .2 x g Aut X is nite. . . . . . . . . . . . . . . x 4. . . . . . . . . . . . . . . . . . . .23 c. . . 238 4. . . . . . . .19 x . . . . . . . . . . x . . . . . . .5. . . . . .5 4. .4. . . . . . . . .5. . . . . . . 240 4. . .35 b.21 IV. . . . . . . . . . . . . . . . . . . . x g 4. .6 Curves In P3 . . . . .4. . . . . . . . . . . . . . .5. . . . . . . . . . . . 235 4. . . . . x . . . . . .37 IV. .8 d. . . .a. . . . . . . . . . . . . . .4.5. . . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . x g . . . . . . . . . . . . . . . 249 4. . . .4. . . . . . . 235 4. . . . . . .8 IV. . . . .4. . . 231 4. . 234 . . . . . . . . . . . . . . 238 4. . . .16 f. . . . . . . . . x . . . . . . . . . . . . . . . . . . . . . . . . . . .4. . .6 x g 4. . . . . . . .4. . .4. . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.6 IV. . . . . .4. . . . . . .4. . . . . . . 239 . . . .1 x g . . .4. . . 230 4. . . . . . .20 x g . . . . 248 IV. . . . . . . . . . . . . . . .14 c. . . . . . . . . . . . . . . . . . . . . . . . . . . . .5. . .1 x g complete intersect is nonhyperelliptic . . . . . . . . . . 237 . .

. . . . . . . 264 . .1. . 262 5. . .6 IV. .5 V.18 b. . . . . . . .2. . . . . . . . .6. . . . . . . . . . . . . . .7 b. . . 263 5.2. . . . . . . . . . . . . . . .18 V. . . . . . . . . .6 x g Grothendieck's Theorem . . . . . . . . . .3 x g:a. . x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x g . . . . . . . . . . . x g 5. . . . . . . . . . . . . . . . . . .4 a x .3 V. . . . . . . . . . . . . . . . .1. . .7 x . . . x . . . . . 255 . .2. . . . . . . . . . . . . 261 5. . .3 V. . . . . . . . . . . . . . . . . . . . . . . . .1. . . . . .1.6. . . .8 x g cohomology class of a divisor . . .4 x g .16 b. . . . . .12 x g Very Ample not numerically equiv 5. . . .13 V. . . . . . . . . . . .2. .2. . .4 x g tangent sheaf not extension of invertibles 5. . . . . .12 V. . . . . . . . 264 5. . . . . 255 . . . .2. . . . . . . . . .8 b. . . . 252 4. . . . . . . . . . . 261 5.1. . . . . . .1. . . . . . . . . . .5 c. . . . . . . . . . . . . . . .6. . . . . . . . . . . . 254 5. . x . . . . . . . . .1. . . 255 5. . . . . . . . . .11 x g 5. . . . . . . . . . . . . . 264 5. 260 . . . . . . . . 256 . . . . . . . . .20 V. . . . x g . . . . . . . . . . .2. . . . . . . . .1 x g Intersection Via Euler Characteristic . . . . . . . .6. . .1. . . . . . . 260 5. x g: with above 5. . . .2. . . . . . . . . . . . . . . . . . . . 266 5. . . . . . . . . . . . . . .9 x g Curves on Quadric Cone . . . . . . . . . . .1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2. . . . . . . 262 5. . . . . . . . . . . . . . .1. . . . . . . . . 265 5. .2. .8. . . . . . . . . . . . . . . . . . . . . . .2. .1 V. . . .1. . . . . . . .1 254 V. . .1. . . . . . . . . .x g . . . . . . . . . . . . . . . . . . . . . . . 256 5. . . . .15 V. . x.11 b. .1. . 259 5. . . . . . .7 x g . .6 V. .3 x g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2. . . . . . . . . 261 . . . . . . . . . .15 c. . . . . . . . . . . . . . . . . . . . . . 263 5. . . . . .1. . . 263 . . . .11 V. . . . . . . . . .2 x g Degree via hypersurface . . . . . . .1. . .1. . . . . x g . . . . . . .7 x Algebraic Equivalence of Divisors . . . . .6. .2 V. . . . . . . . . . . . . . . . . . . 266 5. . . . . . .6.1. . . 268 . . . x . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254 5.5 x g complete intersection doesn't lie on small degree surface .1. . . . . . . . . . . . . . .1 x g . . . . . . . . . . . . . .2. . . . . . . . . .1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1. . . .13 b. . . . . . .10 x g Weil Riemann Hypothesis for Curves . . . . . .2.6. . . . . . . . . . . . x . . . . . . . . . . . .16 V. .10 x . . . . . .2. . . . . . . . . . . . . .9 x g Hodge inequality . . . . . . .1. . . . . . .1 Geometry On A Surface . . . . . 253 4. . .2. .2. . . . . . . . .1. . . . . . . . . . . . .2 Ruled Surfaces .14 b. . . . . . . . . . . .14 c. . . . . . . . . . . 257 5. . . . . . . . . . . . . . . . . . . . . . . . . .9 c. . . . . . . . . . . . . . . . . . .5 x g 5. . . . . . .8 IV. . . . . .1. . .6. . . . . .1. . . . . . . . . . . . . .1. . . . . .6. .2. . . . . . . . . . . . . .9 b. . .12 V. . . 256 P3 . . 260 5. .7 V.6.4 b.1 V. . . . . . . . . . . . . . .1. . . . . . . . . . . . . . . . . . . . . . .6 x g 5. . . . . . . . . . . . . . . . .2. . . . . . . .19 V. . . . . . . . . . . . . . .1. . x . 257 5.1. . . . . . .5 4. . . . . . . . . . . . . . . . . . . .1. . . . . . . . . . . . . 255 . . . . 251 4. . . . . . x g . . . . . . 5. . . .5 IV. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 V Surfaces 5. . . . . . . . . . . . . . . . . . . . . . . .4 IV. . . . .17 V. . . . . . . . . . . . 258 5. . . . . . . . . . . . 259 5. . . . . . . . . . . . . . . . . . . . . . .2 V. . 257 5. . . . . . .1.2. . . . . . . . . . . . . . .2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6.1. 265 5. . . . . . . . . . . . . . .2 V. . . . . . . . . . . . 258 5. . . . . . . . . . . . .6. . . . . . . . . .7 IV. .11 x . . . . . . . . . 251 4.2. 256 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 x g Canonical for a surface in 5. 259 5. . . . . . .2 x . . . . . . . . . .21 b. . . . . . . . .2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x 5. . . . . . . . . . . . . . . . . . . . . .6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265 5. . . . . . .9 (starred) . . . . . . . . . . . . . . . . . .9 IV. .10 d. . .2. . . . . . . . .17 V. . . . . . .a x g decomposable is never stable . . . . . . . . 250 4. . . . x g . . . . . . . .6 b. . . . . . . . . . . .6 x g Projectively normal curves not in a plane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 x . . . . . . . . . . . 267 5. .22 V. . . . . . . . . . . .2.2. . . . .2.b adjunction computational formula 5. . . . . . . . .1. . . . . .4 x g Self intersection of rational curve on surface . . . . . 254 5.8 x g . . . . . . . .2. . . . . . . . . . . . . . . . . . . . . . . . . . . . .6. . . . . .1. . 250 4. . . 261 5. . . .1. . . . . . .10 V. . .8 V.2. . .3 IV. . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . .17* (starred) 5. . . . . . . . .7 V. . . . . . .1 V. . . . .3. . . . . . . . . 287 .3.12 b. . . . . . 272 5. .25 V. . . . . .1 V. . . . . 292 5. . .4. . x g Hodge Index negative denite 5. . . . . . . .2. . . .5 x g Pascal's Theorem 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 V. . . . . . . . . 268 5. . . . . . . .3. . . x g 5. . . . . . . . .4. . . . . .3.2. 295 . . . . .4. . . . .5.22 b. . . . . .5. . . . . . . . .9 V. . . . . . . . . . . . . . . . . .5. . . 275 5. . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . 270 5. . . . . . . . . . . . . . . . . . . . . . 285 5.7 d. . . . . . .24 c. . . . . . . . . . . . . . . . . . . . . . . . . . .8 V. . . . . . . . . . . . . . . . . . . . . . .2 x g . . . . . . . . . . . . . . . . . . . . . . . .4. . . . . . . .5. . . . . .2. . . . . . . . .11 V. . . . .3. 300 . . . . . . . . . . . . . . . . . . . . .26 V. . . . . . . . . x 5. . . . . . . . .20 V. . . .3 Monoidal Transformations . . . . . . . .3. . . . . . . . . . . . . . . . . .18 c. . . . . . . . . . .1 x g Resolving singularities of 5. . . . . . x . . 286 5. . . . .4. . . . . . . . . . . . . . . . . . . . . . . . . .10 x . . . . . . 272 5. . . . 279 5. . . . . . . .3 x g . . . . .5. . . .4. . . .12 x g kodaira vanishing for cubic surface 5. . . . . . . . . 285 5. . . . . . . . . . . . . . . . . . . . . . . 284 5. . . . . . . . . . . . . . . . . . . . . . . . .2.4. . . . . . . . . .b x . . . . . . . . . . . . . . . . . . . . . . . . . . x . . . . . . . . .4. . .2. . . . . . . . . . . . . . . . . . . . . . . . . .3 V. . . . . . . . . . . . . . . . .6 x g 5. . . . . . .5. . . . . . . . . . . . . . . . . . . . . . . . . . .2. . .4. . . . . . . . . 296 5. . . . . . . . . . . . . . . . . . . . . . . . . .4. . . . . . . . . x . . . . . . . .14 x . . . . . . . . x .10 V. . . . . . . . . . . . . 276 5. . . . . . . . . . . .6 c. . . . . . . . . . . . . . . . . .5. . . . . . . . . . . . . . . . . 278 5. . . . . . . . . . .11 V.4. . . . . . . . .5. . . . . . . . . . . . . . . . . . . . . .19 d. . . . . . . . . 276 5. . . . .3. . . . . . . . . . . . . . . . .10 V.4. . . . . . . . . 278 5. . . . .3 x g hodge numbers excercise 5.8 e. . . . 286 5. . . . . .16 x . . . . . . . . . .4. . . . . 290 5. . . . . . . . . . . . . 273 5. . . .4. 298 . . . . . . . 287 . . . . . . .17 b. . . . . . . . . . . . . . . . . . . . . . . . . . . .5 V. . . . . . . . . . .11 x Weyl Groups 5. . . . . . . . . . . . . . . . . . .3. . .2 x g . . . . . .3. . . . . .19 b. . .2 V. . . . . . 293 5. . . . . . x . . . . . .4. .8 V. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297 f 5. .2. . . .4.2 V. . . . . . . . .4. . . . . . x . . . . . .4. . . . . .4. . .6 V. . . 295 5. .4 V. . . . . . . . .5. . . . . . . . . . . . . . . . . . . . . . x 5. .4. . . . . . . . . . . .4 x g hodge index theorem corollary x 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 x . . .2. . . . . . . . . . . .5. . . . . . . . . . . . . . .3. 274 5. .5. . . . .3. . . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . .2. . .15 V. . . . . . . . . . . . . . . . . . . . . . 274 5.5 x g hyperelliptic every genus . . . . . . 269 5. . .4 x Multiplicity of local ring . . . . . . . . x . . 273 V. . . . . . . . . . . . . . . . . . . . .3 x . .13 V. . .4. .6 V. . . . 288 . . . . . . . . . 297 5. . . . . . . . .15 x admissible transformation 5. . . . . . . . . . . . .4. . . . . . . 295 . . . .3 V. .2 x g Castelnuovo Lookalike . . . . . . .5 b. . . . . . . . . . . . .5. . . . . 299 5. . . . . . .5 b. . . . . . . . . . . 286 5. . . . . . . . . . . .4 x g important? . . . .5. . . x g . . . . . . . . . . . . . .5 b. . . . . . . . . . . .3. . . . . . . .5. . . . . . . . . . . . . x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4. .6 x . . . . . . . .3 V. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291 5. . . . . . 292 . . . . . . . . . . . . . . . . . . . . .2 V. . . . . . . . . . . . . . . . . . . . .9 x genus bound for cubic surface. . . . . .7 x . . . . .3. . .1 x g . . .12 x . . . . 299 . . 279 5. . . . .4 Cubic Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2. . . . . . . . . . . .21 b. . . . . .7 V. . . . . . . . . . . . . . . . . . . . . . 274 5. . . . . . . . . . . . .20 V. .3. . . . . . . . . . . . . . . .6 x . . . 268 5. . .5 x g 5. 299 . . . . . . . . . . . . . . . .4. . . . . .3. . . .8a. . . .4. . . . . . .5. . . . . . . . . . . . . . . . . . .16 x Fermat Cubic . . . . . .5. . . . .16 V. . . . . . . . . . . . . . . . 298 . . . . . . . . . .4. . . . . . . . . . . .4. . . . . . . . . . . . .5 Birational Transformations . . . .4. . . . .3. . . . . . . . . . . . . . . . . . . 297 . . . . . . . . . . . . .3. . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . .1 V.4 V. . . . .3. . . . . .3 . . . 288 5. . .4. . . .14 b. . .4. .1 x g P2 blown at 2 points . . . x . . .4. . . . . . 273 5. .2. . . . . . . 291 . . . . . .12 V. . . . .7 x . . . . . .4 V. . 294 . . . . . . .9 V. . . . . . . . . . . . . . . . . x . . . . . . . . . . . . . . . . . . . 273 5. x . . . . . . . . . . . . . . . . . . . . 294 5. . . . . . .23 b.

.1 Ex. .2 V. .1 x g . So we have y -axis. .6. is integral domain. x3 ] ≈ k [x] Thus f or g If Ex. . .6.1 g x We have 1. . x2 . .5. . . . . . . .1.1. .1 x 1. .1. . . .1 V. . . z − x3 and then A (Y ) = . . . . . . . . . . . . . . . . .5. . . . . I. . f. . .10 V. . . . . y] / (y − x2 ) ≈ k [x. .1. . . so one of Clearly dimension 2 1. . . . . . . 300 5. .1. . .3 fg Ex. . . . x2 . . . . . x Surface singularity . . .6. y] / (xy − 1) ≈ k x. . . . . . . . . .1. then f (x. . 18 by y − x2 . . 303 5. . . . it is prime.8. . . . . x3 ) = 0.1 I. 300 V. . . . . . .6 1 V. . . . . . . . . . the ideal is generated k [x. . Varieties 1. . x g   A (z) = k [x.6 Classication Of Surfaces . i. . . .4 k then x=0 and y is anything. . . . . . . .b. x2 ] x2 by plugging in to y. .2 k [x.5. x2 . . g ∈ I (Y ). 303 I. . . . . . . . . x1 1. . . .8 x A surface Singularity . . . . .6. . . . . . . . . . . . . . .5. I.9 5. . k is innite.2 x g . . .2 x g Since we assume algebraically closed. . . . I. z − x3 ) ≈ k [x.3 x g z = 0. . . . since ∈ I (Y ) . . . . . . . 303 5. 1 since parametrized by t. . y. . b.1. .5. . . . . . . then to see irreducible. . . z] / (y − x . . if must be zero. . . . x3 ) g (x. .e. . . . .

B = k [x1 . xn a set of generators. x) x2 = αy for α ∈ k so we have a parabola..1. is f. I (V (J)) = J thus B is coordinate ring of V (I). So closed sets in A × A product topology should be nite unions of horizontal or vertical lines and points. . Assume to contrary that irreducible. and y = 0. 1 1 Now a closde base for the product topology on A × A would be products of sets closed in the Z-topology 1 1 1 on both factors. The diagonal is not such. U = Y1 Y2 . let Then Thus 1. y − kx ) If x = 0. This is z axis.1.. y) If z 6= 0.1. I. X = U c ∪ V c. Let Then but X was supposed irreducible.1. Prime ideal is (x. X2 ⊂ X with Yi = U ∩ Xi . closure irreducible ` Y is irreducible. . .. Then Y = (Y1 ∩ Y ) ∪ (Y2 ∩ Y ) =⇒ Y = (Yi ∩ Y ) for one of i = 1.1. c For X1 .5 then we have Ex. 2 Prime ideal is (z − k. then z is anything. Closed sets in A are just points and the whole space.g. =⇒ U irreducible. 1. but Y = Y1 Y2 .. 2 . no nilpotents. no nilpotents. then (X1 ∪ X2 ) ∪ U = X =⇒ X Contradiction.5 x Clearly if If B is an ane coordinate ring then it's nitely generated. Yi closed.7 B Ex.. Y the smallest closed subset of X containing Y =⇒ Y = Yi =⇒ Y is Suppose 19 irreducible. U is not dense.4 x g he set V (y − x) is the diagonal it's closed in Zariski topology.6 x g U dense Assume to contrary V = X\U . I. xn ] /J where J is reduced since no nilpotents.6 Ex.1.Prime ideal is (z. I.. x1 . =⇒ U dense. 1. U irreducible ` U is not irreducible.

Ex. U1 ∪ U2 .10 U1 . Note irreducible components of Now use to minimal prime ideals of height Ex.1. I.8 Ex.8 x g and below Y ∩ H correspond dim R/p + height p = dim R. I.1. x Apply part (a) to the cover 1.1. .1.1. U1 ∪ U2 ∪ U3 .7 x Since it holds from the equivalent conditions for noetherian modules since ideals correspond to submodules.1.1.1. . 1. .11 d x Argue as in the proof of the baire category theorem. 20 1..9 b. Apply the previous excercise r times.. follows from part (a). I. 1. (c) x.9 x g Each fi ∈ a from i=1 to r denes a hypersurface.

12 x.6. n}.1. {1} . x g and I. 1. I. e. Xi ∩ U is irreducible and dense in Xi so the strict inclusions Thus X0 ∩ U ⊂ . x Consider 1. X . I. d.1.16 Let are maintained.7. 1} with open sets ∅..1.. 21 X so their dimensions are not the . ⊂ Xn ∩ U is a chain and dim (X) ≤ dim U ≤ sup dim Ui . x Y 6= X . I.1. then to any chain in Y we add X to get a longer chain in same. ⊂ Xn is a chain of irreducible closed subset of X Let U 3 X0 . x X be the positive integers with closed sets like {1. By 1.2.11* Ex. (b).13 Y extends to a chain in X.1..15 If X = {0. ..10 x Any chain of irreducible closeds in 1...1. Let {pt} = X0 ⊂ .1. Ex.a sup dim Ui ≤ dim X .14 c..1.12 Ex. By (a) 1.1.

3 I.. but it has two real roots. m Since xi divides monomials of degree m (n + 1).2. 0) so a is √ Assume (ii). The point is that it only factors over C [x.2 x 1.5 c. If a ⊃ Sd then xi ∈ a have no zeros.3. y].a. x trivial 22 either S or ⊕d>0 Sd ..2.1 x homogenous nullstellensatz This follows by looking at the ane cone and using the ane nullstellensatz. If a contains xi then xi ∈ am for all i.1 I. 1. Assume (i). x Trivial..2. d Assume (iii).2.2. 1.2. Then in I.2 I.2. .2 I.2 projective containments x √ n+1 A .2. 1.3 1.4 b. x trivial 1. Z (a) is either empty or (0.2. containments. then Sm(n+1) ⊃ a.Consider y 2 + (x2 − 1) 2 .

1. So we have the reverse containment.2. x Z (I (Y )) ⊂ Y Z (I (Y )) is closed set part 1 containing Y =⇒ Z (I (Y )) ⊃ Y . Y = Y1 ∪ Y2 . then I (Y ) = I (Y1 ) ∩ I (Y2 ) ⊃ I (Y1 ) I (Y2 ). Y ⊂ Z (I (Y )). Note that Z (a) is the set of P where every g ∈ a vanishes.8 b.4 a x By 2.2. f are in √ f ∈ a = ∩P ∈Z(a) (X1 − P1 . 1..3 and 2. x g Since I (Pn ) = 0 which is prime.2. Xn − Pn ). By nullstellensatz. part 2 since some polynomials vanishing on Y don't vanish at I.7 e. then there are ab ∈ I (Y ) with both a. b ∈ / I (Y ). 23 or I (Y2 ).9 c.6 d. x g and below Suppose that I (Y ) is prime. So I (Z (a)) is the set of h which vanish at all P where every g∈a vanishes...2. . such Now let √ r. the smallest closed set containing Y. If If 1.2. . Thus Y is a union of Y ∩ Z (a) and Y ∩ Z (b) and is not irreducible. P.   Suppose p ∈ / Y so I Y ⊃ I Y ∪ P =⇒ P ∈ / Z (I (Y )).2. so use part (b). x I (Z (a)) is the set of f vanishing on the zero set of a.1. Thus I (Y ) = I (Y1 ) Y is not prime.

10.5 (a) x g Irreducible closed chains in Note that 1..b. using the standard ane cover.8 x g 24 to Yi .. Comparing transcendence degrees gives the result. x This follows from the proof of 1. S (Y )xi is the coordinate ring of the cone with xi = 0 which is isomorphic i h =⇒ S (Y )xi = A (Yi ) xi .1. . x By proposition 1.5. 1. So degree 1.3 is noetherian by hilbert basis theorem.2. x1i .7 a.6 x S (Y )(xi ) is the coordinate ring of the cone of if Yi is nonempty.10 I.2.10 and using the ane cone. xn ] Pn corresponds to ascending chains of primes in k [x0 .. xn ] by Ex.11 k [x0 .2.2.14 I.2.2. .. . (b).2.. x g This follows from I. I.12 0 Yi part of I. 2.2.2. 1..13 b. and part (a).

. 1.. x1 . and h (1.. I. So the ideal of Y can have height at most 1 by 1.2.2. then using the n − 1. then β (f ) = f (1.Since irreducible homogeneous polynomial correspond to minimal prime ideals of height height / dimension formula gives Y has dimension Y has dimension n − 1.. Conversely suppose 1.2. then h = β (g) so I Y is generated by β (I (Y )).  2.. x1 . xn ) = g .  If homogeneous h vanishes on Y .15 the height / dimension formula.2..9 x  I Y ⊃ β (I (Y ))  n If f ∈ I Y .. xn ) vanishes on Y ⊂ A =⇒ f ∈ I (Y ).2.10 x 25 . β (I (Y )) ⊃ I Y . =⇒ dim k [Y ] = dim Y + 1 = n (by 1.6).16 I. 1. .

x g is intersection of hyperplanes by (a). x Since they have the same ideal.11 x g (use p2) Trivial. 1. x g By 2.21 x.2.18 c. 26 1.19 I.2.2.17 b.2.2. .2. (prime i the ideal is irreducible was a previous excercise). 1.6.20 Y b. 1. 1. each corresponding to homogeneous primes of height Each additional hyperplane increases the height of the ideal by 1. 1.this is clear.

So ker φ ⊃ I (Im vd ).. Mn ) = 0 so f vanishes on (M0 (a) .2. Mn (a)) f ∈ I (Im (vd )) . =⇒ f (M0 . 1. if 1. then f (x) = 0 for all x ∈ Im (vd ) ..2. Any monomial of degree 1. then f (M0 .. So Z (a) ⊂ Im vd . x0 x1 .. x1 ) (x1 . 3 2 2 3 to get the 3-uple embedding (x0 . . x g and above Take the projective closure of the twisted cubic.2.26 Z (a).b..2.. x0 x1 ..22 c.23 θ I..25 vd c. x By the projective dimension theorem. d. . x g and above and below (part d) f ∈ Ker (φ) . x21 ..2. 1. . Mn ) = 0.by projective dimension theorem.12 x g and below maps into an integral domain so the kernel is prime. . .2. Thus f ∈ ker φ. x31 ) 27 so Im (vd ) ⊂ Z (a). x is an injective isomorphism with image equal to 1.. If Conversely.24 i maps under θ to one of degree d · i which shows that the kernel is homogeneous..

..) (. ak bl .27 I. to P under ψ . Note that ψ is well-dened and injective. Show that the image N of ψ is a subvariety of P . bs ) 7→ (. then we have gets mapped ak bl = (ak bj ) (ai bl ).2.28 for homogeneous f. In ane space with ai bj = 1.15 x g ψ (Pr × Ps ) N subvariety of P .. y 2 . ai bl . . . j ... .. z) = 0 f = g (x .. then there is some point Q in r s P × P mapping to it.e..14 x g ψ : Pr ×Ps → PN dened by (a0 . A curve in 2 2 1..)) where N = (r + 1) (s + 1)−1 = rs + r + s.) = (. .2... ai bj . a0 b1 . It is called the Segre embedding.. So v2 (Z (f )) = V ∩ Y = Z .. ai bj is left of ak bl i i < k or i = k and j < l .. ai bj 6= 0 for some i. Recall that thus a Thus we choose is described by the vanishing of the equations 28 (ai bj ) (ak bl ) − (ai bl ) (ak bj ) and is . ψ (Pr × Ps ) N subvariety of P .... . a1 b1 .). In this manner. a1 b0 ...2.) (..) in lexicographic order (i.. I.2. y. . Since we are in projective space...29 and is thus a Q to have coordinates ak = ak b j and bl = (ai bl ) is described by the vanishing of the equations so that it (ai bj ) (ak bl ) − I.13 x g P2 is dened by f (x. so it's like (a0 b0 ... N Conversely.. z 2 . ak bj . ai bj .. ar )×(b0 . . if P ∈ P with coordinates ai bj satises the above relation. a0 b2 . xy..2. .1.. Let r s Note that any point of ψ (P × P ) satises (. xz. yz) for some g . we know that (ai bl ) (ak bj ) 1.. ...2.....

I. If 1. x g Look at the curve 1.2.34 x2 = yw.31 c. y. you can see the required properties.2.17 x g complete intersections and below 29 . is the origin.2.33 then b. xy = zw L∩C (x.16 x g P = (x. 2 But I (L) + I (C) = (x . y) 6= 1.2. x Lines are given by ψ (P1 × {P }) and ψ ({P } × P1 ).1.2.2.32 x = y. From the picture of a cone.2. x g Looking at the ane picture.30 b. 1. I. w) ∈ Q1 ∩ Q2 =⇒ y 2 w = yx2 = zwx =⇒ w = 0 = x or y 2 = xz . y). z.

3. then Y = ∩Z (fi ) . x g I (Y ) 1.By 1. 0 by the degree genus. x g P2 . If it's generated by q elements.3.4 In k. P1 since the degree is . but not in 30 A2 .1 since ane varieties are isomorphic i coordinate rings are. and taking a point gives an embedding to (You may have to look ahead for this).3 I. d. then the zero set is the intersection of 1. x g The coordinate ring of a proper subset has units not in 1.1 x g By ex 1. the intersection of q -hypersurfaces has q hypersurfaces. starred. x g A conic has genus > 2g = 0.36 is generated by fi . 1. 1. dim at least n−q .3.35 If b. and two lines intersect by ex 3.8.3 c.2 b.7a.1 I.3.2.2.3.3 x 1. 1.

4. Any inverse is a polynomial in Note 1. frobenius not isomorphism. and by part (a). b. Now if ϕ is an isomorphism. t ) → t.4. then topologically it is a homeomorphism. x y 2 3 since (t .2 g bijective.9 k p.5 e. then Then we have a 1.1.3.8 If is a perfect eld.X which is a homomorphism. y) → b.Y to Op.3a.6 I. x g No inverse since it would need f (tp ) = t. the induced map on local rings is an isomorphism. by 1. bicontinuous but not isomorphism. Injectivity follows from denition in characteristic Surjectivity is because 1.3. x g By 3. being algebraically closed. 31 .3. x f f ◦ φ is regular on a neighborhood φ−1 (V ) of p. x The  direction is clear. I. the regular functions on a projective variety are k. map Oφ(p). as it is dened by polynomials. This is only possible for an ane variety if it is a point.3. x is bijective to the cusp and continuous. is regular.3.4 1.3.7 ϕ x and y with (x.3.

k [x. g By the projective dimension theorem. 1.1.3.13 f implies that it is I.11 vanishes on the dense set φ (X) ∩ V . Thus Since 1. I.3.14 g is constant so the regular functions are just k [x.3. d−1 d−1 vd−1 : (x0 : · · · : xn ) 7→ xd−1 0 x0 : x0 x1 : · · · : x0 xn  which is regular.10 If c. I. x 32 . 0) look like with (f ) + (g) = 1.12 f PN − hyperplane is ane and is the same as Pn − H under vd .15 b. x φ∗p (f ) = 0 then Continuity of 1. 0.3.3.3.3.5 x g Note that 1. I.3.4 x g Note that 1. g g can only vanish along f or at (0. y] 6= A2 − {(0.3. 0) and thus has a nite number of zeros.3.6 x g Note that an ane variety should have OX equal to the coordinate ring. 0)} this is a contradiction. f 2 But regular functions on A − (0.7 a x. y].

17 I. 1. f ∈ OX 1.3.3.Also by the projective dimension theorem 1. I.11 x This follows from properties of localization of the coordinate ring.3.8 x g deg(f ) deg(fi ) looks like fi /xi = fj /xj j . y.9 x g S (X) = k [x. z). y. 33 .10 x Note that locally regular implies regular.3.xj then xi |fi so that fi = xi i so the function is constant.16 I. deg(fi ) deg(f ) Since xi . deg(fj ) deg(f ) So fi xj = f j xi i .19 I. k [Y ] = k [x.3.3. y]. It is a 3-dimensional vector space.18 there are no such in S (X).3. but 1. 2 Now look at m/m for m = (x.3. z] / (xy − z 2 ).

we have conrmed the idea was maximal. Quotienting gives the residue eld.12 x g dim X = dim A = dim A/p + ht p = 0 + ht p = dim Op . tu2 .X (since A/p is a eld) I. .1. Since this is a eld. u3 ).3.3. which is the invertible functions on Y. y = t u. 1. Now 1. The set of functions vanishing on Y is the maximal ideal. u3 ) 7→ (t3 .20 I. z = u . 0). 0.13 x g is clearly a ring. t2 u. 3 2 3 We have x = t . Let 1.3. Pn := (x0 6= 0) ⊂ Pn+1 ..23 b.3.3. x g π : (t3 .14 x g (and below) projection from point P = (1..3..3.X . We have 34 at (0 : x1 : · · · : xn ) which is a morphism in a neigh- . t2 u. 2 3 Recall that the cuspidal cubic is given by x z − y .21 OY.22 I. n The line through P and x = (x0 : · · · : xn ) meets P n borhood of P . dim X = dim k [X] /I (Y ) + ht I (Y ) = dim Y + ht I (Y ) = dim Y + dim OY.

. φy to X and Y we can take (φx . Hence f is a trivial relation and the combined generators of X. Y are basis for X × Y . yn ) = 0 gives a relation on the algebraically independent x generators and a similar thing happens on Y .25 b. x The universal property of ber product agrees with the universal property of the tensor product for nitely generated algebras..3.27 Z with morphisms φx . x g f (x1 . Y . . I. 1. x Given a a variety 1..24 t6 u3 − t6 u3 = 0 so this is the cuspidal cupic..26 c. . φy ) : Z → X × Y .3.Note that plugging in gives: 1. Suppose that 35 .. f (−. yn ) = 0 gives a relation on the combined generators xi ..15 x See Gathman's notes.. On X . yk+1 .3. . yk+1 . yi of X. d. 1..3.3.. xk .

16 x g −1 X × Y = (X × Pm ) ∩ (Pn × Y ) = πn−1 (X) ∩ πm (Y ) 1.33 If A d. b. (Use DIRP) 1. For For 1.29 where πm. z] / (z 2 − xy).30 I.4. x is an integral domain integrally closed then so is each localization. Normal is when each local ring is an integrally closed integral domain. Q2 we just need to see that the local ring at the cone point is normal. 1.n : Pn × Pm → Pm . Pn are the projections.3. x g Q1 we can use the jacobian criterion since nonsingular implies normal.3.3.3. k [x.3.3.17 x. y. b.3. 36 .1. Note that xy is square free.12) Thus X is normal.3.31 P2 are isomorphic to P1 which is smooth. The converse also holds (Atiyah- Macdonald 5. x similar. so by II.32 This can be given by c. normal and nonsingular are equivalent. we are done. g By a previous excercise conics in 1.6.28 I. x g For curves.

=⇒ Y 1. b.3. we take all monic polynomials in Corresponding to 0 1. x. Note that k Y In other words. O (1)) = =4<   4+1 5= = h0 (P1 .a x Y projectively normal.  h (P . h i ˜ is nite by 3. x g Another criterion for projectively normal is that But 0  3 3+1 3 H 1 (P3 .3.3. x It is just the image of v4 : P1 → P3 . OY (1)). h i ˜ ˜ the inclusion θ : k [Y ] → k Y we have θ : Y → Y . . IY ) = 0. 37 twisted quartic is image of P1 .1. k [Y ] is integrally closed.9A and unique since the integral closure is unique.37 c.36 is normal. then each local ring is integrally closed. k [Y ] and all their solutions in k (Y ).34 Let e.35 I. Y˜ be the ane variety with h i k Y˜ the normalization of k [Y ]. Assume Then Since the localization of an integrally closed domain is integrally closed (we used this above). 1 1 Also note we have normal since P is nonsingular and 1.3.18. OP1 (4)) = h0 (Y.3.

Note that 1. x Note that 1. and the inverse map is given by b. 38 y 7→ −y . 1. x g Consider a single variable complex function with a pole. 1. x Clear.40 b.21a.42 A1 is a group under addition.3.3.19 x fi ∈ k If any then ϕ would not be surjective.3. . nonconstant polynomial gives a jacobian in SKIP 1.3.39 I.3. Now computing the derivative of a linear.41 I. I. k×.3.20 x g and below Since projective space is proper this works.38 h1 (P1 .3.3. O (1)) = 0.

1 x g Dene a new function which is f on U and g on V. Continue in this manner until you have dened on all such sets.47 I.4. then f : X → A1 \0 and this correspondence preserves multiplication of f ·g for g ∈ O (X)× . x is regular. x0 2 This set isomorphic to A .1.3 x g and below x1 is dened where x0 6= 0 .1 1. x Since we can add morphisms 1. 1.3. e. x f ∈ O (X)× .3.3. f= 39 .3.43 c.45 If d.46 I.4.48 I.2 x see 4.3. I.44 If f ∈ O (X) 1.4.3. then f : X → A1 and adding functions is compatible with this. f is projection to rst coordinate.4 x 1.

0. x1 . t3 ) and an inverse (x.4 x g all parts By I. Dene Note 1. 0] ∈ / P1 . conics are 1.4.4.49 b x.50 I.3. x g t 7→ (t2 .3. (0.51 b.3.4. g (x0 .1 to see birational.3.b. 1) since [0.53 I. 40 . 2 2 2 2 3 2 The inverse is given by (x : y) 7→ ((y − x ) x : (y − x ) y : x ). x between the cuspidal cupic and y A1 c. y) → A1 is birational to P1 . x g (x : y : z) 7→ (x : y).52 P1 .5 x g see V.3. x1 ). We can take the projection Not dened at 1.1. x2 ) 7→ (x0 .3. Y → P . P2 → Y .1. 2 3 3 2 3 That's (y x − x : y − x y : x ) 2 2 2 2 3 (since y z − zx = z (y − x ) = x on the curve) The projection gives 1. and compute the canonicals to see they are not isomorphic.

V . 0. y. ϕ maps U to V : (x. 0) . y. b.56 on on U. ϕ : U → V . 1). picking a point not on |X| ≥ |P1 | = k .6 x g and below Let U. z) 7→ (yz. yzxz) =(x. V be the sets {(x : y : z) |xyz 6= 0}.3. I. and ϕ2 is the identity Then 1. xz. yzxy.2 we have a birational .3. z) 1.55 Let U. c. 1.4. x From V.1.4.3. 1 and projecting to P gives a surjective morphism from so 41 X to . xy) ϕ2 : (x.4.4.58 P.3. 0) .8 x g Base Case: Any curve P1 X is birational to a plane curve so On the other hand. 0.3. z) 7→ (xzxy.7 x The isomorphism of local rings induces an isomorphism of map on neighborhoods of 1. I.54 I.3 we see they are dened on the complment of 1.2. y. Q k (X) ≈ k (Y ) via I.3.57 (1. (0. X |X| ≤ |P2 | = |k| 2 since P is a 2-manifold. x g U = V = {[x : y : z] |xyz 6= 0} . (0.

so a singularity.4. r Then k (X) is a nite algebraic extension of k (P ) generated by xr+1 .. Looking at the equations for y1 . 42 . P By theorem of primitive element.9 x g x0 .4. 1 ˜ ∼ 1 Since the cusp is not isomorphic to A . Using determinant has to be rank one.59 X as a hyperplane in Pn+1 by the primitive element theorem..60 I.Thus |X| = k .→ P \M  Z ( ai xi ) thus k (P ) ..e. then now we have distinct tangent directions in the blow up). xi yj − xj yi = 0 for all i.→ k (Z (F ))  k (X) and there is an open set U ⊂ Pn \M where all bers have cardinality 1 on which the projection is birational... There is only It is thus clear the exceptional divisor meets Y if one such on the cusp.3. Embed 1.. We can nd a linear space generated by 1. .. Inductive Step.10 x g Your blown up thing lives on the blow-up surface which is the set of points and lines through origin in An × Pn−1 (p. i... .→ k (P ) . : yn ) only (x1 . . and the fact that (x1 . r and thus an inclusion of function elds k (X) . yn n−1 each patch of P we can actually see what the blow-up surface looks like. P n r We have X . since any two curves have the same cardinality they are homeomorphic in the nite complement topology. yn ) we see that the blow up surface is described by the conditions  x1 . x g cf example 3.. Y 6= A .e. xn ) is a multiple of (y1 . xn ) is on the line (y1 : .. . if we have some lines intersecting. xr disjoint from X denes a surjective projection to Pr . just restrict to your variety).7. i. .. n n−1 Now the blow up is the blow up surface and the projection π : A × P → An . The projection is a n −1 n−1 birational map onto A − 0. j . (see gure 3) ˜ at points corresponding to singularities of Y ...61 I.. 1.. p) (in other words. (To do the same things for your variety... all minors vanish.3.. k (X) is generated by ai x i .3.. b. xn rank ≤ 2. xn .. The preimage of 0 is π (0) = 0 × P is the exceptional curve.

4. y).a x The jacobian is 1.4. mt) 7→ (x.2 0 at the origin. t2 m · t + tm2 t2 = t4 + m4 t4 So it must be a triple point.1.5 x 1.5. .1.1 I.4 xy = 0 y 2 − x3 . x Cusp since the lowest terms are 1.1. c. 43 we can factor out a line with multiplicity three. The lowest multiplicity term is like 1.b. x g what kind of singularities The jacobian is 0 only at origin. Graphing gives tacnode.4 I.3 so it's normal crossings and thus the node.4. d. I.5.4. x g what kind of singularity By plugging in (t.

5 I. This is double line. x pinch point x Checking the jacobian gives singular points along 1. c.2a. 1.5.4.4.8 I.4. x Check jacobian gives singularity along z .1. so the multilpicity is 1. b.6 so we have the pinch point.7 x-axis 0. 44 f is nonzero.3 x g A nonsingular point is when at least one of the partials of in x or 1 .4.5. x conical double point Check jacobian gives singularity at 1. so f must have a degree 1 term .

then xd + y d + z d . IPr ⊂ (f.12 c.Z)P is nite.4.9 b.13 I.11 (Z.4. 45 P is going to meet Y transversely. Note that multiplicity at a point can be described by the lowest term in an equation for the curve. P r Note that OP /mP has nite length since it has a ltration by powers of Let U By nullstellensatz. b. x g Bezout since the general element of a linear system of lines through 1. As OP = k [U ]a =⇒ mrP ⊂ (f. g . Comparing with bezout's theorem gives 1.4 x g an ane neighborhood where P is the only intersection of f. 1. 1.10 I. To see why this may be so.4. is .1.5. m which are <r and mi /mi+1 nite dimensional. Thus (Y. consider trying to factor out a linear term. x g The multiplicity at the origin is the smallest degree term that appears.5 x If the characteristic doesn't divide d−1 Else.µP (Z). d .4.5. x see b.Y )P ≥ µP (Y ) . g) for some r > 0 . g) .4. xy + yz d−1 + zxd−1 = 0.

2 3 4 2 2 4 On t = 1. p) n n−1 in A × P (in other words.17 d. The strict is nonsingular.6 x g y 2 − x 3 + x4 + y 4 .14 I. One more blowing up gives a smooth one. Again using the jacobian criterion The cusp was 1. 2 1 The blowup surface is dened by xs − yt in A × P . t − y of the strict transform factor linearly. 2 2 2 4 2 On s = 1.16 c. 46 .1. so a singularity. b. x g x2 = x4 + y 4 .15 gives nonsingularity.4. so we get a node.4. 3 3 4 2 Then 1 − yt + y t + y .5. which gives a cusp. we have y = xs and x = x + y so x (1 − x − x s ). 2 2 The lowest degree terms. if we have some lines intersecting. 3 2 3 So x (x − u ) = 0 on t = 1. 2 3 2 4 2 The blop is dened by ys − xt so on t = 1.4. The jacobian criterion shows it's nonsingular. ys = x and y (1 − yt + y t + y ) = 0. x = yt and x = x + y so y (t − y t − y ) = 0 . 2 4 4 2 2 2 4 2 On the patch s = 1. y = xt and x (t − x + x − t x ) = 0. then now we have distinct tangent directions in the blow up) 1.4. Recall the tacnode is dened by 1. x g 2 1 We can use xu = yt for new coordinates in A × P . x g Your blown up thing lives on the blow-up surface which is the set of points and lines through origin (p.

. .9 x f (P ) = g (P ) h (P ) = 0 then taking ∂ ∂ ∂ . (important) x g A3 × P2 is dened by x1 u2 = x2 u1 . 1. u2 . the equations become x1 = x2 u1 . (contradiction) 47 . On u3 = 0. The Jaacobian is just ∂u1 On u1 = 1. and x2 u3 = x3 u2 . 1.5. x2 . u3 ).21 ϕ−1 (P ) by f (1. 1.5. ˜ is dened by f (x2 u1 . x In the above problem we dened 1.22 If I. f (u1 . 1. On u2 = 1 . and x3 = x2 u3 . I.5. x2 u3 ) = 0 = xd2 f (u1 . u3 )..1.18 I. The blow-up hypersurface in 1..4.8 x The hint is the answer.7 x We can just use the jacobian criterion.19 b.4.4. and f (u1 . x1 u3 = x3 u1 . 1) . u2 . and and using the product rule shows all partials must ∂x ∂y ∂z vanish..4. u3 ). x1 u3 = x3 u1 . then X   ∂f ∂f 0 ∂u3 which has rank 1 as f is nonsingular.4.20 c. If X has multiplicity d at 0.

x Note that on the parabola. Let I= I=ideal(x^2 . since the number of independent directions you can go in on the manifold corresponds to the number of tangent directions.4.4. w) 7→ (x.b.10 x g:a. z. 1.y*w) J=ideal(y*z-x*w-z*w) f%I=y^2*z-x*y*w-y*z*w. x Since the tangent space at mϕ(P ) /m2ϕ(P ) → mP /m2P .1. 1.5.4. Let f:y^2*z-x^3+x*z^2. y. A singular point means there are too many tangent directions compared to the one point.23 I.26 P x ∈ m2(x) .c This is clear geometrically. z).x*z .4.24 b.25 is just the dual of mP /m2P c. I. Now f%I / J = y Also f%J / I = -z-x Hence we should have 48 we can just take the dual of the natural map .11 x Dene ϕ : (x.5. f%J=-x^3+x*z^2+x*y*w+x*w^2+z*w^2. 1. y.

w = .4. y. x r=1 it is clear since sum of squares. rr ) − 1 The number of partials taking into account This is because it's the dimension of 1. Ex.27 I. partials are linear. x Singular locus is where partial derivatives are all 0.expand (y · (y · z − x · w − z · w) + (−z − x) · (x2 − x · z − y · w)) = x z 2 + y 2 z − x3 2 3 2 Hence ϕ (Y ) ⊂ Z (y z − x + xz ).29 2.. . but then multiplying the two factors together will create terms of higher than degree 1. Since it's a conic. for larger r. 1.4. y x+z Clearly this is not dened at (1. . −1). xn ] / (x0 .. 49 .4. z. . z) 7→ x.5.. On the other hand.30 d.. yz x2 −xz . any factors must be linear. 1..12 x  n×n Any conic is the determinant of a   x y z   matrix. dim k [x0 . 0. x −xz = x.28 If b. y.. z. x. c.4. y. A B D x  B F C  y  = 0 D C G z Diagonalize this matrix by a lienar transformation. Solving for w in Y gives w = x+z  y   2 yz −1 We dene ϕ : (x. r is n − r − 1.

g by Ox coordinates fi . n 0 Note that Q is made up of such lines since points in P lying on Q must exactly satisfy rst r 2 2 x0 + . then a satises x20 + .... Then Z = sing Q is n − r − 1 dimensional and linear by (c). y]] . r n n Embed P into P to not intersect Z ⊂ P . 1.32 I. A nite intersection of nonempty opens is nonempty open.4. 0 r 2 2 If Q is a cone.4.. The normal locus is therefore nonempty open.33 b.5. is integrally closed i image of fi is in Ox for all i. 2 Now let Z = Sing Q a linear variety as in C (in P it's just that point). . ht Construct fr+1 generate the maximal ideal of 50 k [[x..As an example.. a ∈ Q. b is given by ta + sb with s. i=1..n. then the rst r coordinates are 0 by examining the partials. + x2r = 0. 1 The line between a.. Integral closure is f.31 I. So we are looking at the set of lines between something on the plane and a single point.. b ∈ Sing Q. in P2 this is obvious since the singular locus of the quadric is a point when taking ane coordinates. If If t 1. + xr = 0 and last n − r − 1 coordinates satisfy whatever. 1. = ht gs+1 + gs ht+1 since s + t = r and gs . and rational functions are dened on such. This 2 gives the the cone yx − z = 0. + xr denes a quadric hypersurface Q in P . t ∈ P . then Q = x0 + .14 x g The isomorphism between local rings must map linear terms to linear terms + higher order terms. x g.5. h following hartshorne's example.13 x Assume X ane.4.

(Following Wall . b are weierstrass polynomials and U has nonzero constant term. d (starred) Show any double point of plane curve is analytically isomorphic to singularity at for uniquely determined If r = 2. 1 In P and 3 pairs of lines can be interchanged by a linear transformation. 0 02 If b = 0 then we have C is y = 0.4. r = 3.Plane Curve Singularities. 51 x0n for some x0 a convergent . Now factoring 1.o. y = y − a (x) y + a (x) 2 4 In this way we may use the Weierstrass preparation theorem to write which gives 2 C is y 0 − a (x) y 0 + 14 a (x)2 + a (x) y 0 − 12 a (x)2 + b (x) 2 2 1 1 02 0 which is y − a (x) + b (x) and let b (x) = − a (x) + b (x) the new constant term in y .4.35 f 3 . it's a node. 0 Otherwise b has order k some k ≥ 2 (so it's a polynomial function of x. Etc. x For double points see part (d).1. g = g3 + h. 2 2 1 1 2 0 0 0 Via another linear coordinate change we plug in y = y − a (x). Thus changing coordinates to (x0 .34 c. y) = U (x.o. If r=4 (0. g3 into 3 linear terms we get the result. y) = y m A (y) where A (0) 6= 0. Chapter 2) Change coordinates linearly so that f (0. 0) of curve y 2 = xr it's a tacnode. 0 Now it is a general fact that an order n power series like b may be written as power series in x of order 1. y) {y 2 + a (x) + b (x)} 2 where a. Also the weierstrass preparation gives the constant term of the a (x) are 0. and we already have squared terms and we don't have cancellation). If it's a cusp. r ≥ 2. y 0 ) we are done.t .t. Thus C is given by y + a (x) + b (x). 4 2 02 0 Thus we have reduced to y + b (x) = 0. For triple points write f = f3 + h. f (x. but not for 4 pairs.

1.) which is the localization of a UFD.36 I.1..2 b. A (Y ) = O (Y ) = k [y](t−P1 .3 Y is A1 minus a nite number of points.b Note the homogeneous   monomials of degree There are 1. then it is isomorphic to a proper open subset of .6 x 1.c Y is isomorphic to an open subset of projective space. x Since 1.1 x g:a. 1.4.t−P2 .b.1 As P 1 I.5.4.15 x g:a.5. points in PN with f 6= 0. and partials nonzero are in 1-1 correN spondence with the nonzero locus of a nite set of polynomials dening an open set in P . Y = V (y (x − P1 ) · · · · · (x − Pn )). 52 .5. this is clear. c.37 If If d+2 2 −1 d are like xa y b z c with a + b + c = d.5..5 I. of them so we have a clear correspondence. x f f has no multiple factors. b.6. is reducible. by elimination theory.. x By (b).

5. On the other hand k [x] ⊂ A so k [x] ⊂ A. thus y ∈ k [x]. N (x) = N (a) N (b) = x2 so N (a) Contradiction.4 I. so N (a) ∈ k × . contradiction. x Units a is a unit.5.6. 53 or N (b) is a linear polynomial if a. 1. Not UFD x|y 2 and y 6= ux =⇒ not UFD.5. N (1) = 1. thus y satises z 2 − y 2 . 2 2 3 If a is nonuit.2 x g A nonsingular curve is normal. Thus by the jacobian criterion.1. x Writing f ∈A as y · g (x) + h (x) . b are not in k.5. If irreducible If x = ab is . By easy computation 1. =⇒ g 2 is constant =⇒ a is constant. Thus A ⊂ k [x].6 c. then N (a) = g (x) − f (x) (x − x) . a = yf (x) + g (x) as in (c). x y 2 ∈ k [x]. N (ab) = N (a) · N (b). then N (f ) = (h (x) + yg (x)) (h (x) − yg (x)) ∈ k [x].7 d.5 A is integrally closed. b. If f is nonzero then comparing degrees we see the norm is nonconstant. . then N (a) has inverse N (a−1 ) by (c). As 1.

8 e. then ϕ (P ) is closed by continuity and nite since it's a proper subset of Y with closure Y as ϕ is nonconstant. the inverse matrix gives the inverse of the action.5. I.6. 54 . f induces a map 1.6. 0) to P1 by (x. x Since nontrivial and not UDF then by 6.6.10 A2 \ (0.11 I.5.5 x Note that the image of a projective variety 1.→ k (P ) . so induces k (Y ) . for (b) map P1 \∞ to A1 by (x : y) 7→ x/y . I.9 Y is not rational. 1 1 As f is nonconstant and Y irreducible.6 x g:a.6. I. y) 7→ (x : y) . this matrix therefore has an inverse.5.c ad − bc 6= 0 means the determinant of a 2 × 2 matrix is nonzero .1.5.4 x g Y → A1 by x 7→ f (x) and therefore ϕ : Y → P1 . im (ϕ) = P and ϕ is dominant.12 Since not X under a regular embedding X . 1 −1 If p ∈ P .→ Y is closed. 1.5. Since FLT is a group action.1.3 x For (a) map 1.

13 b. g are linear: f (x) = ax + b . k+s s  55 of degree 2k half xi 's and half yj 's has dim .1. then ad − bc 6= 0. g (x) = cx + d Since f.. x (x) for coprime f. ϕ ∈ Aut k (x) maps x 7→ fg(x) Injectivity of ϕ =⇒ f.6 I.14 (is this a word?) map of P1 . c. y s ] generated by polynomials k+r r+s kr+s = + O (k r+s−1 ). g .1 I. φ ∈ Aut k (x) .15 say.. x r .1 x g  Note dk + n n  = (dk)n n! + O (dn−1 k n−1 ) This is the dimension of the space of monomials of degree 1. The converse is only true for r ≤ 3 by virtue of the fact that mobius transformations can take at most 1 distinct points in P to any other 3 distinct points.6.7 x 1.6. 1. x g k [x0 . x ϕ ∈ Aut (P1 ) .. . φ induces an auto-birational For If 1. I. .7.5. (r+s)! r r The subring of  d.2 b.5.7 x g We can extend any map between the two curves to a map Since Pi must map to Qj thus P1 → P1 .. a nonsingular curve.. r = s.. (f 7→ f ◦ ϕ) ∈ Aut k (x).5.6. y0 . 3 1. g are coprime .

and computing this gives the result.6. b.6. x Note that the hilbert polynomial of Y ×Z is the product of the hilbert polynomials of tensor products multiply dimensions.7. c.6.6 n+k n pa (Pn ) = (−1)r (1 − 1) = 0  1.6.3 I.2 x g arithmetic genus of projective space.5 P n χ (l) = pa (X) = (−1)n−1 (1 − χ (0)) n+l n   − n+l−d n  .d.6.d. 1.6. x g important  As in I. x g By I. g x We have the standard four term exact sequence 0 → S (P3 )l−a−b→ S (P3)l−a  ⊗ S (P3 )l−b → (P3 )l  S so that χ (l) = l+3 l − l+3−a 3 → S (X)  l →0  l+3−b l+3−a−b + . Since the hilbert poly of Then 1. d. we get that Then 1.7 e. PH (0) = 0+2 2   0−d+2 2 −   =1− 2−d 2  2  = 1 − (−1) d−1 2  .4  is χ (OPn (k)) =  .1.6.7. 3 3 − Writing this out and solving gives the solution. 56 Y and Z since .7.

d points. By 7.4 x Using bezout. ) ∈ A2 since P is nonsingular.6.6. y) = y + h. which incidentally is f (x. deg g = d . If we assume substituting 0. 1 2 ∂y ∂z If P is nonsingular.7.1. then the tangent line is given by taking partials and then and so we see the linear term is just the above line.o. multiplicity at P is the x-axis. b. (0.6. a1 .6. a2 ) is given by ∂x ∂f ∂f (y − a ) + (z − a ) = 0 .3 x g ∂f The tangent line to a curve Y dened by a polynomial f at P = (a0 .7. then at least one partial is nonzero so it's well-dend hence unique.9 since and Y transversely at smooth points meet in degree By noetherianness. 0). .8 I. then it can't be irreducible of degree > 1. with 57 deg f = d − 1. I.7. 0. We further assume P = (0.3. y) + g (x.11 the linear part. the singular locus is a nite set of points on 2∗ The lines meeting one of these are a proper closed subset of P Y = Y. y) = 0 .t The only line with higher intersection 1. . P is at zero and the curve is ane.b upper bound on multiplicity x The degree can be read o of the lowest term in the equation if the point is at If all terms have degree 1.10 d.5 x g:a. x Assume Y (x − a0 ) + is dened by f (x. the tangent lines are contained in a proper closed subset of P2∗ ( ⊂ Y × P1 ) I. lines which meet 1.

1. x dim Y = 0. By induction deg X ∩ H ≤ deg Y ∩ H ≤ deg Y = d. x   (t.8 x contained in linear subspace. then by 7.6.7.1) which is projection from a point gives an inverse rational map to A1 . P then Y.7 x Choose a hyperplane PQ Y. map P Q or to the line through A rational inverse takes the line through the vertex of the cone over Note that the cone has dimension 1. r+1 By Ex I. 1.12 If Y I. 7.7.6b.1) =⇒ (x.7.6.6b.7 gives Y is in degee 1 variety of dim r+1 58 . . I. The hilbert polynomial of a linear variety has a leading term which gives linear degree.If t= y . dim Y = r and H is a hyperplane containing P but not Y . If If 1. this is linear. y) 7→ yt. Y is irreducible.14 r+1 Y and Q and the vertex of the cone Q to P Q.7. Y ∩H has degree 1 and is thus linear. b. x n in P .7.13 I. then Y has d points and X has d − 1 lines. −f g(t. If on the other hand Y has degree 1 .6.7. and thus isomorphic to P . ex.6. deg X ∩ H = deg X . then for any hyperplane H not containing Y . If over H in Pn not containing is a line from with endpoints in Y.6 x Linear Varieties x has pure dimension r .15 I.6. then by Thm 7.

2 II.1 x 2. we need to show it's surjective on stalks.1.5 II.1 x II.1.2 x g Filtered colimits (like the stalk) commute with nite limits (like the kernel). x.3x surjective condition. 2. see Vakils notes for this terminology Also since the cokernel is range / image. x This follows from the denition of exactness and part (a). 2. b.1. 2. + Let F the sheaf associated to this presheaf.1 II.1 Constant presheaf F denote the constant presheaf.1.1. (see We have a map Apre → A by taking a∈A to the constant map U → A. then image is the kernel of 2.4 c. x To show ϕ is surjective.3 G → coker ϕ. 59 .1.1. x x These follow from part (a).2 II Schemes 2. is then the maps s : U → ∪P ∈U FP satisfying Let certain conditions. It is easy to check that this is an isomorphism on the stalks.1. + F (U ) .

ti ∈ F (Ui ) using the assumptions of the problem. b says injective / surjective i 60 impre ϕ .1.1 is isomorphism via stalks and Excercise 1.9 x II. 2.1. The converse follows by 1.3.1.6 Let 2.b.2. Note we can write holomorphic functions locally as a logarithm. =⇒ ϕ is surjective on stalks. Globally this doesn't work 2.5 2.8 b.7 II. x Note that sheacation preserves injective morphisms such as 2.→ G .b.1.4 x Use 1.We have a diagram: For some Now nd sp ∈ Gp . x g Surjective not on stalks x F the sheaf of holomorphic functions on consider C\0 ϕ : f → exp (f ). .1.b.1. Since surjective i surjective on stalks. since sheacation preserves stalks. 2.1.2. pull back to s.2.1.

by 1.1.b x In vakil's notes we see that forgetful functor is right adjoint to sheacation. 0 We can also just look at the stalks.6 and 1. is surjective.6.stalks are injective / surjective.1. it preserves colimits and thus surjections.11 II.1.13 II. then ker ϕ = 0 so ker ϕ (U ) = 0 61 .b. 2..10 II.8 x g ϕ 0 → F 0 → F → F 00 .7. 2.6 and 1.1.12 II. 2.6.4.4.1.1.1.2.1. Thus sheacation preserves kernels. F → F /F 0 .1.b x Again. .. 2.a x map to quotient is surjective Sheacation is left adjoint (vakil notes).1.7 x By 1.14 If II.b 2. and so the left most map is injective as presheaf maps. Fp → F /Fp is surjective then using 1.

→ n → n 2.12 x As in the previous.     lim Fn (Uij ) = lim limFn (Uij ) = →n ← ij  ← ij → n lim lim Fn (Uij ) = lim Fn (U ) = We have → n lim ← ij lim Fn (U ).2.1. (See the rst chapter in Vakil's notes) 2.9 x Direct sum of sheaves x g Since the forgetful functor preserves limits and the direct sum is a limit. Ui a nite cover since noetherian.1. and morphisms are inclusions of II.1.1.11 x Suppose Let J U is open.10 x Direct Limits x Since sheacication is left adjoiint and preserves direct limits (see the rst chapter in Vakil's notes) 2.1. 62 restricted to J Ui ∩ Uj in Ui and must be isomorphic to .15 II.1.18 Uij = Ui ∩ Uj . The sheaf axioms is that the limit of the functor limFn (U ) → U 7→ limFn (U ) → . and be the category whose objects are Ui and Uij Uj .1.16 II.17 II.1.

+ Thus s gives a section of F . t ∈ F (U ) such that s|U 0 = t is continuous. 0 0 Let U an open neighborhood of P .19 II. V ⊂ Spe (F ) is open and P ∈ s−1 (V ). Hence the complement of the support is open.2. then s (x) is equal to a germ (t. Conversely if s : U → Spec (F ) is continuous .13 Espace Etale x s ∈ F + (U ) . V is open. vanishes. and t ∈ F (V ) then for x ∈ t−1 (s (U )) we must have s (x) = t (x) so there is an open W with s|W = t|W . then the germ of Thus there is a neighborhood where s s in the stalk at p is zero.1.1. −1 Continuity of s gives s (t (W )) is open (by the same reasoning as above t (W ) is open) 0 so on an open W ⊂ W we have t|W 0 = s|W 0 .1. then t s (U ) is open in U for t ∈ F (U ) =⇒ s (U ) is open in Spe (F ). −1 Since we are using the strongest topology such that t is continuous. −1 −1 Then s|U 0 (V ) = t (V ) is an open neighobrohood of P contained in s−1 (V ) (by srongest topology . Note that 19.1. 63 .) and we need s to be continuous. −1 Thus W ⊂ t (s (U )). gives an an example of non-closed support. If 2. Now if x ∈ U .20 If P II. and thus s locally gives a section of F . s : U → Spec (F ) Suppose we have sending P 7→ sP which is continuous (by strongest topology .. W ) in Fx ...b.14 x g is not in the support. =⇒ s (P ) ∈ FP =⇒ P ∈ U .) −1 Hence each point in s V has an open neighborhood contained in the preimage so s is continuous.

there is an open cover Ui of ti ∈ F (Ui ).1. ti − tj = rij ∈ F (Ui ∩ Uj ) . so we just need to show F (U )  F (U ).16 g Flasque Sheaves x All the restriction morphisms are identity. Let If satises s|Ui = 0 for all i . x 5 lemma 64 t lifts to elements t on Ui ∩ Uj . G |U ). Clearly Hom (F |U .1. extend to by then we have dened a lifting of c. Hence s|U = 0. 0 On Ui ∩ Uj . . then since F is a sheaf.8 gives left exactness.1.21 II.23 b.2. there is f ∈ F (U ) glueing similar. G |U ) Hom (F |U . 00 00 If t ∈ F (U ) . 1. is an abelian group and the obvious restriction maps give a presheaf F : U 7→ identity {Ui } an open cover of U .1. 2.1. 2. x 00 Exc.24 rij 0 ti − rij U on which 0 rij ∈ F 0 (Ui ).22 II.15 Sheaf Hom x: g presheaf Let U open. then by surjectivity of F → F .1. we can  0 ti − rij |Ui ∩Uj = tj |Ui ∩Uj so if we replace ti 2. G |U ) with s (f ) = 0 on U . 0 Since F has surjective restrictions. s ∈ F (U ) = Hom (F |U .

x = 0.25 d. {P }− = A.1. P ∈J FP  x 7→ (P → 7 xP ) . of P such that x|UP = 0.1.  P ∈ /U  P ∈ U .1. 2. then we have an induced map ϕU : limV ⊃f (U ) G (V ) → F (U ). 65 .27 II. is A and taking limits of Q ∈ / {P }− The 2. Now if If 2.1. This inclusion gives a natural categorical surjection For U ⊂X open.2.1. Hence i∗ (A) is the skyscraper sheaf. We show the kernel is trivial. dene F (U ) → G (U ) by Q Q Q∈I FQ .26 Let e.17 x g skyscraper sheaves (important) Taking limits of open sets containing we see the stalk is P . then there is a neighborhood UP UP cover U so by the identity sheaf axiom. If P → 7 xp is zero for x ∈ F (U ) . −1 (U ) = then i ∅ so i∗ (A) (U ) = 0. x g Since continuous maps preserve inclusions.28 II. then i∗ (A) (U ) = A (i−1 (U )) = A {P }− = Γ A. we see the stalk at Q ∈ {P }− 0 elsewhere.18 x g Adjoint property of f^(-1) Suppose that If U ⊂X ϕ ∈ HomX (f −1 G . F ). is open.1. x sheaf of discontinuous sections I⊂J be sets of points in open sets U ⊂ X.

σ : → HomX (f −1 G . then since the stalk .30 If If b.1. x This follows from a.b 1 http://sierra. V ) = (s|V 0 .If W ⊂ Y is open. τ : HomY (G . F ) → HomY (G . g x extending by zero P ∈ /U P ∈U 2. not intersecting Z.1. f∗ (F )) → HomX (f −1 (G ) . (i∗ F ) (V ) = 0.1. then U ⊂ f −1 (V ) so we have maps G (V ) → F (f −1 (V )) → F (U ) The restriction maps and taking the limit over f (U ) ⊂ V give limG (V ) → F (U ) for each U .19 x g Extending a Sheaf by Zero (important?) P ∈ / Z . then we dene ≈ σ (ϕ) : G (W ) 7→ limG (V ) 7→ F (f −1 (W )) = f∗ (F ) (W ). (s. V 0 ) . we see P ∈ V 0 ⊂ U and on FP .1. and (s. then we can nd an open neighborhood V containing P .edu/morandi/notes/sheafcohomology.29 If σ ◦ τ = id and τ ◦ σ = id τ (ψ) : f −1 (G ) 7→ F . then c.31 . . conversely ψ : G → f∗ F a sheaf morphism. Note that 1 2. j : P . f∗ F ). .nmsu.→ Z . II. Let For → Using the sheaf axioms gives a sheaf map . F ) . with f (U ) ⊂ V . then (i∗ F )P = Γ (i∗ j ∗ (i∗ F )) = Γ (j ∗ F ) = FP . Thus If 2. V ) ∈ FP . (j! (F ))P is index by opens containing P . V ⊂ Y open.pdf 66 that it is zero.

If U 2. 2.32 II.20 x Subsheaf with Supports identity.34 II. F (U ∩ V ) = F |U (U ∩ V ) = F |U (j −1 (V )) = j∗ F |U (V ) By (a) and denition of For open open so we obtain the second map on V. let b. and since P was arbitrary. For P ∈ Y ∩ U choose i with P ∈ Ui . f |Ui = fi so that f (P ) = fi (P ) = 0 since fi ∈ IY (Ui ). Thus we know s has support inside Z so s ∈ ΓZ∩U (U. V . Suppose that si ∈ ΓZ∩Ui (Ui .33 is open. So f vanishes at P . F is asque. F |U ). Suppose P ∈ U \Z . x HZ0 (F ).21 g sheaf of ideals x glueing Ui is an open cover of U and fi ∈ IY (Ui ) satisfy fi |Ui ∩Uj = fj |Ui ∩Uj that f |Ui = fi .1. f ∈ IY (U ) . gluing.1. Pick i with P ∈ Ui so that s|Ui = si and thus sP = (si )P = 0.1.1.1. Follows since F is a sheaf.2. F → j∗ (F |U ). As F is a sheaf there is s ∈ F (U ) with s|Ui = si . F |Ui ) satises si |Ui ∩Uj = sj |Ui ∩Uj . injectivity on the left is clear. then hence we have Note that if the restrictions are surjective so we will have surjectivity on the second term. If identity 67 then we can nd f ∈ OX (U ) such . Ui be an open cover.

Q = |g (q) = 6 0. then P = 0. then f is given by a system of regular functions fi on Ui such that fi |Ui ∩Uj = fj |Ui ∩Uj f is a rational function such that f |Ui = fi . Q f f |g (1) = 6 0 . We therefore have an exact sequence 2. then f |Y ∩U gives a section of OY (U ∩ Y ) = i∗ (U ∩ Y ). Iff globally 68 and . and the map given by evaluation We also have OX. On global sections however. P Given f ∈ K (U ) . then quotient denes f ∈ iP (K /OP ) . If P ∈ Y . WLOG assume at 2. On stalks. x g f ∈ OX (U ) .1. Q = 1.Follows since its a subpressheaf of a sheaf (see Vakil's notes) 2.35 b. 1). F (U ) = i∗ OP ⊕i∗ OQ (U ) = 0 for U 63 P.37 d. For P ∈ / Y .P with ϕP (h) = g . On the open set x1 6= 0 . 1) and Q = (1.Q = g g 1 gives an exact sequence 0 → IY.Q → OX. The same thing happens at P . ϕP If U is open. Q. we have 0 → 0 → k → k ⊕ k → 0. x g P = (0. being the induced map on stalks. we have a sequence 0 → OP → K → K/OP → 0 which is clearly exact. then g ∈ (i∗ OY )P has no pole at P and thus there is h ∈ OX. and IY (U ) = OX (U ).Q → FQ → 0. x g ∈ OX (U ).36 0 → IY → OX → i∗ OY → 0 . c. Dene ϕ : OX → i∗ OY to be this restriction.1. (i∗ OY )P is zero. Thus ϕP is surjective.1. This gives exactness away o. The fi dene a section of K (U ). By denition of the skyscraper sheaves. IY. n o n from P. f (1) = 0 .

b 2.39 II.2 x g induced scheme structure. The isomorphism of sheaves is from 2.→ X ij∗ (Fj ) → ijk∗ (Fj |Ujk ).2. x 1 In the case of (d) we have X = P . So it's a scheme.2.2. 2.1. 1 note that H = 0 by cech cohomology.22 x g Glueing sheaves (important) ij : Uj .1 II.2.→ X .2 II.1 x As topological spaces.1. inverse limit.1.2.1. as the primes of f. they are clearly the same. By II. Ui = Spec Ai intersect Ui 's and X. ijk : Uj ∩ Uk .2. 2. 69 Af are the primes of A not containing .2.2 II.2 2. For dene we have morphisms induced by restriction: The inverse limit of the sheaves in the above morphism has morphisms On stalks we have F |Uj → Gj |Uj = Fj F to be the above So the idea is to dene F → ij∗ (Fj ).38 e. then the ane opens are Spec Af . For a cover of we can cover the intersections with ane opens.

(OX )red ) since we have a cover of basic open anes. so (V.3 a.2. induced by the ane ring homomorphisms on global sections to the reduction. Since nil (A) is the intersection of primes of A. OSpec(Ared ) |U ≈ O(Spec A) | since localization commutes with quotient. For the morphism Xred → X . then nd a neighborhood V of P But OX (V ) no nilpotents. For P ∈ X . x reduced is stalk local x (X.2. s) is nilpotent. s) an element of the stalk. Suppose each stalk no nilpotents.5 X = Spec A is ane.2.e. We have a cover of open anes given by OSpec(Ared ) (D (f )) ≈ (A/nil (A))f . b. we can glue the morphisms induced by Ai → Ai /nil (Ai ) from quotienting.3 II. U.4 s=0 s sn and n such that sn = 0 is zero at each point in on V. n > 0. Patching together gives 70 X → Yred → Y . then cover X with open ane schemes Spec Ai . If (U.2. These are unique since they factor uniquely through the reduction. x Dene g : X → Yred by taking the continuous map g=f and the sheaf map OYred (U ) → g∗ OX from (b). then sp Spec (A) = sp Spec Ared . OX ) is reduced =⇒ OX (U ) no nilpotents. then the germ of Then the stalk of But then 2. i. s) = 0. . Thus on each basic open ane U . s|V ) = (U. red U Thus Spec (Ared ) ≈ (X. reduced scheme x Ared = A/nil (A).2. n If s ∈ OX (U ) has s = 0. must vanish at each such point (since no nilpotents) by separatedness. First suppose 2. If X is a scheme. let (U. and let c.

x → k (x) → K = i∗ OSpec K (U ). We have by hypothesis A → Γ (X. Thus we obtain a map Ui ≈ Spec (Γ (Ui .6 II.2.4 x X.7 x g given a point + inclusion gives morphism i : Spec K → X . ] If U is open. By glueing we obtain an inverse map X → Spec A to α. given by identifying the point of Spec K with one x ∈ X gives a continuous morphism.2. Let 2. X.2.2.2.9 0 to any X. OX ). there is a unique morphism of topological spaces to any Since there is a unique trivial map from 2. 2. OUi ).7 Ui an ane cover of II. U 3 x .8 II. and morphisms are in 1-1 correspondence with ring homomorphisms (for ane scheme) we're done. (since prime is maximal in commutative pid) Since rings have unique morphism to Z. then it's initial. OX ) → Γ (Ui .2. II. Since there are no points. OUi )) → Spec A.2.2. i∗ OSpec K is skyscraper sheaf with ring of sections K .6 x g Spec 0 has no points since there are no prime ideals. dene i : OX → i∗ OSpec K by OX (U ) → OX.2. given morphism gives point + inclusion 71 .5 x g Spec Z is (0) (open) and (p) closed. and by restriction Γ (X.

algebra . AG 72 . The closed sets are nite collections of points. we assume X is ane. so k (x) = Ap /pAp → K gives the inclusion. f 7→(df )P 2 → nP /n2P → k denes a derivation. 2 If D : OP → k is a derivation. and thus gives a derivation OP → k .Clearly p = i ((0)) is the point.2. . (0) is the generic point and the other types of points are maximal ideals.10 x g Irreducibles give points corresponding to prime ideals. then 2. − − If U ⊂ Z is open and ζ ∈ U with {ζ} = U then {ζ} = Z since Z is irreducible. f 7→f (P ) c7→c We have k → OP → k is the identity map =⇒ OP ≈ k ⊕ nP .2. Zariski Tangent Space corresponds to derivations First we show that Tx ≈ Der (OP . then α (a) = a0 + Dα (a)  . then if φ : A → K is the p is the kernel of φ . a 7→ a0 is the quotient map OP → OP /m = k . and if p ∈ U = Spec A. f ↔ (f (P ) .9 x g Unique Generic Point (Important) Z = Spec A. As Z is irreducible.11 II. 2..8 x g Dual numbers + Zariski Tangent Space Since the assertion is local.  = X + (X 2 ).2.g. As α is a homomorphism of k -algebras. On the other hand if f : nP /nP → OP . =⇒ nil (a) has a unique minimal prime whose closure is X . f − f (P )) .12 II. then D is zero on k and nP by the product rule. (x − a) for a ∈ R. If On the other hand all derivations arise in this manner. then OP Derivations Correspond to Dual Numbers α : OP → k [X] / (X)2 is a local homomorphism of k -algebras. xn ] /a a f. or (x + ax + b) for an irreducible quadratic.2. corresponding ring homomorphism.. 2 Therefore D denes a k -linear map nP /nP → k . 2 Milne.10 II. the vector space of derivations. A = k [x1 . 2 2. First we reduce to the ane case. k) . Also α (ab) = (ab)0 + Dα (ab)  and α (a) α (b) = (a0 + Dα (a) ) (b0 + Dα (b) ) = a0 b0 +  (a0 Dα (b) + b0 Dα (a)) so that Dα satises the product rule.2. 2 Such are either (0).2..

7c) and I. (descending is noetherian (I.2.2. f 's of degree d is the Fp [x] / (f (x)) ≈ Fpd gives α ∈ Fpn .2.2. f an irreducible monic polynomial. Let X= ` Xi / ∼ with the quotient topology. note that x i ∈ X i . y ∈ Uji ⊂ Xj if and φij x = y .  . clearly satises (1) and (2). 2. then any open If every open U U ⊂X is quasi-compact.2. then U= S Ui U is quasi-compact. . then the isomorphism Conversely. 73 Ui . gives a minimal polynomial of degree d. pd elements.13 x quasicompact vs noetherian. then there are with (4) is similar. Now glue the sheaves Then (X.22 to get OX . 2.12 x g Glueing Lemma Dene an equivalence relation by x∼y x ∈ Uij ⊂ Xi . Thus we count elements of Fp d not contained in any subeld.1. a is noetherian. xj ∈ X j ψ∗ OXi using I.b gives that is covered by a nite subset of Then the chain must stabilize. Qd−1  i=0 x − αp i . The residue eld of a point corresponding to one of the If f is one such. and this number is given by the mobius inversion formula. OX ) For (3). Thus x ∈ ψi Uij =⇒ ψ (Uij ) = ψi (Xi ) ∩ ψj (Xj ). given α ∈ Fpn nite eld with not contained in a subeld.1.11 x g (Spec Fp important) Spec k [x] = {0} ∪ {(f )} . (see Apostol intro analytic number theory). and we have an ascending chain of opens chain of closeds). ψ (Uij ) ⊂ ψi (Xi ) ∩ ψj (Xj ) and conversely that if x ∈ ψi (Xi ) ∩ ψj (Xj ) xi ∼ xj .7.15 If X II..13 II.2..2. U1 ⊂ U2 ⊂ .14 II.

If p⊂U is prime..16 b..2.g.. b. there is Thus there is a homogeneous component This takes closed sets to closed sets. x2 . for any prime ideal so all homogeneous prime ideals contain 2.) . U is −1 For the morphism. 2.19 integers or II. x S {U }i an open cover. then P d≥0 p ∩ Sd ⊂ p is prime. (e. X).2.] .18 k d. .17 c. fi with ϕ (fi ) ∈ / p. x22 .14 x If every element of P roj S = ∅.2.] / (x21 . 2. x Find a space with one prime ideal but which has an increasing chain of ideals.2. f : U → P roj S . Let 2. . and is thus continuous. Suppose S+ 6= 0 . Ui = X =⇒ 1 = ai fi with fi ∈ Ii and the sum is nite. and since U is a union of opens. Thus p ⊂ S. and is nilpotent then s ∈ S+ p ⊃ S+ .2. . These principal opens cover U . Thus there is a principal open p ∈ D+ (ϕ (fi )) ⊂ U . x A decreasing sequence of closed subsets corresponds to an increasing sequence of ideals. Now nilpotent. P Vi = X\Ui dene ideals Ii ⊂ Γ (OX . dene f : p 7→ ϕ (p). If P roj S = ∅ S+ .2. A sheaf morphism is given by Sϕ−1 p → Tp . x f ∈ S+ with ϕ (f ) ∈ /p.2. D+ (s) = ∅ . non-noetherian ane scheme is Spec k [x1 . [x1 . 74 open in P roj T .20 p-adic s =⇒ s since every prime ideal contains every nilpotent... x2 ..

Thus ϕed0 s = t 0 ∈ p which is prime. so it's not in ϕ (S+ ) .21 c. By taking large degrees we can s Let show easily surjective.2. of residue elds k (f (P )) → k (P ).2. let q = P −1 −1 n n by ϕp the image of ϕ. Residue eld has transcendence degree 0 in an algebraically closed so it's k . If f (p) = f (q) . then ϕ p = ϕ q. Let 2. for large enough d. If P ∈ t (V ) has residue eld k . Using similar reasoning. x f ] : OY → f∗ OX gives a morphism k . Thus ϕanm ∈ ϕp so anm ∈ p so a ∈ p.deg ≥ 1 contradiction. we can see that q is prime so in fact we get f is surjective. then irreducible closed subset Z corresponding to P dimension ≥ 1 =⇒ tr. i Now consider the induced map on structure sheaves. Thus in total ϕ−1 q = p . 75 has . Thus ( bi si )m is a polynomial in ϕsi with coecients in cj ∈ S where p lives.→ k (P ) = k . For large enough powers of sm f .10 2.15x g (important) P ∈ t (V ) closed. −1 −1 Next we need injective. Sd → Td sn f m is an isomorphism so s f = 0 so n = 0 so we know S(s) → T(t) is injective. since ϕd is an iso for d ≥ d0 . then there d d is s ∈ S with ϕs = t 0 by assumption. We want that S(si ) → T(t) is an isomorphism. then f −1 D+ (si ) = D+ (t) ⊂ P roj T for some t. For large m. then p can't be contained in P roj T by denition of P roj T as the set of homogenoeus prime ideals not containing T+ = ⊕d>0 Td . x By II.2.23 II. Note that ϕ q ⊃ p .2. If p ∈ P roj S . we can nd s ∈ Sed0 with ϕed0 s=td0 d .2.4. but is not closed. then ϕa ∈ hϕpi so ϕa = bi ϕsi for bi ∈ T and si ∈ p . As p ⊂ T+ . On the other hand if a ∈ ϕ q . p hϕpi the radical of the homogeneous ideal generated Next we show f is surjective. Then ϕs = t 0 ∈ q As q is prime it contains t thus p ⊂ q and by symmetry p = q. thus must contain t. 2.2.24 b. so sm f ∈ ker ϕ. For t ∈ p.22 d. If f 7→ 0 . every monomial in bi is in T≥d0 so we get an isomorphism Td ≈ Sd for P large d. then 0 = tm ϕf = ϕ (sm ) ϕf for large m. If t ∈ Te .→ k (f (P )) . As D+ (s) = D+ (s ) cover P roj S for i ≥ d0 . Let s = si . p ⊃ ϕ (S+ ) . x p ∈ P roj T .

26 this gives regularity.2. is the union of opens. n By the sheaf axioms.28 c. ϕ (Y ) = ϕ (Y ). p ∈ U = spec A0 ⊂ f −1 (U ). f |U 0 : Spec A0 → Spec A is induced by the map on rings. nd an nite open cover of Ui = Spec Ai a|Ui ∩Xf = a|Spec(Ai )f is zero for ecah i. x The natural map is given by ϕ 7→ ϕ∗ . if := ϕ∗ |V . n Thus f i a = 0 in Ai for some ni (by theorem). x By quasi-compactness. then f a = 0 in each Spec Ai .2.27 b. = ϕ (p). then ϕ∗ takes closed points to closed points. ϕ : t (v) → t (W ) . ϕ . II. Then 2. . The maps on schemes over ∗ ϕ For surjectivity.2. ∗ By (b) closed points map to closed points. x 76 . we have Xf = i (Ui ∩ Xf ) We trivially have 2.    ( D f = x ∈ U : f ∈ / x = x ∈ U : fx ∈ / mx ) S Furthermore.2. so ϕ (p) ∗ If Y is an irreducible subvariety. k are extensions of ϕ : V → W and so we get injectivity.25 c.16 x  D f = U ∩ Xf so U ∩ Xj is open in U . an ane open cover {Ui } of X .2. and and thus dene For regularity of 0 Then Thus 2.2. let ϕ (P ) = Q and choose U = Spec A 3 p. n For a large enough n. then f a = 0.

OX ). Together we have a homeomorphism and an isomorphism and gluing along double intersections of the Ui will give an isomorphism of schemes.. is homeo.29 d. x setup f1 = 1 . . and for a large N . Since fi generate A.31 b. Conversely. M Using the sheaf axiom. then fan = 0 on Spec (Aj )fi = Xfi ∩ Xfj in the domain. f (V ) = S f (V ∩ f −1 (Ui )) is open. s − f N +M b restricts to f M bi − f M bi = 0 on Ui ∩ Xf =⇒ f n+M b is the restriction of a global section. For M  mjk all j. x  a|X ϕ : a/f n 7→ f n |nf .2.  nj for all nj so on Xfi fj fk we have bj − bk = − fiN a = 0 . 2. M On the overlaps. . 2. m For a section s on Xf .30 II. we let to Xf ≈ Spec Ai so that by (a). f s is the restriction of a global section. fr generate the unit ideal. f If a fin surjective If a ∈ im (ϕi ) Choose N mjk Thus fi on Xfi .2. suppose f1 . bi − bj |Ui ∩Uj vanishes =⇒ f (bi − bj ) = 0 . fi a = 0 = n .2. OX )fi → Γ (Xfi .17 Criterion for aneness x The condition =⇒ f f −1 (Ui ) ≈ Ui implies for open V ⊂ X .. X → Spec A is an isomorphism. then as  OX Xfi fj = (Aj )fi =⇒ a|Xfj fi = bj n fi j fiN a for bj ∈ Aj . D (fi ) cover Spec A . We need to show that D (fi ) ≈ Spec Af is isomorphic If A is ane. i a n N Hence f j a = 0 in Aj for some nj . OXf be This gives surjectivity. injective ∈ kerϕi .. the map on stalks is an iso.Ui = Spec Ai b|Xf ∩Ui = fbNi for each i. (bj − bk ) = 0 on Xfj fk . there is thus fiM bj 77 in X fj agreeing with fiN +M a .2. If 2. p ∈ Ui Since any some i . k . Consider ϕi : Γ (X. Let ϕ : Aj → Γ Xf . lift f bi on Ui to a global section s on X . n The kernel is trivial since a/f = 0 implies it's zero on Af .

78 ai n fi i ∈ Afi on a principle open set D (fi ) 3 pi .32 a by ϕi d restricting to fiN +M a on X fi . If f is nilpotent.→ B is injective. then it's in all the prime ideals so it vanishes at every 2. If for A ≈ Γ (X. since we have a distinguished base. c. primes of Now D (f ) D (f ) pulls back to to primes of A/I . gives a global section maps to 2. then evaluated on D (a) . and is thus nilpotent. (f∗ OSpec B )p ≈ B ⊗A Ap as it is Thus we have an injective morphism the colimit of OSpec B A p → B ⊗A A p . g x Injective If the sheaf map is injective then A . The map is clearly continuous. OX ) → Γ (X. with f ∈ φ−1 p . This shows that the map is open.2. b.2. g x We have f# is surjective on each stalk so if mapping to the germ of b at each b∈B . By ring theory. 2.2. a ∈ /p. f∗ OY ) ≈ B is injective. Then d f N +M .Using the global generation lemma.2. and checking the stalk gives surjectivity. p ∈ Spec A .34 φf ∈ p for all p ∈ Spec B so φf is in their intersection of φ implies f is nilpotent and thus D (f ) is empty. x g (important) A containing I = ker ϕ correspond D (f + I) in Spec (A/I) .33 p.2. Dominant The largest open set not intersecting the image is covered by For each f Injectivity 2. . there is pi ∈ Spec A.35 d. II.18 g x Since the nilradical is the intersection of the prime ideals. p ∈ Spec B .

. V ((e2 )) covers Spec A . as well as the union of the open sets Spec (Bi )f . −1 Then there is Vi = Spec Bi a covering of Y by open ane subschemes such that f Vi is covered by open anes Spec Aij .By quasi-compactness. . Then Spec A + Spec A/I × A/J so A = A/I × A/J .3. Bi -algebra. then it is q..3. 0). For if is an open cover with each Ui q. e1 e2 = 0 so either e1 ∈ p or e2 ∈ p. and the Spec (Dij )gp cover Spec B . 2.g. then Spec (Ci )g ≈ Spec Bgp . For p ∈ Spec B .2. The intersections Vi ∩ B are open in Bi . e2 = (0.3 II. Spec Bi is an open ane cover of 79 Y and f −1 Spec Bi Ui is q. II. then ϕ Spec (Bi )f = Spec (Aij )f . and thus ij ik each (Aij )f is an f.19 x g ` =⇒ (iii) Suppose Spec A = U V . gi ∈ A and b = P gi fiN b = P 0 gi fiN ai ∈ im ϕ.c. then Vj ∩ Ui is an open cover of Ui which has a nite subcover. and Vj is a cover for X. Now suppose f is quasi-compact. 2.c.1 x f : X → Y is locally of nite type.g.c. 1).c. B -algebra. (iii) =⇒ (ii) for e1 = (1.3 x 2. open sets.2 II. hence an f. p (Dij )gp is an f. Suppose 2. Then V ((ei )) . (ii) =⇒ (i) For any p. V = spec A/J . ik ik Now cover Spec B with open anes Spec Ci whose preimages are covered with open anes Dij . each Aij an f. ` If p ∈ V ((e1 )) ∩ V ((e2 )) then 1 = e1 + e2 ∈ p =⇒ p = A =⇒ Spec A = V ((e1 )) V ((e2 )) (i) . Ci -algebra.3. So p ∈ Bgp ⊂ Spec Ci .36 Spec A = Sn i=1 D (fi ) so 1 = P gi fiN .g.g. (Bi )f -algebra. .g.2 x First note that if a topological space has a nite cover made up of q. each Dij an f.1 II. ij −1 If fij is considered as an element of Aij under Bi → Aij . Bgp -algebra. If gp is associated with its image under B → Ci → Dij .2. for two closed sets U = Spec A/I . and taking the p preimage gives Spec (Dij )g .3. p ∈ Spec Ci for some i.

2.4 b.5 c.g. II. We can cover the intersections U ∩ Ui with distinguished opens D (fij ) = (Bi )f . then (Ai )f is an f. 2. Combining quasicompact covers. the open anes are quasi-compact. (Bi )f -module by basic localization properties.3. If then each intersection Spec Bi ∩ Spec C is covered by basic open anes for By II.3. Bi -module.3 x f Y with Spec Bi and f −1 (Spec Bi ) is covered by a nite Spec Aij are qc. ij ij We have a cover of V by principal open anes Spec Bgi with preimages Spec Ci . let Spec (Ai )g a nite cover of U by principal opens basic in Ui . Bi -module. cover 2. cover each f Spec Bi with a nite number of Spec Aij .6 II. since Spec B is ane.13b. x f −1 (V ) with Ui = Spec Ai . 2.3.3.13b. −1 By quasicompactness of f . it is q.g. then by denition we can cover number of Spec Aij .2. so by the rst paragraph.17.4 x U = f −1 V = f −1 Spec B −1 There is at least one ane cover by Vi = Spec Bi of Y such that each preimage f Vi = Ui = Spec Ai is ane with each Ai an f.b.c.3 If II.g. Then Spec Bi cover X so also Spec C .Spec V ⊂ Y is an arbitrary open ane. there exists a nite subcover of Spec C . gives qc is ft and qc. B -algebra =⇒ A is a nitely generated B -algebra. x By the previous exercises in this section.3.2. Bi -module. Ai an f. By exc. Spec Bi .13. the on the whole space. Using quasicompactness. of the for D (bk ). By a previous excercise.g. Each Ai is an i f − g B algebra. Ci an f. Suppose f is nite.g. By II. The preimage of D (bk ) in Spec Aik j is Spec (Aik j )bk and the union of these is a cover of f −1 Spec C by open anes.2. B -algebra. bk ∈ Bik . we are done. Suppose nite type. by exc 2. =⇒ (Ai )gi = Afi some fi is an f. fij is associated with its image in Ai . Since Ai is an f. Let 80 .3.g. and the preimage of ij D (fij ) = Spec (Ai )fij .2.

2. then there is g ∈ k [Y ] with g (y) = 1 and f (X) ⊂ Z (g) . f Spec Bgi = Uf (gi ) .3. with A an f.B = k [X] . .5 x g Let p a point in Y .. by niteness we assume A ⊗B k (p) is an f.3.g. y ∈ f (X)c . c.g. is closed if the complement is open.3. we want to show the preimage is a nite number of prime ideals. so by ane criterion. then by denition.g.6 x g Function Field U = Spec A is an open ane subset of X. B -algebra. (0) is a prime ..7 P for each i . surjective but not module-nite. A vector space is artinian. Note that if f1 . k [X] annihilates f ∗ (g) Thus we want to show if 2. A ⊗B k (p) .. Bfi -module actually nitely generated as a B -module. t−1 ] ⊕ k t. b.3. = Oη is the fraction eld of OX (U ) . then A is II. then f ∗ (m) B = B so by Nakayama. Thus there are a nite number of Thus −1 2.3. A B -module. so there are a nite number of prime ideals in 2. k (p)-module. contains of By uniqueness. X = Spec A . and a eld-module is a vector space. (t − 1)−1 → Spec k [t] is nite-type.g. fn ∈ B generate the unit ideal and Afi is an f. OU ) generates the unit ideal. Y = Spec B .8 is an f. x g A subset of a topological space is closed i it is closed in every element of an open cover. V (I) (0) i (0) contains I so the closure (0) is the generic point η of X . U = Spec A some f. OX (U )(0) 81 (0) is V ((0)) = Spec A . Y = spec B .10 If II. gi ri = 1 so the image in Γ (U.3. The nullstellensatz gives since y ∈ / f (X). A is an integral domain so ideal. nite bers. Thus we assume f (X) X = spec A . B -module. Let A = k [Y ] . m the maximal ideal of A correspnoding to y .g.9 . x You can check that   Spec k [t. Since the assertion is local.Spec Bgi covering V .

k (B −1 A) = k (A) is nite over k (B).12 II. tn ] . Ab are monic. says there is an integer n such that B A is nite over k (B) [t1 . and V is the inverse image of U ∩ V in V Assume WLOG that . Clearing denominators. X = Spec A. By localizing minimal polynomials we nd. so is k (B) ⊗B A ≈ B A. then image of gi in Bb ..g. B = Cg f. V U0 where is the inverse image of are open anes on some common ane scheme U ∩V in U˜ W = Spec C .. f u ∈ A If ϕ : U0 → V 0 Thus if A˜ is the integral closure of A .3. . and an open ane Now we show for the ane case and leave the patching.. −1 Since A is nitely generated over B . the induced morphism of ane schemes is surjective.11 II. we must nd an isomorphism 0 ˜. V˜ = Spec B of X. . the ai satisfy fi (ai ) = 0 for fi ∈ k (B). and is thus an f. As elements of k (A) . .8 x Normalization Normalization Let Let U and V two open ane subschemes ˜.. Y = Spec B assume {ai } generate A over B . . Clearing denominators of fi gives gi with coecients in B . Choose an open .. u A˜f belongs to integral closure of Af . g ∈ C . tn ] .. −1 Finite-type gives that B A is nite over k (B). h with coecients in Af . −1 Noether normalization. over Bb Hence Ab is integral over Bb . since k (A) is nite over k (B). tn ] is surjective so that n = 0 and B −1 A is integral over k (B). then is integral over u A˜f Af .. −1 Since B A is integral over k (B) [t1 . with U. so is U .g. Spec A = U ⊂ f −1 V such that A is an f.7 x k (X) is a nite eld ane Spec B = V ⊂ Y First we show that extension of k (Y ) as in the hint. If b is the product of the leading coecients of these polynomials. Clearing denominators. U˜ = Spec A˜. A = Cf . Then U = D (b) Let 2. In order to glue. Bb -module.2. is root of monic polynomial is the integral closure of 82 Af .3.g. Spec B A → Spec k (B) [t1 . B -algebra. Since X is irreducible.3. So Ab is f. −1 By going up..3. then l ˜ for some l.. so A is integral.

X of nite type If X is nite type. (g)) ∈ sp Spec k [x] × sp Spec k [y] maps to (f ) and (g) by the projections. y] is sent to (0) by the projections. t] where S is generated by t.3. Now take the induced structure sheaf. y] .3. then we want to show the morphism is nite. Dominant Morphism / factors uniquely If there is a dominant morphism from a normal schemeZ we have a dominant morphism Assume WLOG X. ((f ) . II. g x k (s) × k (t) S −1 k [s. the holomorphic functions with poles along horizontal and vertical lines.13 k -algebra A is a nitely generated But the integral closure of a nitely A-module. Thus. c(s)⊗d(t) Note that is products of irreducible polynomials in s and irreducible polynomials in In other words. d ∈ k [t] .14 b. We want to show that if morphism ZU → U →X im A˜ lies in   ˜ A˜ → f rac B ˜ B as ˜ B . A˜ → B We have a morphism ˜. Note that p = (x − y) ∈ sp Spec k [x. then it is is integrally closed. and as an element of the image is integral over im (A) .3. On the other hand. 83 that aren't in the preimage of the projections (i. elements of k (s) ⊗ k (t) are written as P 1 ( ai (s) ⊗ bi (t)) for ai . c ∈ k [x] and bi .9 x g Topological Space of a Product A1 × A1 ≈ Spec k [x] ⊗ k [x] ≈ Spec k [x. then we have a ˜. generated 2. but (0) 6= (x. y). . and preimage ZU . If 2.→ A˜ induce U˜ → U ⊂ X so there is an . t] poles along horizontal and vertical lines). then for open ane are ane. X get a scheme and the inclusions A . Points of Spec k (s) ⊗k k (t) are points of Spec k [s.U˜ to ˜ X→X Thus we can glue induced morphism ˜. f : A .→ A˜ and and ˜ g : A → B is any ring homomorphism. Z U. integral overB  Thus .e.

generic point. then f (V ) ×Spec A Spec k (y) ≈ Spec (Bi ⊗A k (y)) .11 x g Closed subschemes Y 0 = Y ×X X 0 . Denote 84 . t] / (s − t2 ) ×Spec k[x] Spec k (a) ≈ ≈ Spec (k [s. s = t2 = a so that the elements are a0 + a1 t .15 II. −1 = Spec Bi . 0−1 0 0 0 0 0 0 0 Thus f (U ) ≈ Spec (A ⊗A A/I) ≈ Spec (A /IA ) so that f : Y → X is a closed immersion.  Bp ≈ db |d ∈ / f (p) . 0 0 −1 If U = Spec A and U = Spec A . −1 By ? we have sp Xy ≈ f (p). t] / (s − t2 ) ⊗k k (s) ≈ Spec k (s) [t] / (s − t2 ). s − a). and let g : X 0 → X be any morphism. s − a) ≈ k [t] / (t2 ) which is the dual numbers. 2 2 As t = a . then since f is a closed immersion. q ⊃ f (p)} = f −1 (p). then we have k [s. if a 6= 0 .2. 0 −1 Similarly. s − a) ≈ k ⊕ k by 1 1 1 1 (1. 2.10 x g Fibres of a morphism XyS≈ X ×Y Spec k (y) ≈ f −1 (V ) ×Spec A Spec k (y) S where y ∈ V = Spec A ⊂ Y . 1) ↔ − 2√2 t + 2 . t] / (s − t . Note that −1 If f (V ) 2. d ∈ f (A) =⇒ Spec Bp ={q ∈ Spec B|f −1 (q) ⊂ p}.3. 0 0 0 0 First replace X with an ane open neighborhood U of f (Y ) by basic closed immersion conditions. Note that k ⊕ k has two points.3. ? Now  if y = p ∈ Spec  A =⇒ Spec B ⊗A (A/p)p ≈ Spec (Bp ⊗A A/p) ≈ Spec (Bp /pBp ) . f (U ) ≈ Spec A/I .16 b. k (s) [t] / (s − t2 ) is a eld and s − t2 has degree 2 in t. t] / (s − t2 . −1 Hence Spec (Bp /pBp ) ≈ {q ∈ Spec B|f (y) ⊂ p. assume U ⊂ g (U ) . 0) ↔ 2√a t + 2 and (0. a is zero If a=0 . t] / (s − t2 .3. Xη ≈ Spec k [s. x a is nonzero Xy = Xa ≈ Spec k [s. Each point has residue eld k . where U ⊂ Y is an ane open. t] / (s − t2 ) ⊗ k [s] / (s − a)) ≈ Spec k [s. then k [s.17 II.3.3.

20 d. and the scheme-theoretic image is given by the the closure of image Zred . we have a factorization of f factors through the reduced induced structure on Z → Zred → X f (Z).18 b.3.21 II.12 x Closed subschemes of Proj S ϕ (S+ ) = T+ so that {p ∈ P roj Y |p 6⊃ ϕ (S+ )} = U = P roj T . U open in Y .2. since Y = V (a). then glue together open anes to achieve the result. Clearly 85 to homogeneoeus ideals of .19 c. f : Y → Y → X gives sp (Y ) ≈ sp (Y ) ≈ sp (V (a)) ⊂ sp (X) . 2. (Starred) 2. .3. We have T ≈ S/ker ϕ and homogeneoeus prime ideals of S/ker ϕ correspond S which contain ker ϕ hence f (P roj T ) = f (P roj S/ker ϕ) ≈ V (ker ϕ). x g scheme-theoretic image Suppose Z reduced. Let 2. If X.3. Y are not ane. If Z is non-reduced. By (c). Suppose rstX. 0 0 As a map on topological spaces.3. Y are ane. x f : Y 0 → X a closed immersion. For an open U ⊂ V (a) ⊂ X .3. then OX  f∗ OY 0 extends to OX  f∗ OY 0  f∗ OY .

3. 2.3.b. −1 Then f Ui → Ui is a closed immersion. x 0 ϕ : S/I 0 → S/I = (S/I 0 ) / ⊕di=1 Ii the natural projection homomorphism. Let 2.3. Spec Ai an open open immersion.25 ane cover of c. X. Each Ui is covered by basic open anes D (fij ) ≈ Spec (Ai )f ij Each (Ai )f is a nitely generated Ai algebra. II. x g i : U → X a q. By closedness and surjectivity. II. Thus the sheaf homomor- phism is surjective.3.3.13 x g Properties of Morphisms of Finite Type f : Y → X is a closed immersion. Hence each Bi is an f. Ai -module. ϕd the identity for d ≥ d0 f : P roj S/I → P roj S/I 0 .11. thus Ai → Bi is surjective (See Liu chapter 1).2. 2.14c.24 Let b.g. 86 .a. ij Thus i is locally of nite type By exc II.. f −1 Ui = Spec Bi for some nitely generated algebra Bi . i restricts to open immersions Ui → Spec A .22 Let ϕ b.On stalks. we have a closed immersion. .c.3. is a graded homomorphism of graded rings. II.3.3. x This is follows from the denitions. and by exc. with By exc. Let Ui = Spec Ai be an open ane cover of X .23 ϕ induces an isomorphism . we have S(ϕ−1 (x)) → T(x) induced by ϕ which is surjective since ϕ is . i is nite type. Suppose 2. We have a surjection (Ai )p → (Bi )f −1 (p) on all the localizatons at prime ideals.

x Suppose that X 0 .27 is nite type. X 2.. Since they Ai and g : X 0 → X .g. X ×S Y → Y is nite type. then g −1 Ui ×Ui f −1 Ui is nite type over → g −1 Ui is nite type.3. Each Vi ×X Y is nite type over 0 0 0 cover X . xn ]  A Since this map factors through Thus A is an f. C -algebra.g. . . B [x1 . C -algebra. x C ⊂ Z . and can be factored This is nite type since thus by base extension (d). X ⊃ Spec A ⊂ f −1 (Spec B) nonempty open sets. then let Vi an open ane cover of Y . X. then A ⊗C B is an f. X → S is nite type. Thus g −1 Ui ×X Y 0 of X and thus is nite type.g. g. are ane.. By (c) the composition is nite type..3.. Aij are noetherian =⇒ X is locally For cover of noetherian. B -algebra If if A is an X 0 ×X Y → X 0 f. Now use quasicompactness on a cover 2. . so by the rst case we nish. then B [x1 . X are both ane. Y noetherian =⇒ Each Bi is noetherian =⇒ by Hilbert basis.3c. Spec Ci of Z. as Spec A ⊂ h (Spec C) . xn ]  A. Spec B ⊂ g −1 (Spec C) ..3. then any nite open ane cover Ui ⊂ Y 0 of Y ×X X . 2.g. gives a nite open ane cover 0 is ane. then A is an f.. Y f :Y →X is nite type we want to show that the projection • If • X 0 . xn ] .28 f. by nite type hypothesis. If {ai }ni=1 generate A as a C−-algebra.2. −1 By exc 3. Vi Ui ×X X 0 . there are Uij = Spec Aij a nite f −1 Vi each Aij a nitely gnerated Bi -algebra. So f 0 is nite type on an open cover e. and there is a morphism C → B → A Pick .29 is surjective.. then C [x1 . The second map is nite type by assmption. • If • If X is covered by open anes Ui = Spec g −1 Ui by previous case. and Vi ×X Y is the preimage of Vi then X ×X Y is nite type over X . x Vi = Spec Bi a nite ane cover of Y .26 d. 87 .. B -algebra. x X ×S Y → S X ×S Y → Y → S .3..

Consider a nite open ane cover

{Ui }

of

X.

f 

nite type =⇒ q.c. by exc I.3.3.a.
−1
Thus f
Ui is q.c. exc I.3.2.

{Vj }

is an open cover of Y , taking preimages gives an open cover of
−1
Each such f
Ui is q.c., so there is a nite cover.
If

f −1 Ui

for each

i.

The union of the subcovers is nite and is still a cover.
Thus

2.3.30

X

is q.c., hence noetherian.

II.3.14 x g

We must show every open set in a basis contains a closed point.
Every ane open set contains a closed point, since it's an f.g. algebra.
A closed point is closed in the whole subscheme since closed points correspond to points where 
nite.

example

Spec k [X]() = {0, (x)}

2.3.31

Since

(x)

is a closed point and

II.3.15 x

See Liu, section 3.2.2.

2.3.32

b. x

2.3.33

c. x

88

0

is not.

k (x) /k

is

2.3.34

II.3.16 x g Noetherian Induction

Suppose there are closed subsets where

P

doesn't hold.

X is noetherian, there is a smallest on Z .
Since Z is minimal, there can be no proper closed subsets of Z
But then we have a contradiction as Z it self must satisfy Z .
Since

2.3.35

not satisfying

P.

II.3.17 x Zariski Spaces

Note by 3.1.1

sp (X) is noetherian,

so we need to show each closed irreducible subset has a unique generic

point.
For a closed irreducible

Z

, and open

U

either

U

contains the generic point or does not intersect

Thus the result holds i it holds for an open ane
Suppoes then that

X

Z

.

U.

is ane.

Then irreducible closed subsets correspond to, by the nullstellensatz to prime radical ideals.

p the generic point for V (p).
p, q are two generic points for a

Let
If

2.3.36

closed determined by an ideal

I

, then

p=

p=

I=


q = q.

b. x g

Minimal closed subsets are irreducible, and thus has a unique generic point by denition of a Zariski space.
For another point in the minimal closed set, by minimality, the closure is the whole thing. By uniqueness of
generic point, we are done.

2.3.37

c. x g

x 6= y ∈ X , U = X\ {x}− .
y ∈ U we are done, else y ∈ {x}−

Let
If

, but then

y

is the generic point of

89

{x}−

so

x = y.

2.3.38

d. x g

η∈
/ U , then η ∈ U c ,
So U is empty.
If

2.3.39

Let

closed. By irreducibleness,

e. x g specialization

X = ∪Zi

the expression of

X

as union of maximal irreducible closed subsets.


the generic point of Zi and ηi ∈ {xi } .

Then Zi ⊂ {xi } and since the Zi are maximal,
Let

X = {η}.

ηi

Zi = {xi }−

.

Generic points of irreducible closed subsets are unique by previous

=⇒ ηi

η maximal.
η = ηi .

The also part

Z is a closed subset and z ∈ Z
Z contains z , =⇒ {z}− ⊂ Z .

If

2.3.40

.

is maximal.

Conversely, suppose

Then η ∈ {ηi } and

and

=⇒ ηi = xi

is a point, then since

{z}−

is the smallest closed subset containing

z

,

f. x

noetherian
The lattice of closed subsets of

t (X)

is the same as the lattice of closed subsets of

X

=⇒ t (x)

noetherian.

unique generic points
a generic point in a closed irreducible subset Z of X .

The closure {η} in t (X) is the smallest closed subset of X containing
η is a closed subset of X =⇒ that {η}− = η .


0
0 −
If η is a generic point for {η} then {η} = {η }
=⇒ η = η 0 .

If

η

η

. 

furthermore part
If

X

is a zariski space, then there is 1-1 correspondence between points and irreducible closed subsets.

Hence

α : X → t (X)

is a bijection on udnerlying sets.
90

is

The inverse is clearly continuous .

2.3.41

II.3.18x Constructible Sets

Consider

`

Zi ∩ Ui ⊂ X 

nite disjoint union locally closed.

Suppose that this coproduct satises 1,2,3.

X

F.
F are in F
`
Zi ∩ Ui = ∪Zi ∩ Ui ∈ F .

Conditions 1 and 3 imply closed subsets of

are in

Conditions 2 and 3 imply nite unions of elements of
Thus if

Zi ∩ Ui

are djsoint, then

Consider a collection
1 is since

2

F0

U ∩X =U

.

of such coproducts. We want to show they satisfy 1,2,3.

and

X

is closed.

just take intersections of them and get another similar coproduct

3 is by induction.
0
Thus any such F

=F

.

Fn ⊂ F the collection of subsets of X which can be written
n locally closed, thus ∪n Fn = F .
The itnersection of elements of Fn and Fm is in F as in 2 above.
` c
c
c
c
c
c
If S ∈ F1 , S = U ∩ Z and S = (U ∩ Z) = U ∪ Z = U
(Z ∩ U ) in F .
We proceed by induction showing that S ∈ Fn satises complements.
`
c
∩ S1c .
S ∈ Fn , S = Sn−1 S1 , and S c = Sn−1
c
c
But Sn−1 and S1 are in F by induction, and their intersection is in their by 2.
Thus we have shown that F is the locally closed coproducts as aabove.
On the other hand, let

of

2.3.42

as nite disjoint union

b. x

∈F
S ⊃ {η} = X
Suppose S

is constructible. Let
.

=⇒ S

η

the generic point and suppose it's in

S

is dense.

`
S ∈ F is dense. S = ni=1 Zi ∩ Ui by part (a).
Note that ∪Zi ⊃ S so ∪Zi ⊃ S since the Zi are closed, and the union is nite.
Thus ∪Zi ⊃ S = X .
X irreducible`=⇒
Zi = X some 
i.
`n−1
Thus S = Un
i=1 Zi ∩ Ui .
As X is a zariski space, by exc II.3.17.d the generic point is contained in every
Suppose that

X.
So

Un

contains generic point so

S

contains generic point.

91

nonempty open subset of

2.3.43

c. x

closed

=⇒

constructible and stable under specialization is clear.

`n

i=1 Zi ∩ Ui and is stable under specialization.
if ηi is generic point of irreducible comonent of Zi intersecting Ui nontrivially.
S stable under specialization implies S contains every point in {x}− .

Suppose

S

is constructible, thus by (a)

S=

=⇒ S contains
every point of every irreducible component
S
=⇒ S ⊃ Zi .
Consider x ∈ S . Suppose x ∈ Zi (it is for some i).
S
=⇒ S ⊂ S Zi .
=⇒ S = Zi since we've shown both containments.
=⇒ S is closed as the Zi are .

2.3.44

of each

Zi

.

d. x

`n
A constructive set looks like
i=1 Zi ∩ Ui
`
`
−1
Note f
( Zi ∩ Ui ) = f −1 Zi ∩ f −1 Ui .
−1
By continuity, the preimages f
Zi and f −1 Ui are closed and open respectively.

2.3.45

II.3.19 x

reduce to ane
We want to show that it is possible to reduce (we don't actually have to show the result in this case).
−1
If {Vi } is an ane cover of Y , and {Uij } is an ane cover for each f
(Vi ) , then if f (Uij ) is constructible
each

i, j

, then

f (X) = ∪f (Uij )

is constructible, so we assume

X, Y

are ane.

reduce to irreducible
If

{Vi }

Y , and {Uij } irreducible components of f −1 (Vi ), then if f (Uij )
f (X) = ∪f (Uij ) is constructible, so we can assume X, Y are irreducible.

are irreducible components of

constructible for each

i, j

, then

reduce to integral
WLOG we can assume reduced topologically, so irreducible + reduced gives integral.

reduce to dominant
Suppose

f (X)

is constructible for each dominant morphism.
f 0 : X → f (X) which is also dominant so

We have induced morphism

92

f 0 (X)

is constructible.

is

`
f 0 (X) ⊂ f (X) is Ui ∩ Zi by (a) of previous problem.
f (X) closed =⇒ Zi closed in Y .
Ui = Vi ∩ f (X)`for Vi ⊂ Y `
open under the induced topology.
Then f (X) =
Ui ∩ Zi = Vi ∩ f (X) ∩ Zi is constructible.
Thus

2.3.46

b (starred)

2.3.47

c. x

∃a ∈ A with D (a) ⊂ f (X) .
f (X) ∩ V (a) is constructible in Y , assuming it's nonempty
0
Assume V (a) = Spec (A/ (a)) so we have an induced map f : Spec B/aB → Spec A/ (a) with image
f (X) ∩ V (a) .
A → B injective =⇒ A/ (a) → B/aB injective =⇒ f 0 dominant.
As both rings are noetherian, (a) = ∩pi , pi primary ideals by primary decomposition.
p


pi are prime and (a) = ∩ pi =⇒ V (a) = ∪V (pi ) .


pi B , we have B/qj → Spec A/ pi for qj ∈ Spec B .
For each
Each image contains a nonemprt subset by (b), and hence is constructible in V (pi ) by noetherian induction.
A locally closed subset of V (pi ) is also a locally closed subset of Spec B ,
=⇒ images of Spec B/qi → Spec are constructible in Spec B .
By (b),

We show

2.3.48

d. x g

1
2
We can map A → P by x 7→ (x, 1, 0) .
1
f (Ak ) is neither open nor closed since (x, 1, 0) nor its complement are varieties.

2.3.49

II.3.20 x g Dimension

93

3. x g This follows from Thm 1.3.8A .53 closed subset of largest dimension. X) = codim (spec A/I. x g Since they have the same function elds. we can use (d). 94 with mR = (u). this is 1.dim X = dim A = ht m + dim A/m = ht mAm = dimOP 2. Spec A) = infp⊃I (Spec A/I. then e.3. b. 2. and Z ⊂ Y is an irreducible codim (Y. the nonempty open set Spec R [t]u ⊂ Spec R [t] dimensions are 1 and 2 respectively. If Y is reducible. x g Y is irreducible. (a) Now consider the maximal ideal (ut − 1) . Spec A) = infp⊃I ht (p) = infp∈Y dim Op. 0 dim Y + .X .A.21 x (e). 2.8. x dim X 0 = Dim (X ×k x0 ) = dim X + dim k = dim X 2. X) = dim X .50 .3. X) = dim Z + dim (Z. x codim (Y.55 since a eld has dimension II.54 f.3.52 d.51 c. 2. If 2.3.3.

X 6= dim X II.6. t (V ) ×k t (W ) .56 is a closed point where . If Y = V (P ) 2.3. then 0 + 1 6= 2. is factorial domain.3.3.57 II.22* (Starred) MISS MISS MISS MISS MISS 2.4. so dim OP < dim X dim Y + codim Y. so principal prime ideals have height 1. is separated.3.d. Hence P = (ut − 1) (d). .Then R [t] R [t] / (ut − 1) ≈ Q (R) as t = u−1 modulo (ut − 1). 95 .23 x By II.

Thus a variety.2 Let II. . is integral and nite type. Then Y must be V × W by the universal property Thus t.10. II.  X  f.→ v (V ) R a valuation ring =⇒ R integrally closed.4. consider f Spec A This corresponds to KO o i BO v ~ Ro ? f A u in terms of rings.4 II. Properness is local on the base and f : spec B → Spec A . 96 . is integral (Atiyah Mac p 60) Now the result follows by valuative crit of properness. f.4.4.4 x stopped g'ing here 2. g agree on U =⇒ im (U ) is contained in diagonal.2 x U dense open on / U  / Z  ∆/ U X where f.11 =⇒ closed immersions stable under base extension =⇒ Z → X is closed immersion. So this is an integral sepated scheme of nite type over an algebraically closed eld k.3. g agree. t (V ) ×k t (W ) = t (Y ) for a variety Y . 2. and so topolgocaily the pullback is U . If R is an arbitrary valuation v / Spec K Spec B 8  ? i / u Spec R  f X.4. Y is nite so take ring with quotient eld K ane. of 2.g Y Y ×S Y Y separated =⇒ ∆ is closed immersion.1 II.1 x g Nice example valuative crit Let f :X→Y nite. We have where the middle is pullback of δ .k t (V ) ×k t (W ) is algebraically closed so by 4. u (A) ⊂ R and R integrally closed =⇒ v (B) ⊂ R . A → B nite and B integral over A =⇒ u (A) .

Y be the ane line with two origins.g. algebrae). hence U ∩ V is ane by exc II. g Let If agree on the complement of the origin. (a)X nonreduced counterexample X = Y = Spec k [x.3 x g Consider / U ∩V  δ U ×S V . Hence hence Z|V = Z as schemes.4.4. and let U.4 II. g be the distinct open inclusions of the ane line which agree outside of the double origin.  / X ×S X X separated over S =⇒ δ X is a closed immersion by denition.11. V be the two distinct ane planes . and therefore ≈ Spec A/I . (b) Y is non-separated counterexample. by II. 97 Z→Y is proper so f (Z) is closed.3. =⇒ sp Z = sp X .4.3. into an ane scheme. . Let f. Z|V → X|V = V is a closed immersion.b.8. =⇒ Z|V is ane. the map on global sections disagree.4. By exc II.e. factors through Image is closed subset of which is 0 as A is X is reduced by assumption. U dense. As Spec A/I → Spec A is a homeomorphism. y] / (x2 . Z → X is a closed immersion =⇒ OX → OZ is surjective. 2. closed immersions are stable under base extension.=⇒ U → X Z. U ∩ V → U ×S V is a closed immersion. X. g agree on all of X so they are in fact the same. f. then f.11. then I is contained in the nilradical. xy) f is identity. U ∩V is A2 − {0} which is not ane. However. homeomorphic to V . but send Let the origin to dierent places. As U ×S V is ane (dened by a tensor of f. Hence then U ∩ V → U ×S V is a closed immersion Nonseparated Consider ane plane with two origins. and g maps x to 0. For X ⊃ V = Spec A open ane. =⇒ Z ≈ X .3 II.4. 2.4 x image is closed Note that as Z→S is proper and Y →S is separated. X .

4.e. separated Note the diagonal f (Z) ×S f (Z) Y → Y ×S Y is a closed immerson by separatedness of Y .4. Base extension of Z → f (Z) gives f 0 : T ×S Z → T ×S f (Z) . If x ∈ T ×S f (Z). x see part a. nite-type. so there is a point in T ×S Z mapping to x.4.6 Ox. Thus T ×S f (Z) is closed in T . k (x ) ⊂ k then we have morphisms Spec k → T ×S f (z) and spec k → Z which agree on f (Z) and thus lift to spec k → T ×S Z . by denition. Hence T ×S Z → T ×S f (Z) is surjective. Let 2. By base extension f (Z) → is a closed immersion. the inclusion 2. universally closed. the center at b. and universally closed. Surjectivity 00 00 0 00 of Z → f (Z) implies there is x ∈ Z with k (x ) ⊃ k (x ) .In order to show f (Z) is proper. T → S another morphism. mR lies over mx in Ox. If k (x) .X . we must. 2.  0 −1 If W ⊂ T ×S f (Z) is closed.X ⊂ R ⊂ K is unique.5 Let II.  Spec k Comparing with the valuative criterion of separatedness shows that the diagonal morphism is unique. T = Spec R. f 0 surjective implies   f 0 (f 0 )−1 (W ) = W so that s0 ◦ f 0 (f 0 )−1 W = s0 (W ).4. show separated. then (f ) W is closed and s0 ◦ f 0 (f 0 )−1 W is closed. i.e. Thus we have a diagram: / U  / T .X ⊂ R ⊂ K and K with center at x. i.7 (starred) MISS 98 x is unique. then under the base extension this corresponds to x0 ∈ f (Z) with k (x0 ) ⊂ k (x) . .X where U = Spec K .5 x g center is unique by valuative criterion R Then a valuation ring on Ox. nite type Note that closed subschemes of nite type schemes are nite type.

a ∈ /k. x a ∈ Γ (X. for / Spec K  Spec R / K ⊃ R ⊃ ϕ (B).O OX ) As in the hint. −1 Let R ⊂ K be a valuation ring dominating k [b ](b−1 ) .4. let Let b be the image of z Ro k Then Then b ∈ R so vR (b) > 0 . a ring morphism. and thus A is integral . k is algebraically closed =⇒ b is transcendental over k =⇒ k [b−1 ] is a polynomial −1 Consider k [b ](b−1 ) .4. ϕ : B → A be the map on global K = k (A) . By the valuative criterion of properness. R valuation ring as in 4. II. Ox ). But the valuation of a unit should be 2. we have a unique map Γ (X. we have. a local ring contained in K . b−1 ∈ mR so vR (b−1 ) > 0 . the integral closure of ϕ (B) in K is the intersection of all valuation rings of K containing ϕ (B). sections. a∈k.6 x g f : Spec A → Spec B nite. f is therefore nite. The map B.4. By the valuative criterion of properness.9 Let Let Let ring. 99 in every such valuation ring. OX ) → R : KO o Γ (X. 0 so this is a contradiction which resulted from assuming a∈ /k . mR ∩ k [b−1 ](b−1 ) = (b−1 ) =⇒ b−1 ∈ mR .5. Spec A 8  Spec B By II.8 d.4. As f Spec R → Spec A gives an inclusion of A over is nite type.11A.2.

Suppose we have 100 .4. Then we can glue π (Vij ) → π (Ui ) to get a morphism Y0 → X0 .4.12 being xed by the involution.4.10 II. If π : Y → Y0 is the projection . theorem 16. Y = Spec B . 2. x X0 ane implies X0 ×R C ≈ X is ane.35. σ σ If X = Spec A is ane. . this is ane by (b).2. −1 For each i tlet Vij an open ane cover of f Ui and preserved by σ . Y not ane. A 2. c. then X0 = Spec A . X. then we have an induced morphism A → B This gives X0 → Y0 . then take cover of X by open anes Ui preserved by σ . σ σ If X = Spec A .11 b.4. x f that commutes with σ .7 x R-scheme This follows from Milne AG.

14 e. X × A → Y × A has P . ∞) . By composition. But this is f × g . σ to atk for k ∈ Z .Z] 2 +V 2 +Z 2 ) ≈ P roj (XY +Z 2 ) ≈ PC ≈ X where σ acts by conjugating scalars. We want to show that f × g has P . x ∈ X ≈ P1 a closed point. and  X Y ZY 2  ≈ C [t−1 ] X + +( Y ) X (X ) Y X Y Z −1 −t = Y . −1 Let f send (x. −t = YZ . a) is xed.Y. B changing coordinates to s = ct + d with σs = s . Y × A → Y × B has P . As σ 2 = 1. 1 Writing U = (X + Y ) . If k = 1 . If on the other hand 2. let U = X\ {x. If a is positive.4. x By case II of (e) 2. then k ≈ ±1 . Assume x. V = 2i (Y − X) .8 x g : A → B two morphisms having P .15 σ .Z] 2 +V 2 +Z 2 ) . X × A → Y × A → Y × B has P .Z] 1 Q ≈ P roj (UC[X.Y.4. σx} . The element Thus t is sent under X0 ≈ Q0 ≈ P roj (UR[X. ] C[ X . Now change coordinates from t to −a Y Z C[ . Z ] Z −1 Writing t = . then σ would x C [t](t) so k must be −1. tσt = a is xed by σ . x We proceed by cases. −1 The lift of σ is C -semilinear and σ induces an invertible semilinaer C-algebra homomorphism on C [t. then for . Thus a ∈ R≤0 . t ] If σ has no closed xed points. then a ∈ R .13 d.4. σ restricts to a semilinear automorphism of the 1 complement of the xed point. we have a σ -invariant 1 isomorphism X ≈ PR ⊗R C .4. t ] . we have isomorphisms  Y X ZX 2  ≈ C [−t] . then II. then the ideal (t − 1 −1 √ . σ switches Z and Z and conjugates scalars. which is an open set Spec C [t] ⊂ PC . Also by base change. we get 2 C[X. σx) to (0.2. has at least one xed point. and σx are 0 and ∞ so that U ≈ Spec C [t. Since σ acts √ by conjugation. so the involution becomes t 7→ −t . Since σ is invertible. t gets sent to something of the form at + b.Y. By base change. Let f :X→Y and 101 .

so it has P .used stein factorization X→Y →Z projective. x Consider the ber product: Xred Γfred & id fred Yred /+ Xred ×Y Xred   / Yred  Y Xred → X → Y (top and righ) is a composition of a closed immersion and morphism with P . 2. then fred is a composition of morphisms with property P .16 e.18 Let .4.17 f. x Note this morphism rst in the vertical left hand side of the following diagram: / X   / X ×Z Y Y Y ×Z Y  & Y X &  Z Note top right has it by separated since closed immersion Note the top left is the base change since X ×X X ≈ X and X ×X (Y ×Z Y ) ≈ X ×Z Y .9 x g important . But the graph is the following base change Xred  Γ / Yred  / Yred ×Y Yred Xred ×Y Yred As Yred ×Y Yred = Yred and ∆ = id 2.4. . We have 102 Γ thus has property P. has P .4. then ∆ is closed immersion and II.4.2. Thus Yred ×Y Xred is a base change of morphism with P so has it by assumption. the graph. Note if Γfred . LHS bottom is also base change.

4.4.4.4.4.22 d. P × P × Z r s rs+r+s P ×P ×Z →P →Z factors like ×Z →Z Since segre embedding is closed immersion. part of starred 103 X → Prs+r+s which factors as . g0 and id × g 0 are closed immersions.20 b part of starred 2.19 II. r s Now using segre embedding. then have closed immersion g◦f .X f0 / Pr × Y #  Y id×g 0 r / P × Ps × Z g0 /  Ps × Z g '  Z f 0 . 2. part of starred 2.10 Chow's Lemma (starred) 2.21 c.

.6. . yn ] L over K such that L is nite over K (y1 .. . x By part (a). xn are algebraically independent over K . we only have to consider discrete valuation rings. and so a0 ∈ (t) which must be Thus. x1 .12 x Examples of Valuation Rings R⊂K t ∈ mR mR is principal.. If ... O0 p.4. yn Let transcendental elements of Thus extension of m in O [y1 .4.4. If B (x1 ).11 x y1 . x1 . yn ) . Thus this follows from Thm I.2A the valuation ring s ∈ Pn mR \i (t) .. . K has dimension 1. b. P i−1 ai t i=0 ai t = 0 with a constant term. the m. Localizing the integral closure at a maximal ideal gives a DVR in L dominating O ..2. . Thus ai s = 0 .4. dimension 1. 2..4.. and (t) 6= mR . let is not transcendental then K . since As 104 is a nite algebraic extension of k (t) .1A 2. then s is algebraic over k .. p Let a minimal prime lying over By KPIT.. which is not the whole ring. . xn ](x1 ) is (x1 ).. is not the whole ring so we localize at a prime ideal lying over L is nite eld extension of K . xn a system of parameters for m . then B is a noetherian local domain of dimension By DIRP. K P i s∈ / (t) . . .25 Let For If t II. then a0 = t be a valuation ring. p has height 1...6. Contradiction.24 is the localization of at 1.. and t is transcendental.. then by thm I. Assume WLOG that Let As extension of m in O [x2 . the integral closure of B in L is noetherian.23 II... is discrete.

27 X0 Then (2) x R is DVR. (1) x x1 corresponds to a prime ideal p ⊂ A of height 1 and OX. f induces an inclusion OX. which must have center x1 . By DIRP.→ R . a0 ∈ mR . If 2. a noetherian local ring of dimension 1. As t ∈ mR . Ap is a DVR.4. and a curve so normal then so is A. so R Recall this means R has center x0 . If (s) = mR the we are done. then a0 = 0 or else. Thus t ∈ (s) . then by (1).x . b.x1 ≈ Ap . other wise take 2.28 is smooth.4.4. dominates OX.26 P ai si−1 so f (t) = g (t) s P ai si−1 and thus f (t) ∈ (s) ⊂ an ascending chain and use noetherianness.Thus a0 = s P ai si−1 . . 2. Ap .x0 (3) x This is clear. 105 . Let U = Spec A be an open ane. As X is nonsingular. and a0 = f (t) f (t) so g(t) g(t) =s mR \ (t).

P ) ≈ Hom (M. lefts with colimits (nite direct sums). 2.4 Hom (M ⊗ N. AM p 28 (d) x g Projection Formula We havef∗ n (F ⊗OX f ∗ OYn ) ≈ f∗ (F ⊗OX OX ) ≈ f∗ (F ⊗OX OX )n . ≈ f∗ (F )n ≈ f∗ (F ) ⊗ OYn ≈ f∗ (F ) ⊗OY E .3 (c) x g We take the sheacation of 2. the total space and the generic point.5. OX ) ⊗OX F ] |U by mapping P f 7→ ni=1 e∗i ⊗ f (ei ) . is open and V ⊂ U .2. P )).5. Dene ψ : H omOX (E . ϕ : H omOX E |U . OX (V )) → OX (V ) by evaluation .5. 106 E |Ui is free. F |U ) by f ⊗ a 7→ (x 7→ f (x) a) .5.2 is locally free.2 (a) x Spec R has two nontrivial open subsets.5 II.5. procedd in the same manner on an open cover where the results.5. and then glue . E 2. (b) g x E is gree. Now if E is locally free. Let U be an open set where Dene mapping Now check that these are inverse bijective homomorphisms.1 II.5 x 2.1 g x Let If at s|V If U ϕ : E → H OX (E . then for s ∈ E (U ) dene t ∈ homOX (E (V ) . H om (N. then this is an isomorphism on the stalks. OX ) be dened by evaluation. since right adjoints commute with limits. OX|U ⊗OX (U ) F (U ) → H omOX (E |U .5 II. 2.5. F ) |U → [H om (E .

2. F ) . and the associated sheaf is uniquely isomorphic as a sheaf.4 x First note that locally free implies q. is q. if and a restriction morphism 2. Glueing gives an f = f on X . LR being L considered as an R-module. which is M → Γ (X. so the R-module homomorphism represented by restriction gives the K -module homomorphism of extension M ⊗R K → L by adjunction. given f : M → Γ (X. F ) . Folliwing Eisenbud. and since OX is itself a sheaf. F is injective. . ˜ globally. dene f on the distingushed base. By denition. then clearly . K Basically if we have locally free. M   m ] ] On the other hand.c. ˜ i L ≈ M0 .8 ˜ →F f :M f] . then we know F |Ui ≈ (OX ) some index set K .5.  K ˜K By II.7 . if ρ : M ⊗R K → L is an isomorphism.5.F is an OX (X) = R-module M and there is a OX (U ) = K -module M → LR . we know F ≈ M Thus 2. II. But L ≈ M ⊗R K ≈ M(0) By (a). Let mαα∈A B index the kernel so that 107 ψ γ B A OX|U → OX|U →M →0 is exact. On the other hand. is q.5. Thus f 7→ f ∈ Hom M g Clearly f induces 2. page 17. ˜X OX ≈ O Now suppose F . (This is thm U = Spec A. L F is an OX module. then locally ˜ F ≈M . If f ] induces f .5. OX ≈ O˜X . Restriction and extension of scalars are adjoint.3 x g ˜ (X) → F (X) .6 (b) x If F R has a unique closed point and the neighborhood is the whole space.c. F ≈M Thus L ≈ F (U ) ≈ M(0) ≈ M ⊗R K . so that gives a map on global sections f 7→ f ] is surjective. M ).5.c. Let U II. such that F |U ≈ M a generating set for the OX|U -module M (we can at least take a neighborhood of a point. . D (f ) by f |D(f ) 7→ g   f (m) ] ] ˜ . II.5.5. let the mα to be the elements of ˜ .4).

5.c. then it is a cokernel U .2. Conversely.5.g.5 (a) x g f : X = Spec k (t) → Y = Spec k . Note that OX|U and OX|U are free by above. 1. Ai an f. . . II.c. the same sequence as above.11 Spec (Ai /Ii ) → Spec (Ai ) and each Ai /Ii is f.6 (a) x g p ∈ V (Ann m) =⇒ p ⊃ Ann m =⇒ sm 6= 0 p ∈ Supp m =⇒ sm = 0 some for s∈ / p =⇒ mp 6= 0 =⇒ p ∈ Supp m s∈ / p =⇒ p 6⊃ Ann m =⇒ p ∈ / V (Ann m) 108 . so F |U is q. Note that OSpec k(t) is coherent on k (t) .5. if of locally free sheaves on Now using 2.g. these are 2.5.11.. 2. .5. Bi -module. we have the result in one direction. the cokernel of q. II. (by denition of closed immersion). F coherent..9 II.c.g Bi -module. f∗ F |Spec Bi ≈ (Bi Mi ) .c.7. t2 . U = Spec A is the cokernel of a morphism of free sheaves on U ..a. As F |U . (c) x g {Spec Bi } an open ane cover of Y . and X noetherian =⇒ F ≈ M ∼ By II. the restrictions are closed immersions f : Ui → Ui .c.5. x g closed immersion is nite.12 II. and Bi Mi is an f.g. ˜ i an f.2. As exact functors preserve cokernels. For an open ane cover Ui of X. f :Y →X closed immersion. sheaves on U .3. f nite implies f −1 Spec Bi ≈ Spec Ai . Ai -module. t. ˜ ˜ A B is exact by II. thus it is a cokernel of q. Let 2.b.B A A A /OX|U /ker (γ) ≈ OX|U /im (ψ) ≈ OX|U M ≈ coker (ψ) ≈ OX|U .5. By Ex.5.d.5. is q. We know that locally free implies q. since it contains Thus the pushforward is not coherent. sheaed. Let but k (t) is not nitely generated as a k -module.10 Let b.

5.5. Supp F ≈ Supp M Then Mp 6= 0 i mi 6= 0 in Mp for some i i Ann (mi ) ⊂ p Let 2. 0 0 As HZ (F ) is kernel of q.→ X the inclusion. Note m ∈ ΓZ (F ) i Supp m ⊂ V (a). and thus on X.c gives j∗ F is q. if 0 HZ (F ) |Ui ≈ Γai (Mi )∼ . √ √ a ⊂ Ann m. ˜ ≈ {p ∈ Spec A|Mp 6= 0} . ∼ 0 Γa (M ) ≈ HZ (F ) i Γa (M ) ≈ ΓZ (F ) .c. ˜i F |Ui ≈ M for Ui = spec A. By nullstellants equivaltn to n By noetherian equivalent to a ⊂ Ann m . j : U . then by (b). Ann M = ∩Ann mi . the support is closed on Ui . this is equivalent to V (Ann m) ⊂ V (a).8.5.15 Ui with ˜i F |Ui ≈ M .5. From a previous excercise.16 (e) x Since X is noetherian.14 i V (Ann M ) 3 p (c) x g The support is the union of supports of sheaf of each element.5. (d) x U = X − Z . then by (d)..20b gives 0 → HZ (F ) → F → j∗ F .c. By denition equivalent to m ∈ Γa (m) . HZ (F ) is q. Let 2. sheaves. Z = Spec A/ai 109 an open ane cover.1.2.c. c Let U = V (a) . 0 exc II. Thm I. .13 (b) x g mi a set of generators for M . On an an open ane cover 2.

F = M We have Fx ≈ Mp ≈ Ap x1 + . Gx ⊗ k (x) are dimension 1 . write the image of mi as g i. M is generated by m1 . OX ) ⊗ F ≈ OX via the evaluation morphism is surjective. suppose For 110 . As Fx ⊗ k (x) is a one-dimensional vector space.19 If F (c) x is invertible. Thus Fx ⊗OX. ai.1 ai..x Gx ⊗OX. ˜ h is a nite direct xi linearly independent in Mp =⇒ xi linearly independent in Mh =⇒ F |D(h) ≈ M Let sum. since F is it is an isomorphism.2.. X = Spec A . + Ap xn . + gi.j . 2... Conversely. N are f.n x1 + .g.5. . and an inverse by ms 7→ ms ⊗ n 7→ as . use part (a) 2. so from a set of generators of Fx ⊗ k (x) ≈ Mp ⊗ k (x) we obtain a set of generators for Mp via nakayama. then locally free rank 1. G coherent =⇒ M.5.5.. p ∈ Spec A .5.18 If F b.   x ∈ X .. + Ah xn . and G ≈ N ˜ .n xn . then by denition stalks are free..x . Let U be a set where F |U ≈ M F . In Mp . and Np is generated by n. and xi sections on a principal open set D (f ) . If h = f g .x k (x) ⊗k(x) Gx ⊗OX. then Mh = Ah x1 + . x is locally free. If the stalks are all free. Dene f : Ap → Mp by s This morphism gives Fx ≈ OX. p ≈ mx . a 7→ as m .1 Q Writing g = i. then each point has a neigborhood .. then Fx ⊗OX. F ⊗G ≈O X .17 II..j gi.x k (x) ≈ Fx ⊗OX. H om (F .x k (x) ≈ k (x) . we see that mi are spanned by xi on D (f g) .. Mp is generated by m ∈ M as an Ap -module.7 x ˜ . ˜ .x k (x). As Mp ⊗ Np is therefore generated by m ⊗ n . mn .

8 x Using properties of the induced topology. and since mj generate ni . In Mp . p ∈ Spec A = X . by Nakayama's lemma. and thus Ms is free. mi for Mp . and if s = s . P aij Q Then we can write a nite set of generators ni for M in Mp as ni = mj . a set of n generators. some n such that φ (n) is empty. ni = j ij sij By denition of localization s ∈ / p.5. and choose. by assumption. as F is assumed coherent. thus the whole space or empty since Pick 2. c Let x ∈ Φ (n) so that m = dimk(p) Mp /pMp < n . and p correspond to x. and thus so is Mq /qMq .21 If (b) x F is locally free then every point has a neighborhood where ϕ is constant. is the sum of elements of X n 111 for all such which are therefore in the . But y ∈ ker ϕ As . Now for one larger n. For s = sij . a nitely generated A module.5.5. Let ni a generating set for M . x ∈ U = Spec A . q ∈ D (s). and M = Γ (X. we sij Let have a short exact sequence ϕ 0 → ker ϕ → A⊕n s → Ms → 0 which also holds on Aq . F ) Mp is generated by less than n elements mi ∈ M as well. open. p correpsonds to x. is an open neighborhood of p.. so that p ∈ D (s). By Nakayama's lemma. we only need to check that Φ (n) = {x ∈ X : ϕ (x) ≥ n} is closed on open anes of a cover. so that D (s) ⊂ Φ (n) as q was arbitrary. c c c Thus q ∈ Φ (n) . y is reduced as Mq ⊗ k (q) has dimension is reduced so . then mj generate Mq . but also closed by (a). P a0ij If q ∈ D (s) then s ∈ / q =⇒ s invertible in Aq =⇒ ni = mj there. ker ϕ = 0. then sni = aij mj for some a0ij . Let n = dim Mp ⊗ k (p) . Let M correspond to F |Spec A . φ (n) = {x : ϕ (x) ≥ n} is a union of open sets and therefore X is connected. 2.20 II. ϕ constant q ∈ D (s). Then open. Thus Φ (n) is .22 (c) x x ∈ X .. so k (q) ⊗ ker ϕ = 0 qAs for all q ∈ D (s) . implies each Thus for any nilradical.2.5. Thus Mq is generated by < n elements. s ni generate Mq . Q P 0 P aij m .

15.25 (c) x By (b). As the composite of α ˜ with  is the identity. M(f ) ≈ M(f ) . Thus α gives a morphism of graded modules. 112 . Note that F quasi-coherent implies the (EGA) Note that is a quasi-nitely generated graded counit is an isomorphism by thm II.5.   ˜ is s ∈ Se . Thus we want to show that for q quasi-nitely generated graded for coherent sheaf F . a morphism of quasi-nitely generated modules is a quasi-isomorphism i the morphism on associated modules is an isomorphism. then sα (m) ∈ Γ∗ M Γ (X.5. On such a cover M Suppose rst that is locally equivalent to a coherent sheaf. since n = n+d . and is quasi-nitely generated. Note further that as the twisting functor is exact. so we obtain  ˜ α : M → Γ∗ M  D+ (f ) . Γ∗ (F ) S -module M .23 II.5. M (d) S -module. F M is equivalent to   ˜ Γ∗ M if M is nitely generated. 2. suppose q-f. S -module with M≥d ≈ M>d for some d. M (n)∼ ) . If 2. where ν : F (0) → F is the canonical isomorphism. Let M 0 an f. M (d)∼ ⊗ OX (e)) ≈ (b) x   ˜ Γ X.15. M nitely generated implies M(f ) is nitely generated.5. Γ∗ (F )∼ is isomorphic to F for quasicoherent. m ∈ Md . which is a homomorphism of groups. Dene αn : Mn →   ˜ and Γ∗ (F )∼ → F are adjoint functors with counit Γ (X.g. d mf m 0 0 0 Then for any f ∈ S1 . f f ˜ As S is generated by S1 as an S0 -algebra.5. M (n)∼ ) by m 7→ m/1 . then M(f ) cover X = P roj S for various f . then ˜ M is coherent.generated 0 M is quasi-nitely generated.5. . and thus denes a section on each These sections agree on intersections and give a global section.24 dened as the image of m⊗s in Γ (X. M (d + e)∼ ) . By II.9 x m ∈ Md has degree zero in M (d)(f ) = Γ (D+ (f ) . then α ˜ is the unit. Note that M → Γ∗ M    : Γ∗ (F ) → F given by D+ (s) ms 7→ ν 1s · m|D+ (s) .2.

IY (d + N ) ⊗ O (−N )) ≈ Γ (Ui . ˜ 0 . .28 Let I2 ⊂ I 1 . then by (b) for large enough d.10 x is clearly an ideal. If 2. then d is a section of I (D+ (xi )) . IY (d)) 113 with image s. IY (d + N )) . + sk with each si homogeneous of degree xi homogeneous of degree 1 =⇒ xi sk is homogeneous I homogeneous ideal and xni s ∈ I =⇒ xni sk ∈ I . n+k .26 I II. 2. there is ti ∈ I2 . Md ≈ Γ (X. then M 0 By exactness of twist we get M 0˜(n) .17. Since ti ∈ I2 .suppose coherent F is coherent.c. I2 = I 1 . By thm II. then for each i . sheaf of ideals I on X .. −n N As xi s ∈ Γ (X.5.5.g. F (n) is generated by a nite number of global sections for large enough ˜ ) ≈ F via th II. IY (d)) . n By denition. s If s ∈ I1 is homogeneous of degree d .5. (b) x s xdi X.5. ˜0 ≈ F . Tensoring with O (−n) gives M 0 As M is f. =⇒ I homogeneous. xi As I1 and I2 dene the same ideal sheaf. =⇒ sk ∈ I . Write I 3 s = s0 + . We claim that s|Ui is in Γ (Ui .→ Γ∗ (F n .27 i of degree . F (d)) which shows quasi-nite generation. thus s is in the saturation of I2 . bar stands for saturation.5. by thm II. homogeneous of degree d with = xtid .9. If M 0 is the submodule of Γ∗ (F ) generated by these sections. so is xni i ti = xni i s. Suppose that I1 and I2 dene the same closed subscheme of i hence 2. (c) x s ∈ Γ∗ (IY ) .. Since the operation of saturation is idempotent.5.5. which implies xni i (s − ti ) = 0 for some ni . for each i there is n with xi s Choose N larger than all such n.→ F (n) which is an isomorphism since F (n) is gbgs in M . xi ⊗ xni s ∈ Γ∗ (IY ) ∈ Γ (Ui . they dene the same q.15.

2. Thus P roj S ×A T ≈ X ×A Y .. ..9 to closed subschemes of X. id×j and a closed immersion ψ×id X → Pn1 × Y → Pn1 × Pn2 × Z → PN × Z .5. Then αi ⊗ βj become S1 ⊗A T1 and S ×A T = A [αi ⊗ βj ] .32 (b) x g By assumption. Quasi-coherent sheaves of ideals correspond.5. 0 ≤ j ≤ s D+ (αi ⊗ βj ) ≈ Spec S(αi ) ×A Spec T(βj ) ≈ D+ (αi ) × D+ (βj ) . Consider the composition 114 i   n OPY1 (1) ≈ L .12 x g Nakai moisheazon.31 II.29 (d) x Homogeneous ideals of S correspond to q. αr and β0 . via taking the ideal and Prop II. Or take the segre product of the two closed embeddings..5.5..2. βs be the generators of the A-modules S and T . II. 2.5.. sheaves of ideals via Γ∗ (−) and sheacation. there is a closed immersion j : Y → PnZ2 with j ∗ OPnZ2 (1) ≈ M i:X→ PnY1 with i∗ . Let generators of .5. then The second property follows from the fact that II.12. then by (c) we get a bijection as 2. .11 x g α0 . respectively.5. As there is a unique saturated homogeneous ideal in the preimage of each sheacation by (b).c.30 Γ∗ is the inverse. and the universal property of the cartesian product.5. As S ×A T(αi ⊗βj ) ≈ S(αi ) ⊗A T(βj ) for all 0 ≤ i ≤ r . 1-1..c.

(d) Spec S(f ) ≈ Spec S(f ) .19. then X is reduced.X are integral domains. S is contained in the integral closure of S . As 115 . As 0 0 in thm II. o s If L is the sheaf ⊕n≥0 OX (n) . then Lp = ⊕n≥0 S (n)(p) = ∈ S|deg s ≥ deg f .with N = n1 n2 + n1 + n2 and ψ the segre embedding. 0 Taking global sections is left exact so Γ (X. The identity map fsn → 7 identies S(f ) fn As . Thus the s f ∈ Sd cover both P roj S and P roj S (d) .34 gives P roj S ≈ P roj S (d) . OX (n)) = S . On the other hand only elements which have positive degree can be integral over Lp .5.5. and glueing that O (1) and OX (d) correspond.33 II.14 x Ox. We have   φ∗ OPNZ (1) ≈ L ⊗ f ∗ M which is what we wanted to show.13 x (d) S is generated by S1 over S0 . 2. then S (d) is generated by S1 = Sd over S0 .5. 2. L ) ≈ ⊕n≥0 Γ (X. and S n is a domain. Since X is by assumption normal. which is integrally closed. X is irreducible. so S = S . Then Lp is integrally f closed since an elemeent integral over Lp is integral over Sp which is integrally closed since X is normal.5.5. The same maps give S open sets D+ (f ) (d) and S(f ) so that (d)(f ) ≈ S (d) (1)(f ) so II.

.5. OPr (n) by (a) and this A If Γ (Pr . Nα∼ ≈ i∗ F 0 . By (a) noetherian.5. OX (n)) for all n by (a). OX (n)).2. the integral closure of S (d) . II. d  0 . i∗ F = S Gα . =⇒ Sn = Γ (X. sheaf is a module.. =⇒ S ≈ S 0 .8. i∗ F is q.5. (b) x Using II. (c) x 0 by (b).g. the integral closure.39 . . this union has a maximal F 0 |U ≈ F . since 2. Gα ≈ Nα∼ S element. Snd = Snd (d) (d) is integral over S s∈K S For If .36 S˜ ≈ OX .5.c. xr ] then Tn = Γ PA .35 (b) x By 5.15 x Extension of Coherent Sheaves is ane..  r If T = A [x0 . module. OPr (n))  Γ (X. then a q. it lies in S 0(d) .c. and 116 is a coherent subsheaf of i∗ F . Hence the homogeneous coordinate ring is integrally closed.37 (d) x X projectively normal =⇒ S integrally closed by denition. 2.5. OX (n)) 2. Since X is 0 But then F is a coherent subsheaf of i∗ F .38 If X S = S0 when S Γ (X.c. then surjectcs onto A-module is a union of its nitely generated submodules. Note that an 2.9b. and a coherent sheaf is an f. is normal by (a).5.5.

c. (d) x U1 . Un be an open ane cover of X . Using (b). we apply d to This gives for each open U Now take the union of the and G. As ρ−1 (i∗ F ) |U ≈ F .5. (c). and glue F and F on the open set U1 ∩ U to get a coherent sheaf F on U ∩ U1 ..40 (c) x −1 Consider ρ (i∗ F ) 0 we nd F |U ≈ F . s ∈ F (U ) a coherent subsheaf G . x is a section of F F |U generated by s. 2. 117 of F where s ∈ G (U ) . . Let 2.5. then as in (b).. And repeat.5. . Now do the same thing 0 0 but with X = Ui and U = Ui ∩ (U ∪ U1 ).2. extend F |U1 ∩U to a coherent sheaf F 0 ⊂ G |U1 0 ..42 If s e.. thus is q.41 ⊂G which is the pullback of q.c.

ω ⊗ λ 7→ ω ∧ λ gives an isomorphism OX → Λn F . 2. ∧ en )...5.5. λ F ) → Γ (X.. .. r r r Then the presheaves U 7→ T F (U ) .. r n On the other hand given a morphism of OX -modules Λ F → Λ F ≈ OX . ∧ eir |0 < i1 < . .43 II. Λ F ) ≈ Γ (X. U 7→ S F (U ). and on the cover U .. en be basis elements.. {ei1 · · · eir |1 ≤ i1 ≤ . where jk are elements of {1. ∧ eir ) ej1 ∧ . each of the above presheaves is free with the rank of the basis. we can the ei are basis global sections. n} \ {il } . n−r If λ is a global section of Λ F . then λ denes a morphism Λr F → Λn F ≈ OX by ω 7→ ω ∧ λ . U 7→ Λ F (U ) are free with basis global sections {ei1 ⊗ · · · ⊗ eir |1 ≤ i1 . g Tensor Operations on Sheaves F locally free of rank n implies there are e1 . f 7→ f (e1 ∧ .... OX ) . . The pairing dened by 118 on by .5.2. and {ei1 ∧ ...... Note that free presheaves are sheaves.. en such that on some open cover {U } . ∧ ejn−r . < ir < n} respectively... Then a global section of Λ F is dened P ki (−1) ϕ (ei1 ∧ .. ∗ r n−r The operations dened in the two preceeding paragraphs are inverses so Λ F ≈ (Λ F ) ⊗ Λn F ... i. ir ≤ n} .. . ≤ ir ≤ n} .44 Let (b) xc g e1 . we have a morphism r n n−r global sections ϕ : Γ (X. Each basis can be calculated to have the required dimension.e. F (U ) ≈ OX (U )e1 |U ⊕ · · · ⊕ OX (U ) en |U ..16 xc a..

and assume that we have chosen.5.47 For (e) xc n=0 this is clear. n ∗ ∗ T f F ≈ f F ⊗OX T n−1 f ∗ F by denition. ∗ n ∗ ∗ n By above. 0 00 Suppose that xi are a basis of F |U and yi are a basis for F |U . I is the subsheaf dened 0 → I → T nF → S nF → 0 .46 (d) x As in (c). ∗ n −1 Which is f T F as colimits (f )commute with left adjoints (⊗ Note that ). F are free. Thus tensor algebra commutes with pullback. 0 00 j r 0 If yi + ci . r j−1 Thus the lift of a basis for S F |U /F is independent of splitting.5. Suppose by Pullbacks are left adjoint thus right exact. F i such that F i /F i+1 ≈ S i F 0 |U ⊗ r−i 00 S F |U . F . i ≥ j . −1 That is f F ⊗f −1 OY f ∗ T n−1 F by rules of tensor.45 (c) cx 0 00 be an open set where the sheaves F .5. 2. Dido Λ. and tensor commutes with pullback so we get 0 → f ∗ I → f ∗ T n F → f ∗ S n F → 0. this is 0 → f I → T f F → f S F → 0. then xi ⊗ (yj + cj ) 7→ xi yj + xi cj ≡ xi yj mod F = S F |U .2. and so we dene F to be such a lift. 119 . r+1 Set F = 0 as in the hypothesis. 0 00 Note that F |U ≈ F |U ⊕ F |U implies that S r F |U ≈ ⊕ri=0 (S i F 0 |U ⊗ S r−i F 00 |U ) . n ∗ ∗ n But then S f F ≈ f S F . ci ∈ F |U are another basis for F |U . −1 That is f F ⊗f −1 OY (f −1 F ⊗n−1 ) ⊗f −1 OY OX by induction. Let U 2. −1 That is (f F ⊗f −1 OY OX ) ⊗OX f ∗ T n−1 F by denition.

−1 By denition. so there is a nite subcover of {D (fi )} . −1 Then f |f −1 Vi : f Vi → Vi is induced by φ : Ai → Bi and f −1 |f −1 Vi D (fij ) = D (φfij ) is also ane. 2.48 II. thus the preimage f U = U ×X Y with U ane is Given another open ane subset ane. Thus f |f −1 Vi is ane. restricting −1 to an open ane cover of X gives that Xgi = f D (fi ). Thus f Y.5.5. then V ∩ Vi is covered by −1 Let Ai = Γ (Vi .17 x Ane Morphisms Suppose that f :X→Y is ane. Suppose By (a).5. Furthermore. so ai f ] (fi ) = 1 ∈ Γ (X. it is quasi-compact.2. and so by thm II. V ⊂ Y . each restriction diagonal cover of X → X ×Y X f −1 (Vi ) → Vi is ane. OX ) denote the rings of sections on these ane sets. As open immersions are perserved by base change. Thus i=1 ai fi = 1 and P f −1 D (fi ) are open anes. Thus we have reduced to the case where Y = Spec B is ane and f is ane. Let Vi an open ane f −1 Vi is ane. these restrictions are separated.1. and thus quasi-compact.5. The open sets D (fij ) for a cover of V with ane preimages. where f ] : OY → f∗ OX . Thus we have shown that X is ane. Let {V }i Y such that f −1 Vi is ane for all i D (dij ) which are distinguished on Vi . Bi = Γ (f Vi . By 4.49 (b) x g f : X → Y is ane. and is thus separated. there is an open cover Spec Bi of Y where the preimages f Spec Bi are ane subschemes of X .50 (c) x g (used for stein factorization) We will dene X as the Spec A (U ) glued together. Note by denition that a nite morphism is proper and ane. 2. Pn As Y is ane. an open ane cover of . then morphisms −1 between ane schemes are preserved under base change. is quasi-compact. OX ).1. 120 .4. The factors through the restrictions. Let {D (fi )} be a renement to distinguished open anes with ane preimages. OY ) .

and h : Spec A (V ) → V be the induced morphisms. V . open set U ⊂Y . then U ∩V is covered by open sets that are distinguished in both V. so A is an OY -module which we claim is quasi-coherent as an OY -algebra.If U = Spec A . cover Y with open anes Ui . −1 If U = Spec A ⊂ Y is ane.5. Then X ≈ Spec A by (c).c. Since A is an OY -module. f −1 (U ) ≈ Spec A (U ). then e. then by (a) f (U ) = Spec B is an ane. If V ⊂ U 2. Now we wish to dene f : X → Y . then there is a restriction ρU W : A (U ) → A (W ) . Thus there is a morphism on U. But then glueing 0 morphisms on Spec A (U ) gives an isomorphism X → X . f is ane since for each open ane On the other hand. then f∗ OX (U ) = OX (f (U )) . 2. V ⊂ Y Let as in the hint. We glue the maps A (U ) → U on all open anes since they are −1 compatible on the overlaps.52 . If U ⊂ V ⊂ Y . then we can cover U ∩ V by open sets Wi = Spec Ci distinguished in both U and V . V = Spec B have nonempty intersection. Glueing together gives X . U. then f (V ) → f −1 (U ) is induced from the ring map A (U ) → A (V ) . then this is clearly A (U ). and patching gives an isomorphism Thus A ≈ f∗ OX since it is true on any open set. x U ⊂ Y . where B is an A-module. −1 If U ⊂ Y is open. suppose there is another such scheme X .5. A (W ) is a localization of A (U ) and A (V ). If V ⊂ U is open and ane. then . OX (f −1 (U )) → OX (f −1 (V )) is an OX (U )-module homomorphism. =⇒ C is a localization of A and B . −1 Then OX (f (U ∩ Ui )) ≈ A (U ∩ Ui ). 121 U and . Otherwise. As A |U ≈ B −1 We next claim that X ≈ Spec A . then A is a quasi-coherent sheaf of OY -algebras. If are open and ane .51 d. then f (U ) → f −1 (V ) comes from the restriction morphism ρV U as above. A (V ) along A (W ) . Wi being distinguished in U. if OX (f −1 (U )) f :X →Y A = f∗ OX satises for any OY (U ) → OX (f −1 (U )) so A (U ) has is ane. If U is ane or contained in an open ane. with a morphism f : X → Y . x M q. The isomorphisms given by the distinguished covering of U ∩ V glue together to give an isomorphism −1 g (U ∩ V ) ≈ h−1 (U ∩ V ) and agree on triple overlaps since the restriction maps on a sheaf do. Let g : Spec A (U ) → U . ˜ . Thus we identify A (U ) . A (U ) = OY (U )- the structure of an module. 0 0 0 For uniqueness. and by quasi-coherence.

2. we can glue OX -module M ∼ . −1 the OX (f (U ))-modules M (U )∼ U ∩V gives an and By exc II.5. On the other hand. Let II.18 x Vector Bundles 122 F on an open ane cover. II.22.1.5. OX (f∗ F )∼ ≈ F .The correspondence of sections between elements of localized modules on the intersections isomorphism between M (U ) M (V ) on U ∩ V .5.5. we claim that categories. and so by thm .c.53 Then f∗ M˜ ≈ M ∼ and f∗ (f∗ F )∼ are inverse to each other hand thus give an equivalence of is isomorphic to . F a q. -module. as U ranges over all open anes of Y to get an Following the hint.

then s|V is a section on V so there is a map E (U ) → E (V ) satisfying the requirements for the restriction morphism of a sheaf. then composing with a sections gives a morphism of locally free sheaf which is compatible with restrictions. Conversely.For this excercise I will give the natural 1-1 correspondence between vector bundles and locally-free sheaves over an algebraically closed eld. On an open covering Ui Thus we have trivializations On Ui ∩ Uj where E (Ui ) ≈ (OY (Ui ))r ϕi the transition functions of . Dene a sheaf E by letting E (U ) be the set of sections of E over U. t) |P ∈ Y. We also have a restriction map since if U ⊂ V and s ∈ E (U ).. then EP /mP EP ≈ Ar since k is algebraically closed..5. (OY (Ui ∩ Uj ))r ≈ E (Ui ∩ Uj ) ≈ E (Ui ∩ Uj ) ≈ (OY (Ui ∩ Uj ))r are given by matrices and reducing modulo the maximal ideal gives an element of the general linear group so that −1 we have transition functions ϕj ϕi = (id. φij ) where φij are invertible matrices. where mP is the maximal ideal of the local ring OP . Write a section as s = fi xi where xi r th are the coordinate sections xi : Ui → Ui × A dende by xi : p 7→ (p. Assume E is a vector bundle.. x see part a 2. t ∈ EP /mP EP } . (0. If g is a map of vector bundles...56 d. .5. fr ) gives an isomorphism so that E is locally free. Dene a projection π : E → Y by projecting to the rst coordinate. Dene E to be {(P. 0)) with 1 in the i place.54 b.. x see part a 123 .55 c. P −1 Now let Ui be a covering on which E satises π (Ui ) ≈ Ui × Ar . x see part a 2.. 2.. . . If f is a mpa of locally free sheaves then reduce modulo P on the ber to get a map of bundles. r Then E (Ui ) → (O (Ui )) dened by s → (f1 . This gives a module structure on the ber by adding together sections or multiplying them with functions. 1. suppose that E is a locally free sheaf.5..

Note that j is split by the 1 map k : cl (X) → cl (X × P ) sending ni Zi to ni π1−1 Zi . We claim that i is also split. then Z is π2−1 0. and vY (f ) = 0 for every other prime divisor Y.g (t) .6. π2 : X × P1 → P1 .2 then changing coordinates II. Y a copy of X embedded at the origin or innity which is opposite to Z . If g or h has positive degree. X × P is noetherian integral and separated and further X × A is regular in codimension 1 1 n 1. h(t) g. then X × P satises (*). nZ ∼ 0 exactness. X . As two X × A cover X × P . If nZ ∼ 0 for 1 On X × A . where i : n 7→ nZ . Thus the local ring of Z in K (t) is K [t](t) .6 II. we get vY (f ) = −n .6. 1 As both i and j are split.6 x Divisors 2.6.6. then Z is embedded at the origin as =⇒ ∃f ∈ K (t) with vZ (f ) = n and t . Thus f= tn g(t) . 2.6.6.2.2 (starred) Varieties in Projective Space MISS MISS MISS 124 . If the degrees of g. Thus there 2 is no rational function (f ) = nZ so i is injective. and Z corresponds to π2−1 ∞ ⊂ X × P1 andPj : Cl (XP × P1 ) → Cl (X × A1 ) ≈ cl (X).1 x g 1 1 By thm II.5.1 II.6. We claim that i is injective. j i 1 Now consider the exact sequence Z → Cl (X × P ) → Cl (X) → 0 of thm II. h are 0. h ∈ K [t] t 7→ t−1 . then it has an −2 1 irreducible factor in K [t] corresponding to a prime divisor p2 x for x ∈ P with f vp−1 x (f ) 6= 0 . The map k sends ξ → 7 ξ − kjξ in the kernel of j so in the image of i by n ∈ Z . h (t) . Cl (X × P ) ≈ Cl (X) × Z .

MISS II.3 Cones (starred) MISS MISS MISS MISS 125 .6.

. then h f ∈ /B since f is square-free so that h ∈ B and thus α ∈ A. h f ∈ B ... xn ] i the coecients 2g. K = f rac A we have g+zh g−zh −f h2 Let B = k [x1 . Which 2 happens i 2g.6.6. α ∈ A then 2g. then g ∈ B and thus h f ∈ B . 2 2 If α is integral over B . Thus K = L [z] / (z 2 − f ) a degree 2 extension of L which is galois. xn ] and L = f rac (B) 0 0 0 0 Thus every element of K can be wrriten as g + zh with g ... on the other hand.3 II. (g 2 − h2 f ) are in B .5 (Starred) Quadric Hypersurfaces (starred) part of starred part of starred 126 . If h has a nontrivial denominator. II.4 x 1 g−zh 2 = g2g−zh since z = f in A. If α = g + hz ∈ K ..2. . h f ∈ B so α is integral over B thus A is the integral closure of B In so is integrally closed. then α has minimal polynomial X 2 − 2gX + (g 2 − h2 f ) so that α is integral over k [x1 . h quotients of polynomials. 2 If.6..

3. so Y would be a divisor on the UFD Q. As P has order 2. named Q. if P + Q + R = 0 in the group law. and Y the homogenous ideal would be generated by two elements: codim (x20 + 1 on Q. Assume By L ∼ (z = 0) =⇒ P + Q + R ∼ 3P0 . well Q has codim 1.  so codim of Y is going to be.. (b) x g Assume P has order By exc I. P0 clearly.part of starred • so wts irreducible hypersurface on a quadric is complete intersection with  k [x0 . in group law.7. Q. + x2r ) and the thing from the principal ideal. . so in one additional Thus 2P + Q = 0 since they are on a line.... R} . Let T be the (from bezout) third intersection point with X . 127 point besides P. is a UFD why?? ∗  UFD implies? ∗ ∗ divisors are principle in UFD. xn ] / (x20 + .5 are distinct.6. L ∩ X = {P. + x2r ) r ≥ 4. Thus P + Q + T = 0 in the group law on X so R = −P − Q = T .. R are on the line L.6. TP (X) intersects X in three points by bezout.. then Q = 0.. so the tangent line passes through P0 ..6 x g P.. If P = P0 TP (X) 2 .4 II. and thus principal ideal.  a complete intersection. and by previous...6.6. then tangent line passes through Otherwise.. generated by a single element.10 as 2. . Now arguing as in example II. Q. intersects X at P with multiplicity > 1. Then P + Q + T ∼ 3P0 geometrically. 2. R Consider L through P and Q. Conversely.... Q. bezout.. so codim 2 in total? so . and assume WLOG P. =⇒ (P − P0 ) + (Q − P0 ) + (R − P0 ) ∼ (P0 − P0 ) = 0..

6. let fV map  f −1 (V ) . if 2.8 II. 1) 2. Thus 2P + P0 = 0 so P has order 2. P. 0). OX |f −1 (V ) → .6.6 P is identity on group law so (c) x g an inection point Thus 3P = 0 Conversely.6. Generators are (0. P0 . OY |V Let   f : (X.2.X = 3 by bezout and denition of inection point. (easier) ∗ By II. 0.7 (starred) 2. ˜ ⊗N ˜ ≈ f ∗ M ⊗˜B N ≈ f ∗ (L ⊗ M ) ≈ f ∗ M (M ⊗ B ⊗A)∼ ≈ ((M ⊗A) ⊗ (N ⊗ A))∼ using II. Otherwise. x By mordell-weil theorem.2. OY ) of ringed spaces and . On the other hand. TP (X) intersects X in P. 128 V ⊂Y .e. 0.7 =⇒ TP (X) .X = 3.5. (d). L. This is f M ∗ So f preserves the group structure.6. Locally we have Method 2 (which I will follow for subsequent parts) Notation: For a morphism V.6.9 II. the counit adjunction natural transformation.5. 1). ∗ ˜ ⊗ f∗ N ˜ ≈ f ∗ (L ) ⊗ f ∗ (M ). 0. (1.TP (X) passes through P0 .6. P has order 2 if equal to P0 .8 (a) x g (use easier method) . by part (a) there is an L such that L ∩ X = P . suppose P0 2. 1). 1. 3P = 0. OX ) → (Y. in the group law. (0. Method 1. f takes locally free sheaves to locally free    sheaves  of the same rank. (−1.

−) . f∗ P ) |f −1 (V ∗ ∗ ∗ We thus have a canonical morphism κ : f∗ HomOX (f M. Machinery: • • CaCl (X) .13) To convert from a Cartied divisor D to an invertible sheaf.13. Y are noetherian. Dx must be principal by II. locally factorial scheme then there is a natural isomorphism Cl (X) ≈ P ic (X) . then by II. Div (X) ≈ CaCl (X) and principals correspond to principals are factorial.3. This commutes with the usual adjunction and hence gives For any a natural isomorphism.2. D ∈ Cl (X). Y .11. Using the proof of 6. Theorem II.2.11 in the case of this problem. which by Lang(algebra) XII. We are given X. P )).3.6.1. HomOY (M. Note that on the intersections. it is well-dened by linear equivalence (6. F ≈ F (A). this is ≈ HomOY (N.Let Let η the unit adjunction hA = Hom (A. HomOX (f ∗ M ⊗OX f ∗ N.16 states: If Using II. (a natty ice for short) Now we have HomOY (N. the natural transformations from h to F : C → set (a functor) are in 1-1 correspondence with the elements of F (A). P ). 2. ψ ∈ HomOX (f ∗ M. we have that X. By denition. Notation: Cl (X) := Div (X) /prin (X) the group of divisors mod principal divisors. We nd an open neighborhood Ux so that D and (fx ) has the same restriction to Ux . Then ◦κ : f∗ HomOX (f M. since vY fj on • Cl (X) Cl (X) 3 D = Basically. HomOX (f ∗ M. As they are nite over a eld k = spec (k) . HomOX (f ∗ (M ⊗ N ) . in other words. we can convert a Weil divisor in to a cartier divisor in P{(Ui . CaCl (X) := the group of cartier divisor classes modulo principal diviors (divisors of functions ) by 6. let an open cover of X. Y nonsingular curves. f∗ P )) ≈ HomOY (M ⊗ N. −). integral.5. Also by denition they are regular of dimension 1. A A Yoneda's lemma states that N at h . fi =0.13) To convert from the invertible sheaf L (D) to D 129 . Thus Dx = (fx ).6. f P ) gives a natural transformation. f∗ F ) by adjunction. on X Let can {Ui } taken over prime divisors is Y sum is nite. P )) . P )) ≈ HomOY (N.16 on under this isomorphism. . separated. P ) → HomOY (f∗ f M. f∗ P ) ≈ HomOX (f ∗ (M ⊗ N ) . This is isomorphic to P ic (X) by above.20. X.6. Thus we can use II. HomOX (f ∗ F. f∗ P ).10 (b) x (continuing the notation from part a). By  the  6.11.g. fx ∈ K .6 X. P ∗ HomOY (M. f ∗ (M ⊗ N )). P ) ≈ HomOX (f ∗ N. Then (using the module version of Hartshorne II. f ∗ M ⊗ f ∗ N ) ≈ HomOX (f ∗ (M ⊗ N ) . let fi on Ui be the inverse of a local generator. and I. f∗ HomOX (f ∗ M. For any point in X . P ) |f −1 (V ) where MV is open in Y . Hence N at (HomOX (f ∗ (M ⊗ N ) .6. X . We have natural transformation. f∗ HomOX (f M. G) ≈ HomOY (G. X is a noetherian.c). we have that (fV )∗ ψ ∈ HomOY (f∗ f ∗ M. all the local rings are DVR's. OUi → L (D) |Ui dened by 1 7→ 1 fi is an isomorphism. which are then UFD's. ∗ By adjunction. fi )} vY (fi ) Y In order to convert a Weil divisor to a Cartier divisor. • (6. Y are separated and integral and nite over an algebraically closed eld. we just let denote the cartier divisor.A.6.

D+ (xi ) P ∈D+ (xi ) vP xi Assume 130 . n  o X is not contained in the hyperplane x0 = 0 (this is V ) whose cartier divisor is H = D+ xi . I want to show that f L (D ) ⊗OX L (f D) ≈ OX . (Y − UQ . t ◦ f ) . 1] + v[1. Let fi be a local equation for  Pn for which Yi ∩ Ui 6= ∅.1] ([x.5 or II. Then we can let t ([x. Now let UQ be a neighborhood such that t is only 0 at Q. 2.1 to deal with the poles and to deal with the zeros 1 ). 1) . Let t ∈ OQ be a local parameter at Q. x] so that t (P ) = t ([0. (Ha page 137) Q ∈ Y and then Q considered as a subscheme (using II. 1. xi ∗ The associated sheaf L (H) satises L (H) |D+ (x0 ) = O n| by denition. We can do this since there are only nitely many points where t has a pole or zero. Y ]) = X 2 (X − Y )2 . x]) [1. Let Q = [0. The pullback sheaf f L (H) ≈ xi P D+ (xi ) xi −1 ∗ f L (H) ⊗f −1 OPn OX satises f L (H) |f −1 D+ (xi ) = x0 OX (same logic as in (b) ) associated cartier is n o n o f −1 (D+ (x0 )) .6. 1]P. t) . Y 5 . (Y − UQ ) . f is a nite morphism. Let −1 D−1 ∈ CaCl  (Y )? We  must have it be linearly equivalent D to a principal divisor. The Weil divisor associ  P P x0 ated here is P which is the same as taking f ∗ in the other way.4.2. y]) = [x. 0] which has vP (t) = 1. We can then calculate the associated invertible sheaf L (D−1 ) ∈ P ic (Y ). Then the divisor Ui of V. (f −1 1 ∗ ∗ L (f D) |f −1 (UQ ) which is t◦f OY |f −1 (UQ ) and L (f D) |f −1 (Y −UQ ) ≈ OY |f −1 (Y −UQ ) . t .P ∈f −1 (Y −UQ ) vP (t ◦ f ) · P is the pullback to X . xxii . Let ni = vYi fi where fi is the restriction of fi to Ui ∩ is P ni Yi = P V X on some open set .3.11 (c) x The homomorphisms on divisor class grups in ex 6. xx01 (f is just inclusion in this case ) . and let Yi be the irreducible components of V ∩ X (X is a closed subvariety nonsingular in codimension one). 1]) = [0. 1)} ∈ CaCl is the associated Cartier divisor. 41 .2 is dened as follows: Let V be an irreducible n hypersurface in P which does not contain X .P ∈f −1 (UQ ) vP (t ◦ f ) · P + f (P )=Q. 1]}. P ∈φ−1 (Q)   P  1 D = B [0.• P f ∗ : Cl (Y ) → Cl (X) is dened by f ∗ Q = f (P )=Q vP (t) · P where tis a local parameter (element of K (Y ) with vQ (t) = 1 ).1] ([x. Now we −1 ∗ −1 ∗ must pull back the invertible sheaf L (D ). Y − B [0. i. So we −1 could write it as D = UQ . so this is a closed sum. x]) [0. integral and codimension 1 (every OQ (U ) is integrally closed and the curves are assumed to have codimension 1 local rings). Since f ∗ L (D−1 ) is inverse to L (f ∗ D). 4 . (See Ha I. t is an element of K (Y ) with vQ (t) = 1 . 1] which ramies into φ Q = {[0. Note that −1 restricted to f (Y − UQ )this is clearly the case since we have f ∗ L (D−1 ) ⊗OX L (f ∗ D) ≈ OX ⊗ OX and using a tensor simplication this is just OX .   f : P1 → P1 be dened by φ ([X. 1))} ≈ example from silverman. (Y − UQ .6) is closed. Compute rst for ∗ points Q and then extend by linearity to f : Cl (Y ) → Cl (X). Restricted f −1 OY |f −1 (Y −UQ ) ⊗f −1 OY | −1 f (Y −UQ ) −1 to f (UQ ) (use Grillet Abstract Algebra XI.X  P vYi fi = vYi (fi |Ui ∩X ). f ∗ xx01 = D+ (xi ) ∩ X. 1] + 0 [1.6) we also have −1 f tOY |UQ ⊗f −1 OY |UQ OX ⊗ 1t OX ≈ OX .6. 1]. look at f P P ∗ f D = f (P )=Q. the pullback homomorWhat is phisms correspond. 1] . 1] . ∗ Then f Q = v (t) P = v[0.6. [1. 1] = [0. Then D = {(UQ . 1t . This sheaf is given by L (D−1 ) |UQ = tOY |UQ and L (D−1 ) |X−UQ = OY |X−UQ . and by denition.e. GoLet ∗ −1 ing back to Cartier divisor and then picard group we get: f D ≈ {(f (UQ .

This morphism splits via the rank homomorphism from (b). k [t]⊕n → k [t]⊕m → M → 0 be a presentation of M .13 If II.9 (starred) Singular Curves (starred) MISS MISS 2.s. k [t] -module M . n 7→ nγ (OX ) is surjective. (commutative algebra fact) ⊕n ⊕m → OX →F →0 0 → OX in K (X) we have γ (F ) ≈ (m − n) γ (OX ).e. of coherent sheaves. where the rst morphism Let PID.14 Let (b) x 0 → F 0 → F → F 00 → 0 be an s.6.2.12 II. This gives an exact sequence Thus Thus . as k [t] is a .6.6. 131 is injective. 2.6. then F corresponds to an f. the map Z → K (X) .6.10 x g The Grothendieck Group K(X) F is a coherent sheaf.g.

d). The other half is obvious. Fj |U = I j M .5. Ann M ⊃ I ⊃ I .5. A noetherian implies I is nitely generated. then by exc II. X −Y .6. ξ.6. The functors i and i∗ are adjoint so there is a natural morphism η : F → i∗ i F : Coh (X) → Coh (X).6. Thus γ (F ) is in the image of K (Y ) → K (X) .6. We have √ √ II.11 (Starred) The Grothendieck Group of Nonsingular Curve (starred) 132 . ⊃ Fn = 0 such that each Fi /Fi+1 is the extension by zero of a coherent sheaf on Y . As X is noetherian. We must now show this ltration is nite. Note further that 2. gives an s.. The support of M ∼ is contained in Spec A/I = V (I) so by exc II. Let U = Spec (A) an open set where F |U = M ∼ .b.→ X . Thus η is surjective.Localizing at the generic point..15. ∗ ∗ Let i : Y .15 γ (OX ) 7→ 1 so that rank is surjective.5. F is extendable to Now suppose F is in the kernel of the second map. we may cover X with nitely map U and choose a maximum such N to see the ltration is nite. as well.s. Claim: there is a nite ltration F = F0 ⊃ Fi ⊃ . Let F0 = F and dene Fj inductively as Fj = ker (Fj−1 → i∗ i∗ Fj ) . each Fi /Fi+1 is the extension by zero of a coherent sheaf on Y so we have a ltration satisfying the requried conditions. Then Y ∩ Spec A ≈ Spec A/I so that η is induced by M → M/IM on U . and so dimK Fξ = dimK Fξ00 + dimK Fξ0 and thus the rank homorphism is well-dened. Thus N N there is N with Ann M ⊃ I (as in exc II. (c) x Surjectivity on the right is easy since if F is coherent on 0 0 a coherent sheaf F on X such that F |X−Y ≈ F .e. By denition. Thus we have proven the claim Pn−1 Now we have γ (Fi ) = γ (Fi+1 ) + γ (Fi /Fi+1 ) in K (X) so γ (F ) = i=0 γ (Fi /Fi+1 ) . Thus 0 = I M .

k . and Ann a ⊂ mP so that mP /Ann a = 0 is a submodule of M . This is the opposite inequality. then use an exact sequence and induction. k -ltration of k ni . Now suppose length k ni as a F is a torsion sheaf.starred starred starred 2.→ OPi and OPi /mP ≈ k . (3) is clear by denition of grothendieck group. for rank . γ (F ) ≈ γ (OD ). Thus a ltration of Thus the afore-mentioned stalk has length as an k ni OPi as an OPi -module can module at most equal to k -module.16 as II. for rank n≥2 0. M ≈ Ann A ≈ OPi .6. (1) gives . if a sheaf has rank uniqueness. To see uniqueness. D = P n i Pi OD .12 x g:1st paragraph Let D a divisor. -algebraically closed implies be extended to a it's length as a k -ltration. then it is torsion.6. 133 1. The stalk of at Pi is k ni which has k -module. Such a ltration has simple quotients. M ≈ hai a simple nonzero module. Thus M ≈ mP /Ann a ≈ k so that the ltration has simple quotients as an Now suppose that there is a maximal Let as this OPi -module. As K (X) ≈ P ic X ⊕ Z then and an invertible sheaf corresponds to a weil divisor written P P ni Pi gives an integer ni we have a map K (X) → Z which satises (1) for the degree. so (2) gives uniqueness.

if d ≥ 1. then using example 7. we can m ensure that xi ∈ P pulls back to si thus regaining ϕ via a decomposition φ . where N = n+d d   −1 . P P P P Then ui = aij xi are global sections of O (1) on Pm and φ∗ ui = φ∗ aij xj = aij φ∗ xj = aij tj = si m n If L = Z (u1 .7 II.3. note that ϕ : P n ∗ global sections on P which are si = ϕ (xi ) for L = ϕ∗ OPm (1) . Suppose 2. .. On the other hand.1. If n < m .2.7 x Projective Morphisms 2. ϕ must be constant. then ρ : P \L → P satises ρ ◦ φ = ψ by II.7. A −1 Let f (c) = 1 . Then an inverse to f is given by a 7→ a · c . dim (P ) = dim (ϕ (P )). N n If E is dened by Lj = 0 .7. Note that L is a linear subspace. dimension is preserved from Lj corresponding to the si . the bers must be nite.7.1 x g By another excercise. v (Pn ) is thus generated by n linearly embedding is a closed immersion.2 x P si = aij tj . and ρ is the linear projection. then the projection P − E → P is nite.7. As  dierent Thus we have embedded OPn (d) in OPN the d-uple Pn to PN .3 II.7.7. where ij = d. Id est. II. n n Every ϕ can be obtained in this manner. independent linear forms As the composition of nite morphisms is nite and nite maps have nite bers. which since the sheaves are invertible. si = xi00 · · · ximm . After an isomorphism. Note dimension here comes from fact that there are homogeneous polynomials of degree d in Pn . d+n n (1). and the aij dene an automorphism.7. the sections si are homogeneous polynomials of P d. → Pm is determined by an invertible sheaf L and m + 1 xi homogeneous coordinates of Pm . Thus suggests the d-uple embedding. Then cf (1) = 1 so f (1) = c . f : A → A is a surjective morphism of A-modules.1 II. we can check that f is an isomorphism on the stalks.. n We know all the invertible sheaves on P are If . 134 . i0 im The d-uple embedding v sends the degree d homogeneous polynomial u0 · · · um to the linear form ui0 ···im ∈ degree P N  . We have further that OPn (d).7. d ≤ 0. then as the maps are nite. un ) . are local rings. Choose aij to ll out an invertible matrix.3 x n Following Hartshorne II.2 being a local ring.8. ∀j . d ∈ Z then there are at most only trivial sections so that . Write 2..

b. −1 By II. and thus X → Spec A is separated. As Ln ⊗ Lm ≈ Lm+n we see that there are no ample Ln since large powers will not be gbgs. n > 0 are not gbgs. For N ∈ P ic X . if m<n < n .7. 1 Write Ln for N . M ) ≈ (OA1 . Finally L0 is gbgs.7. P ic U01 = 0 and so LU01 ≈ OU01 and therefore the isomorphism is an automorphism of k [x. Then to give a global sections of (OA1 . N as N is isomorphic to the structure sheaf.16. OA1 ) 1 and g ∈ Γ (A . N is determined by OU0 |U10 ≈ N |U10 ≈ OU1 |U10 . Such automorphisms correspond to the polynomials ax for a unique n ∈ Z. Ln ) it is equivalent to give f ∈ Γ (A .6 X admits closed immersion to some Pn A .4 . then the local ring at the origin of U1 is not generated by a global esction. and we know the number of generating sections for must in this case be constant. x ] n as a module over itself.6.6. As we know all O (d) (not (n + 1)! ). then .5 If II.2. . we know that L is generated by the global sections invertible sheaves are O (d) si = ϕ∗ (xi ) yet now there . The N is determined by this isomorphism so that P ic X ≈ Z .On the other hand. Let Let 135 . Thus f = g and therefore the global sections lie in the −1 −1 intersection which is k [t] for n ≥ 0 or (t ) ⊂ k [t] for n < 0 . x g X admits ample invertible sheaf L Projective space is separated over spec 2. P ic X is then the set of pairs (L . II. Thus an element of k [t] and of t k [t] which −n agree on U0 ∩ U1 . Claim P ic X = Z with elements (OA1 .7. OA1 ) . L (n · 0)). P ic A = 0 so any N ≈ (L . If n < 0 clearly Ln are not gbgs. x X the ane line over k with origin doubled.7.4 a. L 0 ) of invertible sheaves on A1 which have equal restriction to A1 \ {0}. Thus Ln . b. so ϕ are 2. Ln ) such that f |A1 \{0} = g|A1 \{0} . An element of t k [t] must have a homogeneous component of nonnegative degree so if n > 0. 1 By thm II.6. U0 . U1 be the two open copies of ane line. x See above 2.16.

p m n p−m−n Thus F ⊗ (M ⊗ L ) ≈ (F ⊗ L ) ⊗ (M ⊗ L ) ⊗ L As M 2. x g ample large multiple is very ample.11 If ϕL × ϕM in the segre embedding satises ϕ∗ (O (1)) = L ⊗ M which is an immersion since . n Thus M is gbgs. 2. x g L m ⊗ M is gbgs some m.9 is at least coherent. x g ϕL . x Lm is very ample and Ld is gbgs for all d > d0 .7.7. 2. F ⊗ L is gbgs n  0 . x g The geometric denition of ample is that it intersects every curve positively. m For arbitrary F ∈ coh (X).2. then take 136 m + d0 and use part (d) . where ϕ∗L (O (1)) = L and ϕ∗M (O (1)) = . for large enough p. Pm respectively. n For arbitrary F ∈ Coh (X) . m m m Then F ⊗ (L ⊗ M ) ≈ (L ⊗ L ) ⊗ M is gbgs.7. c. The product so is ϕL 2.8 b. . F ⊗ L is gbgs.5 x g Note the tensor of sheaves which are gbgs is gbgs. ϕM the corresponding morphisms to Pn .7.7 II.7.7.10 Let M d.

the morphisms agree. then by exc II. where φ is the hilbert X. and Thus dim |nD| = −1.1. x2 x21 1 1 1 . As in thm II.5. L ) − 1 = dim Sn − 1 = φ (n) − 1 . 2. we function of L with S (1)∼ . we associate the variety. S (n)∼ ) ≈ Γ (X. 137 . Now recall that the hilbert function equals the hilbert polynomial on integers for large enough 2.7 x g Some Rational Surfaces φ : P2 → P5 correspond to the linear system |D| since there are 6 global sections.9.b. x Suppose D is torsion of order there is an eective divisor E r.7.. .n 0 ∼ rnD ∼ rE > 0 contradiction.7. x2 .. b.7.7. If r|n . yy50 y0 7→ x22 x0 x1 x0 x2 x1 x2 .. nD = 0 lienarly equivalent to so nD nD dim |nD| = 0. si the (i + 1) basis vector of |D|. then . is trivial.4. y1 .6 x g The Riemann Roch Problem Using the given embedding. x2 . Thus .13 n. where S is the homogeneous coordinate ring of Sn → Γ (X.. th 2 2 2 dene φ : Ps → D+ (yi ).7. By denition. .12 II.. If s0 = x0 the morphism from Ps is given by 0 ii h h i    Let Spec x1 x2 .. L n ) is an isomorphism for large enough n. n have dim |nD| = dim Γ (X. x0 x0 agrees with → Spec v2 . If r .14 II. As this and on each other open set we have similar agreement.2. Usingexc II. yy05 y0 which on global sections is y1 .7.2.

we have a 1-1 correspondence between sections of 138 π and surjections E → L → 0. For (−a. Two distinct such lines do not meet since the ratio of their coecients are dierent. D+ (x1 ).17 II. X through P can be written as ax0 + bx1 = 0 . y2 ) + V (y0 . b). which is very ample. Assume P = (a. Thus consider two points (a. 2 in U ⊂ P has as closure the line V (ay0 + by2 . −b) we and similarly for other cases. y3 ) + V (y0 . y4 ) which 3. d) = (±a. ˜ is closed and U = π −1 is dense in X ˜ so the closure U is the image of X The image of X Let P ∈ P2 A global section has degree y0 of O (1) corresponds to the divisor V (y0 . 2 2 2 2 Then y − a (1) and z − b (1) separate all points except at the roots can separate points with y − yz − (a − ab) (1) (c. Similarly Thus we assume that the distinct points are not contained in any coordinate hyperplanes. 2. Claim: tangent lines are separated. On 2 2 2. a0 = a1 = 0 . b) . d) on D (x). y − yz. z 2 . x0 xz . then the sections are y 2 . y 2 . x1 x2 . On x = 1 we have 1. x1 i so on D+ (x0 ) . x 2 be given by hx0 . (c. On y = 1 the situation is similar. ˜ . In the last case. then x − y − (a − b) (1) and y 2 − b2 (1) have no tangent lines in common. then xy − y and x − y have dierent tangent lines at the origin. If a 6= 0 . the linear system d 2 2 with basis vectors x0 . z = 1 we have 1.8 x sections vs quotient invertible sheaves By 7. P0 = (a0 : b0 : c0 ) . ay4 + by3 ) Any line of It's image .7. x − y . 2 2 Thus tangent lines are separated.7. . then x2 separates points. yz which generate OP1 (2) Let If If for the other coordinate hyperplanes. and P1 = (a1 : b1 : c1 ) two points. ay1 + by2 .16 c. x Recall closed immersion i separates points and tangent lines. ±b) . If b 6= 0. y2 . y1 . xy − y.7. y1 y3 − y2 y4 ). then x − y − (a − b) (1) and x − a (1) have no tangent lines in common.15 b. if a = b = 0 . y3 .7. x0 x2 maps U homeomorphically onto an open subset of closed subvariety V = V (y2 y3 − y0 y4 . y . z − yz and y − yz and z − yz have dierent tangent lines at the origin. claim: points are separated by the linear system.12. and D+ (x2 ) − P ⊂ P . a0 = 0 and a1 6= 0 .2. x1 . x2 . z .

11 gives isomorphisms Li |Uij ≈ Lj |Uij satisfying the ∗ cocycle condition.11. 2.12. If x ∈ X . As π∗ O (n) is the degree n part of the symmetric algebra.7.9 x g α : P ic X × Z → P ic P (E ) by (L . 0 ∗ 0 ∗ ∗ On the other hand if E ≈ E ⊗ L .7. ni = nj and by ∗ ∗ denition of P (E ) .7.20 II.21 b. x g 139 . f ∗ O0 (1) ≈ O (1) ⊗ π ∗ L for L ∈ P ic X . Glueing gives a sheaf M such that π M ⊗ O (n) ≈ L on any component of X . Claim α is surjective. Claim: α is injective. then by thm II. n) 7→ (π ∗ L ) ⊗ O (n) .7.19 b. we obtain an element Oi (ni ) ⊗ πi Li ∈ P ic Vi ≈ P ic Ui × Z together with transition  ∗ ∗ isomorphisms αij : (Oi (ni ) ⊗ πi Li ) |Vij → Oj (nj ) ⊗ πj Lj |Vji satisfying the cocycle condition. r −1 Let r be the rank of E .18 II. and rank E ≥ 2. so by the projection formula.7. E = π∗ (O (1)) = π∗ (O (1) ⊗ π L ) = π∗ O (1) ⊗ L = E ⊗ L . Using exc projection 0 ∗ 0 0 and thm II. and each Ui is integral and separated. Dene 2.7. x f : P (E ) ≈ P (E 0 ). ∗ For L ∈ P ic P (E ) . then Vi := P (E |Ui ) ≈ Ui × P cover P (E ).b. As in II.7. there is a surjection π E ≈ π E ⊗ π L  ∗ 0 0 O (1)⊗π L .7.7. 2. By II.2. and thus a map P (E ) → P (E ) by thm II. By (a).1.10 x P^n Bundles over a Scheme See Gathmann's notes. Using the projection formula gives π∗ αij : π∗ Oi (ni ) |Vij ⊗ Li → π∗ Oj (nj ) |Vij ⊗ Lj . then n = 0 so L ≈ OX . Oj (n) |Vij = Oij (n) so we have Oij (n) ⊗ πi Li |Vij ≈ Oij (n) ⊗ πj Lj |Vij .6. x ∈ U = Spec A . then this gives a left inverse to Z → P ic P (E ) . π∗ (π L ⊗ On ) ≈ OX .11. If Ui gives an open cover of X on which E is trivial. ∗ ∗ Suppose π L ⊗ O (n) ≈ OP(E ) . As X is regular so are Ui thus satisfy (*) so by exc II.7. As OP(E ) (n) |U ≈ OU (n). where E |U ≈ OX |U then we have π U = Pr−1 so U r−1 r−1 there is an embedding Z ≈ Pk(x) → PU → P (E ).7. An inverse is given by considering E ≈ E ⊗L −1 Suppose that . Tensoring with Oij (−n) and using the projection formula with II. P ic Vi ≈ P ic Ui × Z . (as X is regular. L ⊗ π∗ O (n) ≈ OX . r−1 and ane schemes are separated) .11.7.11.

11 x By exc II.7.. x This follows from (b).7. . and exc II.   ⊕(n+1) −1 Then π Ui ≈ P roj I (E ) (Ui ) ≈ P roj I OUi (Ui )   ⊕(n+1) ≈ I O|U (U ) ≈ P roj A [x0 .7. then as in II.23 II..5. π : P (E ) → X the projection. xn ] ≈ PnU .. by denition of P (E ) we have an automorphisms ψ = ψj ◦ ψi OUn+1 |V coming from the restriction morphisms E (Ui ) → E (V ) ← E (Uj ) .X is covered by open anes Ui = Spec Ai where ⊕(n+1) E ≈ O|Ui . j Let dened via OUn+1 |V ≈ i MISS (starred) 2.24 b. 2. ϕ : T → S is a morphism of graded rings. −1 On V = Ui ∩ Uj .9 140 .5. n Thus we have the trivializations needed for a P -bundle.7.22 d. we have a commutative diagram: / P roj T so the given isomorphism is natural and thus on intersections Ui ∩ Uj we have P roj S Now if  P roj S (d) /  P roj T (d) isomorphisms.7.9. 2.13. we have P roj ⊕n≥0 I (U )nd ≈ P roj ⊕n≥0 I (U )n . (c).7.13. x By lemma II. Now glueing gives the result.

Z ∩ U = Spec A/IZ .. Then Y ∩ U = Spec A/IY ..13 A Complete Nonprojective Variety * MISS MISS 141 . ⊕d≥0 IZ If .25 If c.7.7.. IZ = IZ (U ) . 2. . ⊕d≥0 IZ .7. x f :Z→X Algebraic Geometry 1.7. then II. d d Then the kernel of each of these ring homomorphisms is ⊕d≥0 IY . is birational..27 II. d d d A nonempty intersection gives a homogeneous prime ideal of ⊕d≥0 (IY + IZ ) containing ⊕d≥0 IY . −1 Then π U = P roj d≥0 (IY + IZ )d with IY = IY (U ) . −1 The map π (U ∩ Y ) → π −1 (U ) is given by ⊕d≥0 (IY + IZ )d → ⊕d≥0 ((IY + IZ ) (A/IY ))d and π −1 (U ∩ Z) → π −1 (U ) is given by .7.40).12 x g P ∈ Y˜ ∩ Z˜ . then L π (P ) is contained in an open ane U = Spec A ⊂ X . This is a contradiction. L −1 Then π (U ∩ Y ) = P roj d≥0 ((IY + IZ ) (A/IY ))d ⊂ Y˜ . and π −1 (U ∩ Z) = . 2.26 E = Exc (f ) = Z − f −1 (U ) is a divisor (see Debarre Higher Dimensional Pushing forward to X the corresponding ideal sheaf gives the required I .2..

where the pullback of OPn (1) b.MISS MISS 2. k (B)) → .1 x Injectivity of the rst map is equivalent to surjectivity of homk (m/m2 . x U n > 0. and m  0 . OP (1) ⊗ π ∗ L m is very ample on P m+n relative to X .1 II.5.8 x Dierentials 2. then by thm II.8. k (B) ≈ Derk (B.7. which is not eective.8 II.29 L to Pn .7. π : P → X is projection.12. on X relative to Y . then it would give a closed immersion would be OP1 (−1) .28 II. If O (1) on X was very ample. 142  homk(B) ΩB/k ⊗ k (B) . L n is v.8. contradicting negativity. By exc II.7. for large enough m.a. is ample relative to . for n xed.14 x g X = P (O (−1)) ≈ P1 . OP (1) ⊗ L is very ample on P relative to Y . k (B)). and for If 2.10. Consider 2.7.

8. By thm II. dim Ωk(B)/k = tr. Recalling the proof of thm II.16. tr. λ ∈ k (B) . We have ΩB/k ⊗B K = ΩK/k using thm II. k (B) /k . ∗ Then d is a k (B) derivation and δ (d) = h. If 2. K is separably generated so dimK ΩK/k = tr.8A.6. then F rac A = F rac B and ht p = dim B .8.2A. so b exc II.K/k by thm II.8. At each w ∈ W the stalks are free of rank n . we just need to show that dimK ΩB/k is dim B + tr.Note that if d : B → k (B) is a derivation. then dene b.a. As B = Ap for p ∈ Spec A . K/k .F rac (A/p) /k = dim B + tr. dimk(B) ΩB/k ⊗ k (B) = dim B + tr.8A. If x ∈ U .d. suppose B m/m2 = dim B so B is regular. k (Ap ) /k = dim A = dim X .5.8.7. c ∈ m db = h (c mod m2 ). is ΩAp /k ≈ ΩA/k ≈ ΩX/k 143 .2 (using thm II.6. Thus dimK ΩB/k ⊗ K ≈ tr. 2. by thm I.d. tr. Then similarly. x Suppose ΩB/k is free of rank= dim B + tr.d. Let to 2. there is an open neighborhood W 3 x where ΩX/k |W free of rank n.d.A.x is a regular dim Ap + tr.K/k = dim A = ht p + dim A/p = dim B + dim A/p = dim B + tr. By thm I. ΩX/k is locally free by (c). OX. h ∈ Hom (m/m2 . k (B)) and b = c + λ ∈ B . then δ ∗ (d) is the restriction to m with d (m2 ) = 0 by the product rule. x g U is dense since it contains any open dense subset V ⊂ X which is smooth. k (B) .8. so using δ 0 → m/m2 → ΩB/k ⊗ k (B) → Ωk(B)/k → 0 from (a). By (b).3 c.1. k (B) /k .d. x x ∈ U = Spec A.d.d.d.8.d. k (B) /k .8. is regular.8. so by part (c) w ∈ U .25A). dim Conversely.d.4 local ring i  x is free of rank equal d.4.8.a. By II.d. As k is perfect.8.k (B) /k .

ωX×Y = Λm×n ΩX×Y Now using exc II. then ΩX×S Y /Y ≈ p1 g : Y → S is ΩX/S and also II. you just need to show. and  ΩX×S Y /X ≈ p∗2 ΩY /S .8. ΩX×Y /S ≈ . that the rst sequence splits. Since E is rank r . the cokernel of this map is locally free of rank rk E − 1 . s) |sx ∈ mx Ex } and dene p1 . By exc II. and ∗ is obtained by base extension.8.5.8.8.2. As E is gbgs.11 gives that (writing S for Z . From here.5. then this map is surjective.2 x Z ⊂ X × V by {(x. Dene 2. 2. Using denition of the canonical sheaf. so that p∗1 ΩX/S ⊕ p∗2 ΩY /S . S for Y and Y f 0 : X ×S Y → Y for Y0 ) that if f : X → S is a morphism.8. Thus dim ker = dim V − r .8. using both sequences.b.3 x g Product Schemes II. 144 which is Λmn (p∗1 (ΩX ) ⊕ p∗2 (ΩY )).16. we see that Λmn (p∗1 (ΩX ) ⊕ p∗2 (ΩY )) ≈ Λm p∗1 (Ω) ⊗ Λn p∗2 (Ω) ≈ p∗1 (Λm ΩX ) ⊗ p∗2 (Λn ΩY ).7. p2 : X × V |Z as the projections. and X ×S Y for S )   ΩX×S Y /Y ≈ p∗1 ΩX/S → ΩX×S Y /S → ΩX×S Y /X ≈ p∗2 ΩY /S → 0 and similarly.10 says (writing another morphism.5 II.8.15. and thm II. x g dim (X) = m . X for Y .7 Let b. A ber of p1 over x0 consists of all sections vanishing at x0 which is the kernel of V ⊗k k (x0 ) → Ex0 ⊗Ox 0 k (x0 ) ≈ Ex0 ⊗Ox0 Ox0 /mx0 ≈ Ex0 /mx0 Ex0 .d.8. dim V − dim ker = rkEx0 = r by rank nullity. e.  p∗2 ΩY /S → ΩX×Y /S → p∗1 ΩX/S → 0 are exact sequences. For the second part if s therefore satises sx ∈ / mx Ex then Ox → E is given by ×s . Thus dim Z = dim X + dim V − r . But then as r > n we must have dim Z < dim V so p2 is not surjective. dim (Y ) = n .6 II.

8... .. OY ×Y ) = k pg = dimk Γ (X. If IY = (f1 ..8.. P As S/ (IH1 .8.8 c. . if Y = ∩Hi .. .. x g Y is normal. ..11 c..3 gives This is As Y OY ×Y OY ≈ ωY Y3 pa (Y × Y ) = −1 . Now using ex.8.. Γ (Y × Y. then as Hi are irreducible (IHi ) is a prime ideal in the coordinate ring S = k [x0 .2.b.b.20..9 From (e). .5.4 x Complete Intersections in Pn Y = ∩ri=1 Hi . then (IH1 . x g From exc I..2. Using the unmixedness theorem.. IHr ) .8. By denition geometric genus is II. Hi = Z (fi ) as in chapter 1... thus complete. ωX ) which is 1. II.4. then Sing Y . And thus so is S (Y ) . xn ] of Pn .d to see that which has dimension over 2.. x g 145 . pa (Y ) = 1. ∗ ∗ gives ωY ×Y ≈ p1 OY ⊗ p2 OY we get . On the other hand. Thus the homogeneous coordinate ring S (C (Y )) is integrally closed by thm II. Thus IHi+1 is not a zero divisor mod IH so that (IH1 . . IHr ) are codim ≤ 1 thus J = ∅ so IY = (IH1 .. by denition. Thus Y is projectively normal.23.. is projective. IHr ) has degree deg Hi by bezout.. with k equal to 1. Using (b) . IHr ) = I ∩ J where codim J > 2. fr ) 2. Sing C (Y ) have codimension ≥ 2.. is an intersection of integral hypersurfaces. .8.. we can use exc II. IHr ) forms a regular sequence and the ideal is contained in IY .7. thus proper.10 then b. primary components of (IH1 . As of 2.8.

we get a copy of P which is nonsingular and irreducible and codimension 2. |dH| has a smooth element H of degree d.13 r. on global sections. n 0 0 Recall pg = h (Y. By thm II.H2 ). Suppose where di = deg Hi . thus 0 0 h0 (Pn . ∩ Hn . If we repeat this r times.8.8.H2 corresponds to OH1 (d2 ) and thus in total ωH1 ∩H2 ≈ OH1 ∩H2 (d1 + d2 − n − 1)..14 f.20. n n−1 n−r Now P |H 0 ≈ P . ωY ) ≈ h (Y. 0 Thus. l = 0 . then k  Γ (Y. e. IY (d −n − 1)) + h (Y. Now deg Y = d > d − n − 1 so that there are 0 h (Y. and Now repeat this step by step.e.8. so the number of components is ≤1 . 2.8. x g Y = H1 ∩ . OY (d − n − 1)).12 Y is projectively normal. OY (d − n − 1)) .. OY ) ≤ 1 .Since When 2. x g By adjunction KY ≈ (−n − 1 + d) H . n+d−n−1 . x g Assume that k is algebraically closed so we can use Bertini. we ωH1 ∩H2 ≈ ωH1 ⊗ OH1 ∩H2 (H1 . the general element of a free linear series is smooth. Note that the LHS is no sections vanishing on 146 Y of degree d−n−1 so . By bertini. OPn (d − n − 1))  = h (Y. Taking global sections of 0 → IY → OPn → OY → 0 gives. by part (c) a s.s. if H is a hyperplane. IY (d − n − 1)) = 0. Note that H1 . d. the natural map is surjective by denition. OY ) gives dim Γ (Y. Adjunction formula gives ωH1 ≈ OPn (−n − 1) ⊗ OH1 (d1 ) ≈ OH1 (d1 − n − 1) n OP (−n − 1) represents the canonical of Pn .

ωYy0 ≈ π1∗ ωy ⊗ π2∗ ωY 0 ≈ π ∗ Oy ⊗ π2∗ (f ∗ ωX ⊗ OY 0 (−n − 1)) ≈ Oy ⊗ π2∗ OY 0 (−n − 1) ≈ OYy (−n − 1). ˜ X Thm II. Z → P ic X (The second equality is since exceptional divisors are contracted). By exc II.8.s.15 g. Q) 7→ f P − gQ and the third map is obvious.8.         l+3 l+3−d l+3−e l+3−d−e Now χ (l) = − − + 3 3 3 3       3−d 3−e 3−d−e = + . OX  ˜ (−nY ) ≈ OY 0 (−n) so if OX˜ (nY ) is trivial. 0 ∗ 0 Using adjunction ωY 0 ≈ ωX ˜ ⊗ OY 0 (Y ) ≈ f ωX ⊗ OY 0 ((n + 1) Y ) 0 0 At a closed y ∈ Y . 0 By thm II. ˜ then j ∗ f ∗ is identity on P ic (X) ≈ P ic (U ) . let Yy a ber. P ic X   ˜ → P ic (U ) ≈ P ic (X) → 0.8.6. We have an exact sequence 3 (l−d−e) 3 (l−d) (l−e) (l) 0 → S (P ) → S (P ) ⊕ S (P3 ) → S (P3 ) → S (Y )(l) → 0 where P 7→ (gP. If j : U .e. U = X\Y ˜ 0. x g Note that for a curve we have pa (Y ) = pg (Y ).16 χ (0) = 1 − II. f P ) and (P.17 and thus . then ˜ → P ic (X) → 0 s.8.5 gives is nonsingular so that     ˜ ≈ cl X ˜ . 0 r−1 As Yy ≈ P then ωYy0 ≈ OYy0 (−n) so n = r − 1. 0 → Z → P ic X ∗ This is split by π . L invertible. U = X\Y ˜ 0.b.24.8. 3 3 3 2. b. 2.8.8. By (a) we write 147 n=0 so that we actually have an .3.→ X ∗ Then j (ωX ˜ ) ≈ ωU =⇒ L ≈ ωX .2.24.5 x g Relative Canonicals Important! By thm II. x g ωX˜ ≈ f ∗ (L ) ⊕ OX˜ (nY 0 ).

Thus θ satises the liebniz rule.→ B gives a k -linear morphism.  0 Conversely. I then θ ◦ d : A → I . if θ ∈ HomA ΩA/k.. 0 0 Now add (g (a) g (b) − g (a) g (b)) and factor the θ (ab) into 0 g (b) θ (a) + g (a) θ (b) . then θ lies in I ⊂ B0 θ : 1 7→ 0.I .2. Thus θ is 0 on k . 0 As g − g lifts f − f = 0 .18 II. Thus g + θ is a k -linear homomorphism lifting f .8.6 x Innitesimal Lifting Property 0 Suppose rst g : A → B lifts f . A simple computation shows that g (ab) + θ (ab) = (g (a) + θ (b)) (g (b) + θ (b)) so that g + θ Further k -algebra homomorphism. 148 is actually a .. 0 Then θ (ab) = g (ab) − g (ab) = 0 0 g (a) g (b) − g (a) g (b). 0 0 Note that 0 → I → B → B → 0 exact implies this is in the kernel of B → B .8.

I = 0 for i > 0 . For a ∈ P . Taking the LES associated to the dual gives therefore a surjection  hom ΩP/k ⊗ A. II. x k -homomorphisms h is determined by the images of the xi . I  hom (J/J 2 . and b ∈ J . h0 = h − θ. 0 0 If we dene θ as P → ΩP/k → I → B this gives a k -derivation. There is therefore θ : ωP/k → I with image h by (b).17. 0 For each i . I).2. 0 0 Thus h gives a morphism g : A → B lifting f by (a). 2 2 2 Also if a ∈ J then h (a) ∈ I = 0 so our map descends to h : J/J → I . 0 → J/J 2 → ΩP/k ⊗ A → ΩA/k → 0.8. As h preserves multiplication h is A-linear. x By II. Thus dene h : xi → bi as a k -algebra homomorphism.20 c.8.3A. we have an exact sequence on global sections  A nonsingular implies Exti ΩA/k . then h0 (b) = h (b) − θ (b) = h (b) − h (b) = 0.8. let bi be a lift of f (xi ) in B . if a ∈ J .19 b. so that h (a) ∈ I .8. then h (a) is 0 by commutativity. If also HS def theory 149 . A 2.

22 II.skip II.8. am + a m) . 0 0 The innitesimal lifting property gives a morphism A → A and thus we have A ⊕ M ≈ A by associating We M where + I as an A-module. . m0 ) = (0.9 Formal Schemes . a0 m) since mm0 ∈ I 2 .8.1 MISS 150 on A . m ) 7→ (aa . a ∈ A then (a. and since M ≈I . m) (a .9 II. m) (a0 . then (0.8.19.2. I 2 = 0 . m0 ) = (aa0 .9.8. 0 must show that A ≈ (A ⊕ M. 2.8 x As in II. +) 0 0 0 0 0 is dened by (a. with If If 2. am0 ) by the A-module structure m ∈ M ≈ I .21 Let II. 0) (a0 .7 x A0 /I ≈ A . I ≈ M .8.

2 MISS II.9.3 MISS MISS 151 .9.MISS MISS MISS II.

5 MISS II.4 MISS II.9.6 MISS MISS MISS 152 .9.9.II.

F ) . F ) and we need to show exactness on the right. Z) → Z → Z ⊕ Z → · · · . (see Ex.s.1 no questions.2.s. Z cannot surject to Z2 . (use II.21. there exists s0 mapping to s. 0 We have ΓY (X.e.3 III.2. F ) → Γ (X.3 x Cohomology with Supports 0 → F 0 → F → F 00 → 0 an s.1 III. LES is Since MISS 3. H 1 6= 0.2.MISS 3 III Cohomology 3.2 III.1.e. 3.2. F ) ⊂ ΓY (X. This follows by checking the stalks. We want to show s0x = 0 for x ∈ X\Y . This s gives rise to an element in the kernel of Γ (X.1.2.2 III. Let s be in the kernel of 00 the second map. 0 → ZU → Z → iP Z ⊕ iQ Z → 0.2 x 3.1. Let 153 .2.17) 0 → Γ (X. By left exactness of Γ .2.16) The constant sheaf is asque as Applying the LES in cohomology gives the desired vanishing in cohomology. 3. the sum of skyscraper sheaves.a x g We have an s.2 x asque resolution g By II. K /O is P iP (IP ) .d.1 III. P1 is connected.

we have ax ∈ Fx0 Now nd a small neighborhood Ua and a section a mapping to t|Ua . Find an open cover of such Ua .2.3. Thus choose At 3.c x See the proof of Prop. Flasque Vanishing Theorem 3.5.4 b. So we need to show the RHS is surjective.20. F ). then tx 7→ sx = 0. s ∈ Γ (X. x Flasque Global sections are exact We know from (a) it's left exact. . t − a is zero so t − a is supported on Y .b. t − a 7→ s. 154 mapping to tx . 3.7 Let x F• be an injective resolution of F. 00 00 For s ∈ ΓY (X. and on U .16. x ∈ U = X\Y . F ) in preimage of s.1.2. F Now apply the sequence of part (d) to to get the long exact sequence.2. t restricted to such an open cover agree Using sheaf axioms and asqueness we nd a global a mapping to t on U . III. note that the u.1.2.5 III. We have exactness on open sets by II.2.2. F ) .b. t ∈ Γ (X. so by exactness of stalks.2.6 x easy Apply the global sections functor to the sequence of II.

d. Now take the LES of the top row.2.2.3.2. 155 . then I |V is an injective resolution for F |V . 3.8 excision x ΓY (X. so There is an isomorphism the espace etale of the stated isomorphism gives an isomorphism in cohomology.2.4 x Mayer-Vietoris We have the following diagram: The columns are exact by asqueness and III. i i Now if I is an injective resolution for F . F |V ) (to see this it may be helpful to consider F ) where V is an open subset containing Y . and the bottom row is exact by sheaf axioms. Using the 9 lemma or a spectral sequence gives the top row exact. The middle row is obviously exact.9 II. F ) → ΓY (V.7.

FP ) = FP .10 III. FP ) = lim F (U ) = FP induces a morphism Since ← p∈U f : ΓP (X.5 x H i results from taking cohomology of the global sections functor.3.2. then f factors through fb : R → Ib . Now suppose s ∈ ΓP (XP . PP ) . we claim ΓP (XP . The morphism Γ (X.2. that ΓP (X.2. → If U ⊂ X and R ⊂ ZU . ΓP (XP . Thus f is injective.6 x Note that a sheaf map f :R →I I is injective i for each open set . For each i.11 III. ` R S⊂ ZU . R (Ui ) ⊂ ZU (Ui ) = Z are subgroups generated by si so that nitely many si generate R . → This is a direct system. there exists an extension of f U ⊂ X and for ZU → I to a map every subsheaf R ⊂ ZU and for every . This is essentially the denition of injective in any algebra book. then by noetherianity of X . then is surjective.2. such that sU represents s. by denition of compatible germs. For a map f : R → limIa . Since s = t. so the morphism factors as R → Ib → limIa for some b. FP ).sU |U ∩V = 0 so that sU and 0 glue to give a global section with support in P . F ) ≈ on the localization We just need this to be a bijection.2. Thus f Suppose that f (t) = f (s) in they agree on every stalk. So lim Ia is injective. F ) → Γ (XP . Then if V = X\P . Ib is injective gives fb extends to ZU → Ib and thus we give an extension ZU → Ib → limIa → of f . Ui = Un . If U = Ui is a decomposition of U into connected components. FP ). Thus we can nd a neighborhood U 3 P and sU ∈ F (U ).12 x III. Thus sP = tP since f sends elements to their germs. 3.2.7a g Cohomology of circle 156 . F ) → ΓP (XP . By the hint. Now suppose → 3. If necessary shrink U so that (sU )Q = 0 for Q 6= P . then f (si ) = ti ∈ Iai (Ui ) .

and t ∈ H (S . n 0 1 1 Let s = {(Ui . By thm 3.3. ti )} so t : S → R .3. d d+1 Consider the OXred − module Gd = N · F /N · F .3 x Cohomology of a Noetherian Ane Scheme 3. D) satises r |Ui ∩Ui+1 = ri on Ui ∩ Ui+1 and 0 elsewhere. x D is the sheaf of all real-valued functions. 0 1 Thus r ∈ H (S . Y1 . and 157 . Gd ) = 0 so we have a surjection  1 d+1  F  H 1 X. Gd ) Now let N = H 1 (Xred . Then Xred = Spec (A/N ) .2. III. then each point is in an ane neighborhood so X is ane. then r 7→ r so r is in the image of a.3 III. then we get a LES. By theorem III. If 3. 0 0 1 0 0 a If r ∈ H (S . IY1 ∪Y2 )  H 1 (X.7. OX ).3.13 π1 (S 1 ) is Z. Gd ) = 0 so F is ane. D/R) since S is compact.→ X .7. H X. i : Y2 . D) → H 0 (S 1 .3. H 1 (X. IY1 ∩Y2 ) = 0 so there is a surjection H 1 (X. 1 By induction. N . t = {(Ui . by noetherianity.3.2 x X is ane.1 x X Suppose is ane.1 III. si )}i=1 ∈ H (S .3. Thus s = t − r is in the image of a.7. this is equivalent to X ane. b. every irreducible component is ane. Y2 an irreducible component of X . Thus if r = {(Ui . N k F = 0 for k < d 1 Thus H (X. Y2 is an arbitrary closed subscheme. D) = 0 since D is asque. i∗ IY1 ∩Y2 ) = H 1 (Y2 . Wlog by shrinking assume (Ui ∩ Ui+1 ) ∩ (Ui+1 ∩ Ui+2 ) = ∅ . If X is dimension 0. R) → 0 as H 1 (S 1 . N is the nilradical of A = Γ (X. N d F . N = 0 for d ≥ m. We have by thm III. 1 0 1 Now let ti = si + ri . d be the sheaf of nilpotents on X .2 If X. So Xred = Spec (A/N ). D/R) → H 1 (S 1 . then consider 0 → IY1 ∪Y2 → IY1 → i∗ IY1 ∩Y2 → 0. D/R) so that t is in the image of α . ri − ri+1 = ri . D) since ti |Ui ∩Ui+1 = si + si+1 − si = si+1 = ti+1 |Ui ∩Ui+1 . R) → H 0 (S 1 . a 0 → H 0 (S 1 .b. Continuing in If every irreducible component is ane. ri )} then on Ui ∩ Uj .11.It's a lot easier to just use hurewicz theorem since 3. 0=H 3. H 1 (X. and write ri = si+1 − si extend by zero so ri is dened on Ui . So a is surjective.3. IY1 ). then by exc II. D/R) . Now suppose Xred is ane. Also t is mapped to itself in H 0 (S 1 .

to see that 158 Hai is a quotient of Γa (Ii ).So n ∈ Γa (N ). then a ⊂ Ann m and thus m ∈ Γa (M ) .3. ∼ Thus m ∈ ΓY (X.1 1 this manner step-by-step. X is ane.) .. . t If e ∈ ker (Γa (E) → Γa (M )) . i Let x ∈ Ha (M ). We show that Let 3.. 0 → Γa (N ) → Γa (E) as Γa (P ) ⊂ P for A-mdoules P . n Then a m ∈ / p so an m = 0 and thus m = 0 in Mp ..˜.∪Yn )  H (X. x Γa (·) = ΓY (X.4 b. 3. M √ Ann m ⊃ a by thm II.3. Thus Supp m = V (Ann m) ⊂ V (a) and thus √ ni A is noetherian. we achieve a surjection H (X.. If m ∈ Γa (M ) then an m = 0 for some n.. then p 6⊃ a so there is a ∈ a not in p. IY1 ).3. IY1 ) and thus by thm III. fn ) and fi ∈ Ann m and thus fini ji ∈ Ann m. ∪ is left exact.3. If p ∈ X is not in Y .7.. Next suppose m ∈ ΓY X.3. .1 . (dention of Γa ). M ). 0 N0 Choosing a large enough N . Q If N = ni ji then fiN ∈ Ann m for all i.2.3 Γa Y1 ∪ .5 x Γa (Hai (M )) ⊂ Hai (M ) by denition. x 0 → N → E → M → 0. so a = (f1 . Yn = X so that 0 = H 1 (X. M arbitrary. Take Γa of an injective resolution Ii for M i i Thus Ha (M ) ⊂ Γa (Ha (M )) .   ˜ .3 III. Since global sections are left exact. Consider We have Thus we have shown that the kernel of the second map is contained in the image of the rst map. then there is n 7→ e t Then a n 7→ 0. The opposite inclusion is clear. IY1 ∪Y2 ∪. 3. then a e = 0 some t. IX )  H 1 (X.. Eventually.

Hai (M ) = 0 for i < n + 1 . by exc III.7 b. x n. thus by exc Suppose the statement is true for Choose III. every section of OX over U −P extends to a section of 0 which. Suppose any Now suppose xm 6= 0 Γa (M ) = 0 for M nitely generated. M with deptha M ≥ n + 1.2. computing cohomology. By exc .4 x Cohomological Interpretation of Depth deptha M ≥ 1 then there is x ∈ a such that x is not n..3.3..3.c. a is not in the union of associated of p which is the set of zero divisors of M .3.3. n−1 By induction. Thus an can not annihilate any element so Γa (M ) = 0. then the LES deptha M/x1 M ≥ n − 1 =⇒ deptha M ≥ n Han (M ) = 0 On the other hand if we have Now by induction.3. . III. OX |U ) = 0. OX ) (U.3. a is not contained in any associated prime so by prime avoidance. OX ) ≈ Γ (U − P. OX |U ) ≈ HP1 (U. If x1 .5 x U 3 P . Thus deptha M ≥ 1 .3. Ha (M/x1 M ) = 0. Then primes 3.. xn+1 ∈ a is an M -regular sequence then we have ×x · · · → Han−1 (M/x1 M ) → Han (M ) →1 Han (M ) → · · · . . Hai (M/x1 M ) = 0 for i < n.8 For . Then neither is then there is an xn x ∈ an for with . Thus the induced map ×x1 should be injective.3(e) is the same as HP 159 OX over U i Γ (U. 3.6 III. For a zero-divisor for m∈M and n≥0 M.

5 this is equivalent to and exc III.10 part c. 0 0 To prove injective. So g is injective. fn) is f.3. M ) → n φ(fir ) Note that if g (φ = 0) .. and it's well-dened by calculation. OSpec OP ) ≈ HP1 (Spec OP . Choose one fir  s+r s+r n(s+r)+1 of them.3. given 0 → F → F and F → G we want to lift this to a morphism F → G ...9 G OP ) =0 . Also U is covered by D (fi ) so s ∈ Γ U. ˜ is P mi ai ∈ ⊕Mf . 160 s to work for all . M i fi   P si ˜ On the other hand if a ∈ ⊕Mfi and is sent to 0 by localization at fj .11 III.then = 0 ∈ Mfi so fisi φ (fir ) = 0 ∈ M some si . then it's actually in Γ U. OSpec is equivalent to depthm OP ≥ 2 .2. let fi : Ui . so 1 Thus we have an induced n r g : lim homA (a .3. Thus ˜ fi∗ Ii is injective now the lifts commute with direct sums.3.b.g. III. M fi i   φ(f1r ) φ(fnr ) r .3. A is noetherian. n and let F |Ui  ≈M  i .. Now use exc III. 3.3. . This denes a section. we see injective resolutions are asque.3. If φ : a → M dene f (φ) by fr Note that . 0 0 Note that by injectivity of Ii .6 x ˜ Ui = Spec Ai for i = 1. then for each i. is constructed by covering X with open anes starred 3. Thus if a is generated by fi so that φ fi = 0.. III.3.4 to see this 3. any f : F → fi∗ I˜i gives F |Ui → I˜i which lifts to f : F |Ui → I˜i . If Mi is embedded in an injective Ai -module Ii .→ X and dene G = ⊕fi∗ I˜i .HP0 (Spec OP . . f r ..7 x a = (f1. x Using part (b).

1 x g pushforward cohomology ane morphism Consider Since f {Vi } an ane cover of Y .c. a is generated by fi P P. n n n n Thus b ⊂ a so that HomA (a . (use the given relation).4. f −1 (Vi ) are ane cover of is ane. assume a. . F is q. X. . Thus V (a) ⊂ V (b) =⇒ b ⊂ a. I) is surjective. b are radical. 3.. on the intersections of the preimages..4. This gives surjectivity.. If nd Multiplying both sides by m∈I xn+1 with xn0 (x0 m − 1) = 0.   m01 m0n Check that this is well-dened and the image of φ is = f . mrnn .. x U = X − V (a) and V = X − V (b) . However this ring doesn't have this relation.4 x Cech Cohomology 3.. the pushforwards of such are also of M Now taking cech complexes. f ∈ Γ U.3.. the result follows. then (fi fj ) ij (fi mj − fj mi ) = 0 ∈ M . r+s R For R > n (r + s) . n n ∼ Thus limHomA (a .4 III.13 → III.1 III. f ∈ Γ U. I) so by (a).12 n b.17?) f is ane.. M f f1   s ˜ . As in (a). Since Since ˜.8 x Localization not injective non noetherian. 161 . I → Ix0 is surjective. F |f −1 (Vi )∩.5. shows that xn0 xn+1 = 0. Let U ⊃V with → 3. f denes mr11 .5.4. ≈ M 0 ˜ 's (see ex II..    ˜ . r+s R Dene φ : a → M sending ai fi to ( ai m0i ) |U ..3. r+s f f If 1 3. and applying III. then dene m0i = fis mi and if r > ri then fir+s f − m0i |D(fj ) = 0 so fir+s f = m0i on U . By separatedness and an excercise in (II. I)  lim HomA (b . I is asque. r+s .3.3?) the intersections of the preimages are ane. I) → HomA (b . n+1 Thus x0 m = xn0 . M Now suppose Since  s > sij .

F ) ≈ f∗ G . Suppose that Y is not ane. x (following http://mathramble.wordpress.3 is a q. F ) so that β : H om (f∗ M .5.2 III. F ) → H om (OY .→ X a minimal element of Σ. F ) ≈ r F and H om (f∗ M . . r For M = ⊕ (τj )∗ (Ej ) . F ) which is an isomorphism at the generic point.c. r If F ∈ Coh (X). By exc II.→ X . Thus β : f∗ G → F .→ L. then β is an isomorphism at the generic point. H om (OYr . Y are irreducible. the generators ej of f∗ M give a morphism α : OY → f∗ M which is an isomorphism K r ≈ L at the generic point of Y . H om (f∗ M . and τj : Uj . F ) is q. and let Σ Z .5. . there We have an isomorphism By exc II.2 x Apply H om (.4..4. then f∗ G is acyclic. then (τj )∗ (Ej ) is quasi-coherent since Uj are noetherian. f∗ : Coh (OY ) ≈ Qcoh (f∗ OX ) so r that H om (f∗ M .17. F ) . If Ej is the coherent sheaf sj · OUj . and since f is nite this sheaf is coherent. so by thm III. Z is reduced. F ) to α gives a morphism H om (f∗ M . b. hence F is acyclic so that Y is ane.7. x X.. If D = ker β then D ∈ Qcoh (Supp D). f nite implies there is a basis {e1 .3. G with H om (f∗ M . F ) ≈ f∗ G . Thus as Z is minimal. The morphism f gives an inclusion K . we can assume that proper closed subschemes of Y are ane.. As a nite morphism is ane. As taking H om commutes with taking stalks. r r Now take β = H om (α.3. for f ane.c. er } for L over K . we may assume (by the previous excercise) are not ane. and we are base changing from Y to Z . F ) ≈ F . Note that H om (OY . Taking LES in cohomology of 0 → D → f∗ G → F r → 0 and using induction gives F r .4. G a coherent OX -module. F ) → H om (OYr . Let 3.4 L the function eld of X and c.. f is nite. By a previous excercise. OX ).17. D is acyclic. so WLOG let f := f |f −1 (Z) since nite morphisms are stable under base-change.4.com/2013/03/14/chevalleys-theorem/ ) K the function eld of Y . where ej is represented by a sj ∈ Γ (U. F ) is an f∗ OX -module. then by (b) we can nd G ∈ Coh (X) and β : f∗ G → F which is an isomorphism at the generic point. since 3. Let 162 . Since Z is minimal Supp D is ane.

5 Let III. F )  λi C p (V. y −1 ] → k [x. y.4. After some manipulation of indices. y. = (dλi α)j0 . y.4 x  p ≥ 0 ..3. 0 Then H (U. y −1 ].. Then Uxy = Ux ∩ Uy = Spec k [x. Uy = Spec k [x.. F ) b.7 d d / C p+1 (U. there is an induced morphism from restriction F Uλ(j) → F (Vj ) and hence an induced morphism C p (U. y. y −1 ] → 0 . x−1 . F )  and thus a morphism C • (U. y. α ∈ C (λ dα)j = (dα)λ(j0 ). x−1 . F ) λi+1 / C p+1 (V.λ(jp+1 ) |Vj0 . 1 Thus H is generated by monomials with negative degree 3.4. OU ) ≈ ker d ≈ k [x. y −1 ] .jp+1 = . The cech complex is therefore −1 0 → k [x. III. The dierential is given by d (f1 .. y. F ) → C p (V. y.. x−1 ]..4. This gives a commutative If square C p (U. F ) 3. . . then for each (p + 1)-tuple j ∈ J s. F ) .4. F ) → C • (V. f2 ) = f1 − f2 . x 163 .3 g nice x Ux = Spec k [x.6 in both x and y. x ] ⊕ k [x.. P H 1 is k [x. y].4. x−1 . we nd p i+1 that for any i. y −1 ] / { axi y j | i ≥ 0 or j ≥ 0}..jp+1 .

Let 164 . The λi from (a) are induced by maps of chain complex. In the 1 ∗ ∗ language of cech cohomology. F ). Thus {gab } is a cocycle in C (U. or in multiplicative notation (since we are dealing with O .0 → F → I • be an injective resolution of F which gives a unique up to homotopy map C • (U. Ui an open covering of X on which L is trivial. τ23 τ31 τ12 = 1 1 ∗ 1 ∗ .9 (F )) = 0 ˇ r (X. ˇ 0 (X. F ) → I • . x Weibel 5. F ) → I • is homotopic to C • (U.8. H s (F )) → H r+s (X. H s Hence H 3. F ) ˇ i (V. (You can recall from an excercise in chapter 2 that vector bundles correspond to locally free sheaves so line bundles to invertible sheaves). O ) must be a collection of sections in τij ∈ O (Ui ∩ Uj ) ∗ such that τ23 −τ13 +τ12 |U1 ∩U2 ∩U3 = 0. I will assume L is a holomorphic line bundle. F ) /H H Let &  H i (X. we have ˇ (U.4.8 c. F ) 3. H 1 (F )) = 0 by so we have the required equality.5 x Since this is the case that's important to me. F ) → C • (V. H Now H 0 (X. and by uniqueness. Etale 10.4. F ) → I • and since homotopic maps give the same thing on cohomology. C • (U. O ) and thus gives a cohomology class in H (X. III. ∗ Then gab : (Ua ∩ Ub )×C → (Ua ∩ Ub )×C gives a section of O such that gab gba = 1 and gab gbc gca = 1 .3 gives a spectral sequence Milne. O ) .4. a cocycle in τ ∈ C (U.5.

Recall that to have 0 ∗ ∗ a cech coboundary τ . g − h ∈ (f (x.1. y)) 7→ (gxy . Removing U0 ∩ X X as well.4.y] i > −d − j so that the monomial spanning the image in fk[u. τ is in the image of the boundary map. Write f 0 = f0 + f1 + f2 where the monomial terms xi y j of f0 have i ≤ −d − j .4. O ) represented by a collection σi ∈ O (Ui ) such that δσij = σ (σj − σi ) |Ui ∩Uj ∈ O∗ (Ui ∩ Uj ) and such that δσ = τ or in multiplicative notation we need σji = τij .4.y −1 ] k[u. 3. v) .y.1) Writing u= x0 . v = xx12 . y ] . 0. h) in the kernel.0≤j≤d aij x y with a0d = 1 (so when we plug in (1. 1)) so g − h = f f for f ∈ k [x.v) f (x. 0) to f it doesn't give 0).5 3.4.y. Let σa : Ua → C the restriction of σ . so this is an ane cover. 165 .1) 0 0 −1 For (g. we need σ ∈ C (U.y.v] k[x.y. and of f2 have k[x. and y = xx12 We have a cech complex: k[x. y. Since (1.11 III. On Ua ∩ Ub we therefore gab gab σb · σa = σb or hab =: τab = σa so τab denes a cech coboundary. Take the long exact sequence. y ) − h (x.L ≈ M .4. x1 at the constant term. Closed subschemes from the cover we still have an ane cover.v] of f0 f and the image of f1 f in overlap (u. h (x. and then use III. 0) ∈ / X then P i j f (x. In other words. of f1 have j ≥ 0 . 0.y] −1 −1 ⊕ → where (g (u. then L ⊗ M is trivial. 1) = 0≤i≤d.v) f (x.1) f (x. y) . y. ∗ ∗ So if L ≈ M . f (u.6 x Use the stalks to check exactness. then L ⊗ M ∗ gives a cech coboundary. y.1. Thus two invertible sheaves are the have hab Now we want to show that if same i their dierence is a cech coboundary.7 x The standard cover of P2 consisting of Ui = D (xi ) gives an open cover of of ane schemes are ane.10 III. x = xx02 . so there are transition functions {τab := gab /hab } for L ⊗ M ∗ which give a nowhere vanishing section σ .

→   i i X. G ) → H i+1 (X. Any monomial x y with j ≥ 0 is zero in the cokernel i j i j−i since there is (0. 166 r+1 anes. h0 are constants.5. Thus (g.2. if the are in (f (x. E ) → H i (X. F ) = 0 for i > r. Similarly.I aij x y . h) gives the same element in the kernel 0 as if g. (1. x2 ) = f + a0 x0 for a0 nonzero. x1 .c. OX ) = 1. On the other hand.5. 0. gives that F ∈ Coh (X ) if the theorem holds for locally free..Note that f2 f is in the image of the boundary of the cech complex so either i + d ≤ −j or j ≥ 0. Fa are coherent subsheaves of F . Thm II. Thus P i j d ˜ write f = j<0.12 i≥d . since X is separated.4. . OX ) = 21 (d − 1) (d − 2) . This gives a LES in cohomology · · · → H i (X.4.8 x cohomological dimension cd (X) = n and assume that F is q. we can compute using cech complex of r+1 Then C = 0 and so H i (X.4.18. 0) maps to it and it is likewise 0. y ) mapping to it. E is ite rank locally free. then (u v .a.14 i>n c x. limFa ≈ limH i (X. By Grothendieck's theorem. Assume a0 is 1 since scaling by units doesn't change anything. y.) = 0). E ) → · · · . Hence the kernel is (a. But then g = f0 f + g0 and h = −f1 f + h0 where g0 .. F = limFa . G ) = 0 for i > dim X so H (X. 1)) = x + f˜ (x. Fa ) ≈ lim 0 III. h are constant and thus as if g = h. 1) satises 0 ≤ i < d . 0.15. Note that polynomials with 1 ≤ i < d and −i < j < 0 are not in  the image of the  d boundary map by what we have just said. y. III. 0 ≤ j . 1) . −1 i j Now f in the cokernel is a polynomial in k [x. b. By thm 3. x . then they should have some 3. 0 → G → E → F . Using 4. Thus dim H (X. (Thus when we plug in (1. a) for a ∈ k so that dim H (X. y. F ) → H i+1 (X. x is a quotient of a nite rank locally free sheaf. 1) where f˜ (x. H (X. 0) it won't satisfy f (.4. OX ) ≤ 2 that is how many there are. y.13 → → → for i > n.9 we get H (X. y. if j ≥ i . then the outer terms of above are 0 so the inner terms are i i equal. F ) ≈ H Suppose that By exc II. Thus d we have x = −f˜ (x. Thus dim H 1 (X. Thus we can rephrase the cokernel as the 1 i j 1 (d − 1) (d − 2) since sums of monomials aij x y with 1 ≤ i < d and −i < j < 0 . y ]. 0) is not a point so f (x0 . F ) = 0 for i > n.5. For 3.

x3 .4. x4 . OX ) . OX ) → HY41 ∩Y2 (X. OX ).4. x2 ⊕ · · · ⊕ k x1 .9 x 2 Via exc III. OX ) ≈ HY41 ∩Y2 (X. x4 . OX ) ⊕ HY42 (X. ∪ (X − Hr ) which is covered by r open anes.8.. 0.4. OX−Z ) → HZ (X. x−1 we have H (X − P. Similarly. x4 → · · ·    i j k l −1 −1 −1 3 k x1 .4.2. HY3i (X.8. OX ) = H (X − P. OX ) → H (X − Z. x−1 → 1 4     −1 −1 −1 −1 d1 k x1 . x−1 ⊕· · ·⊕k x1 .5. i+1 i i i+1 we have an exact sequence H (X. x4 . ∩ Hr .e.d. x2 . 0)} then by mayer vietoris. x4 Mayer-vietoris gives 3.4. x3 . Now X − Y1 is a complete intersection of codimension 2 and thus cd (X − Y1 ) ≤ 1 by 3 2 2 (e) of the last exercise.3. OX−Y ) 6= 0 . OX−P ) 6=     d 0 .2. By cech cohomology on the cover Ui = D (xi ) using the complex k x1 . OX−Y1 ) = 0. OX ) → HY41 (X. x3 . Using 3. x4 . and thus X − Y = cd (X − Y ) ≤ r − 1. 1 . Using By exc I. OX−Y2 ) = 1 H 3 (X − Y2 . and same for Y1 . x2 . OX−Y2 ) = 0 .10 (starred) MISS 167 . 0. x3 .15 e. then III.starred 3. k. ane. OX ) → H (X. for each (c). we must show that HP (X. OX ) ⊕ HY32 (X. By exc I. then by exc III. x2 . OX ). OX ) so the middle terms are equal.4. OX ). OX−Y ) ≈ HY3 (X.17. In this manner. x2 . x2 . x2 . H (X − Y2 . x3 . x1 . OX ) = H (X. H (X − Y. OX ) → HY3 (X. 3. x3 . 4 3 If P = Y1 ∩Y2 = {(0. i i+1 2 H (X.3. 0. X − Hi is (X − H1 ) ∪ . l < 0 6= 0 . OX ) = H (X − Y2 . we can show H (X − Y. The mayer-vietoris sequence gives us that HY3 (X. j. Thus HY (X. OX−P ) = x1 x2 x3 x4 : i.16 i.17 III.. x Y = H1 ∩ . x3 . x4 ... If Z ⊂ X is closed.

This direction will p. 168 . the sheaf I i |U is injective and the restriction 0 → F |U → I • |U is an injective resolution of F |U . C q (F . q) = H (X. U) for p = 0 and 0 otherwise.4.1 Let III. F ) = Pn (−1)i (dim kerδ i + dim kerψ i ) = Pi=0 i i i i i P (−1)i (dim kerδ i + dim kerψ i + dim kerφ − dim kerφ ) = P (−1)i (dim kerφi + dim kerψi ) + (−1) (dim kerδ − dim kerφ ) = χ (F 0 ) + χ (F 00 ) .11 x 0 →QF → I • an injective resolution of F .q → E0q+1 give are the usual dierentials on ˇ q (F . F ) by computing Thus the cohomology of the total complex is isomorphic to both H (X. 3. F ) and H rst clockwise and then counterclockwise.. J ) for q = 0 and 0 p • q else.2.q := H q (E1p.q to E0 Next we will compute the spectral sequence going in the counterclockwise direction.5 x Cohomology_Of_Projective_Space 3. E00. H (E1 . I are asque by thm III. For any open U and i. i. the cohomology of F |U .e. U) for p = 0 and 0 otherwise by assumption. U) and thus E2p.q := H p (E0•.4.4.5..3.q ) are C q (F .18 III.• ) is H Let p. P χ (F ) = ni=0 (−1)i dim H i (X.1 x g φi be the map induced on cohomology between F0 and F coboundary.5.<ip I p Ui0 ···iq .5 III. We will compute the spectral sequence associated  = i0 <. so H (E0 ) = Γ (X. F ) for q = 0 and 0 otherwise. • ˇ • (U.• p q p give the cech cohomology of I . and ψi the map for F → F 00 and δi the . On the next page we get . Thus E1p.

OX (1)) such that F (−1) → F is injective. then one of these points contains a nonzero multiple of xn = 1 which contradicts the fact that this point should be a prime ideal of the ring. Thus we are done if the support G is smaller than the support of F . For axn . For such injectivity.S. F (n)) = 0 for i > 0 and n  0. Thus 169 . χ (F (n)) = dim (H 0 (X. Here hyperplanes ∗ are given by elements of V which are k -linear combinations of x0 .5. so we assume k is algebraically closed.E. . and thus innite. and we have an open ane such ∗ ∗ as Spec Sym V with basis x0 .2 III. dim SuppG = dim SuppF − 1 .. is additive on S. Note for example if all hyperplanes pass through nitely many points. xn−1 and xn = 1 . since our eld is innite. we can consider a nitely generated module over a noetherian ring.2.3 b. F (n))) = dim Mn .5. and by primary decomposition theorems.. then P (z) = 0 satises the requirements.. we must avoid nitely many asociated primes to get injectivity (zero divisors are set of union of associated primes).. OX (1)) such that the map F (−1) → F is injective (or just injective on induct on the dimension of the support of Now we want to nd an ane base).. xn ∈ V . By III.. Therefore we achieve x an exact sequence 0 → F (−1) → F → G → 0..5. F (m − 1)) . . x In this manner we achieve an x ∈ Γ (X.2 x First replace F by j∗ F so that we may assume X = Prk . We F . If this dimension is 0. F (m)) = χ (X..5. . xn−1 ] . x x ∈ Γ (X. This gives (since by a previous problem Euler characteristic χ (X. G (m)) + χ (X. But since Gp ≈ Fp /xFp is trivial if x is a unit in x ∈ / p. x g H i (X. that of 3. Thus we want to nd a hyperplane missing all of the nitely many associated primes of the nitely many elements of the open cover. we have supp G = suppF ∩ V (x) so by Hauptidealsatz. a ∈ k if all hyperplanes pass through nitely many points.. xn = ξ gives the spectrum of k [x0 .3. Dimensions of cohmology are preserved under change of base by the at base change theorem.

a) III.3.5. Now just use III.5. 2 III. x g Important genus is birational invariant for curves!! Using. Now using genus degree. OX ) = k (I.5. Thus birational =⇒ isomorphic =⇒ arithmetic genus is the same. a birational map between projective curves gives an isomorphism.5.2.4 III. But a conic.a and c. the valuative criterion.4.4 x g 170 .5. x So note =⇒ cf (−1)r (PY (0) − 1) for the old one. 3.4.6 =⇒ H 0 (X.3.b. x Arithmetic genus Projective 3.5 integral =⇒ variety (II. which is rational by chapter 1 is degree 2 and 2 1 by g = (d − 1) (d − 2) has genus 0. we can extend the map. that χ (F (n)) = P (n) for all n ∈ Z.5. since whereever the map is not dened. g = 21 (d − 1) (d − 2) ≥ 12 (3 − 1) (3 − 2) = 1 (2·) 1 = 1 > 0 for a plane curve of degree ≥ 3.3a.10) (−1)r (χ (OX ) − 1) is the new one .5.3. for instance.

∩ Hs−1 is also H 0 (X.→ Pr . OX (n)) → H 0 (Y. Y = H1 ∩ . Z) = 0 which means only extensions are trivial so that K (Pr ) ≈ Zr+1 which is generated by γ (OLi ) for i = 0. P factors as K (Pr−1 ) → K (Pr ) → Q [z] K (Pr ) is injective. Then Z = H1 ∩ . 3. i i Anything in the kernel of this map must have all ai = 0 by induction on the So P is injective. Z has codimension less than r . OY If Hs has degree d then there is an exact sequence of sheaves: 171 . OX (n)) → H 0 (Y 0 . Assume that Y has codimension r.. i .. 1 By basic properties of projective modules. r . then using II. OZ (n)) → H 0 (Y. K (Pr−1 ) → K (Pr ) → K (Pr − Pr−1 ) → 0 from extension Then exc II. 0 Now suppose that for all complete intersections Y of codimension 0 ≤ i ≤ r − 1 we have shown that H 0 (X. OY 0 (n)) is surjective.5.10.8 III. Since Y is a complete intersection. i = 0. Thus we only have to show that H 0 (Z.. If the codimension of Y is 0.. K (P ) has a subgroup Z with basis OLi .  i:P  P P i+m i+m Now χ (OPi (m)) = so ai γ (OLi ) 7→ ai under P . . Note r = 0 is the trivial base case. OY (n)) is clearly surjective. OX (n)) → (n)) is surjective. . H 0 (Z.. r − 1 which r r r−1 is the kernel of the second map K (P ) → K (P − P ) ≈ K (Ar ) ≈ γ (OAr ) · Z via the method of exc II... x First we show (1) =⇒ (2).8.5.. so that K (Pr−1 ) → r r Thus by assumption of (1) and induction. ∩ Hs where Hi are hypersurfaces..6. OZ (n)) is surjective.7 b.10 gives us a right exact sequence by zero. =⇒ (2) for smaller r. using that (1) coecients.4. which is injective for i<r by induction. r r Thus K (P ) is an extension of Z by Z . Ext (Z.Note that the hilbert polynomial is additivity on short exact sequences.6. Thus the map taking a coherent sheaf to its hilbert polynomial is compatible with the structure of the grothendieck group. 3. For a linear embedding Now we do (1) by induction.5 x g We induct on the codimension r = n − q of Y . then OLi = i∗ OPi . so by the induction hypothesis. Suppose Li satises Li ⊂ Lr−1 for i < r . a complete intersection.5.. then H 0 (X.

4.8.i. OZ (n − d)) → H i (Z.b.... F ). ∩ Hr each Hi has degree di . OY ) =   (−1)q (dimk k − 1) + (−1)2q dimk H q (Y. OY (n)) → . OZ (n − d)) → H 0 (Z.0 → OZ (n − d) → OZ (n) → OY (n) → 0 and taking cohomology gives 0 → H 0 (Z.. OY ) is surjective.5. OX ) → H 0 (Y.5. 0 0 Thus if H (Y. then are hypersurfaces . 3.5.. 3. χ (F ) = (−1)i dimk H i (X. 0 0 0 0 Hence H (Z. OZ (n)) → H (Y. OY ) 6= 0. the cohomology groups are k -modules. x g complete intersection cohomology. ∩ Hi is a c. since the left hand terms are zero.9 b. x g P pa (X) = (−1)r (χ (OX ) − 1) . Then H1 ∩ .5. .i. Assume Y = H1 ∩ .6. OY (n)) → . OZ (n)) → H i (Y. OX ) ≈ k . Recall 172 i . OY (n)) = 0 up to codimension r − 1. n By II. r Then pa (Y ) = (−1) (χ (OY ) − 1) = hP  i q i i (−1) (−1) dimk H (Y.1.a. Hr such is a c. OZ (n)) → H 0 (Y. ∩ Hr .11 d. OY (n)) is surjective so that H (X. and the . OY ) − 1 + (−1)q dimk H q (Y. OX (n)) is surjec- tive.. i Suppose we have proven H (Y. the right hand terms are zero for relevant result follows.10 c. The result is then III.2. H 0 (X. OY )] = dimk H q (Y. OY ) − 1 =   q (−1) (−1)0 dimk H 0 (Y.. 0 The rank of H (Y. Y = Pnk . then H (Y. H 0 (X. OY ).... then Z H1 . of this problem. x g By III. We have the short exact sequence of (a): 0 → OZ (n − d) → OZ (n) → OY (n) → 0 and associated long exact · · · → H i (Z. OY ) = k . OX (n)) → H (X. Y is a complete intersection of codimension r in Pk i there that Y = H1 ∩ . for 1 ≤ i ≤ r . has codim 0. OY ) counts the number of connected components so there By part a. ∩ Hr−1 If Y is only one component.1. so setting Z = H1 ∩ . 0 By (c) all higher cohomology is 0 so the H form a short exact sequence... OY ) = [0 + dimk H q (Y. By III..5...3. sequence: Using induction. then 3.

0)) = k 6= 0 . Thus using exactness we can nish the sequence: δ H 0 : 0 → 0 → k → k ⊕q → δ H 1 : k ⊕(q−1) → 0 → 0 → δ H 2 : H 2 (Q. H 0 (Q. z)) 1 1 Exactness gives H (OQ (a − 1. 0)) = 0 since the only constants in k vanishing on OQ (−q. Y (note 1 H 1 Ex III. a ≤ −2 For (1) (a) → OQ (a) → 0 . Now use . 1 ⊕(−a−1) Thus H (Q.3. then H 2 (Q. OQ (a)) → H (Y. OQ ) = 0.2.5. As Y is a disjoint union of P1 's. OY (a)) follows from considering the degree a part of the coordinate rings for the quadric. y. w] / (xy − zw. OQ ) = 0 . since it's a disjoint union. and from known cohomology of 1 version in stacks). OY ) = 0 since it is generated by monomials with negative degree.b. 0) → OQ → OY → 0.5. −a − n) → OQ (−a) → OY (−a) → 0 P1 . OQ ) = k and thus H 0 (Q.5. OQ (−q. is a disjoint union of copies of 173 .6 x curves on a nonsingular quadric For (3) we will consider the LES associated to 0 0 → OQ (−q. H H 0 (Q. a − 1)) = 0 Let a∈Z anc consider the LES associated to 0 → OP3 (−2 + a) → OP3 P3 . Y = P roj (k [x.12 III. a)) = 0 and similarly H (OQ (a. 1 0 0 Since H (OQ (a)) = 0 and the surjection H (Q. and exactness we nd H 1 (OQ (a)) From the known cohomology of For (2) consider the LES associated to exactness and the fact that Y 0 → OQ (−a. and Y which is writen as a quotient of Q. H (Q. gives H (Q. By chapter 1. OY ) → 0 2 Clearly all H terms are 0.5. z. 0) is the ideal sheaf of Y ) are constants. x. OQ (a. 2 Pn (see the H By serre duality. =0 Now consider the LES associated to 0 → OQ (a − 1. a) → OQ (a) → OY (a) → 0 . OY ) = k ⊕q . 0)) → 0 → H 2 (Q. OQ (−q. and using the that pq (Q) = 0 from exc I.7.

There are no global sections in cohomology by (a) as a. IY (n)) = 0. b > 0 and 1 similarly no H of the rst term by (a). 2 Now χ (OQ (−a − b)) = h (Q.8. Thus we have an LES in cohomology H 1 :0 → h⊕(b−1) → k ⊕a(b−1) → H 2 : H 2 (Q.6. n − b) . WLOG 3.7. then OQ (−a. Q is projectively normal. −b)) = 0 → 0 → k → H 0 (OY ) → 0. H 1 (Q.15 Y looks like Y = Y1 ` III. H 1 (Q. i. −b) (n) ≈ OQ (n − a. OQ (−a. −b) → OY ⊗ OQ (0. we can nd H in P × P which is a nonsingular hyperplane section of the embedding. 1 For |a − b| = |n − a − (n − b)| ≤ 1 .5.3. −b)) = a (b − 1) − (b − 1) = ab − a − b + 1 . 1 ⊕(b−1) 2 As in previous part.5. Thus by II. H (Q. OY ⊗ OQ (0.13 b. By a b Bertini. −b) is associated to a closed immersion P × P = Q → P × P . n ≥ 0 . Pulling back 3 gives a smooth curve Y of type (a. H 1 (Q. −b) (n)) = 0 by (a) so Y is projectively normal. The LES associated to IY has 0. −b) → 0 . since Q is a complete intersection. x Y2 which is a copies of one P1 and b copies of the other P1 making up Q.b) in Q ⊂ P . −b)) → 0 → 0 . Note IY2 is at so we have an SES 0 → IY1 ⊗ IY2 → IY2 → OY1 ⊗ IY2 → 0 which is 0 → OQ (−a. −b)) = k and H (Q. Q ≈ P1 × P1 is locally isomorphic to A1 × A1 ≈ A2 which is normal. OQ (−a. OQ (−a. we have H (Q. By the method of (a). −b)) ≈ k ⊕a(b−1) .2 says that OQ (−a.5. IY ) = H 1 (Q. −b)) = 0 . and by (a). x H 0 (IY ) = 0 . OQ (0. (1). OQ (−a.e. OQ (0. −b) (a) = OQ (0. Note that IY (n) ≈ OQ (−a. −b) → OQ (0.14 c. Consider the LES associated to 0 → IY (n) → OQ (n) → OY (n) . Now use exactness of 1 1 a b (2) Thm II.4. (3) Note that 3.5. Y is ample hence connected by lefschetz hyperplane theorem and thus irreducible by Bertini. OQ ) = k . H 0 (Q. For |a − b| > 1.7 x g 174 . a − b) and now use (a).

i n For an arbitrary coherent sheaf F . F ⊗ i L ) via the projection formula. ample Assume 3.17 c.18 d. x g X = X1 ∪ . [N i F /N i+1 F ] ⊗ 0 for j > 0 and m suciently large. i+1 By decreasing induction using the LES associated to 0 → N F → N i F → N i F /N i+1 F → 0 we j i m nd that H (X. by III. (F /IF ) ⊗ L ) = 0 for j > 0 and m  0 . x g nite pullbacks ampleness X . H j (X. i i+1 The quotients N F /N F are coherent OXred -modules so that since Lred is ample. ∪ Xr be all the irreducible components. The rst term is supported on X2 ∪ . ´ ´ dim W ∗ Then we can write c (f L ) = deg (W → V ) · c (L )dim V by the projection 1 W V 1 Suppose that L is ample on i Nakai-Moishezon we are done. i n i ∗ n i ∗ n Thus 0 = H (X. Rj f∗ (F ⊗ f ∗ L m ) = 0 for j > 0 by ∗ m i m Thus by exc III.1. F ⊗ L ) = 0 for j > 0 and m  0 . then we have an exact sequence 0 → IF → F → F /IF → 0 .. H (X. f∗ F ⊗ L ) = 0 by ampleness of L . 3. For F ∈ Coh (X). ∗ By AMVCCQ it follows that i L is ample.3. IF ⊗ L ) = H (X.8. For F ∈ Coh (X) . 2 r There is a ltration F ⊃ N ⊃ N F ⊃ · · · ⊃ N · F = 0 .5. suppose f L is ample on Y ... j m j m By induction on the number of irreducible components. If F ∈ Coh (X) I is the ideal sheaf of X1 in X . N F ⊗ L ) = 0 for j > 0 and m suciently large. 175 niteness of f.5.. i∗ F ⊗ L ) ≈ H (X. formula. H (X. Lred is ample on Xred . so by . Let and The reverse direction is a consequence of (d).5. H (X. 3.5.16 L x g ample i red is ample =⇒ Lred is ample follows from part (d). ∗ On the other hand. i∗ F ⊗ L ) is 0 for n suciently large. j ⊗m By the exact sequence. then H (X. let N the nilradical of OX . Assume L |Xi is ample for each i. ∪ Xr and the last term on X1 . i∗ (F ⊗ i L )) ≈ H (Y. F ⊗ f L ) ≈ H (Y.

b.6. Neron Models.5. then with a hyperplane in gives an embedding L .c which gives a projective embedding. thus complete. 9.7. 3.7. Completeness follows as in (a) Thus we have projective. O∗ (X1 ) × O∗ (X2 ) →∗ O (X1 ∩ X2 ) argument will give for r > 2. it's separated.20 III.5. is supported at points so has Surjectivity of the map Now an induction Now by (b) the Xi implies our desired surjectiity.5.3.→ Pn and so L By Bertini or some such we can choose this hyperplane missing a nite set of points. X L Pn .2. For the rst part.21 c. Taking cohomology og the ∗ ∗ ∗ ∗ s.s 1 → OX → OX × OX → OX ∩X → 1 gives 1 2 1 2 1 → k ∗ → O∗ (X1 )×O∗ (X2 ) →∗ O (X1 ∩ X2 ) → P ic (X) → P ic (X1 )×P ic (X2 ) → 1 since the skyscraper cd = 0. and pulling back these very ample sheaves gives very ample by 5. x The second part is clear.a x g X is proper. For the next part. are projective so have very ample sheaves. By nonsingularity. x To see nonsingular. see Bosch. 3. OX ). suppose that looks like L (D) L for an intersection of is very ample on ˜ X ˜. note that normal =⇒ regular in codimension 1. Assume ˜ →X f :X is the projection.e.5.22 d.5. then X is projective.8.19 As 3. and thm II. and the singular locus is a nite set of points. x g 1 ∗ Consider two components. From a previous example P ic (X) ≈ H (X. 176 . The existence of m such a very ample L implies that X is projective since it gives an embedding into P .

.8) is false in dimension 3. H 1 (X.3. OX ) ≈ k . The point is that cohomology of X = P2 is known (use duality since it's ω = OX (−3) to get H 2 (X.stanford.10 x g r=3 .9 x g Nonprojective Scheme The only parts of the hint which is not clear is that δ (O (1)) 6= 0 which we need in order to show that δ is injective.5. then we have a short eact sequence. Now use induction hypothesis on both of these sequences. By induction suppose the result holds for r − 1. then if it's nonzero. Now the result follows from Serre vanishing applied to the LES. Essentially we should compute cech cohomology of X w'r't' ω . 177 . 0 Thus we have 0 → P ic X → P ic X → δ k →???.5. hO (1)i is nonzero so we get injectivity.24 If 2. 1 n−2 f We can break 0 → F → · · · → F → F n−1 → Thus we can split into two sequences: 1 n−2 f 0 → F → ··· → F → coker f → 0 and Fn → 0 . t1 . t2 ](t ) d 0 For a possibly easier example of a nonprojective surface see http://math.  Our cech complex looks like 0 → k [t0 . 0 → coker f → F n−1 → F n → 0.5. Since O (1) is a generator. Again by duality.23 III. III. OX (−3)) ≈ H 0 (X. ω) = 0. then P ic X = 0 . 0 So if we show that δ is injective.edu/~vakil/0506-216/216class4 for a much easier example of showing (ex 5.5..

the proof will proceed as follows: (1) use the assumed projectivity U . surely there must be some U with P (U ) → k (x) (U ) not the zero map.x /mx .2 III. x Solution: Suppose to the contrary that P ∈ Mod (X) is projective and there is a surjective map P → OX → 0.3. Clearly k (x) → 0 makes a surjective sheaf morphism.1x I solve this in SummerStudyChallenge2.a. then OX|V (U ) ` on the induced topology for V is the set of functions s : U → p∈U Ap . k (x) denotes the residue eld OX. Thus we will get a contradiction to the assumption that P → OX → 0 is a surjection.→ X the inclusion of an open subset. (2) show that this diagram cannot commute at the level of the abelian group P (U ) whenever P (U ) → k (x) (U ) is nonzero. For any sheaf out of k (x) x ∈ X. Also. Since OX|V is ane.6. j : V .1 III.6.6. As P → OX → 0 is a surjection. if nonzero.6 III. I will abuse the notation and write k (x) for the sheaf ix (k (x)) . is the sheaf dened by ( OX |V (U ) U ⊂ V OV (U ) = 0 U 6⊂ V For where OX |V is the sheaf Recall that OV = j! (OX |V ). If U ⊂ X is arbitrary and x ∈ U . iP (k (x)) (U ) = 0 P ∈ /U P = x. k (x) (U ) will be of P to create a diagram of sheaves involving an arbitrary open set 178 . then we can choose a distinguished open neighborhood V 3 x properly contained in U.pdf under the section Ext (it was a problem given in Dan's class). for j −1 OX = lim OX (V ) . We can make a using the skyscraper sheaf so ( k (x) P ∈ U . 3. First I recall some denitions. V ⊃j(U ) Now that we have made some denitions.6.2.6 x Ext Groups and Sheaves 3.

this contradicts the denition of 3.7. and the denitions of a quotient of we see that ϕ 0 → ker (ϕ) → P → OX → 0 is exact. OX → k (x) → 0 and therefore P → k (x) → 0 are surjective. In the case of P ∈ Qcoh (X) . Also. in the case that as an ascending union of coherent sheaves.x /mx ∼ = Ax /mx .3  k (x) (U ) = k (x) P → k (x) being surjective since Since U has been chosen k (x) 6= 0. which is an ascending union.5. So take the map dened by projection of s to Ap composed with the quotient map (or perhaps in the other order).OX. get an exact sequence 0 → ker (ϕ) → Pi → OX → 0 S i since OX = i ϕ (Pi ) . we note that by II. x ϕ P → OX → 0.7. 179 P = S i Pi and . This denes a morphism of sheafs since we have commutativity of / OV (U )  / OV (V 0 ) k (x) (U )  . We therefore have = OV / P commutative since P  /0 k (x) is projective.6. OX denition. now we take the associated long exact sequence and get: for some 0 → H 0 (ker (ϕ)) → H 0 (P ) → H 0 (OX ) → H 1 (ker (ϕ)) → · · · . By modules is an OX -module Solution: Assume morphism of OX -modules. arbitrarily. P ∈ Coh (X) . In any case. b. Using 1.5. If we evaluate this diagram of sheaves at U. then we can use II. Also. the ker (ϕ) is a subsheaf of P. the kernel of a by denition.15 to write P ker (ϕ) will be coherent. Also. this is OV (U ) = 0 7 / P (U ) so that whenever P (U ) → k (x) (U ) / 0 is nonzero. restriction k (x) (V 0 ) So now we have OV → k (x) → 0 and P → OX → 0 surjective (check the stalks). 1. the diagram will not commute.6.

The N i are f. N )) by denition of ext. N ) i n n i n hi (homA (An• .4 III.a. we can nd n such that H 1 (ker (ϕ) (n)) = H 1 (P (n)) = 0. so since Γ [P (n)] → Γ [OX (n)] is surjective. Since OX → k (x) → 0 is exact (look at the stalks).5 b.N ˜ = Exti (M. N )∼ is q. and A is noetherian.6.ker (ϕ) will be coherent. x by (a) proof.6. Recall that twisting a skyscraper sheaf gives an isomorphic skyscraper sheaf. Thus H (P (n)) → H (OX (n)) is a surjection.g.6 III.2.g. there is a resolution of M by nite rank free A-modules A i . G ) |U = E xtiU (F |U .2.1. we assume to nitely generate A  X = Spec A is ane. Thus E xti M modules M. N ) by exc III. Using III. Since M is ˜.5. G will correspond ˜ i ˜. which is h (N • ) .4.3.6.6. Then ExtA (M.c. ExtiA (M. N ) .2. E xtX M X A Since n i f. since N is so that h (N • )   ˜. we can nd P (n) . we get a commutative diagram: OX (−n − 1) / P via the composition P → OX → k (x) . since 0 0 is a surjection. since twisting is an exact function.7 . Then F .N ˜ is q. N .6. G |U ) by thm III.4 x 180 = = . Thus E xtiX M A 3. this implies that Γ (P (n)) → Γ (OX (n)) Using III.2 and III. G ) is coherent i for every open ane subset U = Spec A E xtiX (F . 3. we get the diagram: O9 X (−1) P (n) The global section s ∈ Γ [P (n)] xn0 in /  k (x) / 0 Γ [OX (n)] is nonzero at x.N ˜ = Exti (M.c. x E xti (F .6. However there are no global sections in OX (−1) so the diagram will not which maps to commute.  /0 k (x) Using projectivity and twisting by n . 3.6.

G ) ≈ E xtn+1 (F . x hd (F ) ≤ n then there is a locally free resolution of F of length n. G ) = 0 for i > 0 and G ∈Mod (X) .10A.g. E xt (F . G ) = 0. G ) → · · · gives. i > n. all i ≥ hdF Ox -modules G . i > 0.6. x E• → F → 0 be a locally free resolution of length n.6. nite rank. J ) = ⊕nj=1 hi (H om (OUi . If i 3. Using thm III. we need only show E xti (L |Ui . The we can reduce to showing that E xti (L . G ) → E xt (F .7. F is a quotient of a locally free sheaf: 0 → H → E → F → 0 .6.b. hd (F ) ≥ supx pdOx Fx . we   • n i n i have E xt ⊕j=1 OUi . n n n+1 The LES · · · → E xt (E . we need to show E xti (·G ) is coeaceable. and all .6.5. G |Ui = h ⊕j=1 H om (OUi . Suppose 3.6. G |Ui ) = 0 for n • a collecton Ui = Spec Ai covering with L |Ui = ⊕j=1 OUi .7 III. G ) = 0. N ) = 0 for all x. i > n−1 . argue as in (a). that E xtn (H .5 x i is locally free nite rank.By thm III. + projective gives free. By thm III. if E xt (F . G ) = 0 for all G ∈ Mod (X).1.5 we can nd E xt (F .8 F b. E xti (H . i Thus E xt (F .6. For an injective resolution 0 → G |Ui → I . Thus for i > n−1 . Exti (Fx .6.6. i Suppose E xt (F .2. Composing this with H → E gives the desired resolution of length n. then by thm III. G ) . H the kernel. G ) = 0 for L locally free. hd (F ) > supx pdOx Fx . Thm III. G ) = 0 for i > n and G ∈ Mod (X). G ) = 0 for i ≥ hdF and OX -modules G which is a contradiction by (b). I • )) = ⊕nj=1 hi (I • ) = 0 . Assume to make things easier that all F ∈ Coh (X) is the quotient of a locally free sheaf of nite rank. G ) → E xt (H . Thus If 181 .5. Thus H has a locally free resolution of length n − 1 . This gives a projective resolution on the stalks of length ≤ n.6. using (a). 3. 1 1 On the other hand. Thus Fx is projective and f. By thm III.3.8 gives 0 = E xt (F . Gx ) on the stalks. L .A. G )x ≈ Ext1Ox (Fx . so Fx is free =⇒ F is locally free by exc II. The inductive hypothesis gives hd H ≤ n−1 .9 Let c. For n = 0.

N ˜ . Using the isomorphisms Exti−1 (M. N )) ≈ homA (M.6.10 III. i > 0 by properties of projective and hom (−.A. Then 182 . Thus M → M factors as M → An → M so M is a direct summand of An so M is projective. and N is an f. A → Ext  (N.6. ·) gives this is also a unii n• versal δ -functor since A-mod has enough injectives and homA (·. i > n. N )) . By induction the same holds for i ≥ 1 .g Using A i gives a nite free of M  resolution       ∼ ˜. faceable.7 x ExtiA (M.12 for i>n−1 implies III. i On the other hand. Note that Ext (M.6.e. M ) is surjective. A-module. A) = 0 . M noetherian and f.6. k M is nitely generated so there is an exact sequence 0 → N → A → M → 0 some k which gives  i−1 k i−1 i → Ext A . A) ≈ Exti (M. i Thus Ext (N.11. A) = 0 pd M ≤ n − 1 .4 is a universal δ -functor. which must therefore be If n = 0. A ) → Ext (M.g. I) is exact and thus h (homA (A . N ) are the left derived functors of hi (homA (An• .6. A-modules N by thm III. and we get a LES in ext. suppose Exti (M. if Ext (M. k A free implies Exti Ak .S. i > dim A for all f.5 we compute E xt. A = 0 . ≤n computes Exti (M. K) = 0 so Ext (M. and projective resolution of length Conversely. A) = 0 for i > 0. N )∼ ) ≈ Let Ani a nite free resolution of M. a projective resolution M of length n.assume 3. i > n − 1 therefore gives pd N ≤ n − 1 by induction. above.N ˜ ≈ hi H om A˜n• gives homA (M. A) for i > 1 .6. For zero . i Note that Ext (M. A) → Ext (M. N ) and h0 (homA (An• . A ) = 0 by computing via n−1 n induction on 0 → A → A → A → 0.6.N ˜ ≈ homA (M. N ) → Exti (M.1. then we have an exact n i n sequence 0 → K → A → N → 0.6. On 0-degree terms we have ExtX M n ˜ so by thm III. Thus N has a projective resolution of length n − 1 which gives by the S. A) . N ) = 0. N ). A) = 0. of Exti (M. K) we see the same is true for i − 1 . N ) ≈ H om M ≈ hi (homA (An• .a x Exti (M.g. Thus E xti M ˜. Otherwise. A) = 0.3. A) We will proceed by induction on n.11 b. A) → · · · in LES of ext. 1 0 n 0 Note that Ext (M. Exactness of homA (An . ·)) are ef  0 ˜. x pd M ≤ N for i > n. M projective implies 3. i > 0 so Exti−1 (N. ∼ is an exact equivalence so by III. then pd M ≤ 0 by (a).

sik mij extends to a global section of L ik ⊗ F . and U an open neighborhood. F ∈ Coh (X) .11 this gives a cartier divisor D given by ∗ (U ) . fi )}. the f1i generate an invertible sheaf OX L (D) and fi fi−1 ∈ Γ (Ui .g. If Z = X − U . nij n By thm II. By thm II. N ))∼ ≈ ExtiA (M. x ˜ i for f. then Ui = Spec Ai covers X with F |Ui ≈ M Thus FUi is generated by a nite number of mij ∈ Mi = Γ (Ui .6. ffji ∈ 3.6.14 b. F |Ui ) .5. Ai -module Mi .6.13. L (D)) glue to give s ∈ Γ (X. nik The global section gives a morphism OX → L ⊗ F .8 x x∈X .Z 0 i Z 0 = Z . and fi ∈ mZ 0 OX. Z = Z1 ∪ · · · ∪ Zn the irreducible components. mij is in the image of L → F which gives surjectivity. twisting gives a morphism to F direct sum of the morphisms gives a morphism to F . As s|Ui ↔ fi then Xs = U so x ∈ Xs ⊂ U . Z is irreducible since we can take the product of sections in each component . By thm II. If 183 and taking a .14. We reduce to the case we can assume {(Ui .13 Let III. 3. Thus Z corresponds to a prime weil divisor.6. −nij On Xsij .hi (homA (An• . sik ∈ Γ (X.6. Now take a renement of the cover Ui given by Xsik ⊂ Ui . N )∼ . Lik ) for some Lij . L (D)) under the f fi−1 ↔ f .

5. X gives that pd Fx ≤ dim OX.10 x Duality for Finite Flat Morphism f∗ OX ≈ N ∼ ∈ Coh (Y ) and G ≈ M ∼ ∈ Qco (Y ). Serre . hdF = supx pdOx Fx .3. M ) ≈ Hom (N. m n Write a presentation OY → OY → F → 0 .6. 3. Let F represent • 184 .11.6. (This functor is left exact) HomX (F .6. G ) as the kernel of a map G n → G m of quasi-coherent sheaves.6.x ≤ dim X via exc III. Applying H omY (−. HomX (N . G ) ≈ H om (N ∼ . M ) . we have Regularity of Thus Thus 3. Thus Coh (X) has . there is a nite locally free resolution of F .16 b.8 gives coherent sheafs are quotients of locally free sheaves of nite rank. x This is given in Borel.6.Theoreme de Riemann-Roch.c.6.6. M ∼ ) ∈ Qco (Y ).17 III.5.9 x Exc III.6. G ) gives H omX (F .5.A hd F = supx pdOx Fx ≤ dim X .15 III. ∼ ∼ ∼ Thus by thm II. enough locally free sheaves so by exc III.

f G → H omY (f∗ F . A modules respectively. −) → HomY (f∗ F .18 b. Therefore we have a natural transformation. 3. −) → ExtY (f∗ F .f ! G → H omY f∗ F . −) which is a universal δ -functor. G ) . H omY (f∗ OX . this  i ! i gives a natural map ExtY f∗ F . By properties / denition of universal δ fucntor. and we have an evaluation map H omY (f∗ OX .3. G ) → H omY (OY . Suppose we ˜ and G = N ˜ . By functoriality. G )) . G ∈ Qcoh (Y ). x   α : f∗ H omX F . N are mapping between open anes Y = Spec A. G ) so β : H omY f∗ F . f∗ f ! G . f∗ −) is that same 0-part. f∗ H ) is natural in H . x F ∈ Coh (X).6. G ) ≈ G . and thus f∗ f ! G ≈ H omY (f∗ OX . G ) . where M.  ! Now δ := α → β → γ gives a natural morphism f∗ H omX F . composed with the 0-part of ExtY (f∗ F . By previous problem. HomA (B. As f∗ OX is an OY -algebra. G ) → G . G )) → H omY (f∗ F . where ψW : F (f −1 W ) → f ! G (f −1 W ) is dened by φf −1 W for open We have a map maps to W ⊂U . f∗ −) is the pushforward f∗ . f∗ f G → ExtY (f∗ F . N ) by φ 7→ (m ⊗ 1 7→ φ (m) (1)) . G ) and composing this with ? gives the map we need. f∗ f ! G which is dened by φ ∈ HomOX |f −1 U F |f −1 (U ) . On the other hand. f ψ ∈ HomOY |U f∗ F .6.  ≈ ∼ and f∗ give an equivalence of categories. −) . this is an isomorphism. f∗ −) ?. and thus γ : H om (f∗ F . N )) → HomA (M ⊗B B. The map from (b) given by f∗ H omX (F . HomY (f∗ F .g. ! By the technique of (b). Thus δ is locally an isomorphism. we have a natural map f∗ f G ≈ H om (f∗ OX . i Note that HomY (f∗ F . B. Composing it with Γ gives HomX (F . H omY (f∗ OX . f∗ −) which is the one desired by the hint ! after plugging in f G . f∗ f ! G → H omY (f∗ F . By assumption 185 .19 c. HomX (F . H ) → H omY (f∗ F . Then δ maps φ ∈ HomB (M. X = Spec B so that F = M are f. −) is the 0part of the derived functor which is the universal delta functor ExtiX (F . i i Thus ? gives a map of δ functors ExtX (F .

3.6.20

d. x

i
!
i
. Then by thm III.6.3.c, ExtX OX , f G ≈ H
X, f ! G .  

i
!
i
!
i
As f is ane, exc III.4.1 implies H
X, f G ≈ H Y, f∗ f G ≈ H (Y, H om 
(f∗ OX , G )). As f∗ OX is

i
i
Y, (f∗ OX ) ⊗ G .
locally free, then by exc II.5.1.b, H (Y, H omY (f∗ OX , G )) ≈ H 


i
i
Y, (f∗ OX ) ⊗ G ≈ ExtY (f∗ OX , G ) . Composing these isomorphisms gives ϕi .
By thm III.6.3, thus H
First assume

F = OX

WLOG we have shown for 

F 

locally free nite rank.

Note Coh (X) has enough locally frees and thus there is a locally free sheaf E and an s.e.s. 0 → R → E →
!
F → 0 for some kernel R . The right
adjoint of f∗ is f and so 0 → f∗ R → f∗ E → f∗ F → 0 is exact by atness 

!
of f∗ . The maps HomX F , f G → HomY (f∗ F , G ) give a morphism of the two LES in ext. By previous,  

1
!
1
!
the degree 0 maps are isomorphisms. Also since E is locally free, the map ExtX E , f G → ExtY f∗ E , f G  

1
!
1
!
is an isomorphism. By the 5-lemma, therefore, map ExtX F , f G → ExtY f∗ F , f G is an isomorphism,  

1
!
1
!
and by similar logic, ExtX R, f G ≈ ExtY f∗ R, f G . We can repeat this argument in higher degrees.

3.7 III.7 x Serre Duality Theorem
3.7.1

III.7.1 x g Special Case Kodaira Vanishing

L ∨ has a global section s.
Let Z be the vanishing set of s.
let C some curve intersecting Z .

Then by ampleness of L , C intersects both L and L
positively.

Suppose to the contrary that

3.7.2

III.7.2 x

f ∗ ΩY → ΩX → ΩX/Y → 0 .
ΩX is locally free rank n and ΩY is locally free rank n so f ∗ ΩY is locally free rank n since f ∗ OY ≈ OX .
Since f is proper, K (X) /K (Y ) is a nite separable extension, which gives ΩK(X)/K(Y ) = 0 (basically since
a separable minimal polynomial has every element β a minimal poly P withP (β) = 0 but dP (β) 6= 0, but
dP (β) = (dβ) · P (β) by the product rule, which is 0 so dβ must be 0. ). Thus ΩX/Y is a torsion sheaf.
This gives (for some kernel) an exact sequence at any P ∈ X : 

0 → KP → OPn → OPn → ΩX/Y P → 0 .
Note KP is torsion, since after tensoring with K (X), we get 

dimK(X) KP ⊗ K (X) = −n + m + dimK(x) ΩX/Y P ⊗ K (X) = 0.
By thm II.8.11, we have

186

n

Since KP ⊂ OP , where OP is a domain, then K = 0. Thus we have 0 → f ΩY → ΩX → ΩX/Y → 0 .
n
∗ n
Hence ΩX ≈ f ΩY ⊗ L (R)
? where R is the ramication divisor ΩX/Y .
0
!
On the other hand, the trace map t : f∗ OX → OY gives t : OX → f OY whose cokernel F ts in an exact
!
sequence 0 → OX → f OY → F → 0 . Taking highest exterior powers and using that the sequence splits since
F is invertible gives f ! OY ≈ L (R) . Thus f ! OY ⊗f ∗ ΩnY ≈ L (R)⊗f ∗ ΩnY . This gives f ! ΩnY ≈ f ∗ ΩnY ⊗L (R).
!
Combining with ? gives that ωX ≈ f ωY .

3.7.3

III.7.3 x Cohomology of dierentials on Pn

n+1
p
r
p
p+1
Consider the ltration F for ∧ O (−1)
from exc II.5.16(d) with F /F
≈ Ωp ⊗ ∧r−p O. Note
r−p
r−p
p
p+1
∧ O ≈ 0 , r − p 6= 0, 1 and ∧ O ≈ O for r − p = 0, 1 then F ≈ F , p 6= r, r − 1 . Thus we
n+1
r
have ∧ O (−1)
⊃ F r ⊃ F r+1 = 0. Then F r /F r+1 = F r is isomorphic to Ωr ⊗ ∧r−r O ≈ Ωr . The
r−1
r
quotient F
/F ≈ ∧r O (−1)n+1 /Ωr is Ωr−1 ⊗ ∧r−(r−1)O ≈Ωr−1 . Thus we have an exact sequence
r

0 → Ω → ∧O (−1)

n+1

⊕
r−1

→ Ω

.

Now

∧L

⊕m

≈ (L

⊗r

)

r
m


(one way of showing this is to take a

trivializing cover, choose a local basis, and then look at the transition morphisms) and so our exact sequence
is

0 → Ωr → O (−r)⊕N → Ωr−1 → 0 for suitable N that we don't care about. This gives rise to a long
i
exact on cohomology. Since H (X, O (−r)) = 0 for i < n or r < n + 1 (by thm III.5.1), we therefore have
i
r
i−1
isomorphisms H (X, Ω ) ≈ H
(X, Ωr−1 ) for 1 ≤ i if r < n + 1 . If r ≥ n + 1 then we still have isomorphisms
but only for 1 ≤ i < n .
0
0
0
Now we know that H (X, Ω ) ≈ H (X, OX ) ≈ k (thm III.5.1) and so using these isomorphisms we see
i
i
n
that H (X, Ω ) ≈ k for 0 ≤ i ≤ n . Again, using thm III.5.1, we know the cohomology of Ω ≈ O (−n − 1),
i
n
i
r
and in particular, that H (X, Ω ) ≈ 0 for i < n. Using our isomorphisms above, this tells us that H (X, Ω )
in the region i ≤ r, 0 ≤ r ≤ n. All that remains to show is the region i > r , 0 ≤ i ≤ n and this follows from
Corollary III.7.13.

3.7.4

III.7.4 (starred)

MISS

187

MISS

MISS

MISS

3.8 III.8 x Higher Direct Images of Sheaves
3.8.1

III.8.1 x g Leray Degenerate Case

0 → F → I • be an injective resolution of sheaves on X .

Then 0 → f∗ F → f∗ I is an injective resolution on Y .
i
By hypothesis, R f∗ (F ) = 0 so this second resolution is exact.

The cohomology of F is the cohomology of the complex Γ (X, I )
Γ (Y, f∗ F • ) , and thus the required isomorphism.
Let

3.8.2

which is isomorphic to the complex

III.8.2 x g

Since

f −1 (af f ine)

is ane, then, using III.8.1, 

H i f −1 (U ) , F |f −1 (U ) = 0 =⇒ Ri f∗ F = 0

188

for

i > 0.

3.8.3

Let

III.8.3 x g Projection Formula derived

0 → F → I•

an injective resolution of

F.

Exc II.5.1.d gives an isomorphism of chain complexes
f∗ (I • ⊗ f ∗ E ) ≈ f∗ (I • ) ⊗ E
?.




As f E is locally free, 0 → F ⊗ f E → I ⊗ f E gives an injective resolution by exc III.6.7.
i

Taking cohomology of LHS of ? therefore gives R f∗ (F ⊗ f E ).
i
i−1
i
Note that R f∗ (F ) is the cokernel of f∗ I
→ ker (f∗ I → f∗ I i+1 ) .
i
Tensoring by locally free E is exact so R f∗ (F ) ⊗ E are cohomology of RHS of ?.

3.8.4

III.8.4 x

Ui = Spec Ai a cover of X on which E is free, we have π −1 (Ui ) ≈ PnAi .
j
Hence H
π −1 (Ui ) , O (l) |π−1 (Ui ) ≈ H j PnAi , O (l) |π−1 (Ui ) which is 0 by the
degrees between 1 and n − 1 .
j
Then R π∗ O (l) = 0 , 0 < j < n by thm III.8.1.
n
Similarl reasoning gives, R π∗ O (l) = 0 , l > −n − 1 .
For

in

3.8.5

known cohomology of

Pn

,

b. x

π ∗ E → OX (1) . This gives an s.e.s. 0 → F →
U = spec A is an open ane subscheme of Y where E is iso-

By thm II.7.11.b, we have a natural surjection

(π ∗ E ) (−1) → OX → 0 after twisting. If
n+1
−1
morphic to OY
, then π
U ≈ PnA and the restriction of this exact sequence is
0 → F |PnA → O (−1)X |PnA → OX |PnA → 0 which is the exact sequence from thm II.8.13. Thus we have
isomorphisms F |Pn ≈ ΩPn /U . These isomorphisms are compatible with restrictions to smaller ane subsets
A
A
and so we obtain a global isomorphisms F ≈ ΩX/Y .
n
∗ n+1
The isomorphism ∧ ΩX/Y ≈ (π ∧
E ) (−n − 1) results from exc II.5.16. If we then cover X with open
n+1
n
−1
subsets of the form Ui = PA , Spec Ai are opens of Y on which E ≈ OY
(and so π
U ≈ PnA ), then restricting
i
to these we get isomorphisms ωX/Y |π −1 (U ) ≈ Oπ −1 (U ) (−n − 1) via the isomorphisms just mentioned. Thus we   


n
n
n
n
n /A
have R π∗ ωX/Y |Spec A ≈ R π∗ ΩX/Y |Pn

H
P
,
ω
≈ A∼ ≈ OSpec A (by thm's III.8.2, III.8.5,
P
A
A
A 

n
and III.5.1. ) Since these isomorphisms are all natural, we obtain the desired isomorphism R π∗ ωX/Y ≈ OY
.
189

3.8.6

c. x

The map

π ∗ E → OP(E ) (1) gives π ∗ E (−1) → OP(E )

which is surjective from a locally free rank

n + 1 sheaf.

Thus we can extend to an exact koszul complex:

n+1

2

E ) (−n − 1) → · · · π (∧ E ) (−2) → π ∗ E (−1) → OP(E ) → 0.
p,q
p
•,q
−i−q
There is a spectral sequence with E2 page, E2 = H (L ) and L
= Rq π∗ (π ∗ ∧i E (−i)) ≈ ∧i E ⊗OS 

Rq π∗ OP(E ) (−i) which converges to 0. As E is locally free, then locally X → Y is identied with PnY → Y
n
−i,q
. By previous parts, and the known cohomology of P we nd L
is 0 except when i = q = 0 or i = n + 1
p,q
0,0
and q = n . Thus E2 = 0 except when (p, q) = (0, 0) and (p, q) = (−n − 1, n) . Thus E2 = π∗ OP(E ) ≈ OY 

−n−1,n
by denition. and E2
≈ ∧n+1 E ⊗OY Rn π∗ OP(E ) (−n − 1) . The lone nontrivial dierential dn+1 gives 

0,0
n+1
an isomorphism dn+1 : OY → ∧
E ⊗ Rn π∗ OP(E ) (−n − 1) which is an isomorphism by convergence of the  

−1
n+1
sequence. Hence ∧
E = Rn π∗ OP(E ) (−n − 1)  . In conjunction with (a), (b), , this gives l ≥ −n − 1.

For l < −n−1 , consider the map π
S −l+n+1 E ⊗OX OX (l) → OX (−n − 1) . The projection formula gives
−1
−l+n+1
a map S
(E ) ⊗Y Rn π∗ (OX (l)) → Rn π∗ OX (−n − 1) = (∧n+1 E ) . Combined with known cohomology
n
n
−l+n+1
and derived cohomology of P , this gives a perfect pairing between R π∗ (OX (l)) and S
(E )⊗∧n+1 (E ).
0 → π (∧

3.8.7

d. x

n
First I attempt to show that pa (X) = (−1) pa (Y ).
n
n
0
1
n
So pa (Y ) = (−1) (χOY − 1) = (−1) (h (OY ) − h (OY ) + ... − h (OY
2n
0
1
2n−1
pa (X) = (−1) (h (OX ) − h (OX ) + ... − h
(OX ) − 1) .
2n
0
1
By exc 8.1 pa (X) = (−1)
(h (OY ) − h (OY ) + ... − h2n−1 (OY ) − 1)
For dimension reasons
pa (X) = (−1)2n (h0 (OY ) − h1 (OY
2n
Note that (−1)
/ (−1)n = (−1)n
n
Thus pa (X) = (−1) pa (Y ).
Next I want to show that
III.8.20.1).

3.8.8

) + ... − hn (OY ) − 1)

pg (X) = 0

.

Recall that geometric genus of projective space is

Recall that geometric genus is dened as dimension of the global sections of

canonical isomorphism
such that

) − 1)

ωX |U

ωX/K |U ≈ ωU/K

X is locally projective
dimΓ (X, ωX ) = 0.

and since

has no sections, thus we see that

e. x

clear from part (c).

190

ωX

space, we can cover

0

(example

.

There is a

X

by sets

Ui

3.9 III.9 x Flat Morphisms
3.9.1

III.9.1 x

The induced morphism

U →Y

is also nite type, so assume

U = X.

Exc II.3.18 gives f (X) is constructible, so closed under generization would imply open.
y 0 ∈ Y is a generization of y ∈ f (X) then we can nd x0 mapping to y 0 .
0
−1
If V = Spec B is an open neighborhood of y , then V 3 y and f
SP ec B → Spec B is nite-type and

WTS if 
at.

x 7→ y , and U = Spec A 3 y , then A is a at B -module by thm II.9.1.A(d).
Let g : B → A the induced ring homomorphism and using going-up theorem for g .
If

3.9.2

III.9.2 x twisted cubic

3 2
2
3
We can write the twisted cubic as (x, y, z, w) = (t , t u, tu , u ) which is projection from (0, 0, 1, 0). This is
3 2
2
3
the projection of the family (t , t u, atu , u ) onto the z 6= 0 plane. The cusp at (0, 0, 0, 1) is on w 6= 0 where
Xa is given by (x, y, z) = (t3 , t2 , at).
3
2 2
2
3
3
2
Eliminating t gives k [a, x, y, z] /I , I = (y − x , z − a y, z − a x, zy − ax, zx − ay ).
3
2 2
At a = 0 we get I0 = (y − x , z , zx, zy) .
3
2
Thus the 0 ber has suppose y − x in Spec k [x, y].
For

p

with

x∈
/p

, then

z∈p

since

xz = 0 ∈ p.

Thus the local rings are reduced.
At

3.9.3

p = (x, y) , z 6= 0

gives a nilpotent.

III.9.3 x g

I want to show nite + surjective + nonsingular gives at. By exc III.10.9 (we don't need this the current
excercise to prove III.10.9) we get that a surjective morphism is at i the bers have the same dimension. So
I need to show the bers have the same dimension. Note that a nite morphism is quasi-nite by exc II.3.5
so all the bers are

3.9.4

0-dimensional.

Thus

f

is at.

b. x

191

f : X → Xred so it's not quite specied. w> = QQ [ ] sage : I = I d e a l ( z ^ 2 .A(f ). the assumption is that Of (x). All other points the jacobian has rank at least k [x. y. z)since since nilpotents don't aect dimension). However. xz − yw) by the nilradical. y . So it's clear that the dimension of the bers changes. Note the bers usually have dimension 2 when x. there. If Ox. then Thus g : Of (x). it has rank one. y. dimension 1. w^ 2 . there are no nilpotents. w . But this contradicts hf .Y ≈ Ox. z) of Multivariate Polynomial Ring in x. w) Another singular point is at there's a singular point. Xred = Spec (Ared ). Then Ox. w2 .1. z. y. x ∗ z−y ∗w) sage : I . the rank is less than 2. (since reduction doesn't change dimension). <x .5 by the isomorhpsim.X /mf (x).X ≈ k . w over Rational Field So an embedded point is a nilpotent element at a singular point. z . But this contradicts thm III. | 0 0 0 {1} | 0 0 0 {1} | 2z w 0 x | {1} | 0 −y | z 2w z Matrix R So clearly at y ∗w) | −w 4 : − I o3 = { 1 } o3 x∗ z | 4 <−−− R (x. Let f = g (1) so that z = hf . when one of them is zero. w ] . z .If x is in the intersection of the two planes.9. zw. sage : P. o2 : Ideal of R i3 : jacobian z ∗w . w] / (z 2 .X is a nite rank free 3. z.Y . z ∗w .Y Ox.Y -module.9. h ∈ Of (x). which gives Ox. radical () Ideal (w . z ∗w .9.Y -modules.Y − module by thm III. c. i2 : I= i d e a l ( z ^ 2 .X as Of (x). dim (X) = 2 since just y is ( Thus no embedded points. y . but probably it's the reduction map f : X → Xred . w^ 2 . There are no nilpotents in that local ring however. y. f is at. x ∗ z−y ∗w) 2 2 o2 = ideal (z .X is nite rank free Of (x). So we need to quotient f be the reduction map. there. and when both are 0 has dimension 0 . 2.10. z. at this point. To nd singular points i1 : R = QQ[ x . Note for ane scheme To see Xred = Y . which is n − dim (X). (x. So let 192 . y not zero. x X = Spec A.

X0 lies on z = 0 so IC(X0 ) has z = 0.6 III. (0 : 1 : 0). These points are only on a line together at t = 0. B an f. then IXt = hxz − txy. x So basically all we need to do is compute relate the hilbert polynomial of polynomial gives 3. let then 3. z − t xyi . and from (b) since the problems cooked up in (b) cannot occur in the case that the dimensions of the graded parts are always constant over T. and M U is open in Spec B . 3 1 2 2 2 2 If Y is the closure of IXt in A × A . yz − txy. c. Consider the at family 3. Y0 has no such linear terms.9. z − t xy. for t = 0. Note that {C (Xt )} is the ber {Yt }. 193 A}.3 which states: A a noetherian ring. .3. X(t) with dimk (Sl /Il ). x y − xy i.5 x Very Flat Families (1 : 0 : 0).9. Set U = {P ∈ Spec B|MP is at over III. and that the hilbert polynomial constant gives atness. A-algebra. yz − txy.8 Xt dened by b.9. On the other hand. Y is the closure of a at family over a smooth by one dimensional base so it is at.g.9 dimk (S/I)l . 2 2 For t 6= 0.9.9.9. x This should be clear from (a).7 a nite B -module. Also just recall the denition of the hilbert polynomial. then IY = hxz − txy. However. we are done. so hilbert Since grading commutes with the quotient in this case. and (1 : 1 : t) in P2 .4 x open nature of atness This follows immediately from Matusumura thm 24.

5 in Migliore. (c).11 0 or equivalent to the one from projective space. the image of I 0 is I . we already get a at family. (b). A/I) and the set of innitesimal deformations of Spec (A/I) /Spec A. and ψ : A → B with ψ −1 J ⊂ I .7 x I ⊂ A .a . III.1.9.3. Then t ∈ I since A [t] /I is a D -algebra. 2 0 Let A be a ring . Finally. the image of I 0 is I since the composition (Spec A [t] /I ) ⊗D k → A/I is an isomorphism.9. (c) give an innitesimal deformation of Spec A/I in Spec A . 3. A/I) . we have a First we do the ane case.9. A/I) . A [t] → A sending t to t 0 . Furthermore. Consider commutative square 194 .11. Note that φ 0 0 0 0 0 0 is A/I linear since for (ax + by) + zt. x + x t and y + y t ∈ I .6 x This is Proposition 5. and φ ∈ homA (I/I . the kernel of the composite 0 t 0 0 morphism A → A [t] /I → A [t] /I is contained in I . Thus Spec A [t] /I is 2 0 an innitesimal deformation of Spec A/I in Spec A i (a) t ∈ I (b) under the map.9. 2 Thus we have an isomorphism between homA/I (I/I . The converse of each of these facts also clearly holds. We know the hilbert polynomials are the same and we know by projectively normal that the higher parts of the hilbert polynomial are all 3. I 0 ⊂ A [t]. Then (a). This choice of φ is well dened by (c) and (b). Suppose Spec A [t] /I 0 is an innitseimal 2 0 0 0 deformation of Spec A/I in Spec A .10 d.9. and (c) the kernel of the composite morphism A → A [t] /I 0 → A [t] /I 0 is contained 0 in I .2.12 III. the set of elements a + bt ∈ I is 2 nonempty. I ⊂ A an ideal. Dene φ : I/I → A/I by φ (a) = b . then ax + ax t and by + by t are in I and thus (ax + by) + zt − (ax + ax0 t) − (by + by 0 t) = (z − ax0 + by 0 ) t ∈ I 0 so z − ax0 − by 0 ∈ I by (b). x By thm 9. by the criterion of thm III.2. Dene I ⊂ A [t] by the set of polynomials a0 +a1 t+· · ·+an tn ∈ A [t] such that a0 ∈ I and φ (a0 ) = a1 or 0 in A/I . pp 129 Intro to Liason theory and deciency modules. Note that for ideals I ⊂ A and J ⊂ B . On the other hand.9. The only dierence here is we are assuming projectively normal varieties instead of just normal. then we can dene 2 0 a morphism φ ∈ homA/I (I/I . Given a ∈ I .note that every 0 0 element of A [t] /I ⊗D (t) is a ⊗ t. by condition (b). if we have an innitesimal deformation of Spec A/I in Spec A .

and A = P/J . A) . yz. then γ is mapped to (0. b3 . A) by giving the value on xz. dz. 0) by the linear transformation. and we can  nd all the other images similar using the liebniz rule. xw. If γ ∈ homA ΩP/k ⊗ A. A → homA (J/J 2 . xw. b2 . then γ is mapped to (z. b3 image of yz . In a similar manner.9 x rigid example  homA ΩP/k ⊗ A. yz. A) is a linear transformation dened by where it sends generators. y. Now by dx.13 III. P = k [x. we can determine the image of By previous excercise we want a surjective morphism other generating morphisms. z. Note that xz. 0. A sends dx to 1 and other 0 generators to zero. Note that d (dx) = zdx + xdz . yw . and b4 image of yw .8 (starred) MISS 3. w. z. ? meaning the innitesimal deformations of that space. J is generated by xz. and so the linear transformation is given by   0 0 z w  .  [Spec (B/J) /Spec B]? In the general space we therefore have nice restrictions and glueing works. and these elements give generators of J/J as an A-module.9. yw 2 2 as a P -module. b4 ) ∈ homA (J/J ) ⊂ A where b1 is the image of xz .9.  2 2 Now J/J → ΩP/k ⊗ A sends f to df ⊗ 1 and any morphism homA ΩP/k ⊗ A.9. dy. yz. xw. yw are zero in A so multiplying by x or y kills terms with z z  0   x 0 w 0 0 x 195 . y 0  0 y 2 4 Consider (b1 . w) . B/J) / ∼ ∼ / [Spec (A/I) /Spec A]? . and homA ΩP/k ⊗ A. A) is 2 determined by the value on on xz. xw. w] . b2 image of xw . yw so we can dene ψ ∈ homA (J/J . dw . A/I)  2 homB/J (J/J .homA/I (I/I 2 . yz. dy . III. A is generated by the duals dx . 0. Thus φ ∈ homA (J/J . If γ sends dy to 1 and all 0 other generators to 0. A → homA (J/J . dw and their images respec ΩP/k and ΩP/k ⊗ A are generated ∗ ∗ ∗ ∗ tively. dz .

9. or 3.9. b4 ) is in the image of homA ΩP/k ⊗ A. then put into P2 using IV.16 For III.w but sends xi y j to xi+1 y j or xi y j+1 respectively. Ok I'm going with it. MISS 3. 0 Use III. w] .9..0 (a. b3 .0 (a. and use the fact that genus is birational invariant. b)·[x2 + yz]+(1 − δ0.0 (a. y) k [x. where δ0. b))·[z 2 ] particular.0 (a. hmm it seems to make sense.11 x interesting g P2 it's clear. Thus b1 = wz b2 + xy b3 .  b4 = xy b2 + wz b3 so that (b1 . TX ). Then pulling back gives V ((1 − δ0. x δ0.14 III. = −2H =⇒ TX = O (2) Innitesimal deformations correspond to Now using the fact that KP1 and computing cohomology using the euler exact sequence gives the result.10 x g H 1 (X. b) · [x + yz]) union of closeds.15 b. 196 Now pa (Y ) = 12 (d − 1) (d − 2)−nodes .9. y]. b2 .3. b)) · [z ]) ∪ V (δ0.10. b3 = xy b1 + wz b4 . w) k [z.9. But does that even give a morphism? so suppose we take the closed 2 2 2 set x = 0 in A . Thus b1 = xy b3 + b01 where b01 ∈ (z.. b1 = wz b2 + b001 with b001 ∈ (x. So in and everywhere else a singular conic. 3.9 and just write our family with a characteristic function. 0). b2 = xy b4 + wz b1 . A) . In total.0 (a.9. b) is the characteristic function of (0. Similarly. else embed into P3 . A → homA (J/J 2 . we get a nonsingular conic at 0.

3.10 III.10 x Smooth Morphisms
3.10.1

III.10.1 x regular != smooth always

R = k [x, y] / (f ) , f = y 2 − xp + t . f denes an irreducible polynomial so X is
p
irreducible and dimension 1 over k . Clearly E = K [x] / (x − t) gives a eld extension of degree p which
th
is inseparable since any k, x, y is a p − th root of some element of E . Note that for τ a p
root of t, then
2
p = (x − τ, y) is not regular over k as dim p/p = 2 > dim R ⊗K K = 1 . However for p > 0 away from p
computing the jacobian shows X/K is smooth. Now use thm 10.2.
Let

3.10.2

X = Spec R

,

III.10.2 x g

This is local so we can assume

x.

X, Y

is ane. Note the map on tangent spaces. Suppose

In this case, the sequence given by II.8.12, is exact on the left at
0 → Ix /Ix2 → j ∗ ΩY /k,x → Ω1X/k,x → 0.

x.

f

is smooth at

I.e. we have

g1 , ..., gn around x with dgi forming a basis around Ω1X/k,x . The gi dene g : U → Ank
for an open set containing x which is etale since the dgi are linearly independent. Thus we f factors as an
etale map together with projection to Y .
There are functions

3.10.3

III.10.3 x

Clearly at and unramied is the same as smooth of relative dimension
Suppose

f

is at and unramied. Then

ΩX/Y ⊗ k (x) = 0

0.

so by Nakayama,

ΩX/Y

is

0

at any stalk. Thus

(iii) implies (ii)
Now suppose (ii). Consider 

mf (x) /m2f (x) 

f ∗ ΩY /k → ΩX/k → ΩX/Y → 0

. At

x

mx /m2x

this is

⊗k(y) k (x) →
→ ΩX/Y,x ⊗ k (x) → 0 where the last term is 0. Thus the middle is
surjective and nakayama gives my OX,x = mx . Since we are looking locally, consider a homomorphism A → B
of f.g. k -algebrais with ΩB/A = 0. If p ⊂ A is prime, this gives a point in Y . For S = B ⊗A k (p), then
ΩS/k(p) = 0 by base extension. The primes of S are the preimages of p by Spec (B) → Spec (A) . If q is in the
preimage then Ω(S/q)/k(p) = 0 . By the logic used in advanced conditions for a closed immersion, k (p) ⊂ (S/q)
is of transcendence degree 0 and separably generated, and because it is nitely generated it must be nite
separable. Thus it is unramied, hence etale. Thus (ii) implies (i).

197

=⇒ ΩX/Y = 0 . Thus we consider preimages.
Y is integral.
Let
X

·
·
·

X
1
n = X the irreducible decomposition of X . If 

ˆi ∪ ... ∪ Xn , then dim (Ui ) = dim (Xi ) and wlog we assume Ui ≈ Spec (B) , and
Ui = X\ X1 ∪ · · · ∪ X
f (Ui ) ⊂ V = Spec (A) . This gives a at extension i : A → B of f.g. k -algebras with ΩB/A = 0 . Then
×s
i is injective since for s mapping to 0, A → A is 0 after tensoring by − ⊗ B contradicting atness. If n is
the nilradical of B,then A → B/n is injective, B/n is a domain and thus as in the previous paragraph, with
p = (0) ⊂ A and q = (0) ⊂ B/n , we nd K (S) /K (A) is a nite extension which gives dim (B/n) = dim (A).
Thus (i) =⇒ (iii)
Now suppose (i). By the previous paragraph, unramied

By base change
we assume 

3.10.4

III.10.4 x

Assume it's etale. We are looking locally so we can assume
get that it's at and unramied. Thus using Matsumura, pp
are isomorphisms.

The reverse is in atiyah macdonals.

X = Spec A , Y = Spec B . From exc III.10.3
n+1
n
74, the induced maps mA /mA
→ mnB /mn+1
B

Thus at the limit, the maps on completions are

isomorpisms. The reverse is in Atiyah Macdonald. Alternatively, use Liu prop 3.26.

3.10.5

For

III.10.5 x etale neighborhood x

x ∈ X , f (x0 ) = x

, let

r = dimk(x) Fx ⊗ k (x)

rk f∗ F at x0 ∈ U .
r
0 → kernel → OX
→ F → 0.

which is

After possibly shrinking U , we get a an exact sequenc
f at implies 0 → f ∗ kernel → OUr → f ∗ F → 0 is exact, and similarly if we localize at

By assumption (f F )x0 is free, thus at so the sequence
r
0 → f ∗ kernelx ⊗ k (x) → OU,x
⊗ k (x) → f ∗ Fx ⊗ k (x) → 0 is exact.
The dimensions on the right are the same so the kernel must be 0 by nakayama.
r
Thus OU → F is an isomorphism.

198

x0 .

3.10.6

III.10.6 x g Etale Cover of degree 2. x

So recall the normalization I think is just like a parabola, (it separates the two branches passing through
2
2
the node) so I'm pretty sure its R [t] with the map t → (t − 1, t (t − 1)). In particular, I compute the  

2
2
2

normalization in Summer Problems 2012. Now we use this form the cover: Spec k [s, t] t − (s − 1)

Spec k [x, y] / (y 2 − x2 (x + 1)), x 7→ (s2 − 1)

y 7→ st . (This is two parabolas joined together).
k [s, t] / (t − (s2 − 1)), and k [s, t] / (t + (s2 − 1)). To

and

The two components are the two parabolas

see that

the cover is etale, check that it gives an isomorphism on the tangent cones. The tangent cone is geometrically
the union of (tangent lines) two branches of

C

at each node. But then geometrically, it is clear this gives an

isomorphism. Thus we have an etale cover.

3.10.7

III.10.7 x Serre's linear system with moving singularities

T denote the base locus consisting of P1 = (1, 0, 0), P2 = (0, 1, 0), P3 = (0, 0, 1), P4 = (1, 1, 0) ,
P5 = (1, 0, 1), P6 = (1, 1, 1), and P7 = (1, 1, 1). A generic cubic in P2 is the zero set of V (f ) of f (x, y, z) =
a1 x3 + a2 x2 y + a3 x2 z + a4 y 3 + a5 xyz + a6 xz 2 + a7 xy 2 + a8 z 3 + a9 y 2 z + a10 yz 2 .
For P1 , P2 , P3 ∈ V (f ) , we must have a1 , a4 , a8 are zero. P4 ∈ V (f ) mean a2 = a7 , P5 ∈ V (f ) gives
a3 = a6 , P6 ∈ V (f ) means a9 = a10 , and P7 ∈ V (f ) means a5 = 0 . Thus a cubic in D looks like
a (x2 y + xy 2 ) + b (x2 z + xz 2 ) + c (y 2 z + yz 2 ). Thus d is generated by 3 cubics Ci with intersection T so there
2
2
2
2
2
2 2
2
is a morphism g : P \T → P given by (x, y, z) 7→ (x y + xy , x z + xz , y z + yz ) .
We can show inseparability locally on one of the anes D+ (x) , D+ (y) , D+ (z) . For example on D+ (z) 
2 

x y+xy 2 x2 +x
we have coordinates s = x/z , t = y/z and g is given by (x, y, 1) 7→
, y2 +y . On function elds, we
y 2 +y
Let

h : k (s, t) ,→ k (x, y) , s 7→ x+y
x , t 7→
y+1
x+y
0 = h (s) + h (s) = h (s) + y+1 x =

have
get

x+1 x
. Note that
y+1 y

y · h (t) + +h (s) =

x+1+x+y
x
y+1

h(s)(y+1)+y 2 h(t)2 +yj(ts)+y 2 h(t)+yh(ts)+h(s)2 +yh(s)
.
y+1
2
Thusy h (t) (h (t) + 1) + h (s) (h (s) + 1) = 0 and we can nd a minimal polynomial
2
of h (s) , h (t) . Then u + c is a minimal polynomial that is inseparable of degree 2.

199

=x

. In

h (s)

we

y 2 + c = 0, c a function

3.10.8

b. x

∂f
∂f
2
2
,0 =
, and 0 =
, which are ay +bz = 0,
0 = ∂f
∂x
∂y√
∂z




ax2 + cz 2 = 0, and
bx2 + cy 2 = 0. Thus ay = bz , ax = cz , and bx = cy . There is a singular point 

√ √ √
S=
c, b, a . Each Pi is singular for only one of the cubics in d. P1 lies on yz (y + z), P2 is singular on
The singular points are given by the partial derivatives

xz (x + z), P3

is singular on

xy (x + z)

, ... . Each of these equations results from choosing

a, b, c

in

F2

which

is the same as choosing a union of three lines. Thus these relations give all the cubics with a singular point
2
in the base locus. Note that two dierent sets of a, b, c give dierent singular points so d → Pk is a bijection
so the singularities of d are moving.

III.10.8 x

geometrically this is fairly obvious.

III.10.9 x

So recall a variety: integral separated scheme nite type over algebraically closed eld

k.

Also atness is

X ⊃ Spec A → Spec R ⊂ Y is at. Or that R → A is at. So A is a local
R-algebra, R is regular by assumption We have dim X = dim Y + dim f iber. So really I just
need to show that dim f iber is dim A/P A. But Xy = Spec (A ⊗ k (P )) = Spec (Ap /P AP ) and dimension
works with localization. So we're good. Now this follows from Eisenbud 18.16.b which says if A is CM then
A is at over R i dim A = dim R + dim A/P A .

local so we just need to show that
noetherian

3.11 III.11 x Theorem On Formal Functions
3.11.1

III.11.1 x g higher derived cohomology of plane minus origin.

n−1
Since R
f∗ OU is the sheaf associated to V 7→ H i f −1 (V ) , OU |f −1 (V ) , we can just compute the cohomoln
−1
−1
ogy of Ak −{0} . WLOG let n = 2 , then take the open cover Ux = Spec k [x, y, x ], and Uy = Spec k [x, y, y
]
. The cech complex is 

200

0 → k [x, y, x−1 ] ⊕ k [x, y, y −1 ] → k [x, y, x−1 , y −1 ] → 0 → ...

.

So the rst cohomology group is linear combinations of monomials of negative degree.
Clearly the bers have dim

3.11.2

0

on the inclusion.

III.11.2 x g

f : X → Y . This result is local on the base, thus we assume Y is ane. Since X is projective,
X ⊂ Y × Pn , and we are projecting to Y . By induction, and blowing up a point of Pn we can assume n = 1.
1
1
If y ∈ Y , then since f has nite bers, we can nd z ∈ y × P , z ∈ Y × P \X . This shows that we can take
a smaller open ane and call that X .
n
Thus assume X is ane, dened by a polynomial in x with coecients in A (Y ), f (x) = an x + · · · + a0 .
n−1
Localizing at an gives a quotient of A [x] / (f ) which is nitely generated by 1, x, ..., x
.
Let

3.11.3

III.11.3 x

(Kleiman)

d

Suppose that
By Bertini,

d

is reducible. I claim that

contains

U,

is composite with a pencil.

has no variable singular points outside the base locus so a general member of

components. Fix a general member
components is

d

d≥2

U

of

d

d

has distinct

with minimal number of components and assume the number of

by contrapositive, so that

d

is disconnected / reducible. Let

P

be a subsystem which

with no xed components, and which is parametrized by a line. A general member of
ith component is a 1-parameter family Pi .

P

has

d

components and the
We can factor

P

so that each factor has coecients which are coordinates of a generic point of a curve in

projective space parametrizing
of hypersurfaces in

P

intersects

x,

Pi

Pi .

Then

Pi

is a linear pencil since the degree of the curve gives the number

which pass through a general point of projective space. But since only one member of

this degree is

1.

Note that each of the Pi must be equal to each other. If U1 is a general member of P1 , and a general
x ∈ U1 , then for each i, x must lie in some Ui ∈ Pi . Thus Ui must be a component of U since P is a pencil.
Thus Ui must equal U1 and thus U1 ∈ Pi . Thus Pi are all the same. Thus all components of U belong to P1
so d system is composite with a pencil.

3.11.4

III.11.4 x Principle of Connectedness

201

Since this question is stable under base change, wlog assume

f∗ OX

torsion free.

one on

U

T

smooth gives

is normalized.

f 

at and projective gives

f∗ OX locally free. f has connected bers over U gives f∗ OX has rank
f∗ OX gives an invertible sheaf. Note that the global sections of f∗ OX
of OX . Thus f∗ OX ≈ OT . By a theorem in this section, the bers are

and thus everywhere. Thus

are the same as the global sections

T

connected.

III.11.5*

Skip

III.11.6 - Skip (formal schemes)

MISS

MISS

MISS

202

F ⊗ mky /mk+1 = 0.12 x Semicontinuity 3. → Oy /mk+1 y  i just have to show that H Xy . F ⊗ Oy /my ) = 0 by assumption.8 x higher derived is 0 in a neighborhood We will show that ∧ (Ri f∗ (F ))y (∧ 0. for all k.12. 203 . then by long exact sequence and induction. The tangent space in this case is the kernel of the linear transformation given by the jacobian matrix of the polynomials in the ideal of Y. since the cohomology commutes with the direct product. the result follows. We also have 0 → mky /mk+1 y we → Oy /mky → 0 and since F is at.7 -Skip (formal schemes) MISS MISS MISS III.12 III. this is equivalent i Note that H (Xy .1 x g upper semi-continuous tangent dimension This is intuitively clear as m/m2 is the number of tangent directions.III.  i to H Xy .11. assume Y is some ane variety.12. y k+1 k is a direct sum of copies of Oy /my . F ⊗ Oy /mky = 0 for completion) is By the formal function theorem. proof: Since this is a local result.1 III. Since my /my 3.11. Since the rank function on matrices is upper-semicontinuous.

xn ] a polynomial of degree d.4 III..12. Fy ) is an isomorphism.12..2 III. we have R [x1 . xn ] / (f ) is a free R [x1 .. If srii ∈ k are n coecients giving the change of coordinates are in R. −1 Now if Ly ≈ My then Ly ⊗ My is trivial on the bers so we are done. Over the ring R0 attained by adjoining a1 si 0 0 d−1 which is therefore to R. X which is trivial on the bers Xy . xn−1 ] module with basis 1. I claim that f∗ F = G is 0 By Grauert.3 x Rational Normal Quartic This question has a typo which makes it not quite work. and k = quot (R/p).2 x f ∈ R [x1 .net/questions/90260/trouble-with-semicontinuity for details. This gives the ane case..12.3. Let 3. . then we can accomplish the change of coordinates on the Q Q open set D (a si ) ⊂ Spec R containing p where a is invertible. . where pj are polynomials in the other variables... xn . p ∈ Spec R . π∗ F is locally free rank 1 (call this G ) and G ⊗ k (y) → H (Xy .. . Let f the reduction of f mod p.. f can be written in weierstrass form with respect to xn : f = axdn + pd−1 xd−1 + · · · + p0 . By a change of coordinates. xn 0 a free R -module.4x Suppose that invertible on Y F is an invertible sheaf on ∗ with f G = F .12.3 III... See http://mathoverow. 3. The ane case n covers P . . and 0 6= a ∈ k .. The ∗ ∗ natural map f G = f f∗ G → G is an isomorphism since it is surjective on the bers.12.12. as free gives us atness of the morphism to T . 204 ..

Xy . Suppose that If m > 0. F invertible so (p∗ OX (m)) ≈ F −1 . Then p∗ OX ≈ OY by thm II. So if rank is > 1. Consider M ∈ P ic (X) . IV Curves 4.7.3.6* MISS MISS 4 m. for m ≤ 0 we're done. riemann-roch we can nd that degree 0 is locally constant so degree the same on each ber. My The restriction to the is an invertible sheaf on Thus for a preimage we consider M ⊗ OX (−m) Pn and thus is OPn (m) for some . Injective: p∗ F ⊗ OX (m) ≈ OX . Now use the previous excercise.12.12 in 3284 book).7. then we have p∗ OX (m) = OY and so F must still be OY (see 11.5 III. we are done for m > 0 since the rank will be too big to have S m (E ) ⊗ F ≈ OY If E is invertible. This follows by semicontinuity and since the euler characteristic is locally constant.1 x Riemann_Roch_Theorem alternative 205 0 By (hirzebruch) + eective gives us a sheaf which is . then  m p∗ OX (m) ≈ S (E ) which has rank n+m+1 n−1  . III.11 ∗ Using projection.1 IV. p∗ (OX (m) ⊗ p F ) ≈ OY So (p∗ OX (m)) ⊗ F ≈ OY .5 x Picard Group of projective bundle We will map F × m ∈ P ic Y × Z to p∗ F ⊗ OX (m) .11. Surjective. By thm II. Note this is eective and the restriction is trivial for every y.12.12.

.1.r gives h (nD) = n + 1 − g ≥ 1. n > max {2g − 2.1. 0 0 So nD − (f ) ∼ D for some eective D . n such that deg (nD) > max {2g − 2.10 Then 206 with no poles except at Pi .1. and let D = 2P − Q. 0 Note since deg (nD) = n thus deg (D ) = deg (nD). ∃f ∈ K (X).4.1.e. 0 0 So D only cancels one pole (since eective). and speciality of nD . h (nD) = n + 1 − g .3 g Nonproper Curve is ane Remark II. 1}.1 g Regular except at a point Q 6= P . nd g which is regular except for a pole at Pi . there is a section f f gives a nite morphism to P1 from X . D = (P1 + . f (A1 ) = X is ane. Embed X in a proper variety over k . − Pr + (r − 1) Q) = (f ) ? 0 Now if D cancels no poles. .10. + Pr − (r − 1) Q). N Then f g will be regular everywhere except at each Pi . D + n (−P1 − . Let Let 4. By 1.4. . 0 Then D − 2nP + nQ ∼ (f ) 0 and so since D is eective. Thus D = nPi .. nD − (f ) = D ≥ 0..R. 0 By denition of linear series.1.. regular except at Pi . X\X = {P1 . embed X as an open subset of a complete curve X . To avoid cancelling.2 (e)...2. 1} . Let Choose 4. Using Excercise IV.1..3 x IV. Pr } since it's closed. 0 i. −1 By niteness of the morphism..6. then (f ) is our function by ? 0 Else D cancels a pole.. 0 r. So embed X in such a complete variety..1.1. f has a pole at P .1 x IV. g. By Excercise IV. 0 By R. says that every variety can be embedded as an open dense subset of a complete variety.2 g Regular Except pole at Points Q ∈ X − {P1 .2 x IV. Pr } .. nd N greater than the order of g at each Pi .1. g.

1. To show base point free. IV. we only need to show for an irreducible component of a reduced scheme. 0 dim |D| = h (D) − 1 = deg D − g + h0 (K − D) Since D is eective. By II. at Pi and regular elsewhere.4 IV.3. then X is ane. since we have a nite surjective 4. we need to g = 0. ˜ X is not proper. it's ≤ deg D − g + h0 (K) 0 0 Note the canonical has h (K) − h (K − K) = 2g − 2 + 1 − g = g. we only need to show for a reduced scheme.1. dim |KX − P | = dim |KX | − 1 for any point. then deg KX = 2g − 2 = 2 and dim |KX | = g − 1 = 1.6.4 x By III.5 x g Dimension less than degree Using riemann roch.1.3. there would be a degree 1 morphism P1 . the normalization By IV.1. By Theorem II. since the image of a proper scheme is proper.2.2. f −1 (∞) consists of the poles Pi .4. (not proper then ane) By III. Chevalley's theorem.4.5 ˜ X with ˜ X ane. is ane. IV.3.1.1.2. P1 By IV. Note that |P | −|KX − P | = 1+1−g = 0 and |P | = 0 since if it was 1.1.1.1.7 If X. = deg D − g + g since the canonical has h0 (K) = g.4.6 IV.1.2. Thus we assume X is integral.8 f is nite. f in k (X) with poles f is nonconstant. 0 0 Note that equality holds when h (K − D) = h (K) so when D = 0 4.6 x g nite morphism to Choose g+1 points Pi on Since 4. then deg f ≤ g + 1. there is a nonconstant function Since or 207 . By III.1.7 x g X has genus 2.

5.9 IV.4.1.10  (b) x g Genus 0 is nonsingular.1. OX ) → H X. OX˜ → 0 . δP = 0 by the formula from the question. deg p2 (P ) = degp2 · deg P so 2 = deg p2 .1.1. Let X 4. P ∈X OP /OP → H (X.1.8. 2) .1. Since deg KX = 2 .  4. This projection is non-constant and thus nite by thm II. for a point. (in this case to dim |KX | = 1 ).7.6. 4. a linear system without basepoints gives a morphism to P . f∗ OX˜ ) ≈ H 1 X. 1 By 3.1. Denote p2 : X → P the 1 1 second projection from Q = P × P .s. H 1 (X.n By remark II.8 x g arithmetic genus of a singular curve ˜ X nonsingular projective by Leray spectral sequence that   H 0 (f∗ OX˜ ) = k . it's P MISS 208 O˜P /OP is a direct . Thus we have an s.8 P1 since b. X   P 0 1 1 ˜ ˜ 0 → H X. Thus local rings are normal.e. x g 1 a curve on the quadric corresponding to divisor of degree (g + 1.3. we have a degree 2 morphism to P1 so X is hyperelliptic.5. As P p∈X ˜ O ˜ . then exc III. k P P a X + P ∈X sum of asque. pa (X) = pa X dim O /O = p δP . By DIRP. By thm ∗ II.8. its nonsingular.6.   P   P ˜ + ˜ ˜ Now by exc III.9.

L (D)) ≈ Ext1 (OX . H (X. L (D) By (b). Choose 4.1.1.11 x g dierence of very amples.7. Then D = B − (m + 1) L is the dierence of very amples. 4.1.9 (star) 4.8.1. ωX )≈ 0 0 (−D)) ≈ H (X.1. ωX ⊗ L 0 Ext0 (L (D) . Now use the Riemann Roch formula from part (a). n > 0 such that D + mL is globally generated where L By II.10 g x 209 X. Alternative riemann-roch note that LCI by II.32 Thus by III.IV. .14 IV.7.5. ωX ⊗ L (−D)).6 1 . Choose a hyperplane which doesn't not intersect the singular locus of 4. (D + mL) + L = D + (m + 1) L = B is very ample.12 x g Invertible sheaves are is some very ample divisor.1. assume very ample cartier.13 x.

Then using Hurwitz. Thus 4. We proceed by induction. it's connected. Assume X is connected.9.1 or something). applied to D + P0 gives an invertible sheaf L (D ) where D 0 Then deg (KX − D − P0 ) = 0 − 1 so it has no sections so dim |D + P0 | = D . 4. g (X) ≥ 0 so this only happens for g (X) = 0 . n Suppose there is a nontrivial etale covering f : X → P .2 IV. For a base case we have 2. By exc IV.7.5. 0 0 By exc IV. So f is an isomorphism. So X f is etale.2 f |H is an isomorphism.9. Now let D be a divisor of degree has support in Xreg . 210 .deg KX = pa − 1 = 0 .3.7. n−1 Now let H ≈ P a hyperplane in Pn . Pulling back H so we apply hurwitz H = Pn−1 a simply ∗ gives f H ample.d. Since Now assume by induction that we know n connected hyperplane in P . and X is proper over k since f Then X is smooth over is nite (exc II.2 x Hurwitz Theorem 4. Suppose f : X k since f is etale is a curve (note connected Pi are all simply connected. (obvious).2 x g classication of genus 2 curves |KX | gives a nite morphism to P1 of degree 2g − 2 = 2. and regular imply irreducible . 2g − 2 = 2 (−2) + deg R so deg R = 6.2. → Pn an etale cover. by III.i'll take your word for it). 1 Let f : X → P an etale cover.9.1x g projective space simply connected 1 First I will retype the case for P since I forget how it goes.2. 0.1.9.2. so it's Let connected by lefschetz hyperplane or thm III. then the ramication divisor R = 0 so 2g (X) − 2 = n (−2). we have have 6 ramication points with ramication index 2 (2 is the degree of the map f ). and n = 1 =⇒ X = P1 . So we want to show no nontrivial etale coverings.4. IV. theorem.c. So f is degree 1. i < n . Thus for any branch point.1. f unramied.1 IV. f is separable. ∗ Pulling back H gives f H ample since f is nite.2.

. x g Projection from Thus there are 6 f onto the x coordinate is ramied (branched) at values of ramication points given by the αi x which have one value of z.5 degree conclusion.x nothing to do here.3 b. 4. Note that Now suppose that there is a divisor giving a degree |D| − |KX − D| = 2 + 1 − 2 = 1 so that |KX − D| has dimension 0. Each case is given .2.2. So Consider OX (D∨ ⊗ KX ) . .4. Are you trying to make me do group theory? 4. Note that |KX − D| − |D| = deg (KX − D) − 1 so that KX − D has KX = D . P3 7→ ∞. x We can just nd linear fractional transformations sending 4. 2 map to P1 .4 in Rudin 14.3. so c. Thus KX − D is trivial. x Clear from parts a-d.2. P2 7→ 1.2. deg R = 6 and the genus of X = 2.6 0. 211 P1 7→ 0. By hurwitz.

7 IV. but this time set L to be the ∂f line at innity. WLOG assume vanishes at n = 0 ∈ A1 0 ∂f | b ∂y Q + . ϕ (0) = 0 so X has no constant ∂y Q ∂x Q 2 2 term. b) is | (x − az) + ∂f (y − bz) = 0 since we take ∂x Q ∂y the projective tangent line this time. so X has a nite number of . y) = ax + by + cx + dxy + ey + . Since ∂f ·y ∂y ∂f ∂x ∂f ∂y +x . which is theintersect of the  ∂f ∂f 1 | : | tangent line and L which is the line at innity.3 x inection points gauss map Suppose rst that For Q = (a. then This gives ramied at so x is ∂f | x ∂x Q is a local parameter at ϕ (0) = 0 A2 and o L is the line y = 0.0)} inection points.. ϕ∗ (t) ∈ m20 ∂f ∂y (0) = 0 which gives 0. x multiple tangents. For Q ∈ X . and ∂f ∂x 6= 0 x= y -axis. . 0) in A2 . Tp be the line x = 0 . Thus ϕ : X → P maps Q 7→ − ∂y Q ∂x X On the other hand if 6= 0 near P so we have ϕ : X → A1 . Note that ∂f | ∂x {(0. 4. TQ . ϕ (Q) can be found by setting y = 0 ∂x Q ∂y Q . TP . . x∈C tangent line to 212 to the coecients of the denition equation of the . If t is the local coordinate at 0 (for the y ). We write f (x. then ∂f ϕ∗ (t) ∈ m20 ⇐⇒ ∂y ∈ m20 ⇐⇒ b + dx + 2ey ∈ m20 |x=0 ⇐⇒ b + 2ey ∈ m20 which means f restricted to x = 0 has degree ≥ 3 in y (so it's only the higher order terms).. then ∂f | b ∂y Q ∂f | ∂x Q +a which is . Note that Hurwitz shows the degree of the ramication divisor is nite.8 b.4. then the tangent at Q = (a. and Tp is x = 0 . then to order ≥2 P is the origin in ∂f | a and dividing by ∂f | gives ∂x Q ∂x Q ϕ∗ (t) = . Then on the · y ∈ m20 .2. given by z = 1. b) ∈ X and solving for If and ϕ t P ∈L x. P ∈ / L. Q is mapped to the slope of its tangent line. So recall the gauss map which takes a point x in P2∨ . Which is the same as intersection mult of f with x = 0 is ≥ 3 or 0 an inection point. Q 7→ − ∂f | / ∂f | . ∂f | (x − a) + ∂f | (y − b) = 0 .2.2. again let P = (0.

Now consider the picture So if P1 and P2 and P3 all map to the same point ∗ ∗ on C . and R is reduced since 0 is not on an inection point or tangent line. this is a cusp. 4. all of the tangents at the point of singularity must be distinct. Thus the number of tangent lines is d (d − 1) = deg R . fy (0. We write this In local coordinates. y) 7→ x/y where U is a x ∗ neighborhood of P . Thus R = d (d − 1). then we have a multiple point on the dual curve. y) 7→ (x : y). Hurwitz gives (d − 1) (d − 2) − 2 = −2d + deg R . Since we don't have any such inection points. 1) ∈ A2 . Let ψ the projection from O. but then at the if there was an inection point. note that an inection point on C Pi there is an intersection.9 c. Thus ψ (P ) = 0. P = (0.2. 0) = 0 fyx fyy 2 2 and fxx (0. or the line x = 0 being tangent to X at P . 0) fyy (0. WLOG assume 213 . On the other hand we have a cusp when fx (0. On the dual curve this would correspond to where two transverse branches become less and less transverse.  fxx fxy as = 0 in local coordinates. t a local parameter y 2 of 0.. x g O = (0. L the line at innity which doesn't contain O. 0) ∈ A2 . i. 0) = 0 for a cusp. Now corresponds to where two tangents come together. 0) p +2pqfxy (0. Thus inection points correspond to cusps. 0) − 4fxx (0. 0) q is the square of a linear factor and a node otherwise. ψ : (x. 0)+fyy (0. If y 6= 0 . This condition  2  2 can be rephrased as the discriminant of the polynomial 2 fxy (0. in small neighborhoods on C corresponding to neighborhoods around each Pi there are distinct branches. 0) = 0 . i. Note that ψ is ramied at P when ψ (t) = ∈ m2P .e. and non-zero otherwise. which is the same as the condition that the above determinant be 0. and since we don't have any we have ordinary multiple points. We can dene ψ : U → D (y) by (x. and nally they are the same. Now since there are only nitely many singular points there must be only nitely many multiple tangents.e. we have an inection point when the hessian curve intersects the curve. ramication is therefore equivalent to x ∈ mP .

X is the normalization of X ∗ . gives (d − 1) (d − 2)+2d−4 = deg R so by allroots ((x − 1) · (x − 2) + 2 · x − 4. y.12 d ∈ TQ (X)}.4. An inection point is where there is multiplicity 3 or greater. Since r = 3. x = 2. x g ψ : X → P1 is projection from O . x g ϕ : X → X∗ is nite and birational. 0. The third ex is therefore x = 0 . 2g−2 = −2d+2+deg (R) . x g 3 · 3 (3 − 2) = 9 inection points. Since P ∈ / L . f.0.2. there are points. = d (d − 1) For P not an inection or on a multiple tangent line. and solving for the number of bitan2 gents gives it. Hurwitz gives that if ramication points by (1) of a.2. Note that x = 0 is the line joining the two Since a plane cubic has degree 3. and inections is 3d (d − 2).10 d. Thus the cubic is yz (ax + by + cz) + dx = 0 by computing the intersection with the hessian. Thus by hurwitz. z = 0 are the tangents through 3 the inection points at (0. 1. x) = [x = −1.13 g. (0. then (c) deg R = 3d2 − 5d.0] we have deg R = (d + 1) (d − 2). Let 4. then by (e). these are ordinary. Now choose coordinates x. Since all are ordinary. The map 4. Then pa (X ∗ ) = 12 (d (d − 1) − 1) (d (d − 1) − 2) . we . 0) . x g −1 Note that ϕ (P ) = {Q ∈ X|P −1 gives ]ϕ (P ) = d (d − 1) so degϕ can add an extra 4. the multiplicity is exactly 3. 2 so by genus degree in P . pa (X ∗ ) = pa (X) + ] inections + ] bitangents 1 Plugging in pa (X) = (d − 1) (d − 2).2.11 O be a point not at any of the nite number of inections or multiple tangents. z such that y = 0. 1) .2. 214 . If e. deg ψ = d−1 (recall X is a curve of degree d). X is normal so by the universal property of normalization.

Thus the xi 's have conjugate stabilizers. f has multiplicity r .4. so every point is an inection point. 4. This is z0 x + x0 y + y0 z = 0 since z0 x0 + x0 y0 + y0 z0 lies on X .2.2. y0 . 3z 2 0 6zx fzx fzy fzz This is 0 in characteristic 3. then these form an orbit of GX . then there is no point where all partials are zero. 0) is not in projective space.2. If y ∈ Y is a branch point. fy = x . Thus the number points in this orbit is the index of the stabilizer which has order |G| /r . i = 1.. .4 x g Funny curve in characteristic p To check singularities. The function eld morphism is thus purely inseparable and nite. z0 ) is fx · (x − x0 ) + fy · (y − y0 ) + fz · (z − z0 ) = 0 . x A plane quartic has degree 4 (we assume nonsingular in this chapter).2. Thus by thm IV. 0.15 IV. 3 Tangent line at P = (x0 . and xi . we compute the hessian:     fxx fxy fxz 6xy 3x2 0  fyx fyy fyz  =  0 6yz 3y 2  . s are the points of y . Thus the gauss map is the frobenius. This is z0 (x − x0 ) + 3 3 3 3 3 3 3 3 x0 (y − y0 ) + y0 (z − z0 ) = 0 . so it's nonsingular. we use jacobian criterion.2.5 x Automorphisms f a curve in genus >= 2 P ∈ X a ramication point. Thus at x. To check inection points. 3 3 3 Partials are fz = y . fx = z .14 h.. Since (0. Now plug in to the formula from (f ). 4.16 IV. Let X of lying over 215 .5? X ≈ X∗ .2. ep = r ..

c. Dene det E = ∧r E ∈ P ic Y . x pa (Y ) ≥ 1 R= P P 1− 1 ri  = 0. x g 216 .e.17 If b. If D is a divisor on X . Thus det f∗ OD = L (−f∗ D) .1 (degenerate leray). we assume X and Y are ane.  pa (Y ) ≥ 1 and R = 1 − r1i > 0. then |G| ≤ pa (X) − 1 .11. If we apply f∗ and take determinants we get f∗ L (D) .e. so 2pa (Y ) − 2 + r ≥ 12 and thus |G| ≤ 4 (pa (X) − 1). Now look at the s. then R ≥ 2 42 42 If if and 4. Using some arithmetic. −1 n If D ≥ 0 . Thus we have det f∗ M ∈ P ic Y .By Hurwitz. of L (−D) we get 0 → f∗ L (−D) → OX → OD → 0 . f∗ M is locally free rank n on Y .b gives det f∗ L (−D) ≈ det f∗ OX ⊗ (det f∗ OD ) .6 x g pushforward of divisors E a locally free sheaf on Y of rank r. R = si=1 1 − r1i > 2 . If D is arbitrary. then the equation from (a) is 2pa (X) − 2 = |G| (−2 + R) so R > 2.2.2.s. so by thm III.  P  1 1 so R − 2 ≥ so |G| ≤ 84 (g − 1) . then since f : X → Y nite. 0 → L (D) → L (−D2 ) → OD1 → 0 .2.8. so 2pa (X) − 2 = |G| (2pa (Y ) − 2) + and rearranging this gives the desired equation. then R ≥ 21 .18 IV. Then from the s.19 b. R1 f∗ L (−D) = 0.s. then thm II. Then L (−D) is q. by (a). 4. Let 4. For an invertible sheaf M on X . If pa (Y ) = 0 . write D = D1 − D2 the dierence of two eective divisors. Then f∗ OD ≈ ⊕i=1 Of∗ D −1 . |G| i=1 ri Ps (ri − 1) pa (Y ) ≥ 2 .2. so det f∗ OD = det Of∗ D = L (f∗ D) .6.

Note that L (D) only depends on linear equvialence class.7 x Etale Covers degree 2 A stalk of f∗ OX is a rank 2 free module over the stalk of OY f∗ OX is isomorphic to the stalk of OY . and L (−B) ≈ (det f∗ OX )2 IV.7. x 217 via the sequence . Thus Now 4.22 . −1 By (a) and (b). d. x branch divisor KX ∼ f ∗ KY + R . 4. 2 Thus L = L (−B) = OY . ∗ Exc III.21 locally free rank n. Thus f∗ f ∗ multiplies by c. Thus f∗ KX ∼ nKY + B .2. ΩY ) = (f∗ OX ) ⊗ ΩY .23 b.2.a.2. . −1 ⊗n Thus L (−B) ≈ ΩY ⊗ L (f∗ KX ) . Thus a stalk of L is invertible. n. .6. gives f∗ ΩX = HomY (f∗ OX . L ≈ det L ≈ det f∗ OX ⊗ (det OY )−1 ≈ det f∗ OX 0 → OY → f∗ OX → L → 0 .10.2.20 divisor. L (f∗ KX ) ≈ det f∗ OX ⊗ det (f∗ ΩX )−1 Note 4.2. det f = n so pullback of a point gives a degree n Furthermore. ∗ Now take the determinants of both sides and note that (f∗ OX ) is Exc III.a gives 4. x f ! ΩY = ΩX .2.

then l (K − D) = 0 so l (D) = 2. then thm 3. exc III.5. Let 218 . deg (K) = 4. Thus l (D) = l (D − P − Q) ≥ 2 . whatever remains. 2 Thus |D| gives a morphism to P so it's a plane curve. dim |L| = 2 since L is determined by 2 points.1 x g D ≥ 5.3 IV.c D = X. By bezout. then thm 3. l (D) ≤ deg (D) + 1 ≤ 2 If deg (D) = 2 then l (D) = l (K − D) + 1 < l (K) + 1 = 2. X is smooth since normal. then by exc IV. however improbable. dim |D| = 6 1 so l (D) 6= 2. 4. The give sequence from (a) has a section Thus f∗ OX ≈ OY ⊕ L so 4. The function eld is a c. so l (K) = 3. must be the truth.3.17.5.3 Embeddings In Projective Space alt: 4. nite type over degree 2 4.b.3.2 gives D is generated. f∗ OX → OY .2 x g :a. X ≈ Spec (OY ⊕ L ) by exc III. If deg (D) = 3 .1 If IV. f is etale.5.17. Thus l (D) > 2 . If deg (D) = 4 then l (D) = 3 .2 IV. Then f is ane and nite k . We have pa (X) = 3. X is a curve.24 extension so by and thus X is integral. x σ 7→ (σ + τ σ) /2 . Then g = 2 If Since we have eliminated the impossible.L .10. 1 (4 − 1) (d − 2) = 3 6= 2 . deg (D) = 4.2. dim X = 1 . Now suppose Now by cases: deg (D) ≤ 1 . D is very ample.2 gives D is very ample.3.1.d. be the morphism separated.3.Let X→Y f given by exc II. Since g = 2.3.

4.3.5 dim |D| = 1 and deg D = 2.3. is the corresponding ring homomorphism. then the kernel is generated by quadrics. is very ample so gives an embedding to l dim |K| = 2 By (a) we assume 4. K is a multiple By exc II. Q.3 b. degree 4. D = P + Q.1. Thus by exc IV. K = P + Q + R + S . 4.1. P2 .d. IV.6 IV.4 . θ P1 is projectively normal. so K is not very ample by exc IV. S . then d-uple is projectively normal. then degree K = 2g − 2 = 2 and L (K) ≈ OX (nH) ≈ OX (n) since is very ample. at 2 other points. Then |K| induces d-uple embedding so K If g = 2.3. Contradiction.3. of hyperplane divisor.3 x g Suppose X is ∩Hi of hypersur.4.4 x g Since If very ample. say R.8. Then Now by riemann roch and above. so 2g − 2 > 0.Thus l (D) = 3. l (K − D) = l (D) + g − deg (D) − 1 = 1 deg (K − D) = 4 − 4 = 0.3. then by bezout it hits X . 219 for some n>0 . P.5 we are done. Then dim |D| = dim |K| − 2 = 2 − 2 = 0 since K is is the line through c. x g Part (b) shows we can't have 4.3.3. x g Let K D eective divisor of degree Note thus If 2 on X .

x By part (b).3. OP2 (1)) = 3 . So it's either rational or elliptic. or else genus is actually < 0 there. degree 1 in P so g ≤ (d − 1) (d − 2) it's ≤ 0.  0 Thus exc II.7 b. Now the only other choice is if it's a cubic in P of genus 0. then it's a plane cubic by IV. Ok well we know genus 4.5. Oϕ(X) (1) ≤ dim Then X Contradiction since 4. then it's a degree 2 in P .3. O (n)) We have Thus Thus 4.11 ϕ (X) ≈ X . OP3 (1)) → H (X. Since deg (D) = dim (D) .3. then by exc IV.6. If genus is 0. x g 1 (d − 1) (d − 2) = 21 (3 − 1) (3 − 2) = 12 2 = 1. 2 3 If genus is 1. OX (1)) is injective. IX (1)) = 0 .5. D corresponds to an (n + 1) dimensional subspace in Γ (P1 .8 . so it's a plane 3 cubic. 2 As ϕ (X) is a complex intersection. and d 6= 0. so contained in a plane.3. so there are no 2 2 nodes.4.a gives that dim H ϕ (X) . 0 Thus dim H (X. b. but then it's a line in P .10 ≤ IV.3. ϕ is an isomorphism.. so H (P . (it's a curve in P ). 0 3 0 Thus H (P . g = 0. OX (1)) ≥ 4. then 1 2 projecting down gives a negative genus. 220 H 0 (P2 . 4.4. then if n = 4.5 x g Suppose to the contrary that ϕ (X) is nonsingular. 0 3 is not contained in a hyperplane. x g Projection from a point preserves degree. Are all of these twisted? yes by denition. deg (D) = d and l (D) = n + 1 ≤ deg (D) + 1 = d + 1 ≤ n + 1 n = d.3. c.5.1. x g dim |D| = n. ..9 d.

X has a lower degree than the normalization.3.e.1.3.12 X0 c. g < 3 .3. n ≥ 4 . 4. Consider the LES associated to 0 → IX (2) → OP3 (2) → OX  (2) → 0.X is the normalization of this projection.r. If g = 2 by exc IV. x g g = 1. then X is rational quartic. 2+3 0 3 As dim H (P .3.6 x g Curves of Degree 4 Suppose X ⊂ Pn . IX (2)) ≥ 2 . 1 (d − 1) (d − 2) = 3. OP3 (2)) = = 1 . x is the curve given by projection as in (b). not just a Suppose component). But this contradicts the fact that for a at family the bers have the same hilbert polynomial.3..3. 2 0 3 then dim H (P .14 is not embedded in P3 contradiction. OX (2)) = h0 (2H) = 8 + 1 − 1 = 8 by r.b. then g (X) = 2 3 2 If n ⊂ P \P . and dim H 0 (X. X is rational normal. b. If n = 2. 2 Once we project enough that we are in P use genus-degree formula. then by exc IV. 4.b. 221 . any divisor of degree 3 is not very ample so X If g = 0. As the intersection of 2 quadrics has degree 4 by bezout this intersection is all of X (i.1.8. By exc IV.3.4. If 4.5.13 IV. then by exc IV.

Thus almost all hypeplanes intersect somewhere not tangent or multisecant.3. this is impossible. IV. By exc IV.6.9 x g Recall the tangent variety is a subvariety of P1 × X and thus has dim ≤ 2 . 0) . 4. tp .15 IV. Any curve projecting to this would have degree 4 and genus 2 by using the degree-genus formula for singular curves. tp .18 X is a line. ≤ 1. Using the trisecant lemma (not in Hartshorne) the dimension of multisecants is 3 ∨ Thus the union of these spaces is a proper closed subset of (P ) .3. w coordinates as (0. Choosing a general point gives a projection to If (1.8 x Computing the tangent line at the point point since char k = p (t.3. x g No strange curves in char 0!!!! char k = 0 X has nitely many singular points. We can also parametrize in Thus gives a line pointing in the direction P P3 . y. xy + x4 + y 4 = 0 . t2p ) . t2 p) x.16 IV.9. direction. Now recall tat a hyperplane intersects at d points i not on a tangent line and three points are collinear i they are not on a multisecant.3.17 at each (t2p−1 .3. 0) (1 : 0 : 0 : 0) is a strangent point. is a strange point we choose an ane cover where P is innity on the x-axis as in thm 3. Thus 4. The tangent at 4.3. b.3.4.3. the image is a point since the map is separable in char 0 . 0) points again in (1.7 x By using the jacobian. 0. 0. 0. As the morphism is unramied at nitely many points. 222 .

n For any point Q = (q1 .20 IV. −1 Thus for H ∈ U − ∪I πI (D) . −1 Thus π (D) is a proper subvariety of U . .. For H in a small neighborhood of H0 .11 x g This is Shafarevich Algebraic Geometry 1. the thus to X. qn ) ∈ X close to π1 (H0 ) we can nd a hyperplane H ∈ U containing Q such that the image of U contains a small opens subset. Since we are in P5 at any rate.21 If b..3.. in } ⊂ {1. .... then the secant variety has dimension at most 4.4. then the intersections of H with X vary smoothly with H ..10 x g X ⊂ Pn not contained in Pn−1 having degree d. × X where H maps to the points of intersection of X with H . . qn ) such that q1 ..... v (P ) v (Q) of X. qn are linearly dependent. d} there is a map πI : U → X n = X × X × . n−1 n Since X is not contained in P ... Let H0 a hyperplane which meets X in p1 .19 IV. theorem 9. D is a proper subvariety of X . 223 . the points of H ∩ X satisfy no n of them are linearly dependent... pd .3.3.3. . . Thus for every multiindex I = {i1 .3. r secant variety clearly has dimension at least 4. Thus if th we choose n − 1 of them.. x g r ∈ P5 is a general point on a secant line lies on the plane spanned by C. On the other hand.. . Let (From GH p 249) 4. the n intersection will not be dependent on the rst n − 1. page 136 4.. n Let D ⊂ X the locus of points (q1 .. then pq maps to a conic Any other line on the plane passing through C in X and is also a secant line to r then C and Thus a general point on a secant line to X lies on a one-dimensional family of secant lines to X ..

1).. product of homogeneous linear poly. in order for there to be a singularity there. check for singularities on the ane patches with Maple once we have a candidate....4.. Use halpen's theorem exa [0 nodes for d=2.... 2.10). Either draw the curve.. by (3. 0.. 3.... This gives r = 1 (d − 1) (d − 2) .... and then project it into P ... we substitute in so that the singularity is at (0.. algebraic curves or Michael Artin's Plane algebraic curves lecture notes from MIT)..   ˜ = 0.... is not the square of a linear polynomial.3. d.... X of in in d = 3. More advanced method 1.1) Note we can also check these with Maple on the ane patches with the algcurves package 224 . Use the following polynomials: 2 4 4 2 3 4 degree 4: xyz + x + y = 0 for char 6= 2 or xyz + x z + y = 0 for char 2 or .. P2 having only nodes.(4..12 x g (just explain the advanced method) Basic Method: 1..5] Use the comment on page 314 about Bertini theorem. 4.5) and (3. Prot. and with g X 2 there are irreducible nonsingular curves of degree d in P2 . Now for each singularity. This is because it is quadratic. We use g = 12 (d − 1) (d − 2) − r to nd curves of a certain degree with the right amount of singularities..3. c.. Check for genus in Sage easily b..22 IV......... We write as poly in z d−2 f2 + z d−3 f3 + . Specically a node is a point where the discriminant of This is when f f2 is nonzero...5] compute partials and nd singularities..3... exa [The maximum amount of notes for For any d......... exa[1 node for d=4.. 4. or note that hessian must be invertible to see that we have nodes.... Verbatim... has two zeros 5... to get a curve For any P1 degree d we can embed d. A double point is where f2 is nonzero (Fulton.. Hartshorne says. Now we: a.4. 2 Pd as a curve of degree d .... 6. 5] Using Comment on Page 314 again..

u*v^4 .v^4*w + u^3*w^2 + u^2*v*w^2 + v^3*w^2 + u^2*w^3 .. 3 compute a genus 4 quintic with discriminant of the z terms nonzero: We compute Next. There is also a method using discriminants.u*v*w^3 + v^2*w^3 Maple likes this one too: theorem Corollary [halphen corollary] There exists a curve section (1) (2) (3) X of degree d and genus g in P3 .degree 5: xyz 3 + x5 + y 5 = 0 for char 6= 5 or xyz 3 + x5 + y 5 + x3 y 2 for char 5. page 314). and 5 nodes for a degree 5 curve] 225 . (pp366.note genus is 1 Maple checks this one out: g = 12 (4 − 1) (4 − 2) − 2 = 3 − 2 = 1 so there can be no more singularities. x*y^3 + x^2*y*z+ x^2*z^2 + x*y*z^2 .. i either g = 0 and d ≥ 1. 4. pp331.u^2*v^3 .u^3*v*w . page 350) corollary [ Any curve is birationally equivalent to. exa[2 nodes for d = 4. so I use a computer to help aid visualization.v^5 + u^4*w . then we should be all set. we in P3 . (Hartshorne 3.u^2*v^2*w .y^2*z^2 . Now if we can argue as in the case for the top dimensional genus. and degree 5 with 4.] a plane curve with at most nodes as singularities. for deg 4 . or g ≥ 2 and d ≥ g + 3. 5] Now I want to do the same sort of thing for 2 nodes as 1 node.u^3*v^2 . exa [degree 4 with 2 or 3 nodes... I -u^4*v . 5. whose hyperplane D is nonspecial.8 exa [3. g = 1 and g ≥ 3. or 6 nodes] We have the following table: degree d Nodes 5 5 r g = 21 (d − 1) (d − 2) − r Halphen Condition 1 ×3×4−4=2 (3) 2 1 ×3×4−5=1 (2) 2 have a curve of degree d and genus listed above Genus: 4 5 By Halphen's Corollary. just the computations are harder. Note genus is a birational invariant by Hartshorne I.11.

u^2*v^2*w .v^5 + u^4*w + u^3*v*w .u^2*v*w^2 + u*v^2*w^2 + u^2*w^3 .u^2*v*w^2 + u*v^2*w^2 + u^2*w^3 .u*v*w^3 + v^2*w^3 v*w + u^2*v^2*w + v^4*w .4.u^4*w + u^3*v*w + u^2*v^2*w + v^4*w . t] .u*v^4 . and (f ) + nP ≥ 0 then f can have poles only at 226 P. 3 We also need discriminant of the z term nonzero.1 x g Dene ϕ : k [x. and dene ϕ on x.v^3*w^2 + u^2*w^3 .1 IV.4 IV. If f ∈ k [x.6. and at u=1 and at v= 1 4.4 Elliptic Curves Note: I will pretty much freely use and quote silverman's books for this chapter's solutions 4. exa [ 4 and 5 nodes for degree 5] We will need genus 2 and 1 respectively. For 5 5 nodes. t] → R = ⊕n≥0 H 0 (X.4. For nodes we will need genus 1.So if we have three nodes. u^5 + u^3*v^2 . .4. y.u*v*w^3 + v^2*w^3. OX (nP )) .u*v*w^3 + v^2*w^3 maple likes: 2. by t 7→ 1 ∈ H 0 (X.u^3*w^2 . OX (P )) in thm IV. y as . the two conditions are going to be we need genus of 1 2 (4) (3) − 3 = 6 − 3 = 3 3 since 12 (5 − 1) (5 − 2)−3 = is the number of singularities. y. we will need genus u^4*v + u^3*v^2 + u*v^4 .u^2*v*w^2 + u*v^2*w^2 .

→ PN the embedding. OPr (n + 1)) Γ (Pr . 227 the bottom is surjective. Therefore H (D − E) is a basepoint free pencil.Considering P to be the point at innity. This gives 0 → L (D) ⊗ L (−E) → L (D) → i∗ OE → 0 . As deg D ≥ 3 > 2g −2 = 0 .7. H 1 (L (D − E)) = H 0 (L (K + E − D)) = 0. Suppose we have a similar surjection when twisted by n. and we can nd an f ∈ A2 ϕ. |D| is bpf and Supp i∗ OD is clearly 0 -dimensional.2 IV. OX (n + 1)) since g is surjective then is projectively normal.4. f it satises A2 . Then degL (D) ⊗ L (−E) = d − deg E = d − d + 2 = 2 .s.4.e. This gives a commutative diagram: / Γ (i∗ OE ) ⊗ Γ (nD) /0 / Γ (D) ⊗ Γ (nD) Γ (−E + D) ⊗ Γ (nD) Denote by  g f  / Γ (D + nD) /  h Γ (i∗ OE + nD) Snake lemma gives us an s. Suppose E is an eective divisor of degree d − 2. Consider the s. 0 → L (−E) → OX → i∗ OE → 0 . y − x (x − t ) (x − λt ) is in the kernel so R is a quotient of the desired ring at least. page 126) to get coker h = 0 so coker f  coker g . 4.2 x ϕ|D| : X . OPr (1)) → Γ (X. so again by BPFPT. OX (nP )) = n and also dimension of the desired ring graded n part is n.4. 2 2 2 On the other hand. coker f = 0 and g is surjective. dim H (X. OPr (n)) Γ (−E + D + nD)  Γ (X. OX (n)) The left is surjective clearly and surjective so by induction. supp i∗ OD = Supp E . such an So for any section of the graded ring R. So they must mapping there under be equal. we have a surjection Γ (P . We can yse the base-point rfee pencil trick (Geom Alg Curves I.e. 0 1 0 R. coker f → coker g → coker h .s. OPr (1)) ⊗ Γ (Pr . OX (1)) ⊗ Γ (X. X /  Γ (X. We have a square / Γ (Pr . Thus the right is .R. By serre duality.8. OX (1)) by ex II. (f ) + nP ≥ 0 lives in some n.r.. gives h (D − E) = deg (D − E) + h (D − E) = d − d + 2 = 2. r Since |D| is complete. 0 By r.

y. 228 (x. v (x) + y · w (x). 2y) and factor to get y2 = .4. ∆ = −b22 b8 − 8b34 − 27b26 + 9b2 b4 b6 . Now suppose that [a (x)] (0.. in f rac (k [x. j = ∆4 .4. with in weierstrass form as . c6 = −b32 + 36b2 b4 − 216b6 . This gives h : expand ( 1 / 4 ∗ s u b s t ( 2 ∗ y . y) with (x. + y ^2. Thus we can write an isogeny as φ (x.4. −c (x) + d (x) y) so that b (x) . b8 = b2 a6 − a1 a3 a4 + a2 a23 c3 24b4 . y. so it has a constant term. 6= 2 .4. −1∗h a4 ∗ x − a6 . a regular function can be written as x and y on a (x) . 1) = 0 and [d (x) y] (0.4 IV. d (x) ∈ K (x) . Now replace g)). b2 = a21 + 4a2 . So clearly d (x) 6= 0. c (x) + d (x) y) for a. c (x) are 0 and thus φ (x. b. If E is dened by y 2 = x3 + ax2 + bx + c. a rational function of 4. g : expand ( 4 ∗ s u b s t ( 1 / 2 i : − − a24 . my summerstudychallenge2 notes).3 x g We can write E f so that if k [E] = k [x. b (x) ∈ K (x) (As in Algebra 1.3 IV. y) = (a (x) . b4 = a1 a3 + 2a4 ..30. Write the tate coecients as c4 = b22 − f : y^2 + a 1 ∗ x ∗ y + a 3 ∗ y If char − − x^3 a2 ∗ x^2 ∗ (y − a1 ∗ x − a3 ) .e.46. (x − e1 ) (x − e2 ) (x − e3 ). c. Thus d (x) is just some constant. But since there are no y terms on rhs. y] / (f ))) can be written as a (x) + b (x) y . E : y 2 = 4x3 + b2 x2 + 2b4 x + b6 . then E (i. c (x) + d (x) y) = (a (x) − b (x) y. NN. y) = (a (x) + b (x) y. a (x) must be linear. f )). b6 = a23 + 4a6 . this means 2 is invertible so we can simplify by completing the square via y 7→ (y − a1 x − a3 ).4 x We'll write the j -invariant via the tate coecients and show this equals Hartshorne's denition of the j -invariant. there can be no x terms in d (x). d (x) y) . y] / (f ) is the ring of regular functions. I will show my calculations and the maxima output even though it looks a tiny bit sloppy. Now in order for the degrees to be correct. 1) = 1. Since this is an isogeny then φ (P ) + φ (−P ) = φ (P − P ) = O hence (a (x) + b (x) y.4.

a 6 : 0 . b4 : 2 ∗ ( a4 ) + a1 ∗ a3 . ok so that we get a j -invariant which agrees with Hartshorne's for algebraically closed. a3 = 0. b8 : a1 ^2 ∗ a6 + 4 ∗ a2 c 4 : b2 ^2 − ∗ 24 ∗ D e l t a : − b2 ^2 b8 j : f a c t o r ( c 4 ^3 a6 − a1 ∗ a3 ∗ a3 ^2 − a4 ^ 2 . We can use the following curves: −1 2 3 2 3 2 If char (k) = 2. then y + y = x . and a2 : − L − 1. then j0 = 123 −j0 . So comparing to Weierstrass equation. j0 6= 0 y + xy + x + x + j0 . then y = x + x. ∆ 6= 0 is another dening j this way will Note that (weierstrass equation. 229 y 2 = x3 +2κX +2κ. b6 : ( a3 ) ^ 2 + 4 ∗ a6 . b4 . a2 = −λ − 1. j0 = 0. and b6 = 0. gives a1 = 0. Now we need c4 and ∆. b4 = 2λ. hartshorne's Now j -invariant of agrees with our j -invariant of 2 We expand y = x (x − 1) (x − expand ( x ∗ (x − ∗ 1) (x − λ) L)). ∗ a4 + a 2 b2 = 4λ. 2 3 3 2 3 3 If char (k) 6= 2. . −1 2 3 2 3 2 If char (k) = 3. j0 = 12 . j0 6= 0 . 3.i is a cubic which will factor into (x − e1 ) (x − e2 ) (x − e3 ) for some ei in the algebraic closure so this gives an equation like y 2 = x (x − 1) (x − λ). 2 Now we want to show that for the curve y = x (x − 1) (x − λ). nonzero discriminant. ∗ b6 ^2 + 9 ∗ b2 b4 Delta ) . a4 = λ a1 : 0 . j0 = 0 . Now we may wish to show the second statement about nding κ j 's. b8 = −λ2 c4 = 16 (−λ − 1)2 − 48λ c6 = 288 (−λ − 1) λ − 64 (−λ − 1)3 ∆ = 16 (−λ − 1)2 λ2 − 64λ3 3 256(λ2 −l+1) j = (λ−1)2 λ2 . j0 = 0. j0 6= 0. Clearly denition for an elliptic curve give a fraction of coecients of the weierstrass equation. 12 . then y = x +1 . We nd the E : y 2 + a1 xy + a3 y = x3 + a2 x2 + a4 x + a6 bi 's b2 : ( a1 ) ^ 2 + 4 ∗ a2 . then y = x +x . a4 : L . a6 = 0. so new x -coordinate is So we have shown our denitions agree. ∗ ∗ b6 . then y = x + x − j0 . c 6 : − b2 ^3 + 3 6 ∗ ∗ / b2 − ∗ b4 − 8 ∗ b4 ^3 216 − 27 ∗ b6 . a3 : 0 .

x  c. E : y = x + ax + bx which looks like E1 above and we has a degree point to x2 2 y (b−x2 ) x2 2.4. (X. 0) know the equation for . We have isogenies of degree  have degree 2 endomorphism has E1 's isogeny. Y ) 7→  2 Y 2 Y (r−X ) . ( ) λ((−1−λ)2 −4λ) Since it's a degree 230 .5 IV. Clearly this is a branch point of x 7→ x . Moving the 2 3 2 gives a weierstrass equation. This calculation is performed on page 110 Silverman. we see there is another branch point at 0.5 we consider the two elliptic curves − 2aX 2 + rX .4. Factoring y 2 = x (x − 1) (x − λ) = x3 + (−1 − λ) x2 + λx these branch points are values of x for which there are one value of y .5 x This goes almost exactly like 4. a kernel with two points.6 b. (Point at innity and 0). 4.4. (x.7 Clearly g : x 7→ y2 x2 E1 ≈ E2 . 2 corresponding to the root at 0.4. we have  X / π P1 2 f g / X  and we know that f. g y2 φ : E1 → E2 . AEC II. x X So from part (a). Now assuming we have such  acurve with x-coordinate. E1 : y 2 = x2 + ax2 + bx connecting the curves:  . and E2 : Y 2 =  . x 2 morphism to P1 . 2 1 1 Note that x 7→ x : P → P is degree 2 so by Riemann-Hurwitz. then by Riemann-Hurwitz. E2 in (b) we need λ2 (−1−λ)2 −4(λ) = 2 . there are 4 branch points.4. 4. Taking π : ∞ → ∞ gives one of the branch points of g . y) 7→ Now if order 2 E P1 (0.4. and φˆ : E2 → E1 . To see the reverse direction.4. look at φˆ given in (b). for E1 ≈ E2 that the 3 3 16((−1−λ)2 +12λ) 256((−1−λ)2 −3λ) j -invariants be equal. by the forms of E1 . Thus. 8X 2 4X 2 . the maps (if coordinates on the second P 1 π and π0 are given by projection of the are given by X := y . as well as the point at innite. and 2 3 2 after taking our equation to the form y = x + ax + bx as in (b). Thus we just need to see (see part (d) for example) when E1 and E2 are isomorphic. π0 By Silverman. Now to see a relation involving the invariant we look at (b) and see that we need. O and an order 2 point. exa III. −2 = 2 · (−2) + R implies there are two branch points.

4 so [ 1 ] [ 2 ] = "" −− i n i t i a l point then p r i n t ( " Durand−K e r n e r " ) p r i n t " Roots of :" peq ( eq ) p r i n t "" end so [ 2 ] = {} so [ 2 ] [ 1 ] = . t a b l e . min for x if there are zero roots do j =1 . t a b l e . . " x " ) end end −− f i n d i n g local for min exp to extract zero roots min = max i =1 . 2) end end if min > 0 and verbose p r i n t ( " Note end eq = ceq ( eq ) −−p e q ( e q ) that then there is a root of multiplicity " . x Expanding the relation (I used maxima since lazy) given above gives the polynomial 9 8 7 6 5504λ − 21696λ − 5760λ + 47008λ − 5760λ5 − 21696λ4 + 5504λ3 − 1808λ2 + 256λ 256λ11 − 1808λ10 + . maxn ( e q ) eq [ j ] [ 2 ] = do s t r i n g . though you would need some of my other code to run it . " x " ) end end −− d i v i d i n g for by i =1 . I use the numerical durandkerner algorithm. min . .4. function d u r a n d k e r n e r ( eq ) local i =1 local numterms = t a b l e . " at zero " . maxn ( e q ) if do −− f i n d max e x p numvars ( e q [ i ] [ 2 ] . " x " ) > max then max = numvars ( e q [ i ] [ 2 ] . maxn ( e q ) local s o = {} if local verbose j =1 so [ 1 ] = {} so [ 1 ] [ 1 ] = . " x " ) < min then min = numvars ( e q [ i ] [ 2 ] . t a b l e . Below is an implementation I did for gma5th (my personal math library)..8 d.9 so [ 2 ] [ 2 ] = local t t a b = {} local n t h t = {} −− f i n d i n g local for max the "i" first time max = 0 i =1 . To nd the roots. sub ( eq [ j ] [ 2 ] .4. maxn ( e q ) if do −− f i n d max e x p numvars ( e q [ i ] [ 2 ] .

t a b l e . " x " ) > max then max = numvars ( e q [ i ] [ 2 ] .max do rootab [ i ] −1) = exeq ( so . ( − 1 ) ^ ( max ) ) db ( " h e r e 1 " ) local nume = {} local denom = {} local r o o t a b = {} local v = verbose verbose = for false i =1 . maxn ( e q ) if −− f i n d do max e x p numvars ( e q [ i ] [ 2 ] . " x " ) if t y p e ( nume ) ~= " t a b l e " then print " error in durker " end denom = c o n s t a n t p o l y ( ) db ( " h e r e for 4") j =1 . t a b l e . " x " ) end end −−t h e n poly needs to be switched depending on max power e q = s e q ( eq . a e q ( r o o t a b [ i ] .local for max = 0 i =1 . maxn ( r o o t a b ) do peq ( r o o t a b [ i ] ) end p r i n t "" end db ( " h e r e 2 local ") n=1 −− l o c a l last = 2341 repeat −− l a s t = rootab [ 1 ] [ 1 ] [ 1 ] db ( " h e r e 3 " ) for i =1 . t a b l e . maxn ( r o o t a b ) db ( " h e r e if do 4") r o o t a b [ i ] == n i l then rootab [ i ] = zeropoly () end nume = e v e q ( eq . i end if verbose then print for "Initial guess " i =1 . t a b l e . r o o t a b [ i ] . s e q ( r o o t end end db ( " h e r e 5 " ) −− p r i n t ( " h e r e 1 " ) . maxn ( r o o t a b ) if i ~= j do then −− p r i n t ( " h e r e 2 " ) if r o o t a b [ j ] == n i l rootab [ j ] then = zeropoly () end denom = c e q ( meq ( denom .

normeq ( r o o t a b [ i ] ) ) end return rootab end Throwing away roots which give singular elliptic curves.9 j = 1728. − 1 ) ) db ( " h e r e 6 " ) −− if t a b l e . s e q ( c o n j d e q ( nume . t a b l e . equation. j = −3375 j -invariant . maxn ( r o o t a b ) do peq ( r o o t a b [ i ] ) p r i n t ( " norm ". denom ) . and plugging into the are left with 4. t a b l e . t a b l e .a.4. denom ) . x g The arithmetic genus. we .4. pa (X) = 12 (d − 1) (d − 2) − r 233 where r is the number of nodes. maxn ( nume ) ) −−p r i n t m a t ( s e q ( c o n j d e q ( nume . maxn ( r o o t a b ) do peq ( r o o t a b [ i ] ) end p r i n t "" end end until n> 15 verbose = v db ( " h e r e if end " ) verbose then print " The roots of this polynomial are approximately : " end for i =1 . maxn ( nume ) > 0 and t a b l e . j = 8000.6.−−p r i n t m a t ( r o o t a b [ −− i ]) p r i n t ( t a b l e . IV.n) i =1 . − rootab [ i ] end dbprint ( rootab [ i ] [ 1 ] [ 1 ] ) for j =1 . maxn ( r o o t a b [ i ] ) if 1/ r o o t a b [ i ] [ j ] [ 1 ] do > 1 e +25 rootab [ i ] [ j ] [ 1 ] then = 0 end end db ( " h e r e 7 " ) end n=n+1 if v then if n % 3 == 0 then p r i n t (" r o o t s for at iteration " . maxn ( denom ) > 0 then = a e q ( r o o t a b [ i ] .

4.math. b.12 |dP0 | .2.pdf which are some notes from Dolgachaev.11.11 c.4.lsa. x This is clear from (a). π gives projection from P . ignore the ramication. the picard groups and jacobians coincide.4.4. which is therefore the degree of If and ϕ. then the plane curve has 4.6 excercises).a. So we just compose the correspondence between jacobian variety and picard group with the pullback on piard group. 234 .7 x g Dual of a morphism Note that the by thm IV.umich.3.10 3d2 − 6d − 6r inection points.4.6 gives that jacobian automatically has a group structure. is rational so it gives a regular map P is a point on the plane curve which is not on a tangent that has an inection or on a multiple tangent. 4. 4. X → P1 . then π induces a map from X to P1 which has degree d.4. this gives d − d − 2r tangents of the plane curve through P .4. x osculating hyperplanes Example 4 of http://www. then P is a hyperosculating point when it is the divisor of P has order dividing d in the group law (see also II. If dP X X has d2 hyperosculating points. As in exc IV. 2 By Hurwitz. Now IV. x g By (b).Dene ϕ ϕ to be the gauss map from the plane curve to a line with no nodes. We know there is an induced homomorphism on picard groups.10.13 b. is embedded via which happens when 4.edu/~idolga/sol. a hyperplane is IV.

x Silverman thm 6. x Silverman thm 6.1. there is an inclusion [n] K (E) ⊂ f K f ∗ λ∗ K (E) = [n]∗ K (E) so that λ ◦ f = [n].strred 4.4. Case 1: n. MISS . x g Silverman.14 c.2.15 e.4.4. thm 6. around page 82 f is separable.2.16 f. page 83 4. then ]ker f = n so every element of ker f ∗ ∗ By Galois theory. Thus Since f has degree n. page 83 235 (E) ⊂ K (E) has order dividing so we can nd a map λ satisfying . ker f ⊂ ker n . Clearly λ = fˆ .4.

9 x g isogeny is equivalence relation.→ K (X) . Thus we have an inclusion of function elds K (X 0 ) .18 IV. x g Every isogeny is a nite map of curves.4. 4.4.4. and degrees come in nonnegative integer sizes.10 x picard of product on genus 1 236 . Since degree 1 would mean an isomorphism.7. The degree of this inclusion is the degree of the eld extension.8 x Algebraic Fundamental Group This is not something that incredibly excites me.4.19 b. we are done.4. Symmetry is by exc IV.4.c. 4.20 IV.4.berkeley 4. (dual isogeny) Reexivity is clear by composition of nites. Here are some notes that have the answer: http://math.17 IV.4.4. Reexivity is clear.

Theorem 2. we have fˆ = α |α|2 = α . 2 Thus deg f = deg (α) = [L : αL] = |α| · A (L) /A (L) .4.a x Write the curve as y Q (τ ) y 2 = x3 + Ax + B . x 2 deg  f2 = |α| .5 Silverman since integral means a nitely generated module contained in 4. Tx (y) = x+y .23 6. where φL is dened by φL (x) is the isomorphism class of Tx∗ L ⊗ L −1 in P ic (X) . the degree of the eld extension is the inverse quotient of the areas.12.24 .2. (page 52??). We know that f ◦ fˆ = |α| by Silverman. x This is theorem VI. Thus our exact sequence implies exactness of the desired sequence.4. The theorem of the square (cor 4) gives us a homomorphism L 7→ φL : P ic (X) → R. Note moreover that L ∈ P ic0 (X) ⇐⇒ Tx∗ L ≈ L for all x ∈ X (by denition) ⇐⇒ m∗ L ≈ p∗1 L ⊗ p∗2 L on X × X . = u3 y 0 for u ∈ C∗ and the substitution gives an automorphism of E i u−4 A = A and u−6 B = B . Abelian Varieties.4. then A (αL) = |α2 | A (L) . For a sublattice. c. Let m : X ×X → X be addition. We therefore have an exact sequence 0 → P ic (X) → P ic (X) → Hom (X. let Tx : X → X be the translation.5.1.22 b. 2 0 Via Milne. Clearly theorem of the cube) m L ⊗ p1 L ∗ −1 it is always trivial on {0} × X and restricts to Ta L ⊗ L on X × {a}. Elliptic Curves.4. Checking 237 . automorphisms xing the point have the form x = u x .21 IV. If A (L) is the area. This last equality follows since by the seesaw theorem (essentially the −1 ∗ ∗ ⊗ p∗2 L −1 is trivial i it is trivial on X × {a} and {0} × X . Then P ic0 (X) is the set of 0 line bundles L where φL is identically 0. IV. thm −1 −1 Thus taking f .11 x g An elliptic curve corresponds to a lattice L.4. P ic0 (X)) → 0. 4.4. we have 4.Following Mumford. 4. By part (a).

if j = 0. just using the fact of corresponding to degree 4. The term we want is (((p−1)/3)!) 3 (xyz) kx +ky +kz =p−1 kx . Basically the idea is we need to nd quadratic extensions which 2. b ∈ Z which we can write as a + If m ≡ 1 mod 4. kz p−1 this coecient 0? What if p − 1 has no factor of 3? Ignore those primes since then automatically (xyz) has Dirichlet's theorem gives that the Dirichlet density of primes in an arthmetic progression 238 . IV. page 149. and then use thm IV. But the method of proof of exc IV. So we are done. a. then OK = a + b + 2 d to compute the 2 2    2 2 2 2 2 b 2 +m 2b = 2 so need expand a + 2b + m · 2b = b 4m + b4 +a b+a2 = norm. So these are clearly the only solutions for have elements of norm m ≡ 2.13 x This is j = 5 mod 13 .11 to connect τ from the lattice to the ring of integers.21 (and some computer algebra software since it's a large polynomial) to see the Hasse invariant is 4. so Q −1 /Q is 2 2 one possibility For example 2 = a + 2b works for a = 0. For example 2 = a + 1b works for a = b = 1. a = 0. x Note that we'll use exc 4.4. it is clear that the lhs will be too large. then If B=0 then writing A. and thus these are the only rings of integers So if we let If m m = 7. b. The uniqueness is Silverman Theorem 4.25 τ λ gives j = 1728. IV.27 0.4. b = 2 then we get under quadratic extensions with elements of norm Now we want to match j with τ.4.b accomplishes this since we are 2.4.26 2. ky .c . In later case we need a + 2 2. 4 is any larger. Taking a trinomial expansion. and then prove a supplementary lemma for the similar case. you can solve the j -invariant equation to get a curve. 1·7 + 14 + 0 + 0 = 2.1. If A=0 then . You can check this is unique. n  √  o  b√ 1+ d b .14 x Fermat Curve and Dirichlet's Theorem a + nb for a. B in terms of Now in part (b). √ √ mb ∈ Q ( m) /Q the norm is given by N (x) = a2 − mb2 . 3 mod 4. we'll see these match the desired values of 4.4.(x + y − z ) =   P p−1 k k k p−1 (p−1)! . The condition that the Hasse Invariant be 0 is that (xyz)p−1 3 3 3 p−1 3 3 3 p−1 has coecient 0 in the expansion of (x + y − z ) .4.5. So given a j -invariant. b coprime has dirichlet density 1/ϕ (b) . The curve is incidentally given by (y^2-x^3+165*x+110)^12. b = 1.4. So for what p is (x3 ) x (y 3 ) y (z 3 ) z .possibilities. AB 6= u = ±1. So we need Note for x = a + √ to solve 2 2 2 2 2 = a − mb with m prime and negative.

8. (The proof follows as in IV.les. c.29 IV.2.4. Otherwise.3. Modulo dirichlet theorem.16 x This is Qing Liu. this ϕ(3) is a larger prime! so we have the whole set. since p 3k + 2 are automatically 1 = 12 .5.x g kernel of frobenius Separability is thm III. Thus all primes like hasse invariant 0 for this.4.word alternatively 239 .coecient 0 since it doesn't appear in the above summation.31 Fq since a∈F lies in Fq when aq = a.2 in the proof of Hurwitz).4.15 x This is silverman V. x I give a proof on page 11. 3.4. IV.5 Silverman. Thus inseparable i ψ ∗ ω = 0 where ω is the invariant dierential of the curve. page 144 4. this gives us a set of density coecient should not be zero mod 4.4.30 b.26a 4. Now note that the xed points of frobenius are the points in 4. 12 of my Fall Break Number Theory Remix Notes from 2012 http://divisibility.4.28 p .1. Now compute the frobenius of invariant dierential using the tate coecients. Essentially we know that F is separable i the pullback on sheaves of dierentials is injective as in II.

(since assume Now if p + 1 − N ≡ 0 (mod p) then pk = (p + 1) − N some k . x Well here is a lua function I made for the group law: • . then clearly Since N ≥ 0. using the previous parts. x The rst assertion is proved in the rst three paragraphs of Silverman theorem V.4. then we are done. x Hasse's Riemann Hypothesis for Elliptic Curves I give a proof on page 7. coefficients pp76 before it runs .com/2012/ Alternatively. 4. a = p − N + 1 ≡ 0 (mod p) ⇐⇒ hasse = 0.4.wordpress.. if q = p) If N = p + 1.34 p ≥ 5.17 a. have hasse = 0.33 e.les.4. see my proof in exc V./gma5th. IV. for + y1 . 4 −x^3 −0∗ x^2 −0∗ x^2 −−1∗x −0∗ x −0 x1 ~= x2 editing equation . Now.1.4. l u a −pv " wpeq ( l e q ( ' h ' ) ) " equation 1 yyy +1yy +−1xxx +1x a1 = 0 a3 = 0 a2 = 0 a4 = −1 a6 = 0 other y^2 j form see hartshorne invariant Group law for 1728 weierstrass may n e e d some function GroupLaw ( x1 .1. alternatively. local x3 = 1 local y3 = 1 if IV . via granville 4. / gma5th .4.8 of my fall break number theory remix notes http://divisibility. following combos y2 ) then lambda = ( y2 − y1 ) / ( x2 − x1 ) 240 or 1c Silverman . 4 .lua -pv "wpeq(leq('g'))" • where h is a le containing andrew@andrew−HP−F o l i o −13− Notebook −PC: ~ $ Printing weierstrass a coeffs for .4.10 and use g = 1. − x2 .32 d.

(67 : −548 : 1) . i n t e g r a l _ p o i n t s ( mw_base=[P . b6 = a3 = 1 .4. ( −3 1) . sage : Q=E ( 2 . 2 ) sage : P=E ( 1 . • sage : E = E l l i p t i c C u r v e (QQ.35 b. (3 − : 1) . b4 = 2 · a4 = −2. 4.18 x • can probably nd these with my calculator sage. 0. x Such a curve is nonsingular i the discriminant is nonzero. Q ] . (1 (5 : : 0. −4 : (31 : 241 : 1) . −172 : 548 ( −2 : (9 : 1) .∗ nu = ( y1 elseif − x2 x1 == x2 y2 ∗ x1 ) / ( x2 − x1 ) then lambda = ( 3 ∗ x1 ^2 + 2 ∗ ( 0 ) ∗ x1 + −1 − ( 0 ) ∗ y1 ) / ( 2 ∗ − ( 0 ) ∗ y1 ) / ( 2 ∗ y1 + nu = (− x1 ^3 + ( − 1) ∗ x1 + 2 ∗ ( 0 ) y1 + ( 0 ) ∗ x1 + ( 0 ) ) ( 0 ) ∗ x1 + ( 0 ) ) end local x3 = lambda ^2 + ( 0 ) ∗ lambda local y3 = return − 1 ∗ ( lambda x3 . (2 : : (67 10]) 1) . : (41 1) .3 Schmitt) 2 2 2 so a3 = 1. (5 : 10 (13 : −46 : 1) . (2 : 1) .4. then use the additional law in silverman. 1) . : 2 : −2 −10 : 1) . ( −2 1) . −7 . 2 ) sage : E . −5248 : −4 : 1) . 1) .4. a4 = −1. − ( 0 ) − x1 − − nu − ( 0 ) + ( 0 ) ) ∗ x3 x2 y3 end 4. : : −2 : : −26 1) . 1) . [ 0 . : (302 172 : ( −1 : : 2 : (9 : 26 : 1) .. 1) .. (13 : 46 (41 : 262 : 1) . Note the discriminant is dened via the Tate coecients (denition 1.36 IV. (31 4 : 1) . so b2 = 0. b8 = −a4 = −1 3 so ∆ = 0 − 8 · 2 − 27 · 1 + 9 · 0 = 8 · 8 − 27 = −64 − 27 = 37 6= 0. b o t h _ s i g n s=True ) [( −3 ( −1 (3 −2 : : : 4 4 : : : 1) . : 1) . (1 : 2 : 1) . : −262 (302 : .

39 • Slight issue? There is a slight error with this problem. a3 = 0. a3 = 0. Now let t ∈ T and [nX ]t : Xt → Xt .b.6. a01 = 0. 4.37 IV. 3  The j -invariant.c Silverman AEC II. then .4.3.4 λ2p (4λp −(−λ−1)2p )  On the other hand. 4. if we let of Xp  Take then the pth j -invariant power of the rst one.4. a02 = (−λ − 1)p .6.4. a'la Ex. Or there is an error on other sources. If [nX ]t is constant. so it always factors! 242 p.7.4 is 256(λ2 −l+1) (λ−1)2 λ2 . a4 = λ. In characteristic multiplication by p is never separable.3.8 / thm IV.17.4. a6 = 0. a6 = 0.5.20 x g Ok here's what we do. a2 = −λ − 1.8.38 • IV. But this doesn't happen by the Criterion of Neron-Ogg-Shafarevich: thm VII. Assume for convenience the curve is in the form 2 3 2 y 2 = x (x − 1) (x − λ) =expand (x · (x − −x λ + x λ + x − x • So a1 = 0. then one of the [pX ]t is constant. Note that nX is obtained by composing multiplication by prime factors p of n. VII.1.1 Silverman AEC I. are the tate coecients.4. The last statement follows by Silverman AEC I. Thus [nX ]t is The fact that consider nite at by thm II. 3 256(3λp −(−λ−1)2p ) will be a'la Ex 4.19 x nX : X → X is dened over T is theorem IV. a4 = λ.4. and factor mod p be the tate coecients gives the second. By thm II. (It helps to use a computer algebra system such as maxima to do the computations for you) 4.4. each such pX is either constant or at.

1. then just take the kernel will be a prime ideal such that is a eld).. So I think there is actually a slight error in this guy. Since I read Milne's Modular forms notes earlier this year.  Also it's in some notes from MIT • By assumption of the problem. Now how does hasse invariant relate to frobenius? H 1 (OX ). Now use Silverman.40 E ≈ Ep p p then is it true that 2 E ≈ Ep p factors as two frobenius morphisms map is separable / inseparable. and if Hasse invariant is nonzero.. If the mult by p In characteristic • hmm. 243 R/p ≈ Fp . we are in characteristic • Thus the multiplication by is either purely inseparable (in which case.4.  if it factors as frobenius -> frobenius. So if λf ∈ / Fp . 145. x Recall that R is the ring of endomorphisms of polynomials with coecients in k. the subgroup of points of order p on X X xing is 0. Since we know it does commute. Note that Fp ⊂ k is the xed point set of frobenius. then both are size p.41 d. then λf must be in Fp ...4. • so if 4.. 4. and it p. X ≈ Xp ? then it factors as 2 E → E (p ) → E . c. Then by group isomorphism theorem since (it's prime since Fp ϕ : R → k. page 144. By exc IV.4. Theorem 3. the multiplication by map in this case multiplication by each • p map p2 ) ( factors as E → E ≈ E so of degree p.15. x Suppose Hasse (X) = 0 . Recall that an isogeny is dened by f commutes with frobenius since all such polynomials ∗ 1 So the frobenius F also gives a map F : H (OX ) → do.•  If it factors as frobenius -> separable. then ∗ ∗ ∗ F (λf ) · F (x) = F (λf x). then both are size p. or it's trivial. Thus it is clear that P0 . Fp are the λf · F ∗ (x) 6= points of order p which will be potential images by what we will see next. then it's separable / inseparable degrees are factors into two things of degree • p p.

42 IV. it's ramied at 2g + 2 g21 so a degree 2 map f : X → P1 . i X K is very ample if it's a complete intersection.3.not algebraic geometry This isn't really algebraic geometry. ∞} are sent.2 |K| is v.5. See for instance. Proof 2 If X is hyperelliptic.4. . X is determined by whether or not it permutes the ramication points.21 x skip .2 x g Aut X is nite.5. is non-hyperelliptic. then it has a By hurwitz. 1.5.5.2 says 4. Proof 1 See Dawei-Chen Notes MT845. X . any nontrivial automorphism has no xed points.22* (starred) MISS 4. proposition 4.4.1 x g complete intersect is nonhyperelliptic 3. points.a. gives us that the canonical bundle But 5. and use Weierstrass points. Dummit and Foote IV.5 IV.4.8. IV. X are therefore determined by automorphisms of P1 permuting all the ramication Any automorphism of By connectedness of Automorphisms of points. Automorphisms of P1 are determined by where the three points 244 {0.1 IV.5 Canonical Embedding 4.4.

so in the 5 3  − 1 = 10 − 1. symmetric matrix.3 d d and g−1 we see that all hyperosculating points cannot lie on a hyperplane Aut (X) so that it is nite. X = (x. OP3 (2)) = P . symmetric matrix below. i. . 1)t . By comparing this number with of degree 4. Nonhyperelliptic So let E a projective bundle parametrizing complete intersections of cubics and quadrics in 3 0 3 9 We have a surjection to the quadrics in P . there are 9 parameters. OQ (3)) 0 3 0 3 And thus dim Eq = H (P . by exc. thus the nite number of hyperosculating points determine IV.6. P3 . g−1 An automorphism of P (where X is embedded) is determined by where it sends g + 1 points not on the Now suppose same hyperplane. So basically I just need to say that the dimension of the space of quadric cones is going to be Note that quadric forms are zero sets of the following matrix: 2 2 Consider P (x. OP3 (2)) = − 1 = 5·4 2 3 2 So now we add 15 + 9 = 24 and quotient by P GL (3) which has dim (3 + 1) − 1 = 15 gives us 24 − 15 = 9. y.4. OP3 (1)) − 1     6 3 4 3 − 1 = 20 − 4 − 1 = 15. Now we must add the dimension of a ber (cubics intersecting each quadric) and then quotient by P GL (3) action. y. A curve with only one g31 corresponds to sublocus where quadric is singular So we counted the quadrics by Note that 10 0  3 dim PH (P . π : E → PH (P . X is non-hyperelliptic.   5 0 3 − 1 = 9. to be singular.So we are permuting them. OP3 (3)) → H 0 (P3 . Aut (X) 2g + 2 points with g ≥ 2.e.5. quadric cone. 245 and 2qy +  2rz + d . OP3 (3)) − H (P .5. These are classied by the hurwitz scheme (see my notes on Severi + Hurwitz scheme) which has dimension 2g − 1. z. OP3 (1)) → H 0 (P3 . Also dim PH (P . we have an exact sequence: Q 0 → H 0 (P3 . 2 Then X has (g − 1) g + gd hyperosculating points. = − = Eq → 0 = Only one g31 .3 x g Moduli of Curves of Genus 4 Hyperelliptic. r  d 8. IV. a f h  f b g A=  h g c p q r  p q  . For the ber over a point Q. and thus since there are only nitely many ways to permute is nite. OP3 (2)) = is also the number of parameters in the Note that a singular quadric is rank In this case we need Q 3. z) = ax + by + 2f xy + 2gyz + 2hzx + 2px + This is a matrix product X t · A · X = 0 .

Since there are only two lines through any given point of Q the image of Now if more than two points are collapsed then 4 X under π can have only two singularities. 2 Projecting from a point P on X gives a morphism to P .5.5. Computing the genus of normalization gives degree 5. Q .2. this is also fairly clear. then by Bezout L must lie on both Q and F .5. hyperplane meeting X Geometrically. eb = 3 as above. Since X is type (3. If ϕ has degree e.5 x g Note X is nonhyperelliptic. Now using 2 the chart around page 506-508 which tells how much a one singularity will drop the genus.4 x g g31 's each giving degree 3 maps to P1 . Each node gives a g3 as in ex 1 IV. and A line through one of the nodes meets in 4 1 other points since it has multiplicity 2 there and using degree 5 and bezout. we must have two nodes.2. 1 Has one singular point. 1 1 In the rst case X0 is smooth and rational and the projections from X0 → P are injective hence the two g3 's Suppose X is nonhyperelliptic with two must coincide so that gives a contradiction.5.5. Then we have 4 = (5 − 1) (5 − 2) − r = 12 4 · 3 = 6 − r so r = 2 .2. By degree genus formula 1 for normalization. By bezout. 1 then ea = 3. geometrically it is clear that we cannot collapse a g31 through a tangent line and similarly. then the projection is birational from X to a plane curve which must have only the one singularity from the trisecant. but also a g3 taking the product gives a map to a quadric with ea = 2. 3) it is a complete intersection of a quadric and cubic. But a quadric and a cubic forming a By example IV. eb = 3 since these are coming from g3 's so either a = b = 1 and e = 3 or a = b = 3 and e = 1. then L must lie on Q. If X0 is the image its on a quadric so it has type (a. P. 4. then it has a g2 . Now suppose we have a plane quintic with degee less than 6 and only nodes. If on the other hand there is a multisecant line L which meets X in more than 3 points.5. we see that the X with a P and tangent at a concave point Q and then project down from a point on the line P Q but not between P. Q is singular. X lies on a unique irreducible quadric surface Q. which was a previous excercise.2. b) for some a. But then the cubic and the quadric both contain lines through 4 points which is a contradiction. Q to the same point. By example IV. Now we want to bound the number of singularities to 2. then the line through P0 . if π maps P. X is the complete intersection of the quadric cone with a cubic surface F . singular point corresponds to a tacnode. X contradiction. Q lies in Q. On the other hand if we have a plane quintic with two nodes then the normalization has genus has genus 4. genus is a birational invariant of a curve so X 4. as Q has degree 2 and L intersects Q in 3 points. points are collinear on X. Let ϕ : X → P1 × P1 be the product morphism. This contradiction shows there must be at most trisecants. genus 4 by assumption. b. Since we are on a quadric. Also by example IV. Thus we have a g3 .2.4 IV.5. Thus we have two g3 's which gives a contradiction.5.2. By Bezout.5. so e = 1 so the product morphism is birational to something with a dierent genus. 1 1 If X was hyperelliptic. complete intersection don't share a line. and as Q has a unique ruling through P . Consider the projection π from a point P0 on X 2 to P . there are no secants with coplanar tangent lines for the same reason. If P lies on a trisecant L the projection is 3 − 1 at some points.4. we need to cut tangent at an inection in one point 246 .

0 ∗ 0 0 Necessarily. 247 dim V5. just need to compute h (X. and thus h (KX − E) = 2 ..5) 0 0 By r.7 b. h (X. x g Has g13 implies plane quintic with one node Suppose X has a divisor 0 Thus h (D) = 2. 5 = pg (X) = 2 Then By R.11 we have  p(t) h0 X. O (2)). (2g-2 .5.r. dim |E| ≤ deg (E) = · 5. By Arbarello. .6 IV. 2 2 7 0 Thus 3 ≤ h (E) = dim |E| + 1 ≤ = 3. By degree-genus. Now . NX/Pr . By the next theorem. NX/Pr − h1 X. Solving gives one node. O (2)) ⊕ h (X. So we need to compute cohomology of the normal bundle of p(t) Hilbr at a point is at least X. however. O (2)) ⊕ h (X.. 5.R.. and OX (E) = f ∗ OP2 (1) .. does this actually work? Hopefully 4. O (2)) − h (X. D which gives a degree 3 embedding to P1 . NX/Pr is an upper bound for the dimension of Hilbr .. I. O (2)) = d + 1 − g = 8 · 2 + 1 − 5 = 17 − 5 = 12 Now multiply that 2 by 3 gives 36.5 x g Curves of Genus 5 There is probably a way to do this via Hartshorne. From Sernesi's book on moduli theory. Plane quintic with one node implies g13 f : X → P2 . χ (D) = 3 + 1 − 5 = −1 so h0 (KX − D) = h0 (D) + 1 = 3 2 2 Thus KX − D is a g5 which maps X to P . so KX − E is a g3 . 1 (5 − 1) (5 − 2) − ]nodes . the dimension of irreducible components of   h0 X.12. so deg (KX − E) = 2 · 5 − 2 − 5 = 3. Irreducible family See my notes on Severi Varieties and Hurwitz schemes: 2 subtract dim P GL (2) = (2 + 1) − 1 − 8 and we're good. O (2)) ⊕ 1 0 1 h (X. Then deg (E) = 5. Note we should have an upper bound on the dimension. Let .5. 0 0 0 1 1 So. 2 0 0 1 Thus h (E) = 3. Now modulo by P GL (4) = (4 + 1) − 1 = 24.5. Well. still get 12 since 36 − 24 = 12.4. O (2)) ⊕ h (X. .5 . (I forget the title) NX/Pr ≈ OX (2) ⊕ OX (2) ⊕ OX (2) . h (E) − h (KX − E) = 5 + 1 − 5 = 1. let sponding to the hilbert polynomial p (t) p(t) Hilbr denote the hilbert space corre- .. Cornalba. and hence by cliord. Geometry of Algebraic Curves II.5 = 3d + g − 1 = 15 + 5 − 1 = 19. O (2)) and h (X. deg (KX − D) = 2 · 5 − 2 − 3 = 5 . h (f OP2 (1)) ≥ h (OP2 (1)) = 3 and so h (KX − E) ≥ 2 1 1 Thus E is special. since these guys are in M5 which has dimension 3g − 3 = 15 − 3 = 12.5.

points. Now apply thm IV. Note that a morphism between nonsingular non-hyperelliptic curves then the pullback of regular dierentials map to regular dierentials.R. Thus such a morphism lifts to a morphism between projective spaces via the canonical embedding. For the second part see Arbarello..9 IV. If C is hyperelliptic it has a unique degree two map Any automorphism preserves f but permutes the 8 f to P1 ramied at 8 points by Hurwitz theorem. 1 (7 − 1) (7 − 2) − ]nodes . • . which gives a degree deg (KX − D) = 2 · 6 − 2 − 3 = 7 .8 g31 is unique.7.a x g Suppose rst that C is non-hyperelliptic. By degree-genus.5.5.5. then the action must be free.MISS. 4. Then By R. 6 = pg (X) = 2 1 Then 6 = · 6 · 5 − ]nodes so 6 = 15 − ]nodes. So a plane curve will have genus Suppose X has a divisor 0 Thus h (D) = 2. 248 .5. 2 So there are 7 nodes.5. . on the bers. Such 1 automorphisms are just given by automorphisms of P permuting 8 points. χ (D) = 3 + 1 − 6 = −2 so h0 (KX − D) = h0 (D) + 2 = 4 2 4 Thus KX − D is a g7 which maps X to P . D 6 by the degree genus formula if it is degree 3 embedding to P1 .6 x g I feel slightly iy about this one. Since the complement of the • 8 . Geometry of Algebraic Curves I. IV.3. Harris. 5. Thus the morphism is determined by the action points on C is connected. already showed 4.

Let D the hyperplane section.5. note 15 + 1 − 0 = 16 d > 2g − 2 h0 (OP3 (3)) − h0 (OX (3)) = by R. dim |2D| − dim |K − 2D| = 8 + 1 − 0 = 9.. nonsingular that X is the n rational normal quartic which has n + 1 linearly independent points in P and thus must be contained in a Denote by X a rational curve of degree nondegenerate quadric. To get f actorial(3)·f actorial(3) 0 thus nonspecial (the degree of the divisor 3P ). If X is contained in two quadrics. at any rate.6.2. then by 1 by Bezout. then 3 0 by exactness.2 x g Always in a cubic f actorial(6) − .solved: MISS 4. X is a complete intersection and by exc II.b.1 IV.6 IV. Consider the LES associated to 0 → IX (2) → OP3 (2) → OX (2) → 0.R.1 x g 4 in P3 . 2 Thus Q is unique.6. Also  deg K = 3+2 2g − 2 = −2 clearly dim |K − 2D| = 0. 4. Thus the ideal sheaf twisted by 3 has some global sections. Thus h0 (OX (2)) = 9 and since h0 (OP3 (2)) = = 10. so h (OX (3)) = 3d + 1 − g = To see it's always in a cubic. Now note that by exc IV.. = 20 − . Exist ones not in any quadric First note that the rational curve of degree V.6. consider the second term.. h (IX (2)) ≥ 1..6 Curves In P3 4.2 IV.4 thus has genus 4 · (2 + 2 − 4) + 1 = 1 contradiction. But nondegenerate quadrics are nonsingular by chapter I.9 (which doesn't use this result) to nding a rational degree a singular quadric.8.6. Thus we can restrict our attention curve not on a smooth quadric since at any rate such a curve won't be on The remainder is Theorem 4 and Proposition 6 of Eisenbuds On Normal Bundles of 249 .3 and since X is. 5 g (X) = 2 5 X cannot be contained in a quadric cone or else by exc which is a contradiction. So at least X lies on a quadric.6.

Then exc IV. X passes through the vertex of the cone. If Q is the product of two hyperplanes. the genus is (9 − 1) (9 − 2) 6= 11 . This curve should be special so 3+2 h0 (OX (2D)) > 8. Then exc II.6.4. 4.9. Alternatively.3 to check it gives an embedding. checking the possible types gives X of type (a. then by the dimension |H − 2K| is a point.2.. So in any case h (OX (2D)) < 10 and thus X lies on a quadric surface Q. b) = (2. By 1 0 Cliord.7.5. then it's clearly contained in a quadric.b.3 x g Contained in aQuadric  .6. then by exc V. then since h (OX (2d)) > 0 . Let H be the hyperplane OX (1) which has degree 5 since X does. then deg (KX ) = 2 and KX spans a line in P3 . This is a degree 5 2 4 4 P1 via can check there are no degree (s : t) 7→ (s5 : s4 t : st4 + as3 t4 : t5 ) for a ∈ k ∗ 2 relations by checking each case (i. gives X is projectively normal.1. 4. we can nd an eective divisor linearly equivalent to 2D . dim |2D| ≤ deg 2D = 9. If it's nonspecial.6. h0 (OX (2)) − h1 (OX (2)) = 5 · 2 + 1 − 2 = 9.3. it must be a cone. Thus H − 2KX has degree 1 but could either be eective or not eective depending As X has genus on how we choose the points (both cases are possible).4 IV. 3). h (OX 250 . H −2KX not eective. By r.14. then it be a line and have genus 0 or it's on a plane so by 1 2 degree genus for P .5.6.1. 3 0 If 2D is special. check and such possibilities. 9 = a + b. So we have rulings meeting at a vertex giving a quadric cone. there are nitely many and I'm too lazy to type them all).6. 11 = ab − a − b + 1 and there are no possible solutions. 9 = 2 · a + 1 2 2 and 11 = a − a so that 11 = 12 which is a contradiction. embed 5 and we (s ) − (s t) (st + as3 t4 ) = 0 . Note 2+3 = 20 = 10. h0 (OP3 (2)) = = 10. If it lies on another quadric. Consider the LES associated to 0 → IX (2) → OP3 (2) → OX (2) → 0.9. then |H − 2KX | = ∅.r.d combined with the LES computation gives X that Q is the unique quadric on which X lies. Then h0 (OX (D)) = 2 and |H − D| = |KX | has dimension 1. then by the proof of exc V. Thus the divisors of D are contained in two planes and therefore spans a line which is the intersection of the planes.4 x g 9 and genus 11 in P3 . (2D)) −h0 (OX (K  − 2D)) = 18 + 1 − 11 = 8. Consider the LES associated to 0 → IX → OP3 → OX → 0. Note that 2 3 h1 (OX (2)) = h0 (OX (KX − 2H)) = 0 since KX − 2 · H has negative degree. two dierent lines don't meet. and so since X is smooth the two vertices must coincide (think of Thus at least X is on a quadric the picture) so they are in fact the same cone. If that quadric is nonsingular.Smooth Rational Space Curves.2. 2. Suppose X has degree 0 By riemann-roch. If Q is a quadric cone. then by rmk IV.3 IV.6. If the quadric is singular.e. So if h0 (IX (2)) = 0 then we that h0 (OP3 (2)) = have a contradiction to exactness. On the other hand if H − 2K ≥ 0 .4. Q. This line is contained in Q since it meets X in 3 points by bezout since it has degree 3 and it's a line. Since |H − 2D| is empty. But For then we have rulings not meeting which gives a smooth quadric. Now use thm II. 2 Now if Q is nonsingular. then by remark IV. Let D = H −KX which has degree 3.6.

we have h (IX (m)) = h (OP3 (m)) − h (OX (m)) =       m+3 m+3 m+3−a − + 3   3  3 m+3−b m+3−a−b + − 3 3       m+3−a m+3−b m+3−a−b = + − . (smooth =⇒ projectively normal) Thus H 0 (OP (l)) → H 0 (OX (l)) is surjective for all l ≥ 0via II. 4. Since χ (l) = h (OX (l)) − h1 (OX (l)) + h1 (OX (l)) − . I can compute h (OP3 (m)) = 3 0 0 We know it's hilbert polynomial.8. 251 1 4 · (72 − 1) − 7 + 1 = 6.6. .5 IV.4) then  for g = 0. it's a plane cubic  for g = 2.  for g = 3.  For g≥7 same reasoning as before. it's rational so in a plane  for g = 1. Via exc II.6 IV. and degree 6 in genus is bounded above by 2.8. 3.  m+3 0 0 Now note if m < min (a.6. Therefore.8.4.4 • ok here is almost everything for  for g = 0.6.5 x g complete intersection doesn't lie on small degree surface X is smooth I guess since we're in chapter 4. ok use castelnuovo's bound (6. b) we want to show that h (IX (m)) = 0. 3 3  3     m+3−a m+3−b m+3−a−b So if m < a and m < b then = 0 and = 0 and = 0 so 3 3 3 We assume =⇒ normal we're good.6 x g Projectively normal curves not in a plane • II.4 We get projectively normal. For (maybe something like canonical embedding) d = 7. 4.??? ∗ • d = 6: see below.. P3 . so that's how we'll compute h (OX (l)).. and h2 (OX (l)) = 0 since X is 0 0 0 dimension 1. we have castelnuovo's bound we have castelnuovo again  Now we just have to check genus • so for genus 2. 1  for g > 9. 4  for g ≥ 5.6.4. but h1 (OX (l)) = 0 by projectively normal.

a = 3. 5) or g = 10 − 7 + 1 = 4 or g = 12 − 7 + 1 . 4.428571428571429  then  d.3.  Also g = ab − 7 + 1. it has a type. wait.  However. the hyperplane section 5 d ≥ 5. g = 2. 3 2 If X ⊂ P \P has degree d then the projection from a general point of 1 smooth plane curve of degree d − 1 with genus (d − 2) (d − 3). so g = 6 − 7 + 1 = 0.c since 4. -should allow us to compute h (OX (1)).3  so hyperplane section is nonspecial. e 1 de (d 2 2= so 6 3 + 36 = 7d or d = 2. this contradicts III. for degree  Then genus 4.5. then g ≥ 21 7 + 1 = 4.1 so |a − b| 6≤ 1.0 =  This is not an integer. 2 252 X 3 gives an isomorphism onto a . b) f actorial(5) = 10 h0 (P3 (2)) = f actorial(3)·f actorial(2) for genus 5 this would not be the case) where so we see a + b = 7. 2. Looking at the picture if X lies on a plane then it is a line or conic since any degree term higher than would have at least one inection and thus a trisecant.7 x g Suppose X is a curve with no multisecants..6. yeah it's 0 so that conrms it's not in a plane.7 so it's nonspecial.5. so it's not a complete intersection. it has a nonspecial. g = 2 case. 3 we get a contradiction same as for using Remark IV. however for it has degree 12 > 2g − 2 h0 (OX (2)) = 12 + 1 − 4 = 13 − 4 = 9.1. h0 (OX (1)) 0 then h (OX (D)) = 4 so that's d = 6. (a.  by halphen..  something's messed up here though.0+36 7. (note  so h0 (OX (D)) = 7 + 1 − g . 3. b = 6. then de = 7.  so we know that it has type (2.  Now let's compute • 7.6.  by 6. we still need to prove the thing.  So if it's on a quadric. a = 1. if it were a complete intersection of surfaces of degree + e − 4) + 1 and d · e = 6 (by bezout) or e = d/6   or 2 = 21 · 6 · d + d6 − 6 + 1 = 3 76d − 6 + 1 so 13 + 6 = 7d 6 5. so it's contained in a thing. very ample divisor of degree 0 is nonspecial. b = 4.  so it's not a complete intersection (also g = 2 =⇒ hyperelliptic =⇒ not a complete intersection by IV. Also that our curve is contained in a quadric. D since we're supposed to have a surjection ! • So for degree 7.1). we'll take 2D still ok.3. a = 2b = 5. IV. if g = 4.b. genus  if hyperplane section is special. but h (OP3 (1)) = 4 so that's actually a contradiction.5 by 6.6. since consider degree degree 1 7.

d ∈ S and so d0 ≤ g + 1 . 4. exists bpf implies d>= g+1 If there exists a base point free. this gives 1 2 (d − 2) (d − 3) ≤ has trisecants which are lines on one of the rulings. We could also have d = 4 so by the genus is 0. then there exists no base point free gd on X . 253 gg1 to get a 3 gg+2 which is .8 IV. Thus we 3 could have d = 3 so X is twisted cubic since it's degree 3 in P and genus 0. S to be the set of all d ∈ N such that there exists base point free gd1 on X . Dene d0 to be the smallest d ∈ S such that m ≥ d implies m ∈ S . 1 Note that if d0 ≤ g + 1 . 1 On the other hand if d < d0 . Thus for any d ≥ g + 1. On the other hand.1(b) contradicting Halphen. 1 depending on nodes.3. Thus d2 − 6d + 8 ≤ 0 so 2 ≤ d ≤ 4. d−g Thus if d ≥ g + 3. the general pencil of the very d−g ample gd gives a base point free so such d ∈ S . 3 Next we have the bpf. nonspecial gg1 . consider a nonspecial very ample gd for any d ≥ g + 3. Thus d − d + 3 ≤ d − 2 2 4 d + 1 d odd d + 1. then X has a nonspecial very ample divisor i d ≥ g + 3 ..6. then we add two points not in very ample by thm IV. 4. d even or 12 d2 − 52 d + 3 ≤ 41 (d2 − 1) − d + 1 if d is odd.( 1 2 d −d+1 4 1 (d2 − 1) − 4 d even 5 1 2 1 2 . and d ≥ g + 1 then there exists a base point free gd on X .9 IV. gg+3 (−P ) where P ∈ C is general which gives g + 2 ∈ S . Such an elliptic would be the complete intersection of two quadrics which has no trisecants.6.8 x g Rephrase X. g ≥ 2. 3) on a smooth quadric so By Castelnuovo. Thus it suces to prove d0 ≤ g + 1 Let K be the eld of rational functions on Dene d >= g+1 implies exists bpf . 3 2 1 We may subtract a further point by gg+3 (−P − Q) = gg+2 (−P ) = gg+1 which is birationally very ample..9 (starred) skip. Halphen gives us that if Clearly these later linear systems are also nonspecial. a rational quartic lives as a type (1.6.6.

Exactness can be checked at the stalks.−c) 0 → OX (−C − D) → are curves meeting transversely.1. −c). write C and D as the dierence of curves all meeting transversally. 2 1 2 2 1 This is P (n) = n H − nH. since c.D 0 → OX (−C) → OX → OC → 0 0 → OC (−D) → OC → OC∩D → 0 and 0 → OX (−D − C) → OX (−D) → OC (−D) → 0 Thus we can use V.−c) (c.2 x g Degree via hypersurface 1 By Riemann-Roch.1 V.KX + 1 + pa . assume C.1. For simplicity. (H + KX ) = 2g − 2 and so H. (nH − KX ) + 1 + pa . (c. then V.1. We must show that the kernel of If 5.1.1 Geometry On A Surface 5.1. then χ of the middle is sum of χ outside. 1 1 2 We need that H. Since the meetings of C.d) OX (−C) ⊕ OX (−D) → OX → OC∩D → 0.5 V Surfaces 5. D Consider the sequence (d. d) → 0. . χ (nH) = (nH) . D are transverse. Making the needed substitutions gives χ (OX ) − χ (OX (−D)) − χ (OX (−C)) + χ (OX (−C − D)) = χ (OC∩D ) = h0 (C ∩ D) Proof 2.1 V. 2 2 2 Note that H. For full generality see mumford chapter 12. As in the proof of V. (c.KX = 2g − 2 − H 254 are relatively prime. af = −cd.1 x g Intersection Via Euler Characteristic Proof 0.d) 0 → Ox → Ox2 → OX → OX / (c. d) is the image of (d. d Thus there is h such that a = hd and b = −hc. 2 2 2 Clearly a = H and c = 1 + pa .5.1 TSADL: We have Since euler characteristic is additive on exact sequences. Ox is a UFD. (d.2 af + cd = 0. # (C ∩ D) = C.KX = H + 1 − π.1. Proof 1.

Plugging these into excercise V.1.7. x g pa (−D) = 21 (−D) (−D + K) + 1 = D2 − 12 D.5 c. so it's H 2.1. (C + D + K) + 1 = 12 C.4 x g Self intersection of rational curve on surface KX X ∼ dH for a hypersurface 3 (since we're in P .3 x g:a. (D + K) + C.2 gives C.D + 1 Note that also 5. x g: with above By part (a). 5. 5. . the degree is dim (X)! times the rst coecient.6 V. (−D − KX ) − 1 − pa + χ (OX ) 2 (note the last term cancels out). First we compute We have 255 .1.H = χ (C.b adjunction computational formula Riemann-roch (for a surface) gives χ (−D) = 12 (−D) (−D − KX ) + 1 + pa .1. Now we have exact sequences 0 → OX (−C) → OX → OC → 0 0 → OX (−C − H) → OX (−H) → OC (−H|C ) → 0. −H|C ).1. KX = (KP3 + dH) |X .3 V.By I. (D + K) + 1 pa (C + D) = 21 (C + D) .4 C.1. Using adjunction. 0. By the exact sequence 0 → OX (−D) → OX → OD → 1 we get χ (OD ) = D. OC ) − χ (C. (C + K) + 21 D. But this is the denition of the degree of the line bundle associated to 5. b. and dim 2).1.

1. ∗ ∗ From exc II.1. KX = p1 KC + p2 KC 0 .Recalling that KP3 = −4H . C 2.8 one point. ∗ ∗ This is (p1 KC + p2 KC ) . Thus 256 . Now we compute 5. 2 2 2 ∗ ∗ ∗ ∗ Thus KX = (p1 KC ) + 2p1 KC . p2 the projections . The fermat hypersurface 5.7 b.8. 2 Now solve for ∆ .∆ by adjunction. x g p1 .H = 2 − d.p2 KC 0 + (p2 KC 0 ) .2 contains the line x− √ n P3 Let 5. xd + y d + z d + w d = 0 V.10 Let V. C. C.1.1. 5.e. Now ∆ ≈ C so deg K∆ = deg KC = 2g − 2 .∆ + ∆.∆ = (2g − 2) + (2g − 2) + ∆2 .1. We can choose H generally enough to meet C at (i. x. KX = (−4 + d) H .9 Let (since canonical in P3 is −4H ) b.H = 1) 2 Then C = −2 + (4 − d) C.5 x g Canonical for a surface in X ∼ dH for some hypersurface.1.3. 0 Then middle term is 2 · (2g − 2) · (2g − 2) and the outer terms disappear. Then KX = (d − 4) H by adjunction 2 2 2 Thus KX = (d − 4) H . C. 2 Now H = d by V. p2 the projections C ×C →C . By adjunction. Note the diagonal is isomorphic to The intersection of the diagonal and a ber is the point. (C + KX ) = 2g − 2 = 0 − 2.6 x g p1 .1. deg K∆ = (KX + ∆) .

13 . Thus 2 (a + c − gc) = 0.1. ∆2 = 2g − 2 by (a)..12 diagonal. Sums 0 = D0 ..1. Thus −D ≡ 0 so it's closed under inverses. Thus a = b = c = 0 by solving the system. an eective divisor on X ×T is at over T when the local equations of the divisor are nonzero when restricted to a ber. V.. D and D + 0 = D + E0 are prealgebraically equivalent..7 x Algebraic Equivalence of Divisors ≡ Write Suppose for algebraic equivalence. l.. .5. the Suppose that a · l + b · m + c · ∆ = 0 for some constants.m = 1 . and m2 = 0.∆ = 1. Dn = D. . . −Dn is a sequence for −D by above. (al + bm + c∆) = 0 =⇒ a + c = 0 ∆. E + Em = D + E is a sequence for D + E Suppose Now 5.1.. Em = E . (al + bm + c∆) = 0 =⇒ a + b + c (2g − 2) = 0. Since the dierence of linearly equivalent divisors is principal. Write equivalent to −D.. . 0 = D0 . where ∆ is again. Then 0 = −D0 ....8.5. Dn = D and 0 = E0 . x g l2 = 0.. D is prealgebraically equivalent to 0.. b. Then l. l.1.. x By II....9. we just need to show that f ∈ K (X). . −D0 . inverses D as the dierence of eective D1 − D2 .. (al + bm + c∆) = 0 =⇒ b + c = 0.. Note that 5. Dn = D is a sequence for D . Then −D = D2 − D1 is prealgebraically equivalent to D is prealgebraically Now suppose that 0 = D0 . 257 (f ) ≡ 0 for . . m. D + E0 . Hence.11 b.

ΩD ) ≈ k By the hint. Note that Thus 5.(tf − u) (f ) = 0. x By bertini we can consider dierences of very ample divisors. the bottom map is degree.H = D .8 x g cohomology class of a divisor By Bertini. 0) and to 0 over (0.H = D . 0 0 Thus we want to show D. 258 by 1. Thus we consider intersections with very ample divisors. 1) .1. 0 If E ⊂ X × T is a divisor with bers E0 = D .H . c/  f H 1 (D. D and very ample H . 5. 0 Thus D. then E is at over T . c. thus wlog smooth curves. so by thm III.9 .E f (c (D)) = D.3. .1.15 V.H and D . write Consider P ic X  P ic D D as a dierence of very ample divisors. ΩX ) . 0 0 n But D.E .H are exactly the degrees of D and D in Pk . E1 = D . c / H 1 (X. H induces an embedding X → Pnk which gives an embedding X × T → PnT . the 0 n degrees of D and D in Pk are equal.9.H for prealgebraically equivalent eective D.14 (f ) restricts to over (1.1. Thus going down and right gives Now going right then down gives L (E) 7→ L (E) ⊗ OD 7→ degD L (E) ⊗ OD = D.

H) = 0. 2 2 2 2 Note as l = 0.1. If D ≡ 0 then D ≡ bl + am and  2 deg (D2 ) = deg (bl + am) = 2ab.5.9 x g Hodge inequality Let a divisor. Dene 259 .D)2 H 2 ≤ 0 =⇒ 2 since H > 0 (nak moish) we factor it out and have 2 2 D H − 2 (H. then 0 > deg (D ) = 16 (deg (D ) − 2ab). 5.H = 0. Thus 2ab > deg (D ). D0 .m = 1 .D) H) ≤ 0 =⇒ H 4 D2 − 2H 2 (H. OX ) and H (X.m) = a+b and deg (D. l. By Nakai Moishezon.D) D.D)2 ≤ 0 =⇒ D2 H 2 ≤ (H.1. then by Hodge index theorem 0 02 2 2 0 if D 6≡ 0. x g 2 determines c1 (OX (D)) ∈ H (X.18 b. Via the exponential sequence. Thus by Then Then 2 hodge index theorem D02 ≤ 0. 0 H is ample thus if D = −4D + 2 (a + b) H − 2 (a − b) E since deg (D0 .E) = a−b.H + (H. (see page 446). m = 0 then deg (E ) = −2.1.1.l)+deg (D. deg (H ) = 2.16 b. 0 2 D = H D − (H. deg (E.D) H . 2 (H D − (H.17 D V.D)2 . 5. x g H = l +m and E = l −m so that deg (D.H) = deg (D. OX ) in the sequence ).D)2 + (H. Z) .H) = 0. This second group 1 ∗ 2 is nitely generated (note it sits between H (X.

x g σ ∈ Aut (C) . r + s) D = rq (2 − 2g) + s (2 − 2g) + 2rsN .21 If b. Similar logic gives ∆. 1) : C × C → C × C . ∆ has type (1. . 1).H = 0 is a very ample hyperplane of deg (H) = 1.19 V. if we can choose innitely many components which each have dierent numerical N um X will not be nitely generated. q since Now 5. Note that.m = 1. and let Γ be its graph.5. s r 2 Now by exc V.H = 0. 1) · ∆2 . 1)∗ (f. Γ meets l = C × pt at f pt × pt.l = 1. then Thus we are 5. Thus Γ has type (q.1.m = 1 .4.10 x g Weil Riemann Hypothesis for Curves Note that Γ is the preimage of ∆ (f. Rearranging gives N ≤ 1+q+ gq+ g s r s r s r for rs > 0 and N ≥ 1+q+ gq+ g for rs < 0.b. s r s r  1 1 2 Note that g qx + is maximized at the same place as g (qx + 1) is maximized which is at x = ± √ and x q   √ √  1 thus we get |N − 1 − q| ≤ g q √ + q =g 2 q . Note that any such Pi .s gq+ g . using the notation of exc V.1.Pj < 0 H D H since if Pi .9. s can be arbitrary we have |N − 1 − q| ≤ supr.1.1. 1)∗ ∆.K ≥ −2 will decrease. Thus D has type (rq + s. Since Frobenius has degree q .∆ = N clearly gives the number of xed points and the xed point set of frobenius are the points ∗ lying in Fq as in exc IV.1. Γ.9.6.1.a and ((f. choosing D from a nite set of a nite number equivalence class. 1).1. ∆) = deg (f.1. the genus. rq (2 − 2g)+s (2 − 2g)+2rsN ≤ 2 (rq + s) (r + s) . this follows from nakai moishezon. ∆.16.20 V. Γ meets m = pt × C at pt × f (pt). On the other hand. then 2g − 2 = C. then Γ. Since r. Thus Γ2 = ∆2 · deg (f ) = (2 − 2g) · q ∆ = 2 − 2g by exc V.l = q . 2 2 Let D = rΓ + s∆ = as in the hint. can only have nitely many components not intersecting and with each additional such component.C + C. 260 of components. under 2 q is the degree of frobenius and Γ.11 x g WLOG assume that For D.

Now use (a).r.3. Since D 's looking like KC + P + Q are parametrized by the set of P.2 Ruled Surfaces 5. gives 2 2 Then Γ < 0 implies Γ. Let H =l+m Γ. then degree is not determined by numerical equivalence on a curve.2.9. 5.2.1 x g Suppose that C 0 × P1 ∼ =X∼ = C 1 × P1 . (since any isomorphism of open sets extends to an isomorphism of the whole curve by the valuative criterions).2. and b of degree 2pa which is not very ample. in other words.12 x g Very Ample not numerically equiv First we examine a curve. let H = |C0 + bf | a linear system b + e are very ample.Doing the hint by contrapositive. Recall from thm h0 (D − P − Q) = h0 (D) − 2 .2 V.2. By Fuentes-Pedreira thm 3. But two curves which are birational are isomorphic.∆ ≥ 0. then for a graph of an automorphism Γ . Thus two graphs of an automorphism are not numerically equivalent.4.2 x Marumaye.H = 2 from exc V.20 261 .1. then r.∆ 6= ∆ and thus Γ 6≡ ∆ . Γ = ∆ < 0 .2 V.9. Remark 1. we know is very ample. But by the coecient on C0 and the degree of the divisor by thm V. 5. 2 2 As in exc 6. thm V. gives 0 h (D) − h0 (KC − D) = 2g + 1 − g = g + 1. Q which is a two dimensional proper subset of the possible divisors D .1.1 V.22 V.2. Case of a Surface pa (C) > 2. If D has degree 2g .4 that D C a curve with is very ample i for any two points.1. pa > 2. when D is not of the form K C + P + Q. Γ 6= Γ which implies Γ. h0 (D − P − Q) − h0 (KC − D + P + Q) = (2g − 2) + 1 − g = g − 1. that we may nd b of degree 2pa |H| which numerical equivalence is only determined X over a curve with is very ample i b and 5.3. Let D a divisor of degree 2g .1. So this holds when KC − D + P + Q has nonpositive degree. Consider a decomposable ruled surface on X.b. By the above. Let IV.1.

1. = this is a contradiction.K is (C0 + (d − e) f ) . D ∼ C0 + (d − e) f ∈ P icX corresponds to a surjection E → L (d) → 0 2 We can also write as D ∼ C0 + (d − e) f ∈ N umX by 2.0.11.K .5 d1 . And thus we have D = 2d. IV. 2g − 2 = D.6.1 5. Plugging in to D.2. Since we cannot solve for 5.4 a x X = P (E) the ruled surface.3 where f = 0 and C0 .8) 262 .11.1. H a hyperplane.5. Let So (We can take the structure sheaf and the one given by Ex. 2 So we have that D = 2d by the above.4 x g tangent sheaf not extension of invertibles P2 is an extension → OX (d2 ) → 0. on X and D with degree 0 on X . (D + K) = D2 + D. Suppose that the tangent bundle on We have 0 → OX (d1 ) → TP2 of line bundles. Then we need D with degree d ≥ g + 1.3 x Miyanishi.2.f = 1.f ).2. we have e = 0. num Now K ≡ −2C0 + (2g − 2 − e) f by 2. Algebraic Geometry Lemma 12.K D2 = 2g − 2 − D.2. d2 . 2 By 2. Then c (TP2 ) = (1 + d1 H) (1 + d2 H). V. 1 Now we note that C × P is the ruled surface with the projection. by 2. By FOIL it's −2e + 2g − 2 − e − 2 (d − e) + 0 and we are left with an e 2 In total: D = 2e + e + 2 (d − e) = 2d + e.2. Let By the euler exact sequence 0 → OP2 → OP2 (1)3 → TP2 → 0 c (TP2 ) = c OP2 (1)3 = c (OP2 (1))3 1 + 3H + 3H 2 .3 V. (−2C0 + [2g − 2 − e] . 2 Let's compute D via adjunction.

0 Now an elementary transform X of a ruled surface blows up a point x. On the other hand. C0 is the new minimum self-intersection curve. mC0 + bf C 0 . and for x ∈ / C ∩ D we have C 0 .5 that I mentioned. and Thus if x ∈ C0 .D − 1. leaving the exceptional divisor as a ber of the new surface (see ex V. if X = P (E0 ) . d = −2 and so we're done.2. C . I get to a decomposable ruled surface.12.2.5. Then if x ∈ / X0 .12.e. Thus starting with the invariant.4.D + nm + n · multx (D) − m · multx (C). For if x ∈ C0 .2. D0 are their elementary transformed curves. closed immersion to P1 of degree 2.7.11. then C 0 . and then blows down the strict transform. and then back to using a nite number of elementary transformations. 0 0 Thus if m.7 V. This is F. we nd that if C.1).D0 = C. This will get us all invariants for this problem Finally. But the genii are dierent.1).5 x g I will do this proof in a way that solves this one and one of the exercises in section 5. Using general rules of monoidal transformations in V.D = C.5. Now consider X an 2 arbitrary indecomposable with invariant e0 . and D0 is another one on X 0 then D02 ≥ D2 − 1 ≥ X02 − 1 = X002 . Thus going high enough. 1 1 If nonhyperelliptic.2. Now subtracting enough points we can get to X × P corresponding to P (OC ⊕ OC ) (example V. x 263 . then we can obtain a new ruled surface with e1 = e0 + 1 . Λ2 E0 ≈ OX (e). D ∈ X .3. 5. we can lower the invariant .2. Let C0 with C0 = −e0 the minimum self-intersection curve.P thm 4. n = 1 . Thus we can go from any X × P1 X to one with high invariant which is decomposable. taking e0 I can arbitrarily lower or raise the reverse transformation. Thus it's an isomorphism. then X is indecomposable with e = e − 1.D0 = C. then for x ∈ C ∩ D .6. embedding in 1 If it has an r = 1 then that would mean we have a degree 1 map from C to P . x ∈ / X1 and P is 0 0 a base point of −e. P2 . x If hyperelliptic. suppose X with the help of thm V. 0 Note that C0 the elementary transform of C0 .5.8 b. is arbitrary decomposable with two sections C0 . Λ2 E00 ≈ OC (e − P ) where x lives in the 1 ber over P .D + 1. Pedreira 4. then by 5. By Fuentes. Thus gives the exercise in V.6 b. 5. . C1 . are nC0 + af .. then the new surface correpsonds to E00 .2.2. then has a Then OC (−1) has g21 i. it has a g3 but not a g2 1 And it is obtained by projecting from a point to P from it's deg 4.

we have shown that E normalized =⇒ for every such E . a nonzero section dening the divisor S . So in particular.6 x g Grothendieck's Theorem (Following Potier) Given a base case of rank 1 so it's already invertible. x By Theorem 1 in Nagata Self-intersection number. LE 6⊂ H . Thus there is E with H (E ⊗ L (−E)) 6= 0. I only need to nd one with invariant −g .11 V.2.9 c. The image of any nonzero element of H (L (E)) by β is a section of H (L (D + K)) corresponding to the eective divisor E + S ∈ |D + K|. 264 . E + S = (D + K − E) + E . 5. x By proof of (a). Now β is dual to the map β 0 : H 0 (L (E)) → H 0 (L (D + K)) also 0 0 0 induced by t.E H 0 (E ⊗ M ) = 0 so the map γ : H 0 (L (D + K − E)) → H 1 (L (−E)) must be injective. α (1) = t. then after twisting.8. where H is the kernel. β (ξ) = 0 contradicting injectivity of γ .10 d. that there exists E such that LE ⊂ H . Thus the 0 1 map γ : H (L (D + K − E)) → H (L (−E)) is not injective and therefore we can nd LE ⊂ ker (ξ) by α using commutativity of the above diagram. 5. If δ H 0 (OC )  α / is normalized. This is given by 3.13 in Maruyama. then H 1 (L (−D))  β / H 1 (L (−E)) H 0 (L (s)) where δ (1) = ξ.s. If i is the largest integer where E (−i) admits a section. therefore image of β as E varies lls up whole H (L (D + K)) . let ξ ∈ H 1 (L (−D)) be the element dening the extension E . By varying E and S .15.. On the other hand. and β is induced from the map OC → L (S) corresponding to t. So suppose to the contrary. so E is not normalized.2. 0 On the other hand suppose E is not normalized. Thus I need to nd a ruled surface with C02 = g .e. we get every divisor in linear 0 0 system |D + K|. I Follow thm IV. writing L (D + K − E) as L (S). then LE 6⊂ H .. So if LE ⊂ |D + K|. In this case. we get an s. 5.2. Then we have a commutative diagram.c and Serre duality. then |D + K − E| + E ⊂ |D + K| On the other hand. Thus we have shown the contrapositive of E is normalized =⇒ for every such E .2. assume locally free sheaves of 1 rank r − 1 all split on P .2. we can nd n  0 such that E (−n) has no sections. By thm III.8.

where F has rank r − 1. Then P f ∼ Qf so L (P ) ≈ L (Q) Note that on an elliptic curve we have the following theorem: 0 If E is an elliptic curve the map E → Div (E) /P rin (E) dened by but then P ∼ Q.15. Then the two exact sequences correspond to dierent points L (P ) and L (Q) P 6= Q . < deg(imϕ) < deg(E rk(E ) rk(imϕ) rk(E ) Thus im ϕ = ϕ so the kernel is 0 since it's locally free. Claim: E stable =⇒ E to itself is an isomorphism.0 → O (i) → E → F → 0. ϕ : E → E as ϕ : E → im ϕ → E . 2 Now for E we have C0 = −e = − (−1) = 1 if we use the notation 2. Thus by be indecomposable.11. Then F ≈ ⊕j O (j)rj by induction.11.2.1.2.2.6. V.15. then by stability.14 b. E must By 2.13 P and Q are the same point. 5.8. for this value of Now according to the defniitions 2.7 x e = −1.2. P → [P − 0] is a bijection denote the equivalence class) so this would mean 5. E → L (d) → 0 point P ∈ C .12 V.1 on the exact sequence E → L (Q) → 0 and similarly. (by 2.12. locally free.6. H om (F . sections correspond to and the exact sequence is 0 → OC → E → L (P ) → 0 2) for some and by 2. O (i))) = 0 so the sequence splits by Weibel 10. But then H (X.a. Factor 5. (brackets . Suppose that C0 ∼ C0 + (P − Q) f .8.x g 265 contradiction. every non-zero morphism of Given this.2. d has degree 1 By 2. with j ≤ i for rj 6= 0 1 or else E (−i − 1) has a nonzero section.1 and exc III.a x g decomposable is never stable A decomposable is a direct sum of two invertible sheaves. 2. f : G ⊕ F → G ⊕ F → f = idG ⊕ 0F is clearly not an iso.6. deg(E ) ) If im ϕ 6= E . there is exactly one ruled surface over ×C e. if we use the notation on the exact sequence E → L (Q) → 0 (maybe we can say it's C00 in Now suppose that where this case). . then clearly Proof of claim. rank By 2.16.9. there is a natural 1-1 correspondence between set of isomorphism classes of indecomposable 2 ∼ E 6= E 0 locally free sheaves of rank and degree 1 on elliptic curve C and set of points of C.

5. and F is invertible and deg F < 0 so by the normalized is indecomposable of rank E F → E. this is clear by exc V. Else.2. Note that (b) follows since µ (G ) < µ (F1 ) for G ⊂ F /F1 .6.15 Let c. χ (f∗ G ) = χ (G ) and s. let f : X → P1 with On pushforward of any subsheaf G of F is a since the euler characteristic is additive in f ∗ OP1 (1) = OX (1) be nite. Since E is non-semistable and F1 is deg L + degF1 = degE . exc III. Note the line has multiplicity 266 . For the uniqueness of this ltration. and the normalized assumption. that deg (F ) ≤ 21 deg (E ). Since necessarily Since the ltration was unique we have this up to isomorphism.e.s.E is (semi) stable of (negative) non-positive OC ⊂ E so deg (E ) (≥) > 0 . −deg L ≤ 2g − 2. see prop 5. then degree . we may therefore assume pa (C) > 0. Otherwise.16. OX ) ≈ H (C. 5. and thm V. The subsheaf of f∗ F . we have a nontrivial extension 0 → F1 → 1 1 ∨ As in thm V.8.9 x g Curves on Quadric Cone So if it's a complete intersection. Thus there is F ⊂ E with deg (F ) > deg (E ) > 0.4. we nd the χ (G ) is bounded. Now suppose E has positive degree and is rank 2 normalized. in the ltration. geometrically this corresponds to some multiplicity of the conic section together with the line. L ) with Now consider the maximal slope subsheaf F1 E → L → 0. then by thm V. we can split the two cases into curves intersecting the vertex and not.4. Since E is rank 2. we can nd F1 ⊂ F with maximal rank among those of maximal slope.2.8. For if (semi)-stability. we need for F ⊂ E . Teixido. F1 is semi-stable.2. E ⊗ ⊗ F ∨ ) = 0 so Assume to the contrary that as there is a map 0 assumption h (E r.8 degree of F would be ≤ 0 so we have the inequality. Claim: E has an increasing ltration by 0 ⊂ F1 ⊂ · · · ⊂ Fk = F where (a) Fi /Fi−1 is semi-stable and (b) µ (Fi /Fi−1 ) < µ (Fi+1 /Fi ) . Since degree is deg (G ) = χ (G ) − rk (G ) · χ (OC ) .12. we must have −deg L > 0. Since the quotient is locally free we can repeat. 2 a − 2a + 1 = (a − 1)2 . and thus by contrapositive we have the if direction. Assume E By exc V.1.2. it has global sections. use 4.2. By Leray.2. by the rst claim. actually it should have no sections. Potier. then by exc II.2. vector bundles: An indecomposable of degree zero is semistable but not stable. maximal slope. deg = 2a.2. and g = 12 2·a (a + 2 − 4)+1 = a (a − 2)+1 = On the other hand. If F is semi-stable then this is clear. this corresponds to nonzero ξ ∈ Ext (L . P1 .16 V. But then 2 ∨ ∨ F has a section.. x g E non-semistable.6. then the degrees are bounded above. Since E E is normalized. contradiction to First note that if has a section. By maximality.2. 1 is not stable. If E is decomposable. then deg F1 > 0 . then E has no sections. since at any rate rk (F ) < 2 . To discern between semistable and stable in this direction. degree d.12(a). Claim: The degrees of coherent subsheaves of F are bounded above.

2. C1 .10 x See example 2. To compute the genus.C1 = C0 (C0 + 2f ) = −2 + 2 = 0 Our original curve which was in the form aC + f . we see the degree is • D 2a + 1 some Y through the vertex will hit q once and p + p twice and a. In particular. Now we attempt to compute .1. . where conic. on the cone will be some multiple of the section A hyperplane intersecting using exc V. achieving a ruled surface C0 is the section of the C1 = C0 + 2f .f = (C0 + 2f ) .f = 1. X over (this is via C0 .11).1 since it's nonsingular and it's going through the vertex.2.1. (else too many tangent directions) Using what we know of ruled surfaces. Thus the hyperplane section on the cone lifts to C12 = C02 + 4 = 2.1 we see that the degree is 2a + 1 since this is the number of intersections.17 V. therefore lifts to genus using adjunction. 5.10 in Kollar's Complex Algebraic Geometry.2. with e = −d.4 the conic. a divisor some multiple of the ruling q.11.  expand 12 · (a · C1 + f ) · (a · C1 + f − 2 · C0 − 4 · f ) + 1 = 2 C 2 a2 − 3 2f − C1 a f − C0 f + 12 − C0 C1 a + 1 2 2 This gives −a − 1 + a + 1 = a − a which is what we wanted.2. In particular. K = −2C0 − 4f V. we blow up the point on the cone via V. 267 aC1 + f .1. Geometrically considering the number of intersections of a plane passing through the vertex (this gives the line) and via exc V.

2. for arbitrary P. If you can just take a tangent along the ber and containing 5.5.1. Since b is nonspecial. From the picture. 5.12 x 268 over P and t ∈ Tx (f ).2. φ very ample on separates tangent vectors. h0 (OX (C0 + bf − (P + Q) f )) = h0 (OC (b − P − Q)) + h0 (OC (b + e − P − Q)) = by thm IV. for if x.2. Further. . then H = |C0 + bf | is bpf on P f (the ber over P ) i h (OX (H − P f )) = h (OX (H))− 2.18 V. y are two points on the same |C0 + bf | is very ample on the ber and thus separates Then ruling then by the second claim in (a) and thm IV.1. Now use exactness. hi (OX (bf )) = hi (OC (b)) and hi (OC0 (C0 + bf )) = hi (OX (b + e)). . Claim: If Proof: P f ≈ P1 b is nonspecial.20 f x. x b−P is bpf since b is very ample and using thm IV. Now let φ be the map determined by |C0 + bf | . b−P is nonspecial as P b. . using the same left exact sequence and noting that since we're on a ruled surface.3.3.1. y φ is injective. since if f x is in the ber so there is a tangent vector through x V. Q on the curve. Claim: |D| has no basepoints Proof:Since there are eective divisors linearly equivalent to point of both. x. Thus |C0 + bf | |C0 + bf | is t∈ / Tx (f ) . h0 (OX (C0 + (b − P ) f )) = h0 (OC (b − P )) + h0 (OC (b + e − P )) = which by bpf is h0 (OC (b)) − 1 + h0 (OC (b + e − P )) = if P is not a basepoint of h0 (OC (b)) − 1 + h0 (OC (b + e)) − 1 so we're done.2. P can't be a base it being bpf. but then h0 (OX (C0 + bf − (P + Q) f )) ≥ h0 (OX (C0 + bf )) − 3 using the rst claim in (a) which is a contradiction. meeting the ber transversally.2. any generic point b. Also h2 on the curve vanishes and h1 (OX (bf )) = 0 by nonspecialness. Thus by b+e b. By the rst claim in (a). then then separates tangent vectors.3. Proof: Consider the LES associated to 0 → OX (−C0 ) → OX → OC0 → 0 . then h (OX (C0 + bf − P f )) ≥ h (OX (C0 + bf )) − 1.11 x hi (OX (C0 + bf )) = hi (OC (b)) + hi (OC (b + e)). b + e is not a base point of the rst claim. 0 0 Claim: If P ∈ C . 0 0 Now if |C0 + bf | had a base point P .1 h0 (OC (b)) + h0 (OC (b + e)) − 4 = by the rst claim in (a) h0 (OX (C0 + bf )) − 4.3. then This is essentially rephrasing thm IV.19 Note .

This would give N = d − 1.. b.15. get H (C. then This map is degree a (mod p) so by (thm V.19 slightly) So by thm V. by thm V.14 x ≥ 2..2. we have nf ) . .2. x Fuentes. L (D)) = H 0 (C. C.4. If char k = p .12.2.2. π∗ L (D)) = .3.) where the last term has dimension d = 2n − e.3) 0 By riemann roche (c. (C0 + nf ) = 2n − e so image has degree so the rst part is done.20.f = C0 .2. L (D)) = H 0 (C.. OC (n) ⊕ OC (n − e)) (c.2..   ˜ (thm IV.f = 1 and f 2 = 0 .2.2. this last inequality is true in any char if g = 0. Futhermore. pp381).8) so On other hand. pa (Y ) ≥ g Y 2pa (Y ) − 2 ≥ α (2g − 2) by getting rid of deg R. we have)   2g Y˜ − 2 ≥ α (2g − 2) + degR where R is (eective ram divisor). pp309. since in any case.13 x (Follow thm V.2.f.b. This is ample by thm V.3.1. V.2. 5. thm pp 319.2. Substituting Y =≡ aC0 + bf and K ≡ −2C0 + (2g − 2 − e) f inequality above we nd that (aC0 + bf ) . pa (Y ) ≥ g .. π∗ L (D)) = H 0 (C. a.21 b.4 V.3. We need V. 2. Pedreira Prop 1. H 0 (X. Predeira prop 1.21 assume genus is Let projection By adjunction. C0 . = H 0 (C.12. we have 2pa (Y ) − 2 = Y. (Y + K) . N.a. that D. exa 3. 1.11). Y˜ the normalization of Y and consider the composition of natural map Y˜ → Y with the π : Y → C .. OC (n)) = n and so this dimension is 2n−1 = d−1.f + nf 2 = 1 so it's an elliptic scroll.This is Fuentes.f. Now just need to nd 2n − e = d.f.. Then using thm V.. Let e = 0 or e = −1 are possible for indecomposable e ≥ 0 are possible for decomposable. and combining with . ([aC0 + bf ] + −2C0 + (2g − 2 − e) f ) ≥ α (2g − 2) 269 from (thm V.2. D = C0 + nf .f = (C0 + D2 = (C0 + nf ) . thm IV.21. H 0 (X. thm V.

suppose that a ≥ p. Since Y is an irreducible curve. degL ≥ degE .22 b. 2 1 And thus we have b ≥ ae + 1 − g .2. Then D. we have b ≥ 0 (in fact ≥ 1 since Y 6= C0 ) so using: a > 0. 5.Y = −ae + ab0 + b > −ae + b ≥ −ae + 12 ae = − 21 ae > 0 since e < 0 by assumption.f = a > 0.C0 = (aC0 + bf ) . C0 .  1 1 2 2 Also D = −a e + 2ab = 2a b − ae > 0. 2 2 p 270 that . Thus (a − deg (π)) > 0.2. then b > 12 ae. By part (a) b ≥ a e and by assumption. 2 D. the same proof as thm V. 1 1 We will have b (a − 1) ≥ ae (a − 1) + (a − deg (π)) (1 − g) 2 2 Now we don't know degπ except it's less than a.Y by thm V. f . 2 In the case a = 1.f > 0 and Suppose that a>0 and     b > 12 a e + p1 (g − 1) = 12 ae + 1a 2p (g − 1) . b > ae.9. 1 Thus b > ae. If a ≥ p. then D2 = (aC0 + bf )2 = a2 C02 + 2abC0 . Suppose that 0 1 0 1 0 0 Next suppose that 2 ≤ a ≤ p − 1. since b − ae > 0 and a > 0. 2 2 1 1 Also D. and e<0 in the assumptions of the problem).f = (aC0 + bf ) .Y > −aa e + aa e + aa e + a − ga + (g − 1) . if D is ample.f = 1.Y > −aa e + aa e + aea = 0. we have ab0 > 0 and thus D. b ≥ a e + 1 − g . So that if p − 1 ≥ a ≥ 2.Y = (a.   2  1 1 0 0 0 To recap.21. 2 p 1 a0 0 0 0 0 1 Thus ab ≥ aa e + a − ga and a b > a a e + (g − 1) .1.f + b (2g − 2 − e) . By part (a). 2 2 1 0 0 0 Finally.2. 2 2 h p i   1 1 aa0 0 0 0 In total.C0 + bf ) . D.f + b2 f 2 − 2aC02 + aC0 . (as is the case when 2 ≤ a ≤ p − 1 ) then we retain the inequality from V. then the term (α − a) (2g − 2) will be (−np) (2g − 2) = np (1 − g). Note g ≥ 2 so (1 − g) < 0. which is b (a − 1) ≥ 21 ae (a − 1). 1 The most that (a − deg (π)) (1 − g) can be is when deg (π) = 1. corresponding to surjective map E → L → 0. x Using thm V. Since E is normalized. Note that when α = a. we have b ≥ 21 ae as required.f + ba0 C0 . b > a e + (g − 1) . (a0 C0 + b0 f ) = aa0 (−e) + ab0 C0 .2.C0 = −ae + b > −ae + ae = − ae > 0 (since a 2 2 Since is positive. Let a0 = 1. 2 2 1 1 0 0 0 In total.a holds: Y is a section.f = −a2 e + 2ab > 0. (2g − 2 − e) f − 2abC0 .f + 0 = −aa0 e + ab0 + ba0 . so b − e ≥ −e and b ≥ 0.Y = −aa e + ab + ba has ab ≥ aa e and ba > aea . But degL = C0 . Then D. 2 2 1 1 0 0 0 0 0 0 0 Thus D. D. g > 2. we also have a > 0.f 2 2 2 we use that C0 = −e.LHS is a2 C02 + 2abC0 . Y an irreducible curve 6= C0 . and f = 0 so that the above LHS becomes: 2 a (−e) + 2ab − 2a (−e) + a (2g − 2 − e) − 2ab.21. and a > 0.Y = −aa e + ab + ba . In the end this becomes ae (1 − a) + 2b (a − 1) ≥ (α − a) (2g − 2). b ≥ 21 ae and that b0 ≥ 1. then since Nakai-Moishezon works in any characteristic.10. D. then by the rst part of the excercise.

OP2 (−4) so we have an exact sequence: 0 → OP (−4) → OP2 → OX → 0. OX ) → H (X. OP (−4p))  f p−1 H 2 (P2 . we are done. OP (−4))  F1∗ H 1 (X p . (since f is a factor /0 OX p−1 p (we can multiply by the equation f to go from the scheme f to f and note that if we start in the thing twisted by −4p (p is the characteristic . (char 3) On other hand.. Thus: / 0 F1∗  / / OP OX F1∗   / 0 F1∗ / OP / OX p /0 OP (−4p) 9 3 9 3 9 3 Namely x y + y z + z x = 0. OX p )  H 1 (X. 1 we have H no (P2 . we get → H (OP2 ) → H (OX ) → H (OP2 (−4)) → H (OP2 ) Via stacks. 2 ∗ If F1 is Frobenius morphism on P .which is 3 ) then we will end in the thing twisted by −4  compute 0 /  is closed subscheme of OP a few examples if it seems mysterious)) Now combining the cohomologies of the above two diagrams gives: ≈ / H 1 (X. OP2 (0)) = 0 H 2 (P2 .1 to get that H (P .2.15 x Funny behavior in char p P2 By degree genus in 3 is genus Let's try and follow thm IV.The rst term is 0 and the second factors to  a (1 − g) 1 − a0 p  ≥ 0. By Nakai-Moishezon. OX ) H 2 (P2 . 1 1 2 2 Taking cohomology.4. OP2 (0 = −r − 1 − (−r − 1))) ≈ H 0 (X.21 1 Calculate H of quartic curve in The ideal sheaf is isomorphic to quartic. 1 and 1 We compute action of frobenius using this embedding. where Xp is subscheme of P2 (since dened by p f = 0. OP (−4)) ≈ H (P . OX ) on the left hand column. OX (−r − 1)) = 0 2 global sections) so we obtain H (OX ) ≈ H (OP2 (−4)). P2 ..   2+1 2 0 Now H (OP2 (−4)) ≈ H (OP2 (1)) of dim = 3 vector space. OP (−4)) 1 1 map H (X. V.. 271 . 2 2 0 2 5. then F1 takes OX to OX p . OX ) ∗ Now F is the We can use ≈ ≈ /  F1∗ / H 2 (P2 . X 3 (f ) ) so we have a commutative diagram: / OP (−4p) / OP / OX p /0 0 0 of / OP (−4) f p−1 /  OP (−4) /  Xp . OP (1)) with dim 3..

lemma 2. . Every P 1 element of H (C.2. −3 p−1 2 6 3 4 4 3 6 2 3 4 6 2 2 −6 −3 −3 Now f is (x z + 2xy z + 2x yz + y z + 2x y z + x y ) and (x yz) is (x y z ). OP (−4)) 2 −1 (xyz )  −p = (xyz 2 ) is (x2 yz) −1 . OC (−1)) 1 will be f 2 · ξp. y. x Y ≡ 3C0 − 3f is inseparable is Miyanishi. 5. and any monomial having nonnegative exponent is 0. OC ) can be represented by a polynomial like aij xi y j with 1 ≤ i < d and −i < j < 0. So since it says plane curve. So Y is some section of To see the curve signicance of C lying on the surface. Thus the image is just 1 times coecient of β . The π : Y → C is that then they will be isomorphic as abstract schemes.   2 subst 1. and the im- −p H 2 (OP (−4)) (bottom right of above diagram) will be (for example) f p−1 ·(x2 yz) = f p−1 (x−6 y −3 z −3 ). Writing this  out gives I can look at f2 in ane coords. and (xyz ) as a free F3 . Now since we have a genus 3 and we are not in characteristic ample to 2.21b since those hypothesis are required by 2.24 1.2. OC (−1)). c. 2 we're in P . x y z ∗ Then F1  2 −1 (x yz)  2 = (x yz) −p ∗ . (xy z) .4. x H sequence: 0 0 Now be hyperplane section. we have a chance of getting a counterex- .5. then −4 < i ≤ −1 and i < −j < 0 which. expand (x3 · y + y 3 · z + z 3 · x) = y 6 + 2 x3 y 4 + 2 x y 3 + x 6 y 2 + 2 x 4 y + x2 . x y x x z x z yz 3 age in −3 z3 xy 6 3 2z + x13 z + y22z2 + yx4 z3 + 2x x2 y 3 y5 y3 2y 2x x3 2 p−1 2 −3 and nally. OP (−3)) is (xyz) . and thus each of the above expressions b.2. -module. F1  −1 2 (xy z)  2 −p = (xy z) ∗ . as in exc III. On the other 1 H (OC (−1)) has basis β . 3 3 2y y 2y z 1 2 We multiply these giving 4 3 + 5 + 6 + 3 2 + + x2 y 2 . and F1  H 2 (P2 . First make the bottom sequence: 0 → OX (−3) → OX → O → 0 put / OX /O 3 /0 / OX (−3) H H3 /  f 3−1 OX (−1) ∗ F (ξ) = ξ p  / OX /  OH and it's image in 0 → OX (−1) → OX → OH → 0 Now top them together: / 0 H (C. or z is ∗ is 0.7 will give a zero image.7 I computed what these elements look like explicitly for a plane curve. Now any monomial having a nonnegative exponent on x. So I need to determine the basis β of H (C. z.−1 2 2 Since a basis for H (P . then we must have a basis for h i 1 1 1 2 −1 2 −1 .21b. Open Algebraic Surfaces. f (xyz ) is x2 z2 + y2 z3 + x3 z4 + z5 + yz6 + xy1 3 .4. If hand. Similarly. Thus F is identically 0. In the exc III.2. β we twist by negative 5.23 Let f p−1 (xy 2 z) is + 0 . 272 0.

12).16 x So suppose they are isomorphic. B ∈ Obj (C) have the same classes in the grothendieck group i there is D ∈ Obj (C) with A ? D ≈ B ? D . 5.15 or 2. so in particular. Then have ses 0 → 0 → E → E0 → 0 so E − E0 − 0 = 0 in K (X) are the same. 1 2 (6) · 6 .2. Note that the invariant is the degree of the the sheaf L (P ) (see for instance Theorem 2.3 V. then A.2.17* (starred) starred 5.3.2. E ≈ E 0.2. Clearly 2D will not be ample.26 V. 273 is calculated from the Euler .77 locally free.1 V.What I need to nd are a > 0.3 Monoidal Transformations 5.25 V. Now by Rotman 7. b > 12 ae where invariant e is invariant of ruled surface is < 0. 5. Since ˜ for some nitely generated projective O (X)module M . since so it's −1 as C0 = −1. K (X). On the other hand suppose class in groth group are the same. If we let a = 6.3.4. it must be M gives that if C is a ?-category. (6C0 − 6f ) = −6 − 6 < 0 and by Nakai moishezon. which says that   ˜ ≈ H i (X) and the fact that pa Hi X characteristic. then −6 > 2 C0 .1 x g This follows from 3. Now use the fact (Rotman 11.118) that stably isomorphic on dedekind so classes in domain (such as a smooth ane curve) is equivalent to isomorphic. b = −6. Note that an ane variety is a smooth curve i it's a dedekind domain (for this denition of a dedekind domain see Bruning coherent sheaves on an elliptic curve).

1.E − E.3.2.7. l So I need to show that Ml = A/m so note M = ⊕d∈Z Md   d d+1 l so then in our case.3. π C = C ˜D ˜ = (π ∗ C − rE) . we have ϕM (l) = dimk Ml where Ml is the lth graded part. anything in ˜ P icX may be written as d>1 in P2 cannot have a point of multiplicity C on X .3. 2.3 x g Ok well I.. via 3.7. Now since D Y of degree is very ample. is the degree of Now we compute ∗ ∗ (2π D − E) (π C − rE) = 2π ∗ D. The second part 5.D E.π C = C. then D.π ∗ C + rE 2 . (π D) + rr E.3.3.π D − r E.π ∗ C = 0 2 and E = −1. Note that 274 C.E = C.5.π C − (rE) . (2π D − E) = 4π ∗ D2 + E 2 > 0 5.C − 2π ∗ D.2) .5. ∗ ∗ 0 ∗ ∗ 0 − rE and π C. then for any other curve Then by V.4 since D2 = 1.4 x Multiplicity of local ring grm A = ⊕d≥0 md/md+1 is a graded ring..6. This generalizes simply to Pn . (π ∗ D − r0 E) = Then C.D − rr0 . So that's the rst part.C π ∗ C − rE.3. and it becomes 2deg (C) − r since E 2 = −1 by V. The middle terms drop out by V. and in I.3.3.2.2: ∗ ∗ π D.3. D follows trivially. But then we know that ∗ ∗ and so we have Further. Ml = m /m so I wts that length A/m = dim ml−1 /ml . ∗ ˜ + rE so C˜ = π ∗ C Usnig thm V. l Now ψ (l) = length A/m in the problem.2 V.1.3. V.2 x g We have the following rules of intersection theory. V. deg (C) > r (2π D − E) (π C − rE) > 0 . says that an irreducible curve d . . note that ∗ ∗ (2π D − E) . (via V.3 ≥ ˜ = π ∗ D − r0 E .

. (note grm A = ⊕d≥0 md /md+1 ≈ A m  l A/m l−1 Thus in our case m /ml becomes dim A [X .5 b. $10 5 275 4 is a eld) PA (l) = ϕM (l) = ψ (l) for l  0.5.. x 3 http://www. Xn ] /m = length A/m . Xn ] and so m   A l l length m [X1 .   l A l Thus must show that length [X . ... Xn ] in n variables.edu/~r-ash/ComAlg/ComAlg6.math. vol 2. X ] /m = length A/m 1 n m A But from the isomorphism.3. $10 gives us that [X1 . xn for m= P Axi exists 3 .. X ] 1 n /m . we have that there is PA (z) ∈ Q [z] such that 5. . vhap VIII.. m Since this is a vector space..7. vhap VIII.. ..uiuc. we have that A ≈ [X1 ..pdf 4 zariski samule. vol 2. so zariski samule.... ..so ok we know that a sop x1 . . 5 Now applying thm I.. .. length and dimension coincide.. ..

5. ch VIII.p (l) = an ln + an−1 ln−1 + .p (r) (Note that Now we know the Hilbert-Samuel function looks like POC. say U = Spec A.f.p / (x. we may assume ane.p /mr ) = limr→∞ (OC.This is given in Zariski samuel vol 2.p /mr ) = limr→∞ POC. change the length. x ( he means n! × the leading coecient of PA ).p /m2 ) = length OC.y] h 7 Now note the local ring is being given by OC. y)2 is a little bigger Assume that Let m C has multiplicity be sheaf of ideals of P on .7 d. .3. so it's n.p /m) = length (OC. and part (a). Thus length (OC. we must have  A degPA = dim Z Ann m [X1 . then the annihilator is just Z (0) = Pn which will just have dimension n. o k[x..3. 6 7 (think gathmanns local ring def) 276 then further quotienting doesn't .p /mr ) is a little bigger but now note that since f ∈ / mr+1 .. Following chapter 9 Zariski-Samuel.p / (x. 6 alternatively.. In particular.. y generate m in some neighborhood U of P .7. Since we are trying to annihilate a vector space. 0.. f n ∈ / mr+1 . length (OC. So 5. x.6 c.. $10. r By denition of multiplicity in chapter. y) a local equation for C on U (shrinking U if necesarry). x c. X. y)) is small. V. Let f (x.  length (OC. + a0 ...6 r at P . length (OC. Xn ] where Z denotes the zero set in Pn of a homogeneous ideal. 5.3. f ∈ m . we have that dimension of A is the length of the system of parameters.p = g ∈ (f ) : g (p) 6= 0 . he is not actually asking any question here.. note by I.

. So I just have to show that the length stabilizes correctly n r i.y)n ) has i  k[x. up to it increases by 1 each time. let's maybe try and gure out a0 ? r = 1 . l → l × n!r + constant. r=3. −1 r = 2.. Here is the picture: So we are localizing at the green point. we have that length Thus I just have to compute the length of we'll here's an example: M = k [x. and then multiply through the n! (for the dimension  it's 1! = 1 in this So the rst coecient is So for a So we'll an = liml→∞ an l n +. deg (M/m7 ) = 28 Now it stabilizes. 0. (fr .p /ml → l × r + constant. y) ]) = r × n + .. Thus localizing won't change the length (This is called the equidimensional case). deg (M/m9 ) = 42 Now note that 35 + 21 = 56 which /8 is an integer 42 + 21 = 63 which /9 is an integer. To compute the length this is how I will argue: k [x. + c0 be the Hilbert Poly of M .. Since localizing commutes with quotients..(x.+a 0 ln case). hilbert thing can also be written as length(M/q k M ) . n so length (M/m ) = a1 × n + a0 so assuming a0 = −21. let cn k + .. here's the idea: Note that µp (C) as dened by the problem (as the leading term of samuel. r = 5 it's −10 for r = 6 it's −15 277 n! times the coecient of the . then a1 × n must be 7 × n. I want to show that length OC. and we just are left with showing that  length OC. y] / (y 7 − x9 ) with deg (M/m) = 1 .p /m divide out the l .p /ml ) = liml→∞  ln where n is the dimension. deg (M/m2 ) = 3 .e.length(OC.(x.y]m (f r .. and we are still in the fuzzy area with blue and red. which makes sense. y) length h k[x. m = (x.. y r ) will be equidimensional...) + (x... since it's the degree we want.y] .y)n ) m = length  k[x.. adding 7 each time deg (M/m8 ) = 35 . n!limk→∞ kd n To see this.y] (f r . length (M/ [(y + . curve (n = 1 ). Ok.3 for r = 4 it's −6.y)n )  .(x. y] / (mn . dim (M ) = 1 ..

8 e.... branch points.5 x g hyperelliptic every genus Singular at innity on y -axis So this is referring to the point (Since it's on y -axis. . + z r P ai  rxr−1 + .e.... yx and . x and yx. we are done.. Now at ∞ r−2 r not in mr−1 where m is the every genus . xn−(r−1) y r−1 plus redundant terms.. z) = P2 .. In total it's y r−1 x. f ∈ m .. .. ∞ on projective space P (x. + genus. yx .11 5.. y r−1 x2 . . y . y r−1 xn . Now note that by projecting to the x-coordinate. This is 5. M/ (y r .. x See Mumford. In our case. This is n × r − 2 2 Since nr is the linear term. (x. and x. .. y . y) |∞ then The projective closure The jacobian at ∞ = (0 : 1 : 0) is Thus it's a singular point.. which is 0.. Algebraic Geometry 1.for r=7 it's so it goes up −21 1 each term we add is like time... we take Sn = n(an −a0 ) so the rst term is 2   r(r−1) which coicidentally is 2 n 2 −1. x = 0 and innity meansP z = 0) Q r−2 2 r r−1 of C is z y = x + x z ai + . 2g + 2 P1 of degree 2. x . formula (IV... + (n − r) r + (n − 1) + (n − 2) + . then the equation gives a cover of By Riemann-Hurwitz. Note that arithmetic sum is − (r − 1) so maybe we have in total. + z r ai . y. 2 3 r−1 2 n−1 2 n−2 Thus a basis: 1 and y.3. The branch points are namely the roots xi ..3.2) such a cover will have places where there is one value of x for the value of y. y)n ) n n−1 ideal on right is gen by x . Proposition 5.... x y. + (n − r) This is (n − 0) + (n − 1) + (n − 2) + . + (n − (r − 1)) Pr−1 Pr−1 Thus it's j=0 n − j=1 j   r (r−1)(r−1+1) = nr − .. + z r−1 2yz r−2 (r − 2) z r−3 y 2 + (x...9 V.y)n   = rn − r 2  for n>r . md + (n + 1) So basically I need to prove that length  M (y r )... y . the last .3.. Compute δP = (0 : 1P : 0) we subs inQ1 for y and the equation becomes: r−1 z = x + x z ai + .. .(x. maximal ideal of the point ∞.. . r−2 Then r − 2 is the multiplicity since it's lowest order term (i. up to 1 + (r − 1) + (n − 1) + (n − 2) + (n − 3) + .

// t h a t ring loaded R=0 .7 x The following code will compute an embedded resolution of singularities in Singular. LIB " r e s g r a p h . C h a r t +++++++++++++++++++++++ 0 in c h a r t −t r e e : c h a r t −t r e e : 1 1 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ==== Ambient Space : _[ 1 ] = 0 ==== I d e a l of Variety : _[ 1 ] = y5+x3 ==== E x c e p t i o n a l Divisors : .3. dp .11 V.10 n = 2g + 2. LIB " r e s z e t a . I=x^2−y ^ 3 .9.3. l i b " . let's do a resolution of Current the charts of x3 + x5 = 0 Current charts : currently chart in . 5. However the example is then given of the two 3 7 3 5 7 curves C1 : y + x = 0 and C2 : y + x y + x = 0 which are equisingular but not analytically isomorphic. some (x . and exceptionals LIB " r e s o l v e .4. For example.3. we obtain hyperelliptic curves of any genus we desire. nd strict transforms. y ] for example 1).3. Singular Points of Plane Curves where he proves that Analytically isomorphic =⇒ puiseux characteristic equivalence in the sense of 3.6 x See Wall's. V. y ) .So choosing 5. ideal libraries // d e f i n e // d e f i n e list L=r e s o l v e ( I . =⇒ same Furthermore he shows that two curves are equisingular i they have the same puiseux characteristic. list c o l l =c o l l e c t D i v ( L ) . a cusp ++++++++++++++ O v e r v i e w number of final Total number of Index of current the r i n g Q[ x . l i b " . l i b " .

empty list ==== I m a g e s of variables of original ring : _[ 1 ] = x _[ 2 ] = y −−−−−−−−−−−−−−−−−−−−−−−−−− Upcoming Center −−−−−−−−−−−−−−−−−−−−− _[ 1 ] = y _[ 2 ] = x −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−− ++++++++++++++ O v e r v i e w Current number of Total number of Index of current the of final charts Current charts : currently chart in C h a r t +++++++++++++++++++++++ 0 c h a r t −t r e e : in c h a r t −t r e e : 2 2 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ==== Ambient Space : _[ 1 ] = 0 ==== I d e a l of Variety : _[ 1 ] = y (1)^3+ x ( 2 ) ^ 2 ==== E x c e p t i o n a l Divisors : [1]: _[ 1 ] = x ( 2 ) ==== I m a g e s of variables of original ring : _[ 1 ] = x ( 2 ) ∗ y ( 1 ) _[ 2 ] = x ( 2 ) −−−−−−−−−−−−−−−−−−−−−−−−−− Upcoming Center −−−−−−−−−−−−−−−−−−−−− _[ 1 ] = y ( 1 ) _[ 2 ] = x ( 2 ) −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−− ++++++++++++++ O v e r v i e w Current number of final Total number of Index of current the charts of Current charts : currently chart in C h a r t +++++++++++++++++++++++ 0 in c h a r t −t r e e : c h a r t −t r e e : 3 3 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ==== Ambient Space : _[ 1 ] = 0 ==== I d e a l of Variety : _[ 1 ] = y (1)^2+ x ( 2 ) ==== E x c e p t i o n a l [1]: _[ 1 ] = y ( 1 ) [2]: _[ 1 ] = x ( 2 ) Divisors : .

==== I m a g e s of variables of original ring : _[ 1 ] = x ( 2 ) ^ 2 ∗ y ( 1 ) _[ 2 ] = x ( 2 ) ∗ y ( 1 ) −−−−−−−−−−−−−−−−−−−−−−−−−− Upcoming Center −−−−−−−−−−−−−−−−−−−−− _[ 1 ] = y ( 1 ) _[ 2 ] = x ( 2 ) −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−− ++++++++++++++ O v e r v i e w Current number of final Total number of Index of current the of charts Current charts : 0 currently chart in C h a r t +++++++++++++++++++++++ c h a r t −t r e e : in c h a r t −t r e e : 5 4 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ==== Ambient Space : _[ 1 ] = 0 ==== I d e a l of Variety : _[ 1 ] = x ( 2 ) + y ( 1 ) ==== E x c e p t i o n a l Divisors : [1]: _[ 1 ] = 1 [2]: _[ 1 ] = y ( 1 ) [3]: _[ 1 ] = x ( 2 ) ==== I m a g e s of variables of original ring : _[ 1 ] = x ( 2 ) ^ 3 ∗ y ( 1 ) ^ 2 _[ 2 ] = x ( 2 ) ^ 2 ∗ y ( 1 ) −−−−−−−−−−−−−−−−−−−−−−−−−− Upcoming Center −−−−−−−−−−−−−−−−−−−−− _[ 1 ] = y ( 1 ) _[ 2 ] = x ( 2 ) −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−− ++++++++++++++ O v e r v i e w Current number of final Total number of Index of current the charts of Current charts : currently chart in C h a r t +++++++++++++++++++++++ 0 in c h a r t −t r e e : c h a r t −t r e e : 7 5 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ==== Ambient Space : _[ 1 ] = 0 ==== I d e a l of Variety : _[ 1 ] = x ( 1 ) ∗ y (0)^ 2+1 ==== E x c e p t i o n a l [1]: Divisors : .

_[ 1 ] = y ( 0 ) [2]: _[ 1 ] = 1 [3]: _[ 1 ] = x ( 1 ) ==== I m a g e s of variables of original ring : _[ 1 ] = x ( 1 ) ^ 3 ∗ y ( 0 ) _[ 2 ] = x ( 1 ) ^ 2 ∗ y ( 0 ) −−−−−−−−−−−−−−−−−−−−−−−−−− Upcoming C e n t e r −−−−−−−−−−−−−−−−−−−−− _[ 1 ] = x ( 1 ) ∗ y (0)^ 2+1 −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−− ++++++++++++++ O v e r v i e w Current number of Total number of Index of current the of final charts Current charts : currently chart in C h a r t +++++++++++++++++++++++ 0 c h a r t −t r e e : in c h a r t −t r e e : 7 6 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ==== Ambient Space : _[ 1 ] = 0 ==== I d e a l of Variety : _[ 1 ] = y ( 1 ) + 1 ==== E x c e p t i o n a l Divisors : [1]: _[ 1 ] = 1 [2]: _[ 1 ] = 1 [3]: _[ 1 ] = y ( 1 ) [4]: _[ 1 ] = x ( 2 ) ==== I m a g e s of variables of original ring : _[ 1 ] = x ( 2 ) ^ 5 ∗ y ( 1 ) ^ 3 _[ 2 ] = x ( 2 ) ^ 3 ∗ y ( 1 ) ^ 2 −−−−−−−−−−−−−−−−−−−−−−−−−− Upcoming Center −−−−−−−−−−−−−−−−−−−−− _[ 1 ] = y ( 1 ) + 1 −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−− ++++++++++++++ O v e r v i e w Current number of final Total number of Index of current the charts of Current charts : currently chart in C h a r t +++++++++++++++++++++++ 1 in c h a r t −t r e e : c h a r t −t r e e : 7 7 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ==== Ambient _[ 1 ] = 0 Space : .

4 . 0 . 1 .3 .0 .0 .3 .0 .3 .3 . 1 .2 .0 .0 .0 .4 (c) 0 .0 .0 .0 .0 . 1 .0 .4 (b) 0 .2 . 0 .3 .2 .0 . 0 .0 .0 .0 . 0 .3 . 1 .==== I d e a l of Variety : _[ 1 ] = y ( 0 ) + 1 ==== E x c e p t i o n a l Divisors : [1]: _[ 1 ] = 1 [2]: _[ 1 ] = y ( 0 ) [3]: _[ 1 ] = 1 [4]: _[ 1 ] = x ( 1 ) ==== I m a g e s of variables of original ring : _[ 1 ] = x ( 1 ) ^ 5 ∗ y ( 0 ) ^ 2 _[ 2 ] = x ( 1 ) ^ 3 ∗ y ( 0 ) −−−−−−−−−−−−−−−−−−−−−−−−−− Upcoming Center −−−−−−−−−−−−−−−−−−−−− _[ 1 ] = y ( 0 ) + 1 −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−− ============= r e s u l t the number of ============= charts will be t e s t e d ========== obtained : result is 2 o.3 .0 .3 .3 .4 .0 .0 .0 .0 .0 .0 .0 . 1 .0 .4 .0 . 1 . 0 .k. 1 .2 .0 . 0 . 1 .0 .0 .2 .0 . 1 .0 .0 .0 .0 .0 .0 .2 .2 .0 . 1 . 0 .2 .0 . 0 .0 . ========== Now let's compare the charts for the exceptionals from (a) 0 .4 .0 .0 .

we use the list L=resolve(I. The multiplicities are given as in the end of 3. we should check the multiplicities of the exceptionals (a).4 .7 which gives 3 charts for (a) yet (4) for (b). 1 . This last set of commands should show that (d) is not equivalent.2 .0 .0 .0 .0 . To see the multiplicity. 0 . and each column tells which exceptional is appearing there.3 . The conguration of innitely near singular points can be seen as in 3. > poly f=x^3+x^4+y ^ 5 .0 . 1 .7. the 1 in the second row rst column means E1 is in the second chart. you have to look at So this would seem to say that (a). > displayMultsequence ( f ) .2 . 0 . 5.4 (e) 0 .0 .0 . 1 . 1 .0 .0 . So for instance in (e) above. 0 . > poly f=x^4+y^5+y ^ 6 . 1 .(d) 0 . (e) are not equivalent to any others.2 . 0 .0 . To see they are not equivalent.0 .2 . > displayMultsequence ( f ) .3 Each row gives one chart. (b). 1) command as in 3. (d) are equivalent and (c).0 .7.3 . Note to compute the multiplicity sequence. 1 .0 .3.3. > displayMultsequence ( f ) . To do this we can use the additional singular commands: > poly f=x^3+y ^ 5 .3 .2 .12 V.0 . 284 . (b).0 .0 . (d) to make sure they are actually equivalent resolutions. However.0 .0 .b x Clearly these two singularities have the same multiplicity by looking at the lowest term (it's 4).8a.3 .0 .

4 Cubic Surface 5. x0 x1 X0 to P3 . 0) ψ of x20 . 0.4. x0 x2 .2. 1. 1) and P2 = (0. 1) and (0. d0 has no base points by thm V. 0. 0 The vector space V ⊂ H (OP2 (2)) corresponding to d is spanned by of all conics passing through (0.a.1 x g P2 blown at 2 points Using the notation of the chapter.4. since r ≤ 4.4. 0).7.1 and by thm II.1 V.1.4.4 V.8. dim d = 3 by thm V. it determines a morphism Let the two blown up points be P1 = (0. x1 x2 .5. You check this by looking . 1. which is the space .

at the dening equations of conics through some points .

 can  2 2 2 .

.

x xy y xz yz z .

.

p2 pq q 2 qr .  ... .

.

.. .

= 0 through points (p, q, r) in the variables x, y, z .

..

..

..
0
3
Now we dene ψ : X → P by where it sends the basis elements.
2
Namely, x0 7→ y0 , and x0 x1 7→ y1 , and x0 x2 7→ y2 , and x1 x2 7→ y3 .
0
Note that for any point in X , the image satises y0 y3 = y1 y2 which is the equation dening the quadric
surface Q (xw = yz ).
Thus the image of ψ is contained in Y .
2
Now let π : Q → P be projection from the point p = (0, 0, 0, 1) ∈ Q to the plane y3 = 0 , i.e. πp :
(x, y, z, t) 7→ (x, y, z) .
Note that we have π ◦ ψ = IdP2 and ψ ◦ π = idQ .
2
Let Γ ⊂ Q × P be the graph of π .
2
Let p = (0, 0, 0, 1) ∈ Q and q = (0, 0, 1), r = (0, 1, 0) ∈ P .

.

 .

 .

 .

 .

 .

 .

 .

.

Fn ].. at • For P2  If X ⊂ Pn with two points blown up: X ⊂ Pm Y ideal with saturation by Y ⊂ X is a subvariety..  Let • For the blowing up of Q at p ˜ = Γϕ ⊂ X × Pn−1 the graph a quasi-projective variety and p ∈ X any point. Let X n−1 ˜ → X is then called the blow-up of X of the projection map of X to P from p. ... while over p the ber is isomorphic to P . We have the following denitions for the blow up of a point and subvariety from Harris: • Blowing up Pn at a point. ϕ : Pn → Pn−1 the rational map given by projection from a point p ∈ Pn and Γϕ the graph. Fn of homogeneous polynomials of the same degree generating an I (Y ) and letting BlY (X) the graph of the rational map ϕ : X → Pn given is a projective variety and by taking a collection [F0 . . n n n The map π : Γϕ → P is the blow up of P at p. .. The map π projects Γϕ isomorphically to P away n−1 from p.. we dene the blow-up of X along F0 . The map π : X  Let p.

x0 x1 . and since this is homogeneous.4. P2 . This is again just Γ. from the three points P1 . xy) = fr1 (yz. z) = fr1 (x. 1. y). r}). y.2 x g C is dened by f (x. xz. Then 5.xz. P3 = (1. P2 = (0. where fi (x. 1) .27 5. xz) (xy) + · · · + fd−r1 (yz. Also a reducible plane cubic is either 3 lines or 1 line and a conic. 0). g (x. Then f (x. 1). xy) dividing out factors of x. Theorem 3. So the terms of lowest degree in x. y. note that 9 points determine a cubic. y. xz. z and doing this twice gets us back to f so Suppose since this agrees with the quadratic transform given in chapter 1. y. y) are homogeneous of degree i. r . z) = i=0 fri +i (y. xy) . 0. 0) . x0 x2 . Thus the graph of the rational map ψ : P → P 2 2 given by [x0 . 0. y. z) → (yz. z factors out.4. we see this is the strict transform. similarly for the other points. xz. Looking at (0.2 is by denition the blow up of Q at V. P3 is then given by g (x. x0 x2 . x1 x2 i generate an ideal with 2 3 saturation the homogeneous ideal of two points I ({q. Note that hx20 . Now g is just f (yz. x0 x1 . x1 x2 ] above is the blowing up of P at the two points q. The strict transform under the quadratic transformation of chapter 1 (x. 0. y. which clearly has degree 2d − r1 − r2 − r3 . x) x y is the lowest degree term there of degree d − r2 − r3 . z) := f (yz.4.4. WLOG assume P1 = (0. z) = 0 which is irreducible thus no x. Now recall that multiplicity is the lowest degree term so looking at Pd−ri d−r1 −r3 d−r1 −r2 d−r2 −r3 y z . y) z + · · · + fd−r1 (x. 5. y have degree r1 xr1 y r2 z r3 d−r1 in x. d−r1 The transform satises f (yz.3 V.4 x g important? Well. xz) since z r1 gets cancelled in g .4 V. x) x −r2 −r3 then fd (y. Suppose P1 has multiplicity r1 on C .Γ = X0 p. y. QQ0 and Let the blue lines be the cubic given by the three lines from Wikipedia: 286 RR0 in the following picture I stole .xy) .4.4.3 x See Algebraic Curves over Finite Fields by Hirschfel. y . P P 0 .

x See Tate. and in the and meets on the outside along the points. L (P → Q → R) and L0 (P 0 → Q0 → R0 ) and the third Now use this Cayley-Bacharach theorem: If two cubics C1 and C2 meet in nine points. 5.6 V. Q. then every cubic that passes through eight of the nine also passes through the ninth.5 x g Pascal's Theorem Consider the cubics Then X∩Y X = AB 0 + BC 0 + CA0 Y = AC 0 + BA0 + CB 0 . R.4.4. 287 inside at P. It's in the rst chapter. Silverman: Rational Points.5 b. conic in 6 points. They even give nice pictures 5. Thus in 9 . This pretty much does it.4.Now consider the cubic determined by the two lines 00 00 00 line L (P → Q ).

This meets the cubics at 8 of the 9 points. Claim: for of degree d d>0 1 3/2 with √ d −d+ 3 on the cubic surface. Thus 2 2  P a = 21 dP+ 23 r − αi . i = 2. Otherwise.7 V. there exists a smooth connected curve +1 Following Hartshorne. bi ≥ bi+1 ≥ 0 nonsingular curve of degree .4. . and and genus g g Given this  2 expand 61 · (d − 1) · (d − 2) = d6 − d2 +  2 expand 16 · (d − 1) · (d − 2) + 23 = d6 −  2 expand 16 · d · (d − 3) + 1 = d6 − d2 + 1 so we have achieved the maximum 1 3 d 2 1 < g ≤ 16 d (d − 3) + 1 .5... s there is a quartic through points 1 through 13 and P.4. But 14 points determine a quartic so this is a contradiction.. Thus if we xed some additional two arbitrary points some t. Q. and αi = r−bi for i = 2.. b1 . master of arithmetic and algebraic geometry.2. For another quartic C3 through those points but not in the pencil then C1 + tC2 + sC3 = 0 is a two P.4...6 x g 4 In exc V. 2 2 288 . 6 . on the cubic. bi = 21 r − αi . The inequalities at the end of the rst paragraph become |α P2 | ≤2 α3 ≤ · · · ≤ α6 ≤ 2 r 3 1 3 2 and −α2 + α3 + · · · + α6 ≤ d − r and the genus becomes g = (r − 1) d − 4 r − αi + 1 .2.Now consider E = conic + P Q .8 V. b1 = a − r. where this hypothesis means that the linear system of curves through the points has the smallest possible for any choice of that number of points.13. Q.. which is a cubic.4. 1 1 Dene for (a.. . Thus C1 + tC2 = 0 13 points determine a 1-dimensional pencil of quartics. bi ∈ Z we need αi ≡ 12 r (mod 1) and 1 d+ 23 r− αi ≡ 0 (mod 2) .4. I prove that there is exactly one curve of degree through 4 · (4 + 3) /2 = 28/2 points in general position.. . then for dimensional family meeting the points. 5.. P13 . C1 ∩ C2 has 3 additional points through which clearly all other quartics in the pencil must pass.a.. Thus the pencil C1 + tC2 = 0 determines all quartics through the P1 .7 x If it's not on a quadric. b6 ) ∈ P ic X r = a−b1 . . as in 4.. this is Harris / Eisenbud. 6 so that r ∈ Z and αi ∈ Z . 5. so by Cayley Bacharach.15. the irreducible P P d = 3a− bi and genus g = 12 (a2 − b2i − d)+1 for a ≥ b1 +b2 +b3 .4. Curves in Projective Spaces Theorem 3. By Bezout. Note for a. proceed as in example IV. Consider.

. n = i=1 xi . then since n < 4d − 6r + 3 and n ≡ 5 (mod 8) . αi to satisfy the |α2 | ≤ α3 ≤ · · · ≤ α6 ≤ 21 r . ≤ α6 ≤ 12 rP. Then m − 1 is a sum of 3 squares. so n is the sum of 4 squares of integers ≤ k. For xi odd. 2 1 If r is even. If m < (k + 1) . n = αi . 2 2 If on the other hand n > 2k . F d = 16 d (d − 3) + 1.. . 1. then we have shown the claim. Assuming we have n < 4 have ? . Else. write n = 2k + m. since P6 1 2 Next assume r is odd. we are 2 a 0 0 done as above.. k > 0Pan odd integer. and in the rst case there is no n ≡ 5 with n ≤ 4d − 6r − 1. 6 (mod 8) so m is the sum of 3 squares. 7 (mod 8). the same argument applies to n − k so either n is the sum of 4 or of 5 squares of integers xi with |xi | ≤ k . If n = x1 + x2 + x3 . Else. m ≡ 0. then −α2 + α3 + . + α6 ≤ d − 32 r is clear. By hypothesis. then = d = 3 so F 3 2   2 1 3 2 expand subst 3 · (d − 2) . 1). r. + α6 ≤ d − 23 r. P P (r − 1) d − 34 r2 − αi2 ≡ 0 (mod 2). n < 3r − 4r + 3 P. We therefore need to check that −α2 + α3 + . Since −x2 + x3 + · · · + x6 ≤ 2d − 3r unless 2d − 3r = 1 or 3. Then m is the sum of 3 squares of integers ≤ k which are odd. and in the second case.. we P 2 3 2 r − r + 56 . 2 2 2 2 Now suppose that n < (k + 1) . After adjusting the order and signs of the αi2 = n ≤ d − 23 r + 43 . m ≡ 7 (mod 8). g = Fd (r) − 21 αi2 . assumption that d + r − 2 Setting −x2 + x3 + · · · + x6 ≤ 2d − 3r . then Fd (r) ∈ Z. 2 0 0 0 2 Then m ≡ 1... 5. Then n = 2 (Fd (r) − g) is an integer. then clearly |xi | ≤ k for all i so we are done. a write n = 4 m. the maximum is attained 3 3 by taking all αi = 0 . n < d − 32 r + 34 . the order and signs of the αi are determined. then n < 3k − 2k + 3 we have If the result. i = 2. Now dene ? The Sum of Five Squares P k ∈ Z+ then any positive integer n < 3k 2 − 2k + 3 can be written as the sum of ve squares n = 5i=1 x2i of integers xi with |xi | ≤ k . P P P 2 Since n = xi . αi = 2 xi . 3 P i = 2. the sum of odd integers with |xi | ≤ r. some r ∈ Z and αi ∈ 21 Z satisfying the required congruences and inequalities above for all g = Fd (r) − 12 αi2 . We want to nd for arbitraryPd.. Thus n ≡ 5 (mod 8) hypothesis. |xi | ≤ k .. of the poly expand · (r − 1) · d − 34 · r2 + 1 = − 3 8r + 2a 2 dr − d2 + 1 as a polynomial in r. 2 2 2 2 If k ≤ n < k + (k + 1) . 2 · (r −1) · d − 4 · r2 + 1 = d6 − d2 + 31 2 2 So Fd (r) has a max at r = d. so d + 34 r − P αi ≡ 0 (mod 2). Then we need ?? integers xi with |xi | ≤ r such that n = xi . 2. 2 2 2 2 Proof: Any such n can be written as 1 + 1 + m. this is true unless (x2 .   2 −b 1 To compute the maximum value of Fd (r). we want to write n = i=2 xi with 8 10 2 xi = 2α . thus Fd (r) ∈ Z. we have n ≤ 4d − 6r − 1. On the other hand. x6 ) = (−1. Since g is an integer. The sign of α2 is given by the 3 αi ≡ 0 (mod 2). 1 + k + m or k + k + m.  1 Proof: Fd (r) − Fd (r − 1) = d −P32 r + 43 so that Fd (r) − g < 12 d − 23 r + 34 . m = 4 (8b − 1). we must verify the congruences and inequalities from the rst paragraph. b ∈ Z.. By 1 2 as above.. we need −x2 + x3 + · · · + x6 ≤ 12 x2i + 21 which is (x2 + 1)2 + 6i=3 (xi − 1)2 − 4 ≥ 0 . If m is the sum of 3 squares and m < (k + 1) . As we hve assumed |α2 | ≤ α3 ≤ . We need ? integers αi . 6 satisfying the requirements so that g = Fd (r) − 12 αi2 . Proof: Recall the sum of three squares theorem from Gauss: A positive integer n is the sum of 3 squares a i it is not of the form 4 (8b − 1) .. and we need αi with |αi | ≤ 21 r 2 √ P 2 such that n = αi . .a. P Fd (r) = 12 (r − 1) d − 43 r2 + 1 . 4. √ 2 Claim: If d > 0 is an integer.. where 0 < m < (k + 1) and m ≡ (mod 8). we use the sum of ve squares theorem. Write n = 2 (k − 1) + m . 1. the congruence −α2 + α3 + . 6 with |αi | ≤ 12 r. ?? Claim: Set 289 . For n = 8 (Fd (r) − g) .. + α6 ≤ d − 32 r. Thus for d ≡ 0 (mod 3). so m = m + 4k − 2 .. 2    2  2(d−1) 2(d−1) If d ≡ 1 mod 3.. Then all positive integers n ≡ 5 mod 8 with n < 3k 2 + 2k + 1 are 5 2 sums of 5 squares. Thus Fd (r) − g < 2  1 d − 32 r + 34 . For other g . 1. since √23 d − 13 ≤ r. √ d ≤ r ≤ 23 d and Fd (r − 1) < g ≤ Fd (r) then we can nd αi ∈ 21 Z . . We need Fd (r) − g = 12 αi2 .

5 P b2i − 1) + 1 > 0. case 3 if P bi < 0.8* 5. • Note • So g = 12 (a2 − P b2i − d) + 1 g > 0 =⇒ g − 1 ≥ 0 so since it's a cubic surface.4. =⇒ a2 ≥ 8 + P b2i ≥ P bi case 1 ≥P 1.• • Existence V. a < 0.9 V.4.. (a2 − P b2i ) ≥ d The bound: LHS I will work downwards from the genus: P P 1 We have g = (a − 1)P (a − 2) −P12 b2i + 21 bi 2 1 2 = 12 (a2 − 3a + 2) − 21 bP bi lets get rid of 12 since on both sides i + 2 P 1 1 a2 − 3a + 2 − b2i + bi get rid of 2 on each side 2 2 P P 1 1 2 2 a − 3a −P2 bi +P2 bi move a2 to other side 1 1 −3a − P b2i + P bi move P 9a here 2 2 1 1 1 2 6a − 2 P bi + 2 bi P move bi down 2 P 2 1 6a − 12 Pb2i bring 6a bi here and move bi up 2 P 6a + 6a bi hmm move 2.9 x genus bound for cubic surface. since we still have 6a bi ≤ 6a2 P P 8a2 + ( bi )2 + 21 b2i If P i bi 290 . then we are done P case 4 if bi < 0..4. then we are done trivially.5 bi here 6a + 6a P P bi 1 2 P g = 2 (a2 − 6a bi ≤ 6a . since Thus bi + 2. a = 0. then a2 ≥ 9 =⇒ a ≥ 3 =⇒ 2a2 ≥ 6a + 1 so we are done. Now. case 2 if P bi = 0. =⇒ . then we are still done.

d = 3a − bi 2 of 2's On the RHS I will work upwards from what we want. 5 with |bi | ≤ a such that bi = a2 − 2d + 8 − b26 . (since d ≥ 8. P 2 Thus we can nd bi .4.4.4.4.5 bi down P P P P 8a2 + (2.5 bi + ( bi )2 + 21 b2i let's move 2. Then all positive integers n < 3k 2 − 2k + 3 is the sum of ve squares 2 = i=1 xi of integers xi with |xi | ≤ k . 2 Choose a ≤ d such that a > d + 2.11 V. see Nagata Rational Surfaces I.11 x Weyl Groups I don't care about group theory.10 x No algebraic geometry here. 291 . . 2 2 2 Since a > d + 4.7) 1 1 2 we want to get min value (d − 6) = (a − b2i − d) + 1 the genus..10 V.. k > 0 an integer. If your interested.5 − 6a) bi + ( bi )2 let's move 6a bi up and bring 12 b2i here P P P P 8a2 − 6a bi + ( bi )2 + 3 bi move 21 bi here P P P 8a2 − 6a P bi + (P bi )2 − 9a + 3 P bi move 9a up 9a2 − 6a P bi + (P bi )2 − 9a + 3 P bi move a2 here 9a2 − 6a bi + ( bi )2 − 9a + 3 bi + 2 schwarz won't help.. and proceed by cases 5.P P P P 8a2 + 2. i = 1. 2 2 2 2 Then 3a − 2a + 3 =≥ a − 2d + 3 + 5 − b6 = a − 2d + 3 − 4. we can get rid 1 (d2 − 3d + 2) let's evict 12P for now 2 1 (d − 1) (d − 2) . this is easy) 2 Let b6 = 3. (see exP 4. Existence: Set n P5 • 5. then 2a > 2d + 2 and so a − 2d + 8 − b6 > 0. 2P 2 5 2 2 2 This is a − 2d + 8 − b6 = i=1 bi .

1. OX ) = H 0 (X.4.4. OX (−D)) is k.4. ωX ) = pg . 292 .→ k thus the kernel of k → H 1 (X.8. this eective divisor should meet at least one line with negative intersection which is silly. ei . OX (−3)) = H 0 (X. OD ) → · · · → 0.ei = 0.e6=i = 0.12 x g kodaira vanishing for cubic surface X ≈ P2 which has known cohomology. H (X.5.11.. OX ) = k .20. 0 By chapter 1. OX (−D)) → 0 H 2 (X. e1 . Suppose there is an eective divisor linearly equivalent to −D.12 b. 0. H 1 (X.4. since −3H is the 2 n canonical for P . e6 given by l2 = 1. H (D. OD ) = k → · · · → H 1 (D. Recall 0 We have a LES 0 → H 0 (OX (−D)) → H 1 (X. OX (−D)) → 0 0 So if H (OX (−D)) = k → H 0 (D. So this is 0.. OD ) = k . So we are done. 5.. . and so by exactness. then we are done as k .4. Using the methods of the chapter we nd: • P ic X ≈ Z6 • generated by intersection pairing on X l. Then by V.13 16 Lines x g So rst we really need to get the intersection theory on the new surface. e2i = −1. x see above. H 2 (X.13 V. By ex II. l. V. OX ) = 0. the dimension of this space is the geometric genus is 0 for P .

(the number of lines through 2 points is a2 = 1. C = a − bi = −1.7) cf 4. P a = 2. OX (−2KX ) ≈ OX (2) ? by tensoring the same thing on both sides. 2 allroots (4 · a − 6 · P a − 4. This Now if a = 1. and since we are just doing monoidal transforms (3. and b i = a2 + 1 We show that only a. (−2KX )+1 = 21 ·2 (2 + 1)·4+1 = 13 since 2 KX =4 by thm V. is irreducible curve on is not one of the Ei .4. then a > 0..8. IX (2)) → H (P . x g X is del-pezzo.. If a = 2. and X is embedded in P4 via the linear system of cubics through the blow up points. So 10 of these.5. P P 2 2 2 Also. as in thm V. and rest are 0.0] so a = 1 or a = 2. OX (−2KX )) = 12 (−2KX − KX ) .7. The rst and second paragraphs of the proof of 27 lines hold exactly by looking at our intersection theory discovered above. and C = −1. We need to solve the quadratic. Thus we claim: • the del pezzo contains 16 lines. only irreducibles negative self-intersection (5 of these) strict transforms P2 containing 5 of the Pi (one of these). 5.4. bi ≥ 0. • The strict transform of conic in Ei −1. P D2 = a2 − b2i . OP4 (−2)) = 0 ).3l −  hyperplane is  canonical class is D ∼ al − bi ei . 293 . number of lines through 4 points is a4 = a3 + 3 = 6. Now suppose If C C 2 P X .1.14 Since is b. number of lines through 3 points is a3 = a2 + 2 = 3. . Note h (P . i P P 2 =⇒ bi = 3a − 1. then bi =P 3 − 1 = 2 and b2i = 1 + P1 2= 2 so 2 one of the Fij . we have OX (1) = −ωX . then bi = 3 · 2 − 1 = 5 and bi = 2 + 1 = 5 so all the bi are 1 and this is Gj . two of the bi are 1 . b1 . OX (−2KX ))−dim H 1 (X. Substituting. 1  genus is 2 ei −h = −3l − P  If  P P ei degree as a curve in (D2 − d) + 1 P3 is d = 3a − P bi by adjunction.7. then above. and number of lines through 5 points is a5 = a4 + 4 = 10. self-intersection • exceptional curves • Fij of lines through Pi and Pj . b5 satisfying all these conditions are those corresponding to Fij and Gj P 2 P Schwarz gives ( bi ) ≤ 5 ( b2i ). deg C = 3a − b = 1 . a) = [a = −0. 1 4 (This ends in 0 since H (P . 0 4 Now we count dimensions.. with deg C = 1. C ∼ al − bi . 0 4 0 4 0 Consider 0 → H (P . then C is one of the 16 listed. OX (2)) → 0 . we get (3 · a − 1)2 = 9a2 − 6a + 1 ≤ 5a2 + 5 2 or 4a − 6a − 4 ≤ 0.4. OP4 (2)) → H (X. OP4 (2)) = 15 Then Note that by Riemann-roch on a surface. dim H 0 (X.

4. But then C is determined For m=1 we can nd a by the choice of the remaining points by the induction hypothesis. I use the method of exc V. then as m (m + 3) = (m − 1) (m − 1 + 3) /2 + m + 1 .16 V. and d = 8.9 (Read that exercise solution rst or this will be nonsense!) 1 We want to get (a2 − b2i − d) + 1 = g for d = 8. 11 For these remaining rest. 9 and d = 8. 294 .4. 9 d = 9.4. g = 8. Now use 5. Now by Bezout.7 we get • just need d = 10. and d = 9. g = 6 2 P 2 2 so a − P5 2 bi =2 2g − 2 + d bi + b6 = 2 − 2g + d + a2 use this line so Pi=1 2 2 2 so P bi2 = a2 − 12 + 22 + 8 − b6 so bi = a − 2 − b6 . g = 10.4.7 • From method of V. with C being of order m − 1 . V.4. ?. g = 8. g = 7.15 x admissible transformation th I will show that there is exactly one curve of m order through m (m + 3) /2 points.4. 1 · 4/2 = 2 points which only contain one curve clearly. L has intersects any curve of m order passing through the rst m + 1 points at least m + 1 times but m · 1 for the degrees don't equal so L 0 0 0 must be a component of such a curve.4. th Suppose C is any curve passing through all the points.15 H 1 (X. P 5.14 x From 4. 1 If the assertion is true for m − 1 .9 and the rest follow similarly. C = L ∪ C . g = 6 and d = 10. we choose m + 1 2 distinct points on a line L and the rest not on the line and in general position. OX (−2KX )) = 0 . Solve for this using sum of 5 squares as in exc V. 1 This is equivalent to there being m (m + 3) points with only one curve passing through those points for 2 any m.Note that by Ramanujam vanishing and Serre duality. g = 7.

There are 21 conics through 5 Now note that cubics through points.13. General position is equivalent to some determinants not vanishing. and that gives as in exc V... 5.17 b.4.18 c.. . we can have exceptional curves (8) Lines through points (8 choose 2 = 28) conics through 5 (8 choose 5 = 56) cubics through 7 with one double point (8 choose 7 = 8) quartics through 8 with three double points (8 choose 3 = 56) quintics though 8 with 6 double points (8 choose 6 = 28) sextics through 8 with 7 double points and one triple point (the rest) 295 Thus if we take .4..5.4.19 d.a.. Pr ) as a point in P2 × . × P2 . 7 points have 1 degree of freedom (by Bezout for intsance). So unless the vanishing set is all of P2 × . Thus the tuples (P1 . Rational Surfaces II. and use (a) of this problem. x at 7 7 exceptional curves.. we nd there are There are 21 lines through the points. and proceeding as in exc V.4. corollary to proposition 9. 7 choose 5. this gives Thus we have 14 + 42 = 56 at 8 Using the same logic.. since no 3 are collinear. x g Consider (P1 . 5.a.. Pr ) in general position form the complement of an the vanishing of these determi- nants.4. one of the points doubled (so make a base point). then it must be proper so we get a zariski open set which is open and dense for the general position points.13..9. By the logic of thm V. × P2 . x This is Nagata.. .4. 21 additional −1 curves. 7 additional −1 curves.

j=k (obvious by description)  so as an example calculation> If we use the parametrization above.20 V. and t is proportional w + y. z) are proportional to − (s + r) t2 + (s2 + 2r2 ) t − s3 + rs2 − 2r2 s − r3 + t3 − (s + r) t2 + (s2 + 2r2 ) t + rs2 − 2r2 s + r3 − t3 + (s + r) t2 − (s2 + 2r2 ) t + 2rs2 − r2 s + 2r3 + (s − 2r) t2 + (r2 − s2 ) t + s3 − rs2 + 2r2 s − 2r3 . not quite obvious. j.10. x unless w + y .16 x Fermat Cubic We can use Elkies parametrization: If (w. y + az = w + ax = 0 7→ (1 : −a0 : −a) . t are proportional to x + y . x.. s. By remark 4. x. with self intersection 15 6 P2 strict transforms of lines containing strict transforms of conics containing −1. w + a0 z = x + a0 y = 0 7→ (−a0 : 1 : 1) .y = y . x + a0 z = y + a0 w = 0 7→ (0 : 1 : −a0 ) . then r. in which case r.4. y.4. x + z are both zero. y + a0 z = w + a0 x = 0 7→ (1 : −a : −a0 ). z) is a rational solution of w 3 + x3 + y 3 + z 3 = 0 . s.1 we have: 3 So it's in P . But then a = a which is false. y. y. w = w 0 0 0 0 0 on the two curves w + az = x + zy = 0 and w + a z = x + z y = 0 where a is a cubic root of 0 unity. t such that (w. A proof is in Hirschfeld Finite projective spaces in three dimensions 20. and the there exist to blow down / points we blow up are x + az = x + ay = 0 7→ (−a : 1 : 1) . k. z = z .5. x + az = y + aw = 0 7→ (0 : 1 : −a). The other's can be similarly veried. 296 . The are incidence relations • so Ei doesn't meet Ej (obvious by description)  Ei meets Fjk  Ei meets Gj  Fij meets Fkl i all  Fij meets Gk i i i i=j i 6= j or i=k (obvious by description) (obvious by description) i. Pi 5 and Pj of the Pi . exceptional lines. y. x .3. it's blow up of • Ei there are • Fjk • Gj 6 there are there are at 6 points. s is propotional to wy − wx + xz + w − wz + z . Automorphism group? E6 .1. and we want to show Ei doesn't 0 0 0 0 then they would meet if there is a point on the line where x = x . l i=k or are distinct. meet Ej . 2 2 where r is proportional to yz − wx.

L (mH + iY )) = 0 for i ≤ k. L (mH + (i − 1) Y )) → H (X. Assume for i − 1. Next observe (mH + kY ) . we know that H (X. 1 Now you have a bunch of copies of P so you can just dene by projection. Now how will further blowing-up make the resulting curves disjoint? Ok.7 Let −m = Y 2 < 0 H on X such that H 1 (X. Lifting this section to H (X. Now the blowing-up surface separates tangent directions at the node . 1. On the other hand. L (H + (i − 1) Y )) = 0. . 1 1 1 Also H (P . L (H)) = 0 by III. OP1 (mk − im)) → 0 1 By the induction hypothesis.Y and assume k ≥ 2.5.. Consider 0 → L (mH + (i − 1) T ) → L (mH + iY ) → OY ⊗ L (mH + iY ) → 0. L (mH + (k − 1) Y )) = 0 by step 1.. further blowing up will make the curves disjoint because you have simle normal crossings to start. 5.5 and so we conclude H (X..5. choose a very ample divisor We get an exact cohomology sequence 1 1 . because M ⊗ IY ≈ L (mH + kY ) ⊗ L − Y ≈ L (mH + (k − 1) Y ) and H 1 (X. → H (X. OP1 (mk − im)) = 0 by III. M ⊗ OY ) is surjective. L (mH + iY )) = 0 for i = 0.1 x g Resolving singularities of Let f Y = div (f ) Take an embedded resolution using V. so OY ⊗ L (mH + iY ) ≈ OP1 (mk − im).5. 1 First we prove that H (X.2 V.5..5. Y ≈ P1 and (mH + iY ) .1 V.Y = mk − mk = 0 and so M ⊗ OY ≈ OP1 which is generated by the global 0 section 1. M) → H 0 (Y. 8 we can change k to the bigger 297 .f.5 V. 8 (H is ample so intersection > 0 and then take a multiple of that)  We will use the invertible sheaf L (mH + kY ) to dene a morphism of X to something. the natural map Step 3. Since away M = L (mH + kY ) is globally generated. k . For i = 0. |mH + kY | has no basepoints away from Y very ample. and using Nakayama lemma.9.5. and the LES of cohomology. Let k = H. M).Y = mk − im.5 Birational Transformations 5.3.5.2. so M is generated by global sections H 0 (X. step 2 Next we show that H is from Y . so f −1 (Y ) is normal crossings. L (mH + iY )) → → H 1 (P1 . it's true by assumption. we see that M is generated by global sections also at every point of Y ..2 x g Castelnuovo Lookalike c.

1). π∗ π ∗ ΩX ) ≈ H i X.7). Let (II. 2 Thus by the hodge index theorem. ΩE ) → 0 → H 0 (X.8. Thus ΩX/X ≈ ΩE .3 Let Y.4 Let is an isomorphism. ˜ Now consider   ˜ Ω ˜ → H 0 (E.5.5. OP1 (−2)) ≈ H 0 (P1 . As h is a birational invariant since genus is. X Note that H 0 (E. Ωq )∨ by serre duality n−p. (II. V. and hence we are done for a rational surface.5. V.p since H q (X. 298 . H 2 (X. ΩX ) → H 0 X. then f1 must map |mK + kY | separates points and not on Y .Y = 0 and (f H) = 0. p. OP1 ) = k ∨ = k by serre duality.3. H M ⊗ OY is 0 (since its Y from points ). X   ˜ Ω ˜ → 0 (the last is 0 since ΩE has support on a one dimensional space). for at other points the rst Now (cf IV. two degree in two variables and 2 If X is P at least. then we are done by exc III. ΩX ) → H 2 X. ΩE ) ≈ H 1 (P1 . Ωp ) = H p (X. ΩE ) ≈ H 0 (P1 . so probably since we are on smooth varieties. the linear system and also separates points of OP1 Y to a tangent vectors onto X 1 − P1 . ΩX ) → H X. Now 0 → π ΩX → ΩX ˜ ˜ → ΩX/X has support equal to the set of ramied points. so f1 is an iso of X − Y is very ample. Y ≤ 0. 2 ∗ ∗ So f H.Therefore M f 1 : X → PN determines a morphism X1 it's image.n−q p.0 and further h  = h  by poincare duality.2) ΩX/X ˜ sheaves are same dimensional so the quotient at the stalks is 0. then we see ˜ Ω˜ H 2 (X.q Otherwise. Note that pullback of ample is ample. X 5.7. ΩX ) → H 2 X.q 1.5. ΩE ) → H (X. ∗ → 0 is the relative cotangent sequence. OP1 (−2)) = 0 since there are no monomials of negative ∨ H 1 (E.4 x g hodge index theorem corollary x H be a very ample divisor on X0 . since away from 5.2.1.   ˜ π ∗ ΩX ≈ H i (X. and since degree of On other hand. H i (X.3 x g hodge numbers excercise F = π ∗ ΩX . X   1 1 ˜ Ω ˜ → H 1 (E. ΩX ) by exc III. ∗ Since f1 O (1) ≈ M point P1 .7. note that the hodge numbers satisfy h = hq.

Ox .5. The center is either the exceptional divisor or point contained in the exceptional divisor. As both R and Ox are valuation rings. Since of a hyperplane not through j. Then on irreducible components we have H2 ≡ H1 + Yi . This corresponds The image of the closed point of If x∈X is the center then R dominates to type (1) of exc II. then by basic facts of symmetric P matrices. On the other hand. 299 . If the center of ν in X1 x has codimension 2. it is thus proper. then since X is projective.12.5. If ν is nontrivial.5. Then hm E . mj Ej i ≤ 0.6 x R is a valuation ring with valuation ν : R → Γ. and 2 H. Note further that hYi . then the dimension of ν . we can't split the Yi 's into two groups which are not connected so the two groups always don't intersect each other. By Zariski's Main theorem.4. Now note that since H1 passes through Ej then i kYi Yj k < 0. Since the matrix i i j j i kYi . 5. mi > 0. If the center has codimension 1.7 If V.5. and so the valuative criterion gives that Spec K (X) → X extends to Spec R → X . then H > 0.a 5. then R = Ox .12. Spec R corresponding to the maximal ideal is the center of the valuation ν. we have kYi . then taking the monoidal transform with center x gives X1 .4.5 x g Consequence of my proof of exc V. then Ox and R must be discrete as R dominates Ox and at any rate a valuation group on a surface is either a one dimensional or two dimension Z-module (see Vaqiue Valuations and local uniformization.12. for each P we have already proved indeniteness.2.5 b.5. m E i = −hH1 . by the general assumptions in chapter 5.5. remark 1.Yj k is symmetric. In the former case we have type (2) of exc II.5.6 V.14).4. Yj i ≥ 0 for i 6= j .Yj k is negative indenite. in the later case we repeat and get a type (3) of exc II. let H1 the strict transform of a hyperplane through P and H2 the strict transform P That Yi2 ≤ 0 P .Yi = 0 so that Yi ≤ 0 by Hodge index theorem. and since we are on a surface this number must be 0 or 1. x g Hodge Index negative denite 2 is easy since if H is strict transform of a hyperplane not through P . these two facts give negative denite. which is corresponds to the transcendence degree of R/mR which corresponds to the dimension of the center.

y = (t ) t y = (t ) v . Since gi ∈ / (z). v.9 V.NN. mn .5. since A [z ] is a UFD. we assume that (maybe X0 5. where fi /z mi are irreducible z m z m1 −1 m in A [z ] .. So I wish to show A/ (z) ≈ k [x. v]. By above. t−1 ]. 300 . a unit u in A. gi ∈ / (z) ..5. u.4 applies since at beginning of surfaces chapter. −1 Now if f ∈ A is nonzero. fi = z i gi for an irreducible gi ∈ A. t−1 ] . t ]. v] then A [z ] ⊂ k [u. Thus z m f = uz ri gi . Step 2: Enter the following commands into Singular > ring > ideal R= 0 . m1 ... then. These are both UFD. So if f is irreducible not in (z). v. then f = z g for an irreducible g ∈ A . Proceed as in Algebra.8 V.pdf. t−1 ] . t−1 = (z −1 ) x.8.b. 3 2 −1 Since t ∈ k [u.5. So z f = uf1 · · · fn P Q r in A. x Surface singularity Step 1: Start your computer.5. . I=x2+y3+z 5 .5. Uniqeness follows in the standard way. we have f1 = u f1 · · · zfmnn for nonnegative m. then f is irreducible −1 in the localization. −1 Geometrically. fn in A . v. g ∈ / (z). ( x . Thus A [z −1 ] = k [u. On the other hand. s = ri − m ≥ 0 so we have factored f .8 x A surface Singularity X is nonsingular isn't) First we claim z is irreducible in A.7 x The answer from 5. Next note that x + y + z = 0 implies −x /z − t /z = z −1 3 3 −1 3 −1 2 2 −1 2 −1 −1 Now we have x = (t ) t x = (t ) u. fi irreducible. we may write this as f1 = zum f11 · · · f1n .30. A [z ] is localizing at things not in z . t−1 ∈ A [z −1 ] so A [z −1 ] ⊃ k [u. summerstudychallenge2.. and f1 . v. Since each z mi is a unit.10 V.5. z ) . y . v. y] / (x2 + y 3 ) is an integral domain. denoting by f1 the localization. dp . my algebra/geometry/analysis solutions.5. then Q Q P s m ≥ ri so f = uz gi . .. 5. (z −1 ) . The converse also holds: If there is a nonzero irreducible element in the localization A [z ] m for f ∈ A . z = t so that A ⊂ k [u. 1 2 3 2 3 5 2 6 3 6 so t = −u − v so t ∈ k [u.

z are . y = zY ) ⇐⇒ (z 2 X 2 + z 3 Y 3 + z 5 . x = yX) strict is y 2 X 2 + y 3 + y 5 Z 5 = y 2 (X 2 + y + y 3 Z 5 ) 2 2 3 5 exceptional is y . xZ = z. but then middle col is nonzero. xY = y) ⇐⇒ (x + x Y + x Z . x = zX. bX = Y . so when rank is ≥ 1. z = cX . xY = y) ∨ (1 + xY 3 + x3 Z 5 . (x + y 3 + z 5 .  2x  3y 2 z  y 3 + 3z 2 but then both y. x = zX. yZ = zY. 0. xY = y) ⇐⇒ (x2 . yZ = zY. xZ = z. X must be zero. y = zY ) ∨ (X 2 + zY 3 + z 3 .  3x2 z 5 + y 3  3xy 2 The jacobian is  3 4 5x z note x 6= 0on the surface if singular then Y. xY = yX) we compute locally. 2 3 If X = 1. yZ = z. If 2 or z is 0 for last col. z =0 S ⊂ A 3 × P2 where n = 3. 1 s ) > list iD=i n t e r s e c t i o n D i v ( L ) . Y c = bz)on a = 1. Alternatively. 2x  3y 2  5z 4 We rst nd sings of  Jacobian is nonsingular when rank is n − dim V x. dim of space. (x + y + 2 3 3 5 5 z 5 . since that is not on the surface. b. =⇒ X + cXb X + c3 X 3 = X 2 (1 + cb3 X 2 + c3 X) jacobian of strict is   2 2   3b cX 3b2 cX 2  b3 X 2 + 3c2 X  =  X (b3 X + 3c2 )  2b3 cX + c3 c (2b3 X + c2 ) from rst row. thus nonsingular. y = zY ) ⇐⇒ If 2  (z 2 . y. 0) New blowup X2 : (X 2 + zY 3 2 3 3 + z 3 . Z = 1. bX = Y a. xZ = z. if you have a few hours to spare you can proceed as follows: X . thus singular when total transform of blow-up at 0 is S : (x2 + y 3 + z 5 . but that points not Y = 1. either y on surface either. xZ = z. X cannot be. xY = y) The latter is strict  transform. strict (X + y + y Z )   2x 2 5 jacobian is  3y z + 1  5y 3 z 4 singular points need x = 0 for rst col. rst row must be 0 so x = 0. (x + y 3 + z 5 . y = zY ) so to be signular. xZ = zX. Z must both be zero. have one singularity at (0. one of c.> list L=r e s o l v e ( I . az = cX. jacobian of RHS strict is and second row also so y or z = 0. x = zX. zero by last row. so this patch is nonsingular. x = zX.

0) so d = 0 Let etc. Thus this patch is nonsingular. we get it's nonsingular. still X or c must be zero. cf = Y e) on d = 1. Y c = bz) on b = 1. ae = cd. 0. 0. X3 :(a2 + cY 2 + Y c3 . X = Y a. a = cd. af = Y so a2 (1 + .. 302 . only singularity is at On the other patch.) and we see it's nonsingular on e = 1.same for If b = 0. c. c = Y e =⇒ Y 2 (d2 + Y e + Y Y e3 ) jacobian on e = 1 is   2cd + f 2  c2 + 2d  2cf so singularity at (0. 0) on f = 1. ae = c.. then Y 6= 0. bX = Y a. jacobian is   2d  2ye3 + e  3y 2 e2 + y so singularity at (0. c = 1. by second row of jacobian. cf = Y =⇒ c2 (d2 + cf 2 + cdc) on f = 1. etc. then by second row. Y =0 Y = 0. a = Y d. Second patch: X2 : (X 2 + zY 3 + z 3 . Y c = z so Y 2 a2 + Y cY 3 + Y 3 c3 = Y 2 (a2 + cY 2 + Y c3 )  2a 2 2 jacobian of strict is  3c y + y  c3 + 2cy clearly a = 0 at a sginularity if c 6= 0. af = Y d. Y is negative 3 we must solve cY (Y + c ) = 0 to be on surface 3 so c = −Y 1 9 4 4 5 last row of jacobian: −Y − 2Y = −Y (Y + 2) so Y = (−2) 5 also by second row of jacobian 6 2 −3Y · y + y = 0 so since this has dierent roots. az = cX.

 Resolution of Singularities Cutrone. Riemann-roch we get that pa (X) = 1 so by thm V. and the canonical subtracts only 7.6. Then d = 2n.1. If n = 3. then we could pick d1 = 2.1. By adjunction.3. it will always be general type. g = n + 1 so d = 2 (g − 1) = 2g − 2. 6. d4 = .  Hilbert's Nullstellensatz Baker and Csirik.6.2.D ≥ d so the second case is impossible since then multiplying by 2 gives 2n + 2 ≤ 2 − 2g + 2d rearranging contradicts d ≤ 2n..1. We could pick n = 4 to have KX ≡ 0 (so by 6. so in this chapter it's smooth.3. Let Sources: Allcock. If n = 6 we need d1 . we could pick d1 . then by thm V. But then 12K ∼ 0 so at least κ = 0. d2 = 3 to get a K3.  Automorphisms groups of Genus 3 Curves Bernd.6. |12K| = ∅ =⇒ X rational or ruled. d3 ..  Space Curves Not Contained in Low Degree Surfaces in Positive Characteristic Bars.6. 2g − 2 = d + 1r H.6 Classication Of Surfaces 5. there is r > 0. n + 1 = pa (H) = 1 + (H 2 + HK) = 1 + 12 (2n + HK) so H.  Rank Two Vector Bundles on Elliptic Curves 303 .3. d2 . Assuming F is not ruled. As H − K is ample by Nakai Moishezon. If n = 4.6.. V. On the other hand.6 V.  Rings of Invariants and Linkage of Determinantal Ideals Commelin. D ≥ 0 with D ∼ rKX . Since X is non-ruled then 2 |12K| 6= ∅ by thm 6. Same for larger 5. X is ruled.6. so by the genus 1 formula.. Using Cliord's theorem and Riemann-Roch either h1 (OH (d)) > 0 and d ≥ 2n.1. the degree of K − H < 0 so l (K − H) = 0 and now using thm V.  Solutions. X is K3.2 x g X span Pn+1 of degree d ≤ 2n since it's not in a hyperplane. then −n − 1 = −4 so we could pick d1 = 2 or d1 = 3 to have K be negative multiple of ample so by 6. Using Serre duality and exc V. d3 = 2 to get a K3 .  An Alternative Approach to Serre Duality for Projective Varieties Ballico. X is K3.5. Let H be the hyperplane section of genus g . g − 1 ≥ n or h1 (OH (H)) = 0 and n + 1 ≤ h0 (OH (d)) = 1 − g + d.K = 0. then KX = OX i=1 di − n − 1 by adjunction since KPn = (−n − 1) H .1 x g  Pn−2 X is a surface. If n = 5 . Daly.2 n.  The structure of [N] Cutkosky. d2 . then h1 (OX (K − H)) = 0 by Kodaira Vanishing so that h1 (OX (H)) = 0 by Serre duality. but since di ≥ 2 .2.1 V. |H|H has degree d and projective dimension n.6.

 Plucker Formulas Eisenbud and Van De Ven.  Semistable Bundles over an Elliptic Curve Vakil's.  Rational Cuspidal Curves Morandi. Rings.  Cohomology of Schemes Stacks Project.  Elliptic Curves over Finite Fields Swanson and Huneke.  Primary Decomposition of Modules Giraldo and Sols.  On the Normal Bundles of Smooth Rational Space Curves Esnault-Viehweg.  Computational Aspects Dolgachaev.  Field of N-Torsion Points of an Elliptic Curve over a Finite Field Weil.1 Murfet.  Lectures on Vanishing Theorems Ghorpade and Verma.  Divisors Class Groups of Singular Surfaces Hartshorne.  Abelian Varieties Murfet.  Formulae for Elliptic Curves with Complex Multiplication Reid.  Surfaces with genus 0 and K2 = 0 Ricol.  Generalized Vector Bundles on Curves Jao.  Bertini and his Two Fundamental Theorems Lipman.  Bertini's Theorem on Generic Smoothness Rosen Zvi and Richard Borcherds.  Isogenies in Quantum in a Quantum World Javanpeykar. and Modules Tu.  Rational Points on Curves Over Finite Fields Harris and Eisenbud.  Algebraic Geometry Hartshorne and Polini.  Commutative Algebra Maruyama.  Arc Valuations on Smooth Varieties Mumford.  Curves in Projective Space .  Field of Denition of a Variety Yildiz.  Grothendieck-Riemann-Roch Theorem With an Application to Covers of Varieties Kedlaya.  Notes on Cartier and Weil Divisors Mathew.  Course Notes Spring-2010 Moe. and Stuhler.  Zariski's Main Theorem and some Applications Mckernan.  Algebraic Geometry Notes MIT Kleiman.  Course Notes of Algebraic Geometry Stacks.  Etale Morphisms of Schemes Sury.  Irregularity of Ruled Surfaces in P3  Hansen.Decker.  Genre des Courbes algebriques dans l'espace projectif  Hida.  Automorphic Forms on Shimura Homann.  Section 2.  Section 2.  Integral Closure of Ideals.  Locally Free Sheaves More.  Foundations of Algebraic Geometry Van Tuyl.8 Piercey.  Vector Bundles Rajadwe.7.  Algebraic Geometry Notes Siegel.  Curves in Projective Spaces Hartshorne.  Boundedness of semi-stable sheaves of small ranks Mathew.  An Equivalence for Modules over Projective Schemes Murfet. Jahnel.  Picard Groups of Ane Curves Potier.  Unique Factorization in Complete Local Rings Madapusi.