SECTION 14.

1 FUNCTIONS OF SEVERAL VARIABLES |||| 865

It’s very difficult to visualize a function f of three variables by its graph, since that
would lie in a four-dimensional space. However, we do gain some insight into f by exam-
ining its , which are the surfaces with equations f s x, y, zd ­ k, where k is
level surfaces

a constant. If the point s x, y, zd moves along a level surface, the value of f s x, y, zd remains
fixed.

z ≈+¥+z@=9 EXAMPLE 15 Find the level surfaces of the function

≈+¥+z@=4
f s x, y, zd ­ x 2 1 y 2 1 z 2

SOLUTIONThe level surfaces are x 2 1 y 2 1 z 2 ­ k, where k ù 0. These form a family of

concentric spheres with radius sk . (See Figure 20.) Thus, as s x, y, zd varies over any
sphere with center O, the value of f s x, y, zd remains fixed. M

y
Functions of any number of variables can be considered. A function of n variables
x is a rule that assigns a number z ­ f s x 1, x 2 , . . . , x n d to an n-tuple s x 1, x 2 , . . . , x n d of real
numbers. We denote by R n the set of all such n-tuples. For example, if a company uses n
≈+¥+z@=1 different ingredients in making a food product, ci is the cost per unit of the i th ingredient,
and x i units of the i th ingredient are used, then the total cost C of the ingredients is a func-
FIGURE 20 tion of the n variables x 1, x 2 , . . . , x n :
3 C ­ f s x 1, x 2 , . . . , x n d ­ c1 x 1 1 c2 x 2 1 ? ? ? 1 cn x n
The function f is a real-valued function whose domain is a subset of R n. Sometimes we
will use vector notation to write such functions more compactly: If ­ k x 1, x 2 , . . . , x n l ,
x

we often write f s d in place of f s x 1, x 2 , . . . , x n d. With this notation we can rewrite the

x

function defined in Equation 3 as

f s xd ­ c ? x
where ­ k c1, c2 , . . . , cn l and ? denotes the dot product of the vectors and in Vn .
c c x c x

In view of the one-to-one correspondence between points s x 1, x 2 , . . . , x nd in R n and

their position vectors ­ k x 1, x 2 , . . . , x n l in Vn , we have three ways of looking at a func-
x

tion f defined on a subset of R n :

1. As a function of n real variables x 1, x 2 , . . . , x n
2. As a function of a single point variable s x 1, x 2 , . . . , x n d
3. As a function of a single vector variable ­ k x 1, x 2 , . . . , x n l
x

We will see that all three points of view are useful.

14.1 EXERCISES

1. In Example 2 we considered the function W ­ f sT, vd, where
W is the wind-chill index, T is the actual temperature, and v is
the wind speed. A numerical representation is given in Table 1.
(a) What is the value of f s215, 40d? What is its meaning?
(b) Describe in words the meaning of the question “For what
value of v is f s220, vd ­ 230 ?” Then answer the question.
(c) Describe in words the meaning of the question “For what
value of T is f sT, 20d ­ 249 ?” Then answer the question.
(d) What is the meaning of the function W ­ f s25, vd?
Describe the behavior of this function.
(e) What is the meaning of the function W ­ f sT, 50d?
Describe the behavior of this function.
866 |||| CHAPTER 14 PARTIAL DERIVATIVES

2. The temperature-humidity index I (or humidex, for short) is the TA B L E 4

perceived air temperature when the actual temperature is T and Duration (hours)
the relative humidity is h, so we can write I ­ f sT, hd. The fol-
lowing table of values of I is an excerpt from a table compiled √
t 5 10 15 20 30 40 50
by the National Oceanic & Atmospheric Administration.
10 2 2 2 2 2 2 2
Apparent temperature as a function 15 4 4 5 5 5 5 5

Wind speed (knots)

TA B L E 3
of temperature and humidity
20 5 7 8 8 9 9 9
Relative humidity (%)
h
30 9 13 16 17 18 19 19
T
20 30 40 50 60 70
40 14 21 25 28 31 33 33
Actual temperature (°F)

80 77 78 79 81 82 83
50 19 29 36 40 45 48 50
85 82 84 86 88 90 93
60 24 37 47 54 62 67 69
90 87 90 93 96 100 106
95 93 96 101 107 114 124 6. Let f s x, yd ­ lns x 1 y 2 1d.
100 99 104 110 120 132 144 (a) Evaluate f s1, 1d. (b) Evaluate f se, 1d.
(c) Find and sketch the domain of f .
(d) Find the range of f .
(a) What is the value of f s95, 70d? What is its meaning? 7. Let f s x, yd ­ x 2e3 xy.
(b) For what value of h is f s90, hd ­ 100? (a) Evaluate f s2, 0d. (b) Find the domain of f .
(c) For what value of T is f sT, 50d ­ 88? (c) Find the range of f .
(d) What are the meanings of the functions I ­ f s80, hd
and I ­ f s100, hd? Compare the behavior of these two 8. Find and sketch the domain of the function
functions of h. f s x, yd ­ s1 1 x 2 y 2 . What is the range of f ?
3. Verify for the Cobb-Douglas production function 9. Let f s x, y, zd ­ esz2x 22y 2 .
(a) Evaluate f s2, 21, 6d. (b) Find the domain of f .
PsL, K d ­ 1.01L 0.75K 0.25 (c) Find the range of f .
discussed in Example 3 that the production will be doubled 10. Let ts x, y, zd ­ lns25 2 x 2 2 y 2 2 z 2 d.
if both the amount of labor and the amount of capital are (a) Evaluate ts2, 22, 4d. (b) Find the domain of t.
doubled. Determine whether this is also true for the general (c) Find the range of t.
production function
11–20 Find and sketch the domain of the function.
PsL, K d ­ bLaK 12a
11. f s x, yd ­ sx 1 y
4. The wind-chill index W discussed in Example 2 has been 12. f s x, yd ­ sxy
modeled by the following function:
13. f s x, yd ­ lns9 2 x 2 2 9 y 2 d
WsT, vd ­ 13.12 1 0.6215T 2 11.37v 0.16 1 0.3965Tv 0.16
14. f s x, yd ­ sy 2 x lns y 1 xd
Check to see how closely this model agrees with the values in 15. f s x, yd ­ s1 2 x 2 2 s1 2 y 2
Table 1 for a few values of T and v.
16. f s x, yd ­ sy 1 s25 2 x 2 2 y 2
5. The wave heights h in the open sea depend on the speed v
of the wind and the length of time t that the wind has been sy 2 x 2
17. f s x, yd ­
blowing at that speed. Values of the function h ­ f sv, td are 1 2 x2
recorded in feet in Table 4.
(a) What is the value of f s40, 15d? What is its meaning? 18. f s x, yd ­ arcsins x 2 1 y 2 2 2d
(b) What is the meaning of the function h ­ f s30, td? Describe 19. f s x, y, zd ­ s1 2 x 2 2 y 2 2 z 2
the behavior of this function.
(c) What is the meaning of the function h ­ f sv, 30d? Describe 20. f s x, y, zd ­ lns16 2 4 x 2 2 4 y 2 2 z 2 d
the behavior of this function.
 SECTION 14.1 FUNCTIONS OF SEVERAL VARIABLES |||| 867

21–29 Sketch the graph of the function. 32. Two contour maps are shown. One is for a function f whose
21. f s x, yd ­ 3 22. f s x, yd ­ y
graph is a cone. The other is for a function t whose graph is a
paraboloid. Which is which, and why?
23. f s x, yd ­ 10 2 4 x 2 5y 24. f s x, yd ­ cos x
I y II y
25. f s x, yd ­ y 1 1
2
26. f s x, yd ­ 3 2 x 2 2 y 2
27. f s x, yd ­ 4 x 2 1 y 2 1 1
28. f s x, yd ­ s16 2 x 2 2 16y 2
x x
29. f s x, yd ­ s x 2 1 y 2

30. Match the function with its graph (labeled I–VI).Give reasons
for your choices.
(a) f s x, yd ­ | x | 1 | y | (b) f s x, yd ­ | xy | 33. Locate the points A and B in the map of Lonesome Mountain
(c) f s x, yd ­
1 (Figure 12). How would you describe the terrain near A?
(d) f s x, yd ­ s x 2 2 y 2 d2 Near B?
1 1 x2 1 y2
(e) f s x, yd ­ s x 2 yd2 (f) f s x, yd ­ sin(| x | 1 | y |)
34. Make a rough sketch of a contour map for the function whose
I z II z graph is shown.
z

x y x y

III z IV z

y
y x
x
x y
35–38 A contour map of a function is shown. Use it to make a
V z VI z rough sketch of the graph of f .
35. y 36. y
14
13 _8
12
11
_6

y x _4
x x y

31. A contour map for a function f is shown. Use it to estimate the x

values of f s23, 3d and f s3, 22d. What can you say about the 8
shape of the graph?
y y
37. y 38.
5
4
3 _3
_2
_1
2
0
1
2
1
70 60 50 40 3 x
1 0 0 4 5
3
2
0 1 30 x 1
0
20
x
10
868 |||| CHAPTER 14 PARTIAL DERIVATIVES

39– 46 Draw a contour map of the function showing several level 61–64 Describe the level surfaces of the function.
curves. 61. f s x, y, zd ­ x 1 3y 1 5z
39. f s x, yd ­ s y 2 2 xd2 40. f s x, yd ­ x 3 2 y
62. f s x, y, zd ­ x 2 1 3y 2 1 5z 2
41. f s x, yd ­ y 2 ln x 42. f s x, yd ­ e yyx
63. f s x, y, zd ­ x 2 2 y 2 1 z 2
43. f s x, yd ­ ye x 44. f s x, yd ­ y sec x
64. f s x, y, zd ­ x 2 2 y 2
45. f s x, yd ­ sy 2 2 x 2 46. f s x, yd ­ yys x 2 1 y 2 d

65–66 Describe how the graph of t is obtained from the graph

47– 48 Sketch both a contour map and a graph of the function of f .
and compare them.
65. (a) ts x, yd ­ f s x, yd 1 2 (b) ts x, yd ­ 2 f s x, yd
47. f s x, yd ­ x 2 1 9 y 2 (c) ts x, yd ­ 2f s x, yd (d) ts x, yd ­ 2 2 f s x, yd
48. f s x, yd ­ s36 2 9x 2 2 4y 2 66. (a) ts x, yd ­ f s x 2 2, yd (b) ts x, yd ­ f s x, y 1 2d
(c) ts x, yd ­ f s x 1 3, y 2 4d
49. A thin metal plate, located in the xy-plane, has temperature
Ts x, yd at the point s x, yd. The level curves of T are called
isothermals because at all points on an isothermal the temper- ; 67–68 Use a computer to graph the function using various
ature is the same. Sketch some isothermals if the temperature domains and viewpoints. Get a printout that gives a good view of
function is given by the “peaks and valleys.” Would you say the function has a maxi-
mum value? Can you identify any points on the graph that you
Ts x, yd ­ 100ys1 1 x 2 1 2y 2 d might consider to be “local maximum points”? What about “local
minimum points”?
50. If Vs x, yd is the electric potential at a point s x, yd in the
67. f s x, yd ­ 3x 2 x 4 2 4y 2 2 10 xy
xy-plane, then the level curves of V are called equipotential
curves because at all points on such a curve the electric 68. f s x, yd ­ xye2x 2y
2 2

potential is the same. Sketch some equipotential curves if

Vs x, yd ­ cysr 2 2 x 2 2 y 2 , where c is a positive constant.

; 51–54 Use a computer to graph the function using various ; 69–70 Use a computer to graph the function using various
domains and viewpoints. Get a printout of one that, in your opin- domains and viewpoints. Comment on the limiting behavior of
ion, gives a good view. If your software also produces level the function. What happens as both x and y become large? What
curves, then plot some contour lines of the same function and happens as s x, yd approaches the origin?
compare with the graph. x1y xy
69. f s x, yd ­ 70. f s x, yd ­
51. f s x, yd ­ e2x 1 e22 y
2 2
x2 1 y2 x2 1 y2

f s x, yd ­ s1 2 3 x 2 1 y 2 de12x 2y
2 2
52.
; 71. Use a computer to investigate the family of functions
53. f s x, yd ­ xy 2 2 x 3 (monkey saddle) f s x, yd ­ e cx 1y . How does the shape of the graph depend
2 2

on c?
54. f s x, yd ­ xy 3 2 yx 3 (dog saddle)
; 72. Use a computer to investigate the family of surfaces
55–60 Match the function (a) with its graph (labeled A–F on 2
2y 2
z ­ sax 2 1 by 2 de 2x
page 869) and (b) with its contour map (labeled I–VI). Give
reasons for your choices. How does the shape of the graph depend on the numbers a
55. z ­ sins xyd 56. z ­ e x cos y and b?
57. z ­ sins x 2 yd 58. z ­ sin x 2 sin y ; 73. Use a computer to investigate the family of surfaces
x2y z ­ x2 1 y 2 1 cxy . In particular, you should determine the
59. z ­ s1 2 x 2 ds1 2 y 2 d 60. z ­ transitional values of c for which the surface changes from
1 1 x2 1 y2
one type of quadric surface to another.
 SECTION 14.1 FUNCTIONS OF SEVERAL VARIABLES |||| 869

Graphs and Contour Maps for Exercises 55–60

A z B z C z

y
x y
x y
x

D z E z F z

x
y
x y x
y

I y II y III y

x x x

IV y V y VI y

x x x
870 |||| CHAPTER 14 PARTIAL DERIVATIVES

; 74. Graph the functions ; 75. (a) Show that, by taking logarithms, the general Cobb-
Douglas function P ­ bLaK 12a can be expressed as
f s x, yd ­ sx 2 1 y2 f s x, yd ­ e sx 21y 2
P L
ln ­ ln b 1 a ln
f s x, yd ­ ln sx 2 1 y2 f s x, yd ­ sin( sx 2 1 y2 ) K K

1 (b) If we let x ­ lnsL yK d and y ­ lnsPyK d, the equation in

and f s x, yd ­ part (a) becomes the linear equation y ­ a x 1 ln b. Use
sx 2 1 y2
Table 2 (in Example 3) to make a table of values of
In general, if t is a function of one variable, how is the graph lnsLyK d and lnsPyK d for the years 1899–1922. Then use a
of graphing calculator or computer to find the least squares
f s x, yd ­ t (s x 2 1 y 2 ) regression line through the points slnsLyK d, lnsPyK dd.
(c) Deduce that the Cobb-Douglas production function is
obtained from the graph of t? P ­ 1.01L0.75K 0.25.

14.2 LIMITS AND CONTINUITY

Let’s compare the behavior of the functions

sins x 2 1 y 2 d x2 2 y2
f s x, yd ­ and ts x, yd ­
x2 1 y2 x2 1 y2
as x and y both approach 0 [and therefore the point s x, yd approaches the origin].
TA B L E 1 Values of f s x, yd TA B L E 2 Values of ts x, yd
y 21.0 20.5 20.2 0 0.2 0.5 1.0 x
y 21.0 20.5 20.2 0 0.2 0.5 1.0
x
21.0 0.455 0.759 0.829 0.841 0.829 0.759 0.455 21.0 0.000 0.600 0.923 1.000 0.923 0.600 0.000
20.5 0.759 0.959 0.986 0.990 0.986 0.959 0.759 20.5 20.600 0.000 0.724 1.000 0.724 0.000 20.600

20.2 0.829 0.986 0.999 1.000 0.999 0.986 0.829 20.2 20.923 20.724 0.000 1.000 0.000 20.724 20.923

0 0.841 0.990 1.000 1.000 0.990 0.841 0 21.000 21.000 21.000 21.000 21.000 21.000

0.2 0.829 0.986 0.999 1.000 0.999 0.986 0.829 0.2 20.923 20.724 0.000 1.000 0.000 20.724 20.923

0.5 0.759 0.959 0.986 0.990 0.986 0.959 0.759 0.5 20.600 0.000 0.724 1.000 0.724 0.000 20.600

1.0 0.455 0.759 0.829 0.841 0.829 0.759 0.455 1.0 0.000 0.600 0.923 1.000 0.923 0.600 0.000

Tables 1 and 2 show values of f s x, yd and ts x, yd, correct to three decimal places, for
points s x, yd near the origin. (Notice that neither function is defined at the origin.) It
appears that as s x, yd approaches (0, 0), the values of f s x, yd are approaching 1 whereas the
values of ts x, yd aren’t approaching any number. It turns out that these guesses based on
numerical evidence are correct, and we write
sins x 2 1 y 2 d x2 2 y2
lim 1 and lim does not exist
x 1y
2 2
x2 1 y2
­
s x, yd l s0, 0d s x, yd l s0, 0d

In general, we use the notation

lim
s x, yd l s a, bd
f s x, yd ­ L