Translation and Rotation of Conics

A translation of the axes gives a new set of axes parallel to the old ones and in a rotation the axes are rotated about the
origin. Both yield a cleaner equation on the new set of axes (or the new plane).

Translation

Theorem 4.1 A point (𝑥, 𝑦) is on the parabola with focus (ℎ + 𝑐, 𝑘) and directrix 𝑥 = ℎ − 𝑐 if and only if it satisfies the
equation
(𝑦 − 𝑘)2 = 4𝑐(𝑥 − ℎ).

A point (𝑥, 𝑦) is on the parabola with focus (ℎ, 𝑘 + 𝑐) and directrix y= 𝑘 − 𝑐 if and only if it satisfies the equation
(𝑥 − ℎ)2 = 4𝑐(𝑦 − 𝑘).

Theorem 4.2 A point (𝑥, 𝑦) is on the ellipse with center (ℎ, 𝑘), vertices (ℎ ± 𝑎, 𝑘) and covertices (ℎ, 𝑘 ± 𝑏) if and only
if it satisfies the equation

(𝑥 − ℎ)2 (𝑦 − 𝑘)2
+ =1
𝑎2 𝑏2

The foci are (ℎ ± 𝑐, 𝑘), where 𝑐 2 = 𝑎2 − 𝑏 2 .

A point (𝑥, 𝑦) is on the ellipse with center (ℎ, 𝑘), vertices (ℎ, 𝑘 ± 𝑎) and covertices (ℎ ± 𝑏, 𝑘) if and only if it satisfies
the equation

(𝑦 − 𝑘)2 (𝑥 − ℎ)2
+ =1
𝑎2 𝑏2

The foci are (ℎ, 𝑘 ± 𝑐), where 𝑐 2 = 𝑎2 − 𝑏 2 .

Theorem 4.3 A point (𝑥, 𝑦) is on the hyperbola with center (ℎ, 𝑘), vertices (ℎ ± 𝑎, 𝑘) and foci (ℎ ± 𝑐, 𝑘) if and only if it
satisfies the equation

(𝑥 − ℎ)2 (𝑦 − 𝑘)2
− =1
𝑎2 𝑏2

where 𝑏 2 = 𝑐 2 − 𝑎2 .

A point (𝑥, 𝑦) is on the hyperbola with center (ℎ, 𝑘), vertices (ℎ, 𝑘 ± 𝑎) and foci (ℎ, 𝑘 ± 𝑐) if and only if it satisfies the
equation

(𝑦 − 𝑘)2 (𝑥 − ℎ)2
− =1
𝑎2 𝑏2

where 𝑏 2 = 𝑐 2 − 𝑎2 .
Theorem 4.4 Every conic with axis (or axes) parallel to or on a coordinate axis (or axes) may be represented by an
equation of the form

𝐴𝑥 2 + 𝐶𝑦 2 + 𝐷𝑥 + 𝐸𝑦 + 𝐹 = 0,

Where A and C are not both zero. Furthermore, 𝐴𝐶 = 0 if the conic is a parabola, 𝐴𝐶 > 0 if it is an ellipse, and 𝐴𝐶 < 0
if it is a hyperbola.

Example 1 Find an equation of the hyperbola containing (7, −2), (−1, −2), (8,1) and (-2, -5).

It is easier to use the general form 𝐴𝑥 2 + 𝐶𝑦 2 + 𝐷𝑥 + 𝐸𝑦 + 𝐹 = 0 from theorem 4.4 rather than the standard form from
theorem 4.3. Since we know the four points are on the hyperbola they must each satisfy the equation from theorem 4.4.
Thus we have

Using (7, −2): 49𝐴 + 4𝐶 + 7𝐷 − 2𝐸 + 𝐹 = 0

and (−1, −2): 𝐴 + 4𝐶 − 𝐷 − 2𝐸 + 𝐹 = 0
and (8,1): 64𝐴 + 𝐶 + 8𝐷 + 𝐸 + 𝐹 = 0
and (−2, −5): 4𝐴 + 25𝐶 − 2𝐷 − 5𝐸 + 𝐹 = 0

There is some really cool techniques from Matrix Theory that would help us solve this, but this one is easily handled
using basic techniques from linear algebra.

Subtract the second equation from the other three, eliminating F.

48𝐴 + 8𝐷 =0
63𝐴 − 3𝐶 + 9𝐷 + 3𝐸 = 0
3𝐴 + 21𝐶 − 𝐷 − 3𝐷 = 0
adding the last pair, we get:
48𝐴 + 8𝐷 = 0
66𝐴 + 18𝐶 + 8𝐷 = 0
subtract the first from the second
18𝐴 + 18𝐶 = 0
or
𝐶 = −𝐴
Now, substituting back, we get:
𝐷 = −6𝐴
𝐸 = −4𝐴
𝐹 = −11𝐴
and the desired equation is
𝐴𝑥 2 − 𝐴𝑦 2 − 6𝐴𝑥 − 4𝐴𝑦 − 11𝐴 = 0
or
𝑥 2 − 𝑦 2 − 6𝑥 − 4𝑦 − 11 = 0

Translation of General Equations

Another method of translating will be illustrated in the following examples.

Example 2 Translate axes to eliminate the constant term and the x term of

𝑥 2 − 2𝑥𝑦 + 𝑦 2 + 4𝑥 − 6𝑦 + 10 = 0.
The xy term prohibits completing the square and getting one of the forms in Theorems 4.1 – 4.3. We don’t
know what h and k need to be before translating, so we use the following translation
 𝑥 = 𝑥′ + ℎ 𝑎𝑛𝑑 𝑦 = 𝑦′ + 𝑘

and determine the values of h and k needed to eliminate the desired terms. Substituting, we have

(𝑥 ′ + ℎ)2 − 2(𝑥 ′ + ℎ)(𝑦 ′ + 𝑘) + (𝑦 ′ + 𝑘)2 + 4(𝑥 ′ + ℎ) − 6(𝑦 ′ + 𝑘) + 10 = 0

2
𝑥′2 − 2𝑥 ′ 𝑦 ′ + 𝑦 ′ + (2ℎ − 2𝑘 + 4)𝑥 ′ + (−2ℎ + 2𝑘 − 6)𝑦 ′ + (ℎ2 − 2ℎ𝑘 + 𝑘 2 + 4ℎ − 6𝑘 + 10) = 0

Remember, we need to eliminate the constant term and the x term, so we want the constant term and the
coefficient of 𝑥′ (the x term) to be zero. So, we must determine h and k so that

ℎ2 − 2ℎ𝑘 + 𝑘 2 + 4ℎ − 6𝑘 + 10 = 0 𝑎𝑛𝑑 2ℎ − 2𝑘 + 4 = 0.

Solve for h in terms of k in the second equation and substitute into the first, then

ℎ =𝑘−2
(𝑘 − 2) − 2(𝑘 − 2)𝑘 + 𝑘 2 + 4(𝑘 − 2) − 6𝑘 + 10 = 0
2

−2𝑘 + 6 = 0
𝑘=3
ℎ =𝑘−2=1

Hence, the translation becomes

𝑥 = 𝑥′ + 1 𝑎𝑛𝑑 𝑦 = 𝑦′ + 3

Which, in the new coordinate system (the 𝑥′𝑦′ plane), gives us the equation
2
𝑥′2 − 2𝑥 ′ 𝑦 ′ + 𝑦 ′ − 2𝑦 ′ = 0.

The graph showing both sets of coordinate axes is in Figure 1.

Figure 1
Example 3 Translate axes to eliminate the first degree terms of
𝑥 2 − 4𝑥𝑦 + 3𝑥 − 2𝑦 + 4 = 0.

We can do the same thing we did in example 1 and the translation becomes

1 1
𝑥 = 𝑥′ − 𝑎𝑛𝑑 𝑦 = 𝑦′ +
2 2

Which, in the new coordinate system (the 𝑥′𝑦′ plane), gives us the equation
11
𝑥′2 − 4𝑥 ′ 𝑦 ′ + = 0.
4

The graph showing both sets of coordinate axes is in Figure 2.

Figure 2

Both of these translations were carried out on a general second-degree equation

𝐴𝑥 2 + 𝐵𝑥𝑦 + 𝐶𝑦 2 + 𝐷𝑥 + 𝐸𝑦 + 𝐹 = 0

and it should be noted that the values of A, B and C did not change. Also, when 𝐵 ≠ 0, we can not use AC to
determine the type of conic section we have.
Rotation

We now consider rotation of the axes about the origin. If the axes are rotated through an angle 𝜃, then each point in the
plane has two representations: (𝑥, 𝑦) in the original coordinate system and (𝑥 ′ , 𝑦′) in the new coordinate system.

Likewise, every vector v in the plane has two representations: 𝒗 = 𝑥𝒊 + 𝑦𝒋 in the original coordinate system and 𝒗 =
𝑥 ′ 𝒊 + 𝑦′𝒋 in the new coordinate system (see Figure 1).

Figure 3
We want to find the relationship between (𝑥, 𝑦) and (𝑥 ′ , 𝑦′). To develop this relationship we will look at the relationship
of i and j with 𝒊′ and 𝒋′ – unit vectors for each of the two coordinate systems. From Figure 2,

Figure 4

we see that
𝒊′ = cos 𝜃 𝒊 + sin 𝜃 𝒋
′
𝒋 = − sin 𝜃 𝒊 + cos 𝜃 𝒋

This gives us
𝑣 = 𝑥 ′ 𝒊′ + 𝑦 ′ 𝒋′ (in the new coordinate system)

= 𝑥′(cos 𝜃 𝒊 + 𝑠𝑖𝑛𝜃𝒋) + 𝑦′(− sin 𝜃 𝒊 + cos 𝜃 𝒋)

= (𝑥 ′ cos 𝜃 − 𝑦 ′ sin 𝜃)𝒊 + (𝑥 ′ sin 𝜃 + 𝑦 ′ cos 𝜃)𝒋

= 𝑥𝒊 + 𝑦𝒋 (in the original coordinate system)

Thus, we have developed the relationship

𝑥 = 𝑥 ′ cos 𝜃 − 𝑦 ′ sin 𝜃
and
𝑦 = 𝑥 ′ sin 𝜃 + 𝑦 ′ cos 𝜃

called the equations of rotation.

Let’s look at a couple of simple examples.

Example 4 Find the new representation of
𝑥 2 − 𝑥𝑦 + 𝑦 2 − 2 = 0
after rotation through an angle of 45°

Since sin 45° = cos 45° = 1⁄ , the equations of rotation are

√2
𝑥 ′ − 𝑦′ 𝑥 ′ + 𝑦′
𝑥= 𝑎𝑛𝑑 𝑦=
√2 √2
and substituting into the original equation, we get
𝑥′2 𝑦′2
𝑥′2 + 3𝑦′2 = 4 𝑜𝑟 + =1
4 4⁄
3

Figure 5
Example 5 Find the new representation of
𝑥 2 + 4𝑥𝑦 − 2𝑦 2 − 6 = 0
after rotation through an angle 𝜃 = 𝑡𝑎𝑛−1 (1⁄2).

Since 𝜃 = 𝑡𝑎𝑛−1 (1⁄2) , we know 0 ≤ 𝜃 < 90° and

1 2
sin 𝜃 = 𝑎𝑛𝑑 cos 𝜃 =
√5 √5
giving us equations of rotation
2𝑥 ′ − 𝑦′ 𝑥 ′ + 2𝑦′
𝑥= 𝑎𝑛𝑑 𝑦=
√5 √5

Substituting into the original equation, we get

2
2 𝑥′ 𝑦′2
2𝑥′2 − 3𝑦 ′ = 6 𝑜𝑟 − =1
3 2

Figure 6

Rotation of General Equations

A conic section with axes parallel to the coordinate axes can be represented by a second-degree equation

𝐴𝑥 2 + 𝐵𝑥𝑦 + 𝐶𝑦 2 + 𝐷𝑥 + 𝐸𝑦 + 𝐹 = 0

with B = 0; and, any second-degree equation with B = 0 represents a conic or degenerate conic with axes
parallel to the coordinate axes. We extend this to conic sections in any position.
A conic or degenerate conic given by a second-degree equation with 𝐵 ≠ 0 rotated through some positive angle
less than 90° should give us a conic with axis (or axes) on or parallel to the coordinate axes – such a rotation
should eliminate the xy term.
Assume 0° < 𝜃 < 90°, 𝐵 ≠ 0 and
𝐴𝑥 2 + 𝐵𝑥𝑦 + 𝐶𝑦 2 + 𝐷𝑥 + 𝐸𝑦 + 𝐹 = 0.
Substituting the equations of rotation,
𝑥 = 𝑥 ′ cos 𝜃 − 𝑦 ′ sin 𝜃
𝑦 = 𝑥 ′ sin 𝜃 + 𝑦 ′ cos 𝜃

we have
𝐴(𝑥 ′ cos 𝜃 − 𝑦 ′ sin 𝜃)2 + 𝐵(𝑥 ′ cos 𝜃 − 𝑦 ′ sin 𝜃)(𝑥 ′ sin 𝜃 + 𝑦 ′ cos 𝜃)
+𝐶(𝑥 ′ sin 𝜃 + 𝑦 ′ cos 𝜃)2 + 𝐷(𝑥 ′ cos 𝜃 − 𝑦 ′ sin 𝜃)
+(𝑥 ′ sin 𝜃 + 𝑦 ′ cos 𝜃) + 𝐹 = 0

and, after carrying out the multiplication and combining similar terms, we find that the coefficient of 𝑥′𝑦′ is

(𝐶 − 𝐴)2 sin 𝜃 cos 𝜃 + 𝐵(𝑐𝑜𝑠 2 𝜃 − 𝑠𝑖𝑛2 𝜃) = (𝐶 − 𝐴) sin 2𝜃 + 𝐵𝑐𝑜𝑠 2𝜃

We need to choose 𝜃 properly so that this coefficient is zero. Set it equal to zero and determine 𝜃.

(𝐶 − 𝐴) sin 2𝜃 + 𝐵𝑐𝑜𝑠 2𝜃 = 0

Case 1: If A = C, then
𝐵𝑐𝑜𝑠 2𝜃 = 0
2𝜃 = 90°
𝜃 = 45°

Case 2: If 𝐴 ≠ 0, then
(𝐴 − 𝐶) sin 2𝜃 = 𝐵𝑐𝑜𝑠 2𝜃
(𝐴 − 𝐶) tan 2𝜃 = 𝐵
2 tan 𝜃
(𝐴 − 𝐶) =𝐵
1 − 𝑡𝑎𝑛2 𝜃
which gives us the quadratic equation in tan 𝜃
𝐵𝑡𝑎𝑛2 𝜃 + 2(𝐴 − 𝐶) tan 𝜃 − 𝐵 = 0.

Using the quadratic formula, we have

(𝐶 − 𝐴) ± √(𝐶 − 𝐴)2 + 𝐵2
tan 𝜃 =
𝐵

There are two possible values of tan 𝜃, but since we are assuming 0° < 𝜃 < 90°, we want the positive value of tan 𝜃.

Now, since
√(𝐶 − 𝐴)2 + 𝐵2 > √(𝐶 − 𝐴)2 = |𝐶 − 𝐴|

we can always be sure that tan 𝜃 is positive as long as the sign on the radical is chosen to agree with the sign of B. Once
we have tan 𝜃 it is easy to find sin 𝜃 and cos 𝜃 and substitute them into the equations of rotation. Thus we are always
able to rotate axes to eliminate the xy term and the resulting equation must represent a conic or degenerate conic.

Theorem 4.5 Any conic section can be represented by the second-degree equation

𝐴𝑥 2 + 𝐵𝑥𝑦 + 𝐶𝑦 2 + 𝐷𝑥 + 𝐸𝑦 + 𝐹 = 0

where A, B, and C are not all zero. Any second-degree equation represents either a conic or a degenerate
conic.

And summing up, if 𝐵 ≠ 0, then the axes may be rotated to eliminate the xy term following Case 1 or Case 2
above.
Example 6 Rotate axes to eliminate the xy term of 𝑥 2 + 4𝑥𝑦 − 2𝑦 2 − 6 = 0.

(−2 − 1) + √(−2 − 1)2 + 42 −3 + √9 + 16 1

tan 𝜃 = = =
4 4 2

1 2
sin 𝜃 = 𝑎𝑛𝑑 cos 𝜃 =
√5 √5

Thus the equations of rotation are

2𝑥 ′ − 𝑦′ 𝑥 ′ + 2𝑦′
𝑥= 𝑎𝑛𝑑 𝑦=
√5 √5

Substituting into the original equation, we get

2
2 𝑥′ 𝑦′2
2𝑥′2 − 3𝑦 ′ = 6 𝑜𝑟 − =1
3 2

Figure 7
Example 7 Rotate axes to eliminate the xy term of 2𝑥 2 − 𝑥𝑦 + 2𝑦 2 − 2 = 0.

𝐴=𝐶 ⟹ 𝜃 = 45°

Thus the equations of rotation are

𝑥 ′ − 𝑦′ 𝑥 ′ + 𝑦′
𝑥= 𝑎𝑛𝑑 𝑦=
√2 √2

Substituting into the original equation, we get

2
2 ′2
𝑥′ 𝑦′2
3𝑥′ + 5𝑦 =4 𝑜𝑟 − =1
4⁄ 4⁄
3 5

Figure 8
Example 8 Rotate axes to eliminate the xy term of

21𝑥 2 + 50𝑥𝑦 − 99𝑦 2 + 6√26𝑥 − 82√26𝑦 − 494 = 0

(−99 − 21) + √(−99 − 21)2 + 502 −120 + 10√122 + 52 1

tan 𝜃 = = =
50 50 5

1 5
sin 𝜃 = 𝑎𝑛𝑑 cos 𝜃 =
√26 √26

Thus the equations of rotation are

5𝑥 ′ − 𝑦′ 𝑥 ′ + 5𝑦′
𝑥= 𝑎𝑛𝑑 𝑦=
√26 √26

Substituting into the original equation, we get

2
𝑥′2 − 4𝑦 ′ − 2𝑥 ′ − 16𝑦 ′ − 19 = 0

and then by completing the square

(𝑥 ′ − 1)2 (𝑦 ′ + 2)2
− =1
4 1

Figure 9
Theorem 4.6 If the equation

𝐴𝑥 2 + 𝐵𝑥𝑦 + 𝐶𝑦 2 + 𝐷𝑥 + 𝐸𝑦 + 𝐹 = 0

is transformed into the equation

2 2
𝐴′𝑥 ′ + 𝐵′𝑥′𝑦′ + 𝐶′𝑦 ′ + 𝐷′𝑥′ + 𝐸′𝑦′ + 𝐹′ = 0

by rotating the axes, then

2
𝐵2 − 4𝐴𝐶 = 𝐵′ − 4𝐴′𝐶′.

Theorem 4.7 The equation

𝐴𝑥 2 + 𝐵𝑥𝑦 + 𝐶𝑦 2 + 𝐷𝑥 + 𝐸𝑦 + 𝐹 = 0

represents a hyperbola, ellipse, or parabola (or a degenerate case of one of these) according to whether 𝐵2 − 4𝐴𝐶 is
positive, negative, or zero, respectively.