3.5.

TANGENT PLANES AND LINEAR APPROXIMATIONS 149

3.5 Tangent Planes and Linear Approximations
In the same way that tangent lines played an important role for functions of one
variables, tangent planes play an important role for functions of two variables.
We rst learn how to derive them. We will then explain why they are important.
3.5.1 Tangent Planes
Goal: given a surface . = ) (r, j) where ) has continuous rst partial deriva-
tives, and a point 1 (r
0
, j
0
, .
0
) on the surface, we wish to nd the equation of
the plane tangent to the surface at the given point. First, we need to ask our-
selves how this plane can be dened. For this, we let C
1
be the curve obtained
by intersecting the surface and the plane j = j
0
. Similarly, let C
2
be the curve
obtained by intersecting the surface and the plane r = r
0
. Note that these two
curves intersect at 1. Let T
1
be the line tangent to C
1
at 1. Let T
2
be the line
tangent to C
2
at 1. Then, the slope of T
1
is )
x
(r
0
, j
0
) and the slope of T
2
is
)
y
(r
0
, j
0
). This is illustrated in gure 3.5.1. C
1
is the black curve, C
2
is the
red curve. The plane tangent to the surface at the point where the two curves
intersect is also shown.
2 2
0
2
0
0
-2
x y
z
-2
-2
-4
4
-4 -4
4 4
Tangent plane to . =
1
4

r
2
+j
2

+ 4 at the point

1, 2,
11
4

Denition 250 The plane tangent to the surface at 1 is the plane containing
T
1
and T
2
.
150 CHAPTER 3. FUNCTIONS OF SEVERAL VARIABLES
Remark 251 At this point, it is important to understand that the technique
to nd tangent planes discussed in this section only applies to surfaces given
explicitly, that is with an equation of the form . = ) (r, j). Tangent planes of
level surfaces, that is surfaces given by an equation of the form 1 (r, j, .) = C
will be discussed in the next section.
Now, we want to derive the equation of this plane. The general form of this
equation is
(r r
0
) +1(j j
0
) +C (. .
0
) = 0 (3.7)
Dividing each side by C gives us
. .
0
= a (r r
0
) +/ (j j
0
) (3.8)
where
a =

C
and
/ =
1
C
The intersection of this plane with the plane j = j
0
should be T
1
. To compute
it, we simply set j = j
0
in equation 3.8. This gives us

. .
0
= a (r r
0
)
j = j
0
We should recognize the equation of a line (in point-slope form) in the r.-plane
with slope a. But the slope of T
1
is )
x
(r
0
, j
0
). Thus, a = )
x
(r
0
, j
0
). Similarly,
we can see that / = )
y
(r
0
, j
0
). Therefore, we have:
Theorem 252 Suppose that a surface is given by . = ) (r, j) where ) has
continuous rst partials. Let 1 = (r
0
, j
0
, .
0
) be a point on the surface. An
equation of the tangent plane to the surface . = ) (r, j) at 1 is
. .
0
= )
x
(r
0
, j
0
) (r r
0
) +)
y
(r
0
, j
0
) (j j
0
) (3.9)
Remark 253 We can also derive this result using a slightly dierent approach.
Recall that the equation of the plane through (r
0
, j
0
, .
0
) with normal ha, /, ci
is a (r r
0
) + / (j j
0
) + c (. .
0
) = 0. We already have our point on the
tangent plane, it is (r
0
, j
0
, .
0
). To nd a normal to the tangent plane, we
nd two vectors on the tangent plane and take their cross product. Using the
notation above, A vector on the tangent plane will be a vector parallel to T
1
.
Since the slope of T
1
is )
x
(r
0
, j
0
) and T
1
is in the r.-plane, a vector parallel
to T
1
is h1, 0, )
x
(r
0
, j
0
)i. Similarly, a vector on the tangent plane, parallel to
T
2
will be h0, 1, )
y
(r
0
, j
0
)i. Hence, a vector normal to the tangent plane is
h1, 0, )
x
(r
0
, j
0
)i h0, 1, )
y
(r
0
, j
0
)i = h)
x
(r
0
j
0
) , )
y
(r
0
, j
0
) , 1i. Hence, the
equation of the tangent plane is
)
x
(r
0
j
0
) (r r
0
) )
y
(r
0
, j
0
) (r r
0
) +. .
0
= 0
or
. .
0
= )
x
(r
0
, j
0
) (r r
0
) +)
y
(r
0
, j
0
) (j j
0
)
3.5. TANGENT PLANES AND LINEAR APPROXIMATIONS 151
-4
-2
-10
y
-4
x
0
-2 0
0
4
10
20
z
40
30
50
2
2
4
Figure 3.11: Graph of . = r
2
+j
2
and its tangent plane at (.5, .5, .5)
Example 254 Find the equation of the plane tangent to . = ) (r, j) = r
2
+j
2
at (.5, .5, .5).
First, we compute the rst order partials.
)
x
(r, j) = 2r
so
)
x
(.5, .5) = 1
Also
)
y
(r, j) = 2j
so
)
y
(.5, .5) = 1
Therefore, the equation of the tangent plane is
. .5 = r .5 +j .5
or
. = r +j .5
A picture of ) (r, j) = r
2
+j
2
and its tangent plane are shown in gure 3.11.
Example 255 Find the equation of the plane tangent to . = ) (r, j) = c
x
2
y
2
at (1, 1, 1).
First, we compute the rst partials.
)
x
(r, j) = 2rc
x
2
y
2
152 CHAPTER 3. FUNCTIONS OF SEVERAL VARIABLES
Figure 3.12: Graph of c
x
2
y
2
and its tangent plane at (1, 1, 1)
So
)
x
(1, 1) = 2
Also
)
y
(r, j) = 2jc
x
2
y
2
So
)
y
(1, 1) = 2
It follows that the equation of the plane is
. 1 = 2 (r 1) + 2 (j + 1)
. = 2r + 2j + 1
The graph of ) and the tangent plane appear in gure 3.12
3.5. TANGENT PLANES AND LINEAR APPROXIMATIONS 153
3.5.2 Linear Approximations
You will recall that one of the interpretations of the tangent line is that it
approximated a curve at the point of the tangent. More precisely, you remember
that given a dierentiable curve, if one zooms in close enough at one point, the
curve appears to be at (like a line). The slope of that line is precisely the
derivative of the curve at the given point. A similar analogy exists for tangent
planes. Using the notation of the previous section, if we zoom in close enough
to 1, the surface . = ) (r, j) will appear to be at. It will be like the tangent
plane at 1. This means that as long as we are close enough to 1, the function
. = ) (r, j) can be approximated by the tangent plane whose equation is given
in 3.9.
More precisely, we know from 3.9 that the equation of the tangent plane to
. = ) (r, j) at (a, /, ) (a, /)) is
. = ) (a, /) +)
x
(a, /) (r a) +)
y
(a, /) (j /) (3.10)
Denition 256 The linear function dened by
1(r, j) = ) (a, /) +)
x
(a, /) (r a) +)
y
(a, /) (j /) (3.11)
is called the linearization of ) at (a, /).
Denition 257 The approximation
) (r, j) 1(r, j)
is called the linear approximation of ) at (a, /).
Remark 258 The approximation ) (r, j) 1(r, j) is only valid for points
(r, j) close to (a, /). For the linearization to be useful, we should be able to
compute ) (a, /), )
x
(a, /) and )
y
(a, /) easily.
Example 259 Using example 255, approximate ) (1.02, .9). Compare it with
the exact answer.
) (1.02, .9) = 2 (1.02) + 2 (.9) + 1
= 1.24 (note we can do the computation by hand)
The exact value is
c
1:02
2
(:9)
2
= 1.259
Example 260 Find the linearization of ) (r, j) = sinr + cos j at (, ) and
use it to approximate ) (3.1, 2.9).
We use the approximation ) (3.1, 2.9) 1(3.1, 2.9) where
1(r, j) = ) (, ) +)
x
(, ) (r ) +)
y
(, ) (j )
We begin by computing rst order partials.
)
x
(r, j) = cos r
154 CHAPTER 3. FUNCTIONS OF SEVERAL VARIABLES
so
)
x
(, ) = cos
= 1
)
y
(r, j) = sinj
so
)
y
(, ) = sin
= 0
Also
) (, ) = sin + cos
= 1
Hence
1(r, j) = 1 (r )
First, looking at gure , we can verify that at the point (, ), ) (r, j) 1(r, j).
2
4
y
2
2
x
z
0
0
-1
1
0
-2
4
. = sinr + cos j and its tangent plane at (, )
Now,
1(3.1, 2.9) = 1 (3.1 )
= 0.958 41
Note that a calculator gives sin(3.1) + cos 2.9 = 0.929 38.
3.5. TANGENT PLANES AND LINEAR APPROXIMATIONS 155
3.5.3 Dierentiability
Loosely speaking, being dierentiable means that locally (at each point), a
function can be linearized. In other words, its graphs near a point, if we zoom
in close enough, will look like a plane. This means that the function can be
approximated by its tangent plane. This denition can be made much more
precise, but we will not do it here. Instead, we give a theorem which provides a
sucient condition for dierentiability.
Theorem 261 If the partial derivatives )
x
and )
y
exist near a point (a, /) and
are continuous at (a, /), then ) is dierentiable at (a, /).
Example 262 Show that ) (r, j) = rc
xy
is dierentiable at (1, 0).
We need to compute )
x
and )
y
and show that they exist near (1, 0) and are
continuous at (1, 0).
)
x
(r, j) =
0r
0r
c
xy
+r
0c
xy
0r
= c
xy
+rc
xy
0 (rj)
0r
= c
xy
+rjc
xy
This is continuous at (1, 0), in fact it is continuous everywhere (why?). It is
also dened at (1, 0), )
x
(1, 0) = 1.
)
y
(r, j) = r
0c
xy
0j
= rc
xy
0 (rj)
0j
= r
2
c
xy
This is also continuous at (1, 0), in fact it is continuous everywhere (why?). It
is also dened at (1, 0), )
y
(1, 0) = 1. Since both partial derivatives satisfy the
condition, it follows that ) is dierentiable at (1, 0).
3.5.4 The Dierential
Given j = ) (r), the dierential of ), denoted dj or d) was dened to be
dj = )
0
(r) dr
The equivalent for functions of several variables is:
Denition 263 Let . = ) (r, j). the dierential of ), denoted d. or d) is
dened to be
d. = )
x
(r, j) dr +)
y
(r, j) dj
=
0.
0r
dr +
0.
0j
dj (3.12)
The dierential tells us how small changes in r and j produce changes in ..
156 CHAPTER 3. FUNCTIONS OF SEVERAL VARIABLES
Example 264 Find the dierential of the function . = ) (r, j) = r
2
+rj
d. =
0.
0r
dr +
0.
0j
dj
= (2r +j) dr +rdj
We have similar formulas for functions of three or more variables. For ex-
ample, if n = ) (r, j, .), then
dn = )
x
(r, j, .) dr +)
y
(r, j, .) dj +)
z
(r, j, .) d.
=
0n
0r
dr +
0n
0j
dj +
0n
0.
d.
3.5.5 Assignment
1. Find an equation of the plane tangent to . = 4r
2
j
2
+ 2j at the point
(1, 2, 4).
2. Find an equation of the plane tangent to . = j cos (r j) at the point
(2, 2, 2).
3. Find an equation of the plane tangent to . = ln

r
2
+j
2

at the point
(1, 0, 0).
4. Graph both . = r
2
+ rj + 3j
2
and its tangent plane at (1, 1, 5). Zoom
in close enough until the surface and the tangent plane become indistin-
guishable.
5. Explain why ) (r, j) = r
p
j is dierentiable at the point (1, 4), then nd
the linearization 1(r, j) at the given point.
6. Explain why ) (r, j) = tan
1
(r + 2j) is dierentiable at the point (1, 0),
then nd the linearization 1(r, j) at the given point.
7. Find the linear approximation of ) (r, j) =
p
20 r
2
7j
2
at (2, 1) and
use it to approximate ) (1.95, 1.08).
8. Find the dierential of . = r
3
lnj
2
9. Find the dierential of 1 = c,
2
cos