14.

5  Directional Derivatives and Gradient Vectors 821 822 Chapter 14: Partial Derivatives

through P0(x0, y0), parametrized with the arc length parameter s increasing in the direction y (Figure 14.29). The derivative of ƒ at (2, 0) in the direction of v is therefore
of the unit vector u = u1 i + u2 j. Then by the Chain Rule we find
∇f = i + 2j (Duƒ)(2,0) = ∇ƒ 0 (2,0) # u Eq. (4) with the (Du ƒ)P0 notation
2
dƒ 0ƒ dx 0ƒ dy
a b = a b + a b Chain Rule for differentiable ƒ 3 4 3 8
ds u, P0 0x P ds 0y P ds = (i + 2j) # a i - jb = - = -1.
0 0 5 5 5 5
1
0ƒ 0ƒ From Eqs. (2), dx>ds = u1
= a b u1 + a b u2 and dy>ds = u2
0x P 0y P Evaluating the dot product in the brief version of Equation (4) gives
0 0
x
0 1 P0 (2, 0) 3 4
0ƒ 0ƒ Duƒ = ∇ƒ # u = 0 ∇ƒ 0 0 u 0 cos u = 0 ∇ƒ 0 cos u,
= c a b i + a b j d # c u1 i + u2 j d . (3)
0x P 0y P u = 3i − 4j
0 0 where u is the angle between the vectors u and ∇ƒ, and reveals the following properties.
(++++)++++* (++)++* −1 5 5
Gradient of ƒ at P0 Direction u

Equation (3) says that the derivative of a differentiable function ƒ in the direction of u Figure 14.29 Picture ∇ƒ as a vector Properties of the Directional Derivative Duƒ = ∇ƒ # u = 0 ∇ƒ 0 cos u
at P0 is the dot product of u with the special vector, which we now define. in the domain of ƒ. The figure shows a
1. The function ƒ increases most rapidly when cos u = 1 or when u = 0 and u
number of level curves of ƒ. The rate at
is the direction of ∇ƒ. That is, at each point P in its domain, ƒ increases most
which ƒ changes at (2, 0) in the direction
rapidly in the direction of the gradient vector ∇ƒ at P. The derivative in this
Definition The gradient vector (gradient) of ƒ(x, y) at a point P0(x0 , y0) is u is ∇ƒ # u = - 1, which is the component
direction is
the vector of ∇ƒ in the direction of unit vector u
(Example 2). Duƒ = 0 ∇ƒ 0 cos (0) = 0 ∇ƒ 0 .
0ƒ 0ƒ 2. Similarly, ƒ decreases most rapidly in the direction of - ∇ƒ. The derivative in
∇ƒ = i + j
0x 0y this direction is Duƒ = 0 ∇ƒ 0 cos (p) = - 0 ∇ƒ 0 .
obtained by evaluating the partial derivatives of ƒ at P0 . 3. Any direction u orthogonal to a gradient ∇f ≠ 0 is a direction of zero change
in ƒ because u then equals p>2 and
Duƒ = 0 ∇ƒ 0 cos (p>2) = 0 ∇ƒ 0 # 0 = 0.
The notation ∇ƒ is read “grad ƒ” as well as “gradient of ƒ” and “del ƒ.” The symbol ∇ by
itself is read “del.” Another notation for the gradient is grad ƒ. Using the gradient notation,
we restate Equation (3) as a theorem.
As we discuss later, these properties hold in three dimensions as well as two.

theorem 9—the Directional Derivative is a Dot product If ƒ(x, y) is differ- example 3 Find the directions in which ƒ(x, y) = ( x2 >2 ) + ( y2 >2 )
entiable in an open region containing P0(x0 , y0), then
(a) increases most rapidly at the point (1, 1), and
dƒ (b) decreases most rapidly at (1, 1).
a b = (∇ƒ)P0 # u, (4)
ds u, P0
(c) What are the directions of zero change in ƒ at (1, 1)?
the dot product of the gradient ∇ƒ at P0 and u. In brief, Du ƒ = ∇ƒ # u.
z z = f(x, y) Solution
x2 y2
= + (a) The function increases most rapidly in the direction of ∇ƒ at (1, 1). The gradient there
2 2
example 2 Find the derivative of ƒ(x, y) = xey + cos (xy) at the point (2, 0) in the is
direction of v = 3i - 4j. (∇ƒ)(1,1) = (xi + yj)(1,1) = i + j.
(1, 1, 1)
Solution Recall that the direction of a vector v is the unit vector obtained by dividing v Its direction is
by its length: i + j i + j 1 1
1 u = = = i + j.
y 0i + j0 2(1)2 + (1)2 22 22
v v 3 4 1 −∇f (1, 1)
u = = = i - j. Zero change
0v0 5 5 5 (b) The function decreases most rapidly in the direction of - ∇ƒ at (1, 1), which is
Most rapid in f
decrease in f
The partial derivatives of ƒ are everywhere continuous and at (2, 0) are given by x 1 1
Most rapid ∇f = i + j -u = - i - j.
increase in f 22 22
fx(2, 0) = (ey - y sin (xy))(2,0) = e0 - 0 = 1
Figure 14.30 The direction in which (c) The directions of zero change at (1, 1) are the directions orthogonal to ∇ƒ:
fy(2, 0) = (xey - x sin (xy))(2,0) = 2e0 - 2 # 0 = 2.
ƒ(x, y) increases most rapidly at (1, 1) is 1 1 1 1
The gradient of ƒ at (2, 0) is the direction of ∇ƒ 0 (1,1) = i + j. It corre- n = - i + j and -n = i - j.
22 22 22 22
sponds to the direction of steepest ascent
∇ƒ 0 (2,0) = ƒx(2, 0)i + ƒy(2, 0)j = i + 2j on the surface at (1, 1, 1) (Example 3). See Figure 14.30.
 14.5  Directional Derivatives and Gradient Vectors 823 824 Chapter 14: Partial Derivatives

gradients and Tangents to Level Curves The gradient of ƒ at (-2, 1) is

If a differentiable function ƒ(x, y) has a constant value c along a smooth curve x
r = g(t)i + h(t)j (making the curve part of a level curve of ƒ), then ƒ(g(t), h(t)) = c. Dif- ∇ƒ 0 (-2,1) = a i + 2yjb = -i + 2 j.
2 (-2,1)
ferentiating both sides of this equation with respect to t leads to the equations
The tangent to the ellipse at (-2, 1) is the line
d d (-1)(x + 2) + (2)(y - 1) = 0 Eq. (6)
ƒ(g(t), h(t)) = (c)
dt dt
x - 2y = -4.
0ƒ dg 0ƒ dh
+ = 0 Chain Rule
0x dt 0y dt If we know the gradients of two functions ƒ and g, we automatically know the gradients of
0ƒ dh their sum, difference, constant multiples, product, and quotient. You are asked to establish
a (5)
0ƒ # dg
i + jb a i + jb = 0. the following rules in Exercise 40. Notice that these rules have the same form as the cor-
0x 0y dt dt
(++)++* (++)++* responding rules for derivatives of single-variable functions.
dr
∇ƒ
dt

Equation (5) says that ∇ƒ is normal to the tangent vector dr > dt, so it is normal to the curve. Algebra Rules for Gradients
The level curve f (x, y) = f (x 0 , y 0 )
1. Sum Rule: ∇(ƒ + g) = ∇ƒ + ∇g
2. Difference Rule: ∇(ƒ - g) = ∇ƒ - ∇g
(x 0 , y 0 )
At every point (x0 , y0) in the domain of a differentiable function ƒ(x, y), the gra- 3. Constant Multiple Rule: ∇(kƒ) = k∇ƒ (any number k)
∇ f (x 0 , y 0 )
dient of ƒ is normal to the level curve through (x0 , y0) (Figure 14.31). 4. Product Rule: ∇(ƒg) = ƒ∇g + g∇ƒ Scalar multipliers on left
ƒ g∇ƒ - ƒ∇g of gradients
5. Quotient Rule: ∇ agb =
g2
Equation (5) validates our observation that streams flow perpendicular to the contours
Figure 14.31 The gradient of a dif-
in topographical maps (see Figure 14.26). Since the downflowing stream will reach its
ferentiable function of two variables at a
destination in the fastest way, it must flow in the direction of the negative gradient vectors example 5 We illustrate two of the rules with
point is always normal to the function’s
from Property 2 for the directional derivative. Equation (5) tells us these directions are
level curve through that point. ƒ(x, y) = x - y g(x, y) = 3y
perpendicular to the level curves.
This observation also enables us to find equations for tangent lines to level curves. ∇ƒ = i - j ∇g = 3j.
They are the lines normal to the gradients. The line through a point P0(x0 , y0) normal to a We have
vector N = Ai + Bj has the equation 1. ∇(ƒ - g) = ∇(x - 4y) = i - 4j = ∇ƒ - ∇g Rule 2
A(x - x0) + B( y - y0) = 0 2. ∇(ƒg) = ∇(3xy - 3y2) = 3yi + (3x - 6y)j
(Exercise 39). If N is the gradient (∇ƒ)(x0, y0) = ƒx(x0 , y0)i + ƒy(x0 , y0)j, the equation gives = 3y(i - j) + 3yj + (3x - 6y)j g∇ƒ plus terms . . .
the following formula. = 3y(i - j) + (3x - 3y)j simplified.
= 3y(i - j) + (x - y)3j = g∇ƒ + ƒ∇g Rule 4

Tangent Line to a Level Curve Functions of Three variables

ƒx(x0 , y0)(x - x0) + ƒy(x0 , y0)(y - y0) = 0 (6) For a differentiable function ƒ(x, y, z) and a unit vector u = u1 i + u2 j + u3 k in space,
y
we have
∇ f (−2, 1) = −i + 2j x − 2y = −4
0ƒ 0ƒ 0ƒ
∇ƒ = i + j + k
x2 + y2 = 2 0x 0y 0z
"2
4
1 example 4 Find an equation for the tangent to the ellipse and
(−2, 1)
x x2 0ƒ 0ƒ 0ƒ
−2 −1 0 1 2 2" 2 + y2 = 2 Duƒ = ∇ƒ # u = u + u + u.
4 0x 1 0y 2 0z 3
(Figure 14.32) at the point (-2, 1). The directional derivative can once again be written in the form
Duƒ = ∇ƒ # u = 0 ∇ƒ 0 0 u 0 cos u = 0 ∇ƒ 0 cos u,
Figure 14.32 We can find the tangent Solution The ellipse is a level curve of the function
to the ellipse ( x2 >4 ) + y2 = 2 by treating so the properties listed earlier for functions of two variables extend to three variables. At
the ellipse as a level curve of the function x2 any given point, ƒ increases most rapidly in the direction of ∇ƒ and decreases most rap-
ƒ(x, y) = + y2.
ƒ(x, y) = ( x2 >4 ) + y2 (Example 4). 4 idly in the direction of - ∇ƒ. In any direction orthogonal to ∇ƒ, the derivative is zero.
 14.5  Directional Derivatives and Gradient Vectors 825 826 Chapter 14: Partial Derivatives

example 6
(a) Find the derivative of ƒ(x, y, z) = x3 - xy2 - z at P0(1, 1, 0) in the direction of exercises 14.5
v = 2i - 3j + 6k. Theory and examples
Calculating gradients
(b) In what directions does ƒ change most rapidly at P0 , and what are the rates of change In Exercises 1–6, find the gradient of the function at the given point. 29. Let ƒ(x, y) = x2 - xy + y2 - y. Find the directions u and the
in these directions? Then sketch the gradient together with the level curve that passes values of Du ƒ(1, -1) for which
through the point. a. Du ƒ(1, -1) is largest b. Du ƒ(1, - 1) is smallest
Solution 1. ƒ(x, y) = y - x, (2, 1) 2. ƒ(x, y) = ln ( x2 + y2 ) , (1, 1) c. Du ƒ(1, -1) = 0 d. Du ƒ(1, -1) = 4
(a) The direction of v is obtained by dividing v by its length: x2 y2 e. Du ƒ(1, -1) = -3
3. g(x, y) = xy2, (2, - 1) 4. g(x, y) = - , 1 22, 1 2
0 v 0 = 2(2)2 + (-3)2 + (6)2 = 249 = 7 2 2 (x - y)
30. Let ƒ(x, y) = . Find the directions u and the values of
5. ƒ(x, y) = 22x + 3y, (- 1, 2) (x + y)
v 2 3 6
u = = i - j + k. 1 3
0v0 7 7 7 2x Du ƒa- , b for which
6. ƒ(x, y) = tan-1 y , (4, - 2) 2 2
The partial derivatives of ƒ at P0 are In Exercises 7–10, find ∇f at the given point. 1 3 1 3
a. Du ƒa- , b is largest b. Du ƒa- , b is smallest
7. ƒ(x, y, z) = x2 + y2 - 2z2 + z ln x, (1, 1, 1) 2 2 2 2
ƒx = ( 3x2 - y2 ) (1,1,0) = 2, ƒy = -2xy 0 (1,1,0) = -2, ƒz = -1 0 (1,1,0) = -1.
3 2 2 -1 1 3 1 3
8. ƒ(x, y, z) = 2z - 3(x + y )z + tan xz, (1, 1, 1) c. Du ƒa- , b = 0 d. Du ƒa- , b = - 2
The gradient of ƒ at P0 is 2 2 2 2
9. ƒ(x, y, z) = (x2 + y2 + z2)-1>2 + ln (xyz), (- 1, 2, -2)
∇ƒ 0 (1,1,0) = 2i - 2j - k.
10. ƒ(x, y, z) = ex + y cos z + (y + 1) sin-1 x, (0, 0, p>6) 1 3
e. Du ƒa- , b = 1
The derivative of ƒ at P0 in the direction of v is therefore 2 2
Finding Directional Derivatives 31. Zero directional derivative In what direction is the derivative
2 3 6 of ƒ(x, y) = xy + y2 at P(3, 2) equal to zero?
(Duƒ)(1,1,0) = ∇ƒ 0 (1,1,0) # u = (2i - 2j - k) # a i - j + kb In Exercises 11–18, find the derivative of the function at P0 in the
7 7 7
direction of u. 32. Zero directional derivative In what directions is the derivative
4 6 6 4 11. ƒ(x, y) = 2xy - 3y2, P0(5, 5), u = 4i + 3j of ƒ(x, y) = ( x2 - y2 ) > ( x2 + y2 ) at P(1, 1) equal to zero?
= + - = .
7 7 7 7 2 2 33. Is there a direction u in which the rate of change of ƒ(x, y) =
12. ƒ(x, y) = 2x + y , P0(- 1, 1), u = 3i - 4j
(b) The function increases most rapidly in the direction of ∇ƒ = 2i - 2j - k and x - y x2 - 3xy + 4y2 at P(1, 2) equals 14? Give reasons for your answer.
13. g(x, y) = , P0(1, - 1), u = 12i + 5j
decreases most rapidly in the direction of - ∇ƒ. The rates of change in the directions xy + 2 34. Changing temperature along a circle Is there a direction u in
are, respectively, 14. h(x, y) = tan-1 (y>x) + 23 sin-1 (xy>2), P0(1, 1), which the rate of change of the temperature function T(x, y, z) =
u = 3i - 2j 2xy - yz (temperature in degrees Celsius, distance in feet) at
0 ∇ƒ 0 = 2(2)2 + (-2)2 + (-1)2 = 29 = 3 and - 0 ∇ƒ 0 = -3. P(1, -1, 1) is -3°C>ft? Give reasons for your answer.
15. ƒ(x, y, z) = xy + yz + zx, P0(1, -1, 2), u = 3i + 6j - 2k
35. The derivative of ƒ(x, y) at P0(1, 2) in the direction of i + j is
16. ƒ(x, y, z) = x2 + 2y2 - 3z2, P0(1, 1, 1), u = i + j + k
The Chain rule for paths 2 22 and in the direction of - 2j is -3. What is the derivative of
17. g(x, y, z) = 3ex cos yz, P0(0, 0, 0), u = 2i + j - 2k ƒ in the direction of -i - 2j? Give reasons for your answer.
If r(t) = x(t) i + y(t) j + z(t) k is a smooth path C, and w = ƒ(r(t)) is a scalar function 18. h(x, y, z) = cos xy + eyz + ln zx, P0(1, 0, 1>2), 36. The derivative of ƒ(x, y, z) at a point P is greatest in the direction
evaluated along C, then according to the Chain Rule, Theorem 6 in Section 14.4, u = i + 2j + 2k of v = i + j - k. In this direction, the value of the derivative is
In Exercises 19–24, find the directions in which the functions increase 2 23.
dw 0w dx 0w dy 0w dz
= + + . and decrease most rapidly at P0 . Then find the derivatives of the func- a. What is ∇ƒ at P ? Give reasons for your answer.
dt 0x dt 0y dt 0z dt
tions in these directions. b. What is the derivative of ƒ at P in the direction of i + j ?
The partial derivatives on the right-hand side of the above equation are evaluated along the
19. ƒ(x, y) = x2 + xy + y2, P0(- 1, 1) 37. Directional derivatives and scalar components How is the
curve r(t), and the derivatives of the intermediate variables are evaluated at t. If we express
this equation using vector notation, we have 20. ƒ(x, y) = x2y + exy sin y, P0(1, 0) derivative of a differentiable function ƒ(x, y, z) at a point P0 in the
21. ƒ(x, y, z) = (x>y) - yz, P0(4, 1, 1) direction of a unit vector u related to the scalar component of
(∇ƒ)P0 in the direction of u? Give reasons for your answer.
22. g(x, y, z) = xey + z2, P0(1, ln 2, 1>2)
The Derivative Along a Path 38. Directional derivatives and partial derivatives Assuming
23. ƒ(x, y, z) = ln xy + ln yz + ln xz, P0(1, 1, 1) that the necessary derivatives of ƒ(x, y, z) are defined, how are
d 24. h(x, y, z) = ln (x2 + y2 - 1) + y + 6z, P0(1, 1, 0) Di ƒ, Dj ƒ, and Dk ƒ related to ƒx , ƒy , and ƒz? Give reasons for
ƒ(r(t)) = ∇ƒ(r(t)) # r′(t). (7) your answer.
dt
Tangent Lines to Level Curves 39. Lines in the xy-plane Show that A(x - x0) + B( y - y0) = 0
In Exercises 25–28, sketch the curve ƒ(x, y) = c together with ∇ƒ is an equation for the line in the xy-plane through the point (x0 , y0)
and the tangent line at the given point. Then write an equation for the normal to the vector N = Ai + Bj.
What Equation (7) says is that the derivative of the composite function ƒ(r(t)) is the tangent line.
“derivative” (gradient) of the outside function ƒ “times” (dot product) the derivative of the 40. The algebra rules for gradients Given a constant k and the
25. x2 + y2 = 4, 1 22, 22 2 gradients
inside function r. This is analogous to the “Outside-Inside” Rule for derivatives of com-
posite functions studied in Section 3.6. That is, the multivariable Chain Rule for paths has 26. x2 - y = 1, 1 22, 1 2 0ƒ 0ƒ 0ƒ 0g 0g 0g
∇ƒ = i + j + k, ∇g = i + j + k,
exactly the same form as the rule for single-variable differential calculus when appropriate 27. xy = - 4, (2, - 2) 0x 0y 0z 0x 0y 0z
interpretations are given to the meanings of the terms and operations involved. 28. x2 - xy + y2 = 7, (- 1, 2) establish the algebra rules for gradients.
 14.6  Tangent Planes and Differentials 827 828 Chapter 14: Partial Derivatives

To find an equation for the plane tangent to a smooth surface z = ƒ(x, y) at a point
14.6 Tangent planes and Differentials P0(x0 , y0 , z0) where z0 = ƒ(x0 , y0), we first observe that the equation z = ƒ(x, y) is equiva-
lent to ƒ(x, y) - z = 0. The surface z = ƒ(x, y) is therefore the zero level surface of the
In single-variable differential calculus we saw how the derivative defined the tangent line to function F(x, y, z) = ƒ(x, y) - z. The partial derivatives of F are
the graph of a differentiable function at a point on the graph. The tangent line then provided
for a linearization of the function at the point. In this section, we will see analogously how 0
Fx = (ƒ(x, y) - z) = fx - 0 = fx
the gradient defines the tangent plane to the level surface of a function w = ƒ(x, y, z) at a 0x
point on the surface. In the same way as before, the tangent plane then provides for a linear- 0
ization of ƒ at the point and defines the total differential of the function. Fy = (ƒ(x, y) - z) = fy - 0 = fy
0y
0
Tangent planes and normal Lines Fz = (ƒ(x, y) - z) = 0 - 1 = -1.
0z
∇f
v2 If r(t) = x(t) i + y(t) j + z(t) k is a smooth curve on the level surface ƒ(x, y, z) = c of a The formula
differentiable function ƒ, we found in Equation (7) of the last section that
v1 Fx(P0)(x - x0) + Fy(P0)( y - y0) + Fz(P0)(z - z0) = 0
P0 d
ƒ(r(t)) = ∇ƒ(r(t)) # r′(t).
f (x, y, z) = c
dt for the plane tangent to the level surface at P0 therefore reduces to
Since ƒ is constant along the curve r, the derivative on the left-hand side of the equation is
ƒx(x0 , y0)(x - x0) + ƒy(x0 , y0)( y - y0) - (z - z0) = 0.
0, so the gradient ∇ƒ is orthogonal to the curve’s velocity vector r′.
Figure 14.33 The gradient ∇ƒ is
Now let us restrict our attention to the curves that pass through P0 (Figure 14.33). All
orthogonal to the velocity vector of every
the velocity vectors at P0 are orthogonal to ∇ƒ at P0 , so the curves’ tangent lines all lie in
smooth curve in the surface through P0.
the plane through P0 normal to ∇ƒ. We now define this plane.
The velocity vectors at P0 therefore lie in a Plane Tangent to a Surface z = ƒ(x, y) at (x0 , y0 , ƒ(x0 , y0))
common plane, which we call the tangent
The plane tangent to the surface z = ƒ(x, y) of a differentiable function ƒ at the
plane at P0 . Definitions The tangent plane at the point P0(x0 , y0 , z0) on the level surface point P0(x0 , y0 , z0) = (x0 , y0 , ƒ(x0 , y0)) is
ƒ(x, y, z) = c of a differentiable function ƒ is the plane through P0 normal to
∇ƒ 0 P0 . ƒx(x0 , y0)(x - x0) + ƒy(x0 , y0)( y - y0) - (z - z0) = 0. (3)
The normal line of the surface at P0 is the line through P0 parallel to ∇ƒ 0 P0 .

From Section 12.5, the tangent plane and normal line have the following equations: example 2 Find the plane tangent to the surface z = x cos y - yex at (0, 0, 0).
z
The plane
x + z− 4 = 0 Solution We calculate the partial derivatives of ƒ(x, y) = x cos y - yex and use
Tangent Plane to ƒ(x, y, z) = c at P0(x0 , y0 , z0) g(x, y, z) Equation (3):

ƒx(P0)(x - x0) + ƒy(P0)(y - y0) + ƒz(P0)(z - z0) = 0 (1) ƒx(0, 0) = (cos y - yex)(0,0) = 1 - 0 # 1 = 1
ƒy(0, 0) = (-x sin y - ex)(0,0) = 0 - 1 = -1.
Normal Line to ƒ(x, y, z) = c at P0(x0 , y0 , z0)
z x = x0 + ƒx(P0)t, y = y0 + ƒy(P0)t, z = z0 + ƒz(P0)t (2) The tangent plane is therefore
The surface 1 # (x - 0) - 1 # (y - 0) - (z - 0) = 0, Eq. (3)
x2 + y2 + z − 9 = 0
P0(1, 2, 4)
∇g or
example 1 Find the tangent plane and normal line of the level surface
The ellipse E x - y - z = 0.
ƒ(x, y, z) = x2 + y2 + z - 9 = 0 A circular paraboloid
Normal line (1, 1, 3)
at the point P0(1, 2, 4). ∇f × ∇g ∇f

Solution The surface is shown in Figure 14.34. example 3 The surfaces

Tangent plane The tangent plane is the plane through P0 perpendicular to the gradient of ƒ at P0 . The
gradient is ƒ(x, y, z) = x2 + y2 - 2 = 0 A cylinder
1 2
∇ƒ 0 P0 = (2xi + 2yj + k)(1,2,4) = 2i + 4j + k. y
and
y The tangent plane is therefore the plane The cylinder
x
x x2 + y2 − 2 = 0
2(x - 1) + 4( y - 2) + (z - 4) = 0, or 2x + 4y + z = 14. g(x, y, z) = x + z - 4 = 0 A plane
f(x, y, z)
Figure 14.34 The tangent plane and
normal line to this level surface at P0
The line normal to the surface at P0 is Figure 14.35 This cylinder and plane meet in an ellipse E (Figure 14.35). Find parametric equations for the line tangent to E at
(Example 1). x = 1 + 2t, y = 2 + 4t, z = 4 + t. intersect in an ellipse E (Example 3). the point P0(1, 1, 3).