3 Quadrics

3.1 Quadratic forms

The projective geometry of quadrics is the geometrical version of the part of linear
algebra which deals with symmetric bilinear forms – the generalization of the dot
product a · b of vectors in R3 . We recall:

Definition 8 A symmetric bilinear form on a vector space V is a map B : V ×V → F

such that

• B(v, w) = B(w, v)

• B(λ1 v1 + λ2 v2 , w) = λ1 B(v1 , w) + λ2 B(v2 , w)

The form is said to be nondegenerate if B(v, w) = 0 for all w ∈ V implies v = 0.

P P
If we take a basis v1 , . . . , vn of V , then v = i xi vi and w = i yi vi so that
X
B(v, w) = B(vi , vj )xi yj
i,j

and so is uniquely determined by the symmetric matrix βij = B(vi , vj ). The bilinear
form is nondegenerate if and only if βij is nonsingular.

We can add symmetric bilinear forms: (B+C)(v, w) = B(v, w)+C(v, w) and multiply
by a scalar (λB)(v, w) = λB(v, w) so they form a vector space isomorphic to the space
of symmetric n × n matrices which has dimension n(n + 1)/2. If we take a different
basis X
wi = Pji vj
j

then X X X
B(wi , wj ) = B( Pki vk , P`j v` ) = Pki B(vk , v` )P`j
k ` k,`

so that the matrix βij = B(vi , vj ) changes under a change of basis to

β 0 = P T βP.

23
Most of the time we shall be working over the real or complex numbers where we can
divide by 2 and then we often speak of the quadratic form B(v, v) which determines
the bilinear form since
B(v + w, v + w) = B(v, v) + B(w, w) + 2B(v, w)
Here we have the basic result:

Theorem 10 Let B be a quadratic form on a vector space V of dimension n over a

field F . Then
P
• if F = C, there is a basis such that if v = i zi vi
m
X
B(v, v) = zi2
i=1

• if F = R, there is a basis such that

p q
X X
B(v, v) = zi2 − zj2 .
i=1 i=j

If B is nondegenerate then m = n = p + q.

Proof: The proof is elementary – just completing the square. We note that changing
the basis is equivalent to changing the coefficients xi of v by an invertible linear
transformation.

First we write down the form in one basis, so that

X
B(v, v) = βij xi xj
i,j

and ask: is there a term βii 6= 0?. If not, then we create one. If the coefficient of xi xj
is non-zero, then putting yi = (xi + xj )/2, yj = (xi − xj )/2 we have
xi xj = yi2 − yj2
and so we get a term βii0 6= 0.
If there is a term βii 6= 0, then we note that
1 X
(βi1 x1 + . . . + βin xn )2 = βii x2i + 2 βik xk xi + R
βii k6=i

24
where R involves the xk with k 6= i. So if

yi = βi1 x1 + . . . + βin xn

then
1 2
B(v, v) = y + B1
βii i
where B1 is a quadratic form in the n − 1 variables xk , k 6= i.

We now repeat the procedure to find a basis such that if v has coefficients y1 , . . . , yn ,
then m
X
B(v, v) = ci yi2 .
i=1
√
OverpC we can write zi = ci yi and get a sum of squares and over R we put
zi = |ci |yi to get the required expression. 2

Example: Consider the quadratic form in R3 :

B(v, v) = x1 x2 + x2 x3 + x3 x1 .

We put
y1 = (x1 + x2 )/2, y2 = (x1 − x2 )/2
to get
B(v, v) = y12 − y22 + x3 (2y1 ).
Now complete the square:

B(v, v) = (y1 + x3 )2 − y22 − x23

so that with z1 = y1 + x3 , z2 = y2 , y3 = x3 we have p = 1, q = 2.

3.2 Quadrics and conics

Definition 9 A quadric in a projective space P (V ) is the set of points whose rep-
resentative vectors satisfy B(v, v) = 0 where B is a symmetric bilinear form on V .
The quadric is said to be nonsingular if B is nondegenerate. The dimension of the
quadric is dim P (V ) − 1.

25
A quadric in a projective line is either empty or a pair of points. A quadric in a
projective plane, a one-dimensional quadric, is called a conic.

Note that bilinearity of B means that B(λv, λv) = λ2 B(v, v) so that the set of points
[v] ∈ P (V ) such that B(v, v) = 0 is well-defined. Also, clearly B and λB define the
same quadric. The converse is not true in general, because if F = R and B is positive
definite, then B(v, v) = 0 implies v = 0 so the quadric defined by B is the empty
set. A little later we shall work over the complex numbers in general, as it makes life
easier. But for the moment, to get some intuition, let us consider conics in P 2 (R)
which are non-empty, and consider the intersection with R2 ⊂ P 2 (R) defined by the
points [x0 , x1 , x2 ] such that x0 6= 0. Using coordinates x = x1 /x0 , y = x2 /x0 this has
the equation
β11 x2 + 2β12 xy + β22 y 2 + 2β01 x + 2β02 y + β00 = 0.

Examples:
1. Consider the hyperbola xy = 1:

In P 2 (R) it is defined by the equation

x1 x2 − x20 = 0

and the line at infinity x0 = 0 meets it where x1 x2 = 0 i.e. at the two points
[0, 1, 0], [0, 0, 1].
Now look at it a different way: as in Theorem 10 we rewrite the equation as
1 1
( (x1 + x2 ))2 − ( (x1 − x2 ))2 − x20 = 0
2 2

26
and then if x1 + x2 6= 0, we put
x1 − x2 2x0
y1 = , y2 =
x1 + x2 x1 + x2

and the conic intersects the copy of R2 ⊂ P 2 (R) (the complement of the line x1 +x2 =
0) in the circle
y12 + y22 = 1.
The original line at infinity x0 = 0 meets this in y2 = 0:

So a projective transformation allows us to view the two branches of the hyperbola

as the two semicircles on each side of the line at infinity.

2. Now look at the parabola y = x2 .

The equation in homogeneous coordinates is

x2 x0 = x21

and the line x0 = 0 meets it in one point [0, 0, 1]. In projective space it still looks like
a circle, but the single branch of the parabola is the complement of the point where
the line at infinity meets the circle tangentially:

27
Thus the three different types of conics in R2 – ellipses, hyperbolas and parabolas –
all become circles when we add in the points at infinity to make them sit in P 2 (R).

3.3 Rational parametrization of the conic

Topologically, we have just seen that the projective line P 1 (R) and a conic in P 2 (R)
are both homeomorphic to a circle. In fact a much stronger result holds over any
field.

Theorem 11 Let C be a nonsingular conic in a projective plane P (V ) over the field

F , and let A be a point on C. Let P (U ) ⊂ P (V ) be a projective line not containing
A. Then there is a bijection
α : P (U ) → C
such that, for X ∈ P (U ),the points A, X, α(X) are collinear.

Proof: Suppose the conic is defined by the nondegenerate symmetric bilinear form
B. Let a ∈ V be a representative vector for A, then B(a, a) = 0 since A lies on the
conic. Let x ∈ P (U ) be a representative vector for X ∈ P (U ). Then a and x are
linearly independent since X does not lie on the line P (U ). Extend a, x to a basis
a, x, y of V .
Now B restricted to the space spanned by a, x is not identically zero, because if it
were, the matrix of B with respect to this basis would be of the form
 
0 0 ∗
0 0 ∗
∗ ∗ ∗

which is singular. So at least one of B(x, x) and B(a, x) is non-zero.

28
Any point on the line AX is represented by a vector of the form λa + µx and this lies
on the conic C if

0 = B(λa + µx, λa + µx) = 2λµB(a, x) + µ2 B(x, x).

When µ = 0 we get the point X. The other solution is 2λB(a, x) + µB(x, x) = 0 i.e.
the point with representative vector

w = B(x, x)a − 2B(a, x)x (4)

which is non-zero since the coefficients are not both zero.

We define the map α : P (U ) → C by

α(X) = [w]

which has the collinearity property of the statement of the Theorem. If Y ∈ C is

distinct from A, then the line AY meets the line P (U ) in a unique point, so α−1 is
well-defined on this subset. By the definition of α in (4), α(X) = A if and only if
B(a, x) = 0. Since B is nonsingular f (x) = B(a, x) is a non-zero linear map from V
to F and so defines a line, which meets P (U ) in one point. Thus α has a well-defined
inverse and is therefore a bijection. 2

Remark: There is a more invariant way of seeing this map by using duality. The
line Ao in P (V 0 ) is dual to the point A. Each point Y ∈ A0 defines a line Y o in P (V )
through A which intersects the conic C in a second point α(Y ). What we do in the
more concrete approach of the theorem is to compose this natural bijection with the
projective transformation A0 → P (U ) defined by Y 7→ Y 0 ∩ P (U ).

29
Example: Consider the case of the conic

x20 + x21 − x22 = 0.

Take A = [1, 0, 1] and the line P (U ) defined by x0 = 0. Note that this conic and the
point and line are defined over any field since the coefficients are 0 or 1.
A point X ∈ P (U ) is of the form X = [0, 1, t] or [0, 0, 1] and the map α is

α([0, 1, t]) = [B((0, 1, t), (0, 1, t))(1, 0, 1) − 2B((1, 0, 1), (0, 1, t))(0, 1, t)]
= [1 − t2 , 2t, 1 + t2 ]

or α([0, 0, 1]) = [−1, 0, 1].

This has an interesting application if we use the field of rational numbers F = Q.

Suppose we want to find all right-angled triangles whose sides are of integer length.
By Pythagoras, we want to find positive integer solutions to

x2 + y 2 = z 2 .

But then [x, y, z] is a point on the conic. Conversely, if [x0 , x1 , x2 ] lies on the conic,
then multiplying by the least common multiple of the denominators of the rational
numbers x0 , x1 , x2 gives integers such that [x, y, z] is on the conic.

But what we have seen is that any point on the conic is either [−1, 0, 1] or of the form

[x, y, z] = [1 − t2 , 2t, 1 + t2 ]

for some rational number t = p/q, so we get all integer solutions by putting

x = q 2 − p2 , y = 2pq, z = q 2 + p2 .

For example, p = 1, q = 2 gives 32 + 42 = 52 and p = 2, q = 3 gives 52 + 122 = 132 .

One other consequence of Theorem 11 is that we can express a point (x, y) on the
general conic
ax2 + bxy + cy 2 + dx + ey + f = 0
in the form
p(t) q(t)
x= , y=
r(t) r(t)

30
where p, q and r are quadratic polynomials in t. Writing x, y as rational functions of t
is why the process we have described is sometimes called the rational parametrization
of the conic. It has its uses in integration. We can see, for example, that
Z
dx
√
x + ax2 + bx + c
√
can be solved by elementary functions because if y = x + ax2 + bx + c then

(y − x)2 − ax2 − bx − c = 0

and this is the equation of a conic. We can solve it by x = p(t)/r(t), y = q(t)/r(t)

and with this substitution, the integral becomes
Z 0
r (t)p(t) − p0 (t)r(t)
dt
q(t)r(t)

and expanding the rational integrand into partial fractions we get rational and loga-
rithmic terms after integration.

3.4 Polars
We used a nondegenerate symmetric bilinear form B on a vector space V to define
a quadric in P (V ) by the equation B(v, v) = 0. Such forms also define the notion of
orthogonality B(v, w) = 0 and we shall see next what geometrically this corresponds
to. First the linear algebra: given a subspace U ⊆ V we can define its orthogonal
subspace U ⊥ by
U ⊥ = {v ∈ V : B(u, v) = 0 for all u ∈ U }.
Note that unlike the Euclidean inner product, U and U ⊥ can intersect non-trivially
– indeed a point with representative vector v lies on the quadric if it is orthogonal to
itself. Note also that U ⊥ is the same if we change B to λB.

Orthogonal subspaces have a number of properties:

• U = (U ⊥ )⊥

• if U1 ⊆ U2 , then U2⊥ ⊆ U1⊥

• dim U ⊥ + dim U = dim V

31
These can be read off from the properties of the annihilator U 0 ⊆ V 0 , once we realize
that a nondegenerate bilinear form on V defines an isomorphism between V and its
dual V 0 . This is the map β(v) = fv where

fv (w) = B(v, w).

The map β : V → V 0 defined this way is obviously linear in v and has zero kernel since
β(v) = 0 implies B(v, w) = 0 for all w which means that v = 0 by nondegeneracy.
Since dim V = dim V 0 , β is therefore an isomorphism, and one easily checks that

β(U ⊥ ) = U o .

Definition 10 If X ∈ P (V ) is represented by the one-dimensional subspace U ⊂ V ,

then the polar of X is the hyperplane P (U ⊥ ) ⊂ P (V ).

At this stage, life becomes much easier if we work with the field of complex numbers
F = C. We should retain our intuition of conics, for example, as circles but realize
that these really are just pictures for guidance. It was Jean-Victor Poncelet (1788-
1867) who first systematically started to do geometry over C (he was also the one to
introduce duality) and the simplifications it affords are really worthwhile. Poncelet’s
work on projective geometry began in Russia. As an officer in Napoleon’s army, he
was left for dead after the battle of Krasnoe, but was then found and spent several
years as a prisoner of war, during which time he developed his mathematical ideas.

We consider then a complex projective plane P (V ) with a conic C ⊂ P (V ) defined

by a non-degenerate symmetric bilinear form B. A tangent to C is a line which meets
C at one point.

Proposition 12 Let C be a nonsingular conic in a complex projective plane, then

32
 • each line in the plane meets the conic in one or two points

• if P ∈ C, its polar line is the unique tangent to C passing through P

• if P 6∈ C, the polar line of P meets C in two points, and the tangents to C at

these points intersect at P .

Proof: Let U ⊂ V be a 2-dimensional subspace defining the projective line P (U )

and let u, v be a basis for U . Then the point [λu + µv] lies on the conic if

0 = B(λu + µv, λu + µv) = λ2 B(u, u) + 2λµB(u, v) + µ2 B(v, v) (5)

Over the complex numbers this can be factorized as

0 = (aλ − bµ)(a0 λ − b0 µ)

giving the two (possibly coincident) points of intersection of the line and the conic

[bu + av], [b0 u + a0 v].

Suppose the point P lies on C, and let u be a representative vector for P , so that
B(u, u) = 0. Then any line through P is P (U ) where U is spanned by u and v. Then
from (5) the points of intersection are given by

2λµB(u, v) + µ2 B(v, v) = 0.

If the only point of intersection is [u] then µ = 0 is the only solution to this equation
which means that B(u, v) = 0. Since any vector w ∈ U is a linear combination of u
and v and B(u, u) = B(u, v) = 0 this means B(u, w) = 0 and P (U ) is the polar line
of X.

From the above, if P does not lie on C, its polar must meet C in two distinct points
with representative vectors v1 , v2 . We then have

B(u, v1 ) = 0 = B(v1 , v1 ) (6)

Since B(u, u) 6= 0 and B(v1 , v1 ) = 0, u and v1 are linearly independent and span a
2-dimensional space U1 . From (6) P (U1 ) is the polar of [v1 ] ∈ C and hence is the
tangent to C at [v1 ]. Similarly for [v2 ]. 2

33
The picture to bear in mind is the following real one, but even that does not tell the
full story, since if P is inside the circle, its polar line intersects it in two complex
conjugate points, so although we can draw the point and its polar, we can’t see the
two tangents.

Quadrics are nonlinear subsets of P (V ) but they nevertheless contain many linear
subspaces. For example if Q ⊂ P (V ) is a nonsingular quadric, then P (U ) ⊂ Q if and
only if B(u, u) = 0 for all u ∈ U . This implies

2B(u1 , u2 ) = B(u1 + u2 , u1 + u2 ) − B(u1 , u1 ) − B(u2 , u2 ) = 0

and hence
U ⊂ U ⊥.
Since
dim U + dim U ⊥ = dim V
this means that
2 dim U ≤ dim U + dim U ⊥ = dim V.
In fact, over C the maximum value always occurs. If dim V = 2m, then from Theorem
10 there is a basis in which B(v, v) can be written as

x21 + x22 + . . . + x22m = (x1 + ix2 )(x1 − ix2 ) + . . . + (x2m−1 + ix2m )(x2m−1 − ix2m )

and so if U is defined by

x1 − ix2 = x3 − ix4 = . . . = x2m−1 − ix2m = 0,

then dim U = dim V /2 and U ⊆ U ⊥ so U = U ⊥ . Over R this occurs when p = q = m

and the form can be reduced to

x21 + . . . + x2m − x2m+1 . . . − x22m .

34
We can see this in more detail for the quadric surface

x21 − x22 = x23 − x24

in P 3 (R). This intersects the copy of R3 ⊂ P 3 (R) defined by x4 6= 0 in the hyper-

boloid of revolution
x2 + y 2 − z 2 = 1
where x = x2 /x4 , y = x3 /x4 , z = x1 /x4 . This is the usual “cooling tower” shape

There are two one-parameter families of lines in the quadric given by:

λ(x1 − x2 ) = µ(x3 − x4 )
µ(x1 + x2 ) = λ(x3 + x4 )

and

λ(x1 − x2 ) = µ(x3 + x4 )
µ(x1 + x2 ) = λ(x3 − x4 ).

In fact these two families of lines provide “coordinates” for the projective quadric:
the map
F : P 1 (R) × P 1 (R) → Q
defined by

F ([u0 , u1 ], [v0 , v1 ]) = [u0 v0 + u1 v1 , u1 v1 − u0 v0 , u0 v1 + u1 v0 , u1 v0 − u0 v1 ]

is a bijection.

3.5 Pencils of quadrics

The previous sections have dealt with the geometrical interpretation of a symmetric
bilinear form. Now we look at the theory behind a pair of bilinear forms and we shall
see how the geometry helps us to determine algebraic properties of these.

35
We saw in Theorem 10 that over C, any quadratic form B can be expressed in some
basis as
B(v, v) = x21 + . . . + x2n .
In particular the matrix of B is diagonal (actually the identity matrix). If we have a
pair A, B of symmetric bilinear forms we ask whether we can simultaneously diago-
nalize them both. The answer is:

Proposition 13 Let α, β be symmetric n × n matrices with complex entries, and

suppose α is non-singular. Then if the equation det(λα − β) = 0 has n distinct
solutions λ1 , . . . , λn there is an invertible matrix P such that
P T αP = I, P T βP = diag(λ1 , . . . , λn ).

Proof: From Theorem 10 we can find an invertible matrix Q such that QT αQ = I.

Write β 0 = QT βQ, so that β 0 is also symmetric. Then
det QT det(λα − β) det Q = det(QT (λα − β)Q) = det(λI − β 0 )
and so the roots of det(λα − β) = 0 are the eigenvalues of β 0 . By assumption these
are distinct, so we have a basis of eigenvectors v1 , . . . , vn with eigenvalues λ1 , . . . , λn .

If v` = (x1 , . . . , xn ) and vk = (y1 , . . . , yn ) then

X X X X
λ` xi yi = βij0 xj yi = 0
βji x j y i = λk xi yi
i i,j i,j i

and since λ` 6= λk , we have X

xi yi = 0.
i

Thus vk and v` are orthogonal if k 6= `. We also must have (vi , vi ) 6= 0 since otherwise
vi is orthogonal to each element of the basis v1 , . . . , vn , so we can write
1
wi = p vi
(vi , vi )
and obtain an orthonormal basis. With respect to this basis, β 0 is diagonal so if R
is the invertible matrix defining the change of basis, RT β 0 R = diag(λ1 , . . . , λn ) and
RT R = I. Putting P = QR we get the result. 2

Now let us try and set this in a geometric context. Let A, B be symmetric bilinear
forms on a complex vector space V which define different quadrics Q and Q0 . Then

36
Definition 11 The pencil of quadrics in P (V ) generated by Q and Q0 is the set of
quadrics defined by λA + µB where (λ, µ) 6= (0, 0).

Another way of saying this is to let SV denote the vector space of symmetric bilinear
forms on V , in which case a pencil of quadrics is a line in P (SV ) – a family of quadrics
parametrized by a projective line. The singular quadrics in this pencil are given by
the equation det(λα +µβ) = 0. If α is nonsingular then for a solution to this equation
µ 6= 0, and so under the hypotheses of the proposition, we can think of the points
[λi , −1] ∈ P 1 (C) as defining n singular quadrics in the pencil. The geometry of this
becomes directly visible in the case that P (V ) is a plane. In this case a singular
quadric has normal form x21 or x21 + x22 = (x1 − ix2 )(x1 + ix2 ) and so is either a double
line or a pair of lines.

Theorem 14 Let C and C 0 be nonsingular conics in a complex projective plane and

assume that the pencil generated by C and C 0 contains three singular conics. Then

• the pencil consists of all conics passing through four points in general position

• the singular conics of the pencil consist of the three pairs of lines obtained by
joining disjoint pairs of these four points

• each such pair of lines meets in a point with representative vector vi where
{v1 , v2 , v3 } is a basis for V relative to which the matrices of C and C 0 are
simultaneously diagonalizable.

Proof: The proof consists of reducing to normal form and calculating. Since by
hypothesis there are three singular conics in the pencil, Proposition 13 tells us that

37
there is a basis in which the two conics are defined by bilinear forms A, B with
equations:
x21 + x22 + x23 = 0, λ1 x21 + λ2 x22 + λ3 x23 = 0
and we directly find the points of intersection are the four points
p p p
[ λ2 − λ 3 , ± λ3 − λ1 , ± λ 1 − λ2 ]

where we fix one choice of square root of λ2 − λ3 . To show that these are in general
position we need to show that any three are linearly independent, but, for example
√ √ √ 
√ λ 2 − λ 3 √λ 3 − λ 1 √ λ 1 − λ 2 p p p
det  √λ2 − λ3 −√λ3 − λ1 λ
√1 − λ 2
 = 4 λ 1 − λ 2 λ 2 − λ 3 λ3 − λ 1
λ2 − λ3 − λ 3 − λ 1 − λ 1 − λ2

and since the λi are distinct this is non-zero.

Now clearly if [u] is one of these four points A(u, u) = 0, B(u, u) = 0 and so (λA +
µB)(u, u) = 0, and every conic in the pencil passes through them. Conversely by
Theorem 3 we can take the points to be

[1, 0, 0], [0, 1, 0], [0, 0, 1], [1, 1, 1].

The conics which pass through these points have matrices βij where

β11 = β22 = β33 = β12 + β23 + β31 = 0 (7)

The vector space of symmetric 3 × 3 matrices is of dimension 6 spanned by

          
1 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0
0 0 0,0 1 0 , 0 0
  0   1 0 0 , 0
  0 0,0 0 1
0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 1 0

and the four equations (7) are linearly independent, so define a 2-dimensional space
of bilinear forms spanned by A and B – this is the pencil.

We need to understand the singular quadrics in the pencil, for example

λ1 (x21 + x22 + x23 ) − (λ1 x21 + λ2 x22 + λ3 x23 ) = (λ1 − λ2 )x22 − (λ3 − λ1 )x23 = 0

But this factorizes as

p p p p
( λ1 − λ2 x2 − λ3 − λ1 x3 )( λ1 − λ2 x2 + λ3 − λ1 x3 ) = 0.

38
These two lines, by belonging to the pencil, pass through the points of intersection,
but since those points are in general position, each line passes through only two of
them, so these are the required pairs of lines.
√ √ √ √
The intersection of λ1 − λ2 x2 − λ3 − λ1 x3 = 0 and λ1 − λ2 x2 + λ3 − λ1 x3 = 0
is [1, 0, 0], which is the first basis vector in which the two conics are diagonalized.
Similarly the other two pairs of lines give the remaining basis vectors. 2

This geometrical approach tells us when two symmetric bilinear forms in three vari-
ables can and can not be simultaneously diagonalized. If the values λi are not all
distinct, and the two forms can be simultaneously diagonalized, then they are either
multiples of each other or are of the form

x21 + x22 + x23 = 0, λx21 + x22 + x23 = 0.

In this case the two conics intersect where x1 = 0, x2 = ±ix3 , i.e. at two points:

The non-diagonalizable case is where the intersection is at an odd number of points:

3.6 Exercises

1. Which of the following quadratic forms defines a non-singular conic?

39
 • x20 − 2x0 x1 + 4x0 x2 − 8x21 + 2x1 x2 + 3x22

• x20 − 2x0 x1 + x21 − 2x0 x2 .

2. Take five points in a projective plane such that no three are collinear. Show that
there exists a unique non-singular conic passing through all five points.

3. Let C be a non-singular conic in a projective plane P (V ). Show that if X ∈ P (V )

moves on a fixed line `, then its polar passes though a fixed point Y . What is the
relationship between the point Y and the line `?

4. Let τ : P 1 (R) → P 1 (R) be a projective transformation and consider its graph

Γτ ⊂ P 1 (R) × P 1 (R)

i.e. Γτ = {(X, Y ) : Y = τ (X)}. Using the one-to-one correspondence between

P 1 (R) × P 1 (R) and a quadric surface in P 3 (R), show that Γτ is the intersection of
the quadric surface with a plane.

5. Prove that if L ⊆ V and M ⊆ V are vector subspaces of the same dimension then

dim(L ∩ M ⊥ ) = dim(L⊥ ∩ M ).

6. Show that the two quadratic forms

x2 + y 2 − z 2 , x2 + y 2 − yz

cannot be simultaneously diagonalized.

7. Let P 5 (R) = P (R6 ) be the space of all conics in P 2 (R). Show that the conics
which pass through three non-collinear points form a projective plane P (V ) ⊂ P 5 (R).
Show further that the conics parametrized by this plane and which are tangent to a
given line form a conic in P (V ).

8. Prove that the set of tangent lines to a nonsingular conic in P (V ) is a conic in the
dual space P (V 0 ).

40