C H A P TE R 3 0

Applications of Integration II: Volume

A solid of revolution is obtained by revolving a region in a plane about a line that does not intersect the region. The line about which the rotation takes place is called the axis of revolution. Let f be a continuous function such that f (x) 0 for a x b. Consider the region under the graph of f, above the x axis, and between x = a and x = b. (See Fig. 30-1.) If is revolved about the x axis, the resulting solid is a solid of revolution. The generating regions for some familiar solids are shown in Fig. 30-2.

Fig. 30-1

Disk Formula
The volume V of the solid of revolution obtained by revolving the region of Fig. 30-1 about the x axis is given by
V = ( f ( x))2 dx = y 2dx
a a b b

(disk formula)

Fig. 30-2

244
   ! "#%$ & ' '()%"* +,- * .' / 0121 )+304+,.

CHAPTER 30 Applications of Integration II: Volume

245

See P roblem 9 for a sketch of the proof of this formula. Similarly, when the axis of rotation is the y axis and the region that is revolved lies between the y axis and a curve x = g( y) and between y = c and y = d (see Fig. 30-3), then the volume V of the resulting solid of revolution is given by the formula
V = (g( y))2 dy = x 2dy
c a d b

(disk formula)

Fig. 30-3 E X A M P L E 3 0 .1 : Consider the solid of revolution obtained by revolving about the x axis the region in the first q uadrant bounded by the parabola y2 = 8x and the line x = 2. (See Fig. 30-4.) B y the disk formula, the volume is
2 2 V = y 2 dx = 8 x dx = (4 x 2 ) 0 = (16 0) = 16 0 0 2

Fig. 30-4

E X A M P L E 3 0 .2 : Consider the solid of revolution obtained by revolving about the y axis the region bounded by the parabola y = 4x2 and the lines x = 0 and y = 16. (See Fig. 30-5.) To find its volume, we use the version of the disk formula in which we integrate along the y axis. Thus,

V = x 2 dy =
0

16

16

y dy = y 2 = (256 0) = 32 4 8 0 8
16

246

CHAPTER 30 Applications of Integration II: Volume

Fig. 30-5

W ash e r M e th od
Assume that 0 g(x) f (x) for a x b. Consider the region between x = a and x = b and lying between y = g(x) and y = f (x). (See Fig. 30-6.) Then the volume V of the solid of revolution obtained by revolving this region about the x axis is given by the formula
V = [ ( f ( x))2 (g( x))2 ] dx
b a

(washer formula)

Fig. 30-6

The justification is clear. The desired volume is the difference of two volumes, the volumes ( f ( x))2 dx a of the solid of revolution generated by revolving about the x axis the region under y = f (x) and the volume b (g( x))2 dx of the solid of revolution generated by revolving about the x axis the region under y = g(x).
a

A similar formula
V = [ ( f ( y))2 (g( y))2 ] dy
d c

(washer formula)

holds when the region lies between the two curves x = f ( y) and x = g( y) and between y = c and y = d, and it is revolved about the y axis. (It is assumed that 0 g( y) f ( y) for c y d.)

The word washer is used because each thin vertical strip of the region being revolved produces a solid that resembles a plumbing part called a washer (a small cylindrical disk with a hole in the middle).

CHAPTER 30 Applications of Integration II: Volume

247

Fig. 30-7 E X A M P L E 3 0 .3 : Consider the solid of revolution obtained by revolving about the x axis the region bounded by the curves, y = 4x2, x = 0, and y = 16. (The same region as in Fig. 30-5.) H ere the upper curve is y = 16 and the lower curve is y = 4x2. H ence, by the washer formula,
5 V = [ 162 (4 x 2 )2 ] dx = [ 256 16 x 4 ] dx = ( 256 x 16 5 x ) 0 = 512 2 2 2 0 0

512 2048 = 5 5

C y lin d ric al S h e ll M e th od
Consider the solid of revolution obtained by revolving about the y axis the region in the first q uadrant between the x axis and the curve y = f (x), and lying between x = a and x = b. (See Fig. 30-7.) Then the volume of the solid is given by
V = 2 xf ( x) dx = 2 xy dx
a a b b

(cylindrical shell formula)

See P roblem 10 for the justification of this formula. A similar formula holds when the roles of x and y are reversed, that is, the region in the first q uadrant between the y axis and the curve x = f ( y), and lying between y = c and y = d, is revolved about the x axis
V = 2 yf ( y) dy = 2 yx dy
c c d d

E X A M P L E 3 0 .4 : R evolve about the y axis the region above the x axis and below y = 2x2, and between x = 0 and x = 5. B y the cylindrical shell formula, the resulting solid has volume

2 xy dx = 2 x(2 x 2 ) dx = 4 x 3dx = ( x 4 ) ]0 = 625

5 5 5 5 0 0 0

N ote that the volume could also have been computed by the washer formula, but the calculation would have been somewhat more complicated.

Diffe re n c e of S h e lls Formula

Assume that 0 g(x) f (x) on an interval [ a, b] with a 0. Let be the region in the first q uadrant between the curves y = f (x) and y = g(x) and between x = a and x = b. Then the volume of the solid of revolution obtained by revolving about the y axis is given by
V = 2 x( f ( x) g( x)) dx
a b

(difference of shells formula)

248

CHAPTER 30 Applications of Integration II: Volume

This obviously follows from the cylindrical shells formula because the req uired volume is the difference of two volumes obtained by the cylindrical shells formula. N ote that a similar formula holds when the roles of x and y are reversed.
E X A M P L E 3 0 .5 : Consider the region in the first q uadrant bounded above by y = x2, below by y = x3, and lying between x = 0 and x = 1. W hen revolved about the y axis, this region generates a solid of revolution whose volume, according to the difference of shells formula, is

2 x( x 2 x 3 ) dx = 2 ( x 3 x 4 ) dx = 2
0 0

1 1 1 1 x x ) = 2 ( ) = (4 5 4 5 10
1 4 5 0

C ross-S e c tion Formula (S lic in g Formula)

Assume that a solid lies entirely between the plane perpendicular to the x axis at x = a and the plane perpendicular to the x axis at x = b. For each x such that a x b, assume that the plane perpendicular to the x axis at that value of x intersects the solid in a region of area A(x). (See Fig. 30-8.) Then the volume V of the solid is given by
V = A( x) dx
a b

(cross-section formula)

For justification, see P roblem 11.

Fig. 30-8 E X A M P L E 3 0 .6 : Assume that half of a salami of length h is such that a cross-section perpendicular to the axis of the salami at a distance x from the end O is a circle of radius x . (See Fig. 30-9.) H ence, the area A(x) of the cross-section is ( x )2 = x. So, the cross-section formula yields

V = A( x) dx = x dx =
0 0

2 h2 x = 2 2 0
h

Fig. 30-9

This formula is also called the slicing formula because each cross-sectional area A(x) is obtained by slicing through the solid.

CHAPTER 30 Applications of Integration II: Volume

249

S O LV E D P R O B L E M S
1. Find the volume of a cone that has height h and whose base has radius r. The cone is generated by revolving about the x axis the region between the line y = r x and the x axis, h between x = 0 and x = h. (See Fig. 30-2(a).) B y the disk formula, the volume of the cone is
2 2 h h 2 x 3 ) = r2 ( 1 h3 = 1 r 2h y 2 dx = r 2 x 2 dx = r2 ( 1 0 h 3 ) 3 0 0 h h 3 h

2.

Find the volume of the cylinder of height h and radius r. The cylinder is generated by revolving about the x axis the region between the line y = r and the x axis, between x = 0 and x = h. (See Fig. 30-2(b).) B y the disk formula, the volume of the cylinder is
h h 2 V = y 2 dx = r 2 dx = r 2 x 0 = r h. 0 0 h

3.

Find the volume of a sphere of radius r. The sphere is generated by revolving about the x axis the region between the semicircle y = r 2 x 2 and the x axis, between x = r and x = r. (See Fig. 30-2(c).) B y the symmetry with respect to the y axis, we can use the part of the given region between x = 0 and x = r and then double the result. H ence, by the disk formula, the volume of the sphere is
3 r r r 4 3 3 3 V = 2 y 2 dx = 2 (r 2 x 2 ) dx = 2 ( r 2 x 1 = 2 r 3 r = 2 ( 2 3 r ) = 3 r 3 x ) 0 0 0 3

4.

Let be the region between the x axis, the curve y = x3, and the line x = 2. (See Fig. 30-10.) (a) Find the volume of the solid obtained by revolving about the x axis. (b) Find the volume of the solid obtained by revolving about the y axis.

Fig. 30-10

(a) The disk formula yields the volume

2 2 2 128 V = y 2 dx = ( x 3 )2 dx = x 6 dx = x 7 0 = 7 0 0 0 7 2

25 0

CHAPTER 30 Applications of Integration II: Volume

(b) (First solution) The cylindrical shells formula yields the volume
2 2 2 64 5 V = 2 xy dx = 2 x( x 3 ) dx = 2 x 4 dx = 2 ( 1 5 x ) = 0 0 0 5 0 2

(Second solution) Integrating along the y axis and using the washer formula yields the volume
8 2 V= 2 0

3 ( y) dy = [ 4 y ] dy = ( 4 y 5 y
3 2 8
2 3

32 = 64 ) = (32 ( 3 5 5) )
8 0

5.

Find the volume of the solid obtained by revolving about the y axis the region in the first q uadrant inside the circle x2 + y2 = r2, and between y = a and y = r (where 0 < a < r). See Fig. 30-11. (The solid is a polar cap of a sphere of radius r.)

Fig. 30-11

Integrating along the y axis, the disk formula yields the volume
3 V = x 2 dy = (r 2 y 2 ) dy = ( r 2 y 1 3 y ) a = a a r r r

2 3

(2r 3 3r 2 a + a 3 ) 3 r 3 ( r 2a 1 3a ) = 3

6.

Find the volume of the solid obtained by revolving about the y axis the region in the first q uadrant bounded above by the parabola y = 2 x2 and below by the parabola y = x2. (See Fig. 30-12.)

Fig. 30-12

CHAPTER 30 Applications of Integration II: Volume

25 1

The curves intersect at (1,1). B y the difference of cylindrical shells formula, the volume is
1 1 1 1 4 2 V = 2 x((2 x 2 ) x 2 ) dx = 4 ( x x 3 ) dx = 4 ( 1 2 x 4 x ) = 4 ( 2 0 0 0 1 1 4

)=

7.

Consider the region bounded by the parabola y = 4x2 and the lines x = 0 and y = 16. (See Fig. 30-5.) Find the volume of the solid obtained by revolving about the line y = 2. To solve this problem, we reduce it to the case of a revolution about the x axis. R aise the region vertically upward through a distance of 2 units. This changes into a region * that is bounded below by the parabola y = 4x2 + 2, on the left by the y axis, and above by the line y = 18. (See Fig. 30-13.) Then the original solid of revolution has the same volume as the solid of revolution obtained by revolving R* around the x axis. The latter volume is obtained by the washer formula:
V = (182 (4 x 2 + 2)2 ) dx = (256 16 x 4 16 x 2 4) dx
0 0 16 3 5 = ( 252 x 16 5 x 3 x ) 0 = ( 504 2 512 5 2 2

5384 128 3 )= 15

8.

As in P roblem 7, consider the region bounded by the parabola y = 4x2 and the lines x = 0 and y = 16. (See Fig. 30-5.) Find the volume of the solid obtained by revolving about the line x = 1.

Fig. 30-13

To solve this problem, we reduce it to the case of a revolution about the y axis. M ove the region to the right through a distance of 1 unit. This changes into a region * that is bounded on the right by the parabola y = 4(x 1)2, above by y = 16, and on the left by x = 1. (See Fig. 30-14.) The desired volume is the same as that obtained when we revolve * about the y axis. The latter volume is got by the difference of cylindrical shells formula:
V = 2 x(16 4( x 1)2 ) dx = 2 x(16 4 x 2 + 8 x 4) dx
1 1 2 3 = 2 (16 x 4 x 3 + 8 x 2 4 x) dx = 2 (8 x 2 x 4 + 8 3 x 2 x ) 1 3 1 8 = 2 (72 81 + 72 18) (8 1 + 3 2 ) = 112 3 3 3 3

25 2

CHAPTER 30 Applications of Integration II: Volume

Fig. 30-14

9 . J ustify the disk formula: V = ( f ( x))2 dx.

a

D ivide the interval [ a, b] into n eq ual subintervals, each of length x = b a . (See Fig. 30-15.) n Consider the volume Vi obtained by revolving the region i above the ith subinterval about the x axis.

If mi and Mi are the absolute minimum and absolute maximum of f on the ith subinterval, then Vi lies between the volume of a cylinder of radius mi and height x and the volume of a cylinder of radius Mi V and height x. Thus, mi2 x Vi M i2 x and, therefore, mi2 i M i2 . (W e have assumed that the x volume of a cylinder of radius r and height h is r2h.) H ence, by the intermediate value theorem for the V 2 continuous function ( f (x))2, there exists xi* in the ith subinterval such that i = ( f ( xi* ) ) and, therefore, x 2 Vi = ( f ( xi* ) ) x. Thus,
V = Vi = ( f ( xi* ) ) x
2 i =1 i =1 n n

Letting n + , we obtain the disk formula.

Fig. 30-15

CHAPTER 30 Applications of Integration II: Volume

10 . J ustify the cylindrical shells formula: V = 2 xf ( x) dx.
a b

25 3

D ivide [ a, b] into n eq ual subintervals, each of length x. (See Fig. 30-16.) Let i be the region above the x +x ith subinterval. Let xi* be the midpoint i 1 i of the ith interval. The solid obtained by revolving the region 2 t about the y axis is approximately the solid obtained by revolving the rectangle with base x and height yi* = f ( xi* ). The latter solid is a cylindrical shell, that is, it lies between the cylinders obtained by revolving the rectangles with the same height f ( xi* ) and with bases [0, xi1] and [0, xi]. H ence, it has volume

xi2 f ( xi* ) xi21 f ( xi* ) = f ( xi* )( xi2 xi21 )

= f ( xi* )( xi xi 1 )( xi + xi 1 ) = f ( xi* )(2 xi* )( x) = 2 xi* f ( xi* )( x)
* * Thus, the total V is approximated by 2 x1 f ( x1 ) x which approaches 2 x f ( x)dx as n + . i =1 a n b

11. J ustify the cross-section formula: V = A( x) dx.

a

D ivide [ a, b] into n eq ual subintervals [ xi1, xi], and choose a point xi* in [ xi1, xi]. If n is large, x is small and the piece of the solid between xi1 and xi, will be close to a (noncircular) disk of thickness x and base area

A( xi* ). (See Fig. 30-17.) This disk has volume A( xi* ) x. So V is approximated by A( xi* ) x, which approaches
b a

A( x) dx as n + .

i =1

Fig. 30-16

Fig. 30-17

25 4

CHAPTER 30 Applications of Integration II: Volume

12 . A solid has a circular base of radius 4 units. Find the volume of the solid if every plane section perpendicular to a particular fixed diameter is an eq uilateral triangle. Take the circle as in Fig. 30-18, with the fixed diameter on the x axis. The eq uation of the circle is x2 + y2 = 16. The cross-section ABC of the solid is an eq uilateral triangle of side 2y and area A(x) = 3 y2 = 3 (16 x2). Then, by the cross-section formula,
3 V = 3 (16 x 2 ) dx = 3 (16 x 1 3 x ) 4 = 4 4 4 256 3

13 . A solid has a base in the form of an ellipse with major axis 10 and minor axis 8. Find its volume if every section perpendicular to the major axis is an isosceles triangle with altitude 6. 2 2 Take the ellipse as in Fig. 30-19, with eq uation x + y = 1. The section ABC is an isosceles triangle of base 25 16 2y, altitude 6, and area A( x) = 6 y = 6 4 25 x 2 . H ence, 5

V=

24 5

25 x 2 dx = 60

Fig. 30-18

(N ote that

25 x 2 dx is the area of the upper half of the circle x2 + y2 = 25 and, therefore, is eq ual to 25/2.)

Fig. 30-19

S U P P LE M E N TA R Y P R O B LE M S
14 . Consider the region bounded by the parabola y2 = 8x and the line x = 2. (See Fig. 30-4.) (a) Find the volume of the solid generated by revolving about the y axis. (b) Find the volume of the solid generated by revolving about the line x = 2. (a) 128 ; (b) 256 5 15

Ans.

CHAPTER 30 Applications of Integration II: Volume

25 5

15 . Find the volume of the solid generated by revolving the region between the x axis and the parabola y = 4x x2 about the line y = 6. Ans.
1408 15

16 . Find the volume of the torus (doughnut) generated by revolving the circle (x a)2 + y2 = b2 about the y axis, where 0 < b < a. Ans. 22ab2

17 . Consider the region bounded by y = x2 3x + 6 and x + y = 3. Find the volume of the solid generated by revolving about: (a) the x axis; (b) the line x = 3. Ans. (a) 1792 ; (b) 256 15 3

In P roblems 18 26, find the volume generated when the given region is revolved about the given line. U se the disk formula. 18 . The region bounded by y = 2x2, y = 0, x = 0, x = 5, about the x axis. ns. 25 0 0 19 . The region bounded by x2 y2 = 16, y = 0, x = 8, about the x axis. Ans.
256 3

2 0 . The region bounded by y = 4x2, x = 0, y = 16, about y = 16. (See Fig. 30-5.) Ans.
4096 15

2 1. The region bounded by y2 = x3, y = 0, x = 2, about the x axis. ns. 4 2 2 . The region bounded by y = x3, y = 0, x = 2, about x = 2. Ans.
16 5

2 3 . The region within the curve y2 = x4(l x2), about the x axis. Ans.
4 35

2 4 . The region within the ellipse 4x2 + 9y2 = 36, about the x axis. Ans. 16

2 5 . The region within the ellipse 4x2 + 9y2 = 36, about the y axis. ns. 24

25 6

CHAPTER 30 Applications of Integration II: Volume

2 6 . The region within the parabola x = 9 y2 and between y = x 7 and the y axis, about the y axis. Ans.
963 5

In P roblems 27 32, find the volume of the solid generated by revolving the given region about the given line. U se the washer formula. 2 7 . The region bounded by y = 2x2, y = 0, x = 0, x = 5, about the y axis. ns. 6 25 2 8 . The region bounded by x2 y2 = 16, y = 0, x = 8, about the y axis. Ans. 1 28 3

2 9 . The region bounded by y = x3, x = 0, y = 8, about x = 2. Ans.

144 5

3 0 . The region bounded by y = x2, y = 4x x2, about the x axis. Ans.

32 3

3 1. The region bounded by y = x2, y = 4x x2, about y = 6. Ans.

64 3

3 2 . The region bounded by x = 9 y2, y = x 7, about x = 4. ns.

153 5

In P roblems 33 37, find the volume of the solid generated by revolving the given region about the given line. U se the cylindrical shells formula. 3 3 . The region bounded by y = 2x2, y = 0, x = 0, x = 5, about x = 6. ns. 3 7 5 3 4 . The region bounded by y = x3, y = 0, x = 2, about y = 8. ns.
320 7

3 5 . The region bounded by y = x2, y = 4x x2, about x = 5. Ans.

64 3

3 6 . The region bounded by y = x2 5x + 6 and y = 0, about the y axis. Ans.

5 6

CHAPTER 30 Applications of Integration II: Volume

25 7

3 7 . The region bounded by x = 9 y2, y = x 7, x = 0, about y = 3. Ans.

369 2

In P roblems 38 42, find the volume generated by revolving the given region about the given line. U se any appropriate method. 3 8 . The region bounded by y = e x , y = 0, x = 0, x = 1, about the y axis.
2

ns. (1 e1 ) 3 9 . The region bounded by y = 2x2, y = 2x + 4, about x = 2. ns. 27 4 0 . The region bounded by y = 2x, y = 0, x = 0, x = 1, about the y axis. ns.
4 3

4 1. The region bounded by y = x2, x = y2, about the x axis. ns.

3 10

4 2 . The region bounded by xy = 4, y = (x 3)2, about the x axis. ns.

27 5

4 3 . Find the volume of the frustum of a cone whose lower base is of radius R, upper base is of radius r, and altitude is h. Ans.
1 h(r 2 + rR + R 2 ) 3

4 4 . A solid has a circular base of radius 4 units. Find the volume of the solid if every plane perpendicular to a fixed diameter (the x axis of Fig. 30-18) is: (a) a semicircle; (b) a sq uare; (c) an isosceles right triangle with the hypotenuse in the plane of the base. Ans. (a) 128 ; (b) 1024 ; (c) 256 3 3 3

4 5 . A solid has a base in the form of an ellipse with major axis 10 and minor axis 8. Find its volume if every section perpendicular to the major axis is an isosceles right triangle with one leg in the plane of the base. Ans.
640 3

4 6 . The base of a solid is the first-q uadrant region bounded by the line 4x + 5y = 20 and the coordinate axes. Find its volume if every plane section perpendicular to the x axis is a semicircle. Ans.
10 3

4 7 . The base of a solid is the circle x2 + y2 = 16x, and every plane section perpendicular to the x axis is a rectangle whose height is twice the distance of the plane of the section from the origin. Find its volume. ns. 1 0 24

25 8

CHAPTER 30 Applications of Integration II: Volume

4 8 . The section of a certain solid cut by any plane perpendicular to the x axis is a circle with the ends of a diameter lying on the parabolas y2 = 4x and x2 = 4y. Find its volume. Ans.
6561 280

4 9 . The section of a certain solid cut by any plane perpendicular to the x axis is a sq uare with the ends of a diagonal lying on the parabolas y2 = 4x and x2 = 4y. Find its volume. Ans.
144 35

5 0 . A hole of radius 1 unit is bored through a sphere of radius 3 units, the axis of the hole being a diameter of the sphere. Find the volume of the remaining part of the sphere. Ans.
64 2 3

C H A P TE R 3 1

Techniques of Integration I: Integration by P arts

If u and v are functions, the product rule yields
Dx (uv) = uv + vu

which can be rewritten in terms of antiderivatives as follows:

uv = uv dx + vu dx

N ow, uv dx can be written as u dv, and vu dx can be written as v du. Thus, uv = u dv + v du and, therefore,

u dv = uv v du
The purpose of integration by parts is to replace a difficult integration v du.
E X A M P L E 3 1 .1 :

(integration by parts)

u dv

by an easy integration

Find

x ln x dx .

In order to use the integration by parts formula, we must divide the integrand x ln x dx into two parts u and dv so that we can easily find v by an integration and also easily find v du. In this example, let u = ln x and dv = x dx. Then 1 dx . So, the integration by parts formula yields: 2 we can set v = 1 2 x and note that du = x

x ln x dx = u dv = uv v du = (ln x)(
2 =1 2 x ln x 1 2

1 2

2 1 x2 ) 1 dx 2 x x

( )

x dx =

1 2

2 x 2 ln x 1 4 x +C

2 =1 4 x (2 ln x 1) + C

Integration by parts can be made easier to apply by setting up a rectangle such as the following one for E xample 1.
u = ln x dv = x dx 1 2 du = dx v = 1 2 x x

uv dx = u dv, where, after the integration on the right, the variable v is replaced by the corresponding function of x. In fact, by the Chain R ule, Dx u dv = Dv u dv Dx v = u v . H ence, u dv = uv dx. Similarly, v du = vu dx .

56789 : ;< =>? @@ABCAAABCAA@BCAD?E 8F<GHI!J9 KL%M N: O O(56P%7KQ: GR,BS Q IT5O : IU<G9 GV1692=1G9 PR36VWR,GT

25 9

26 0

CHAPTER 31

Techniques of Integration I: Integration by P arts

In the first row, we place u and dv. In the second row, we place the results of computing du and v. The desired result of the integration parts formula uv v du can be obtained by first multiplying the upper-left corner u by the lower-right corner v, and then subtracting the integral of the product v du of the two entries v and du in the second row.
E X A M P L E 3 1 .2 :

Find xe x dx.

Let u = x and dv = ex dx. W e can picture this in the box below.

u = x dv = e x dx du = dx v = e x

Then,

xe dx = uv v du = xe
x

e x dx = xe x e x + C

= e x ( x 1) + C
E X A M P L E 3 1 .3 :

Find e x cos x dx.

Let u = ex and dv = cos x dx. Then we get the box

u = ex dv = cos x dx du = e x dx v = sin x

So,

cos x dx = uv v du = e x sin x e x sin x dx

(1)

H owever, let us try to find e x sin x dx by another integration by parts. This time, let u = ex and dv = sin x dx.
u = ex dv = sin x dx du = e x dx v = cos x

N ow we have the problem of finding e x sin x dx, which seems to be just as hard as the original integral e x cos x dx.

Then,

sin x dx = e x cos x e x cos x dx = e x cos x + e x cos x dx

Substituting in formula (1) above, we get:

cos x dx = e x sin x e x cos x + e x cos x dx = e x sin x + e x cos x e x cos x dx

Adding e x cos x dx to both sides yields 2 e x cos x dx = e x sin x + e x cos x. So,

e
W e must add an arbitrary constant:

x x cos x dx = 1 2 (e sin x + e cos x )

x x cos x dx = 1 2 (e sin x + e cos x ) + C

N otice that this example req uired an iterated application of integration by parts.

CHAPTER 31

Techniques of Integration I: Integration by P arts

26 1

S O LV E D P R O B L E M S
1. Find x 3e x dx. 2 Let u = x2 and dv = xe x dx. N ote that v can be evaluated by using the substitution w = x2. (W e get
2

v=

1 2

1 x w dw = 1 2 e = 2 e .)
2

u = x2 dv = xe x dx x2 du = 2 x dx v= 1 2e
2

H ence,

x e

3 x2

2 x dx = 1 xe x dx 2 x e
2 2

2 x x =1 1 +C 2e 2 x e
2 2

x 2 =1 2 e ( x 1) + C
2

2.

Find ln( x 2 + 2) dx. Let u = ln (x2 + 2) and dv = dx.

u = ln( x 2 + 2) dv = dx du = 22 x dx v = x x +2

So,

ln( x

+ 2) dx = x ln( x 2 + 2) 2

x 2 dx x2 + 2

= x ln( x 2 + 2) 2 1

2 dx x2 + 2

= x ln( x 2 + 2) 2 x + 4 tan 1 x + C 2 2 = x (ln( x 2 + 2) 2) + 2 2 tan 1 x + C 2

3.

Find ln x dx. Let u = ln x and dv = dx.

u = ln x dv = dx du = 1 dx v = x x

So,

ln x dx = x ln x 1 dx = x ln x x + C
= x (ln x 1) + C

4.

Find x sin x dx. W e have three choices: (a) u = x sin x, dv = dx; (b) u = sin x, dv = x dx; (c) u = x, dv = sin x dx. (a) Let u = x sin x, dv = dx. Then du = (sin x + x cos x) dx, v = x, and

x sin x dx = x x sin x x(sin x + x cos x) dx

The resulting integral is not as simple as the original, and this choice is discarded.

26 2

CHAPTER 31

Techniques of Integration I: Integration by P arts

2 (b) Let u = sin x, dv = x dx. Then du = cos x dx, v = 1 2 x , and

x sin x dx =

1 2

2 x 2 sin x 1 2 x cos x dx

The resulting integral is not as simple as the original, and this choice too is discarded. (c) Let u = x, dv = sin x dx. Then du = dx, v = cos x, and

x sin x dx = x cos x cos x dx = x cos x + sin x + C

5. Find x 2 ln x dx. 3 Let u = ln x, dv = x2 dx. Then du = dx , v = x , and x 3 x3 x 3 dx x 3 x 2 ln x dx = 3 ln x 3 x = 3 ln x

1 3

x dx =
2

x 3 ln x 1 x 3 + C 9 3

6.

Find sin 1 x dx. Let u = sin1 x, dv = dx.

u = sin 1 x dv = dx 1 du = dx v = x 1 x2

So,

sin

x dx = x sin 1 x = x sin 1 x +
1 2

x dx 1 x2

(1 x

2 1/ 2

(2 x )dx

2 1/ 2 = x sin 1 x + 1 2 (2(1 x ) ) + C

(by Quick Formula I)

= x sin 1 x + (1 x 2 )1/ 2 + C = x sin 1 x + 1 x 2 + C

7.

Find tan 1 x dx. Let u = tan1x, dv = dx.

u = tan 1 x dv = dx 1 du = dx v = x 1 + x2

So,

tan

x dx = x tan 1 x

x dx = x tan 1 x 1 + x2

1 2

1+ x

2 x dx 2

2 = x tan 1 x 1 2 ln(1 + x ) + C

(by Quick Formula II)

8.

Find sec3 x dx. Let u = sec x, dv = sec2x dx.

u = sec x dv = sec 2 x dx du = sec x tan x dx v = tan x

CHAPTER 31

Techniques of Integration I: Integration by P arts

26 3

Thus,

sec

x dx = sec x tan x sec x tan 2 x dx = sec x tan x sec x(sec 2 x 1) dx = sec x tan x sec3 x dx + sec x dx = sec x tan x sec3 x dx + ln |sec x + tan x |

Then, H ence, 9.

2 sec3 x dx = sec x tan x + ln |sec x + tan x |

sec

x dx = 1 2 (sec x tan x + ln |sec x + tan x |) + C

Find x 2 sin x dx. Let u = x2, dv = sin x dx. Thus, du = 2x dx and v = cos x. Then

x
N ow apply integration by parts to

sin x dx = x 2 cos x 2 x cos x dx = x 2 cos x + 2 x cos x dx

x cos x dx, with u = x and dv = cos x dx, getting

x cos x = x sin x sin x dx = x sin x + cos x

H ence,

sin x dx = x 2 cos x + 2( x sin x + cos x) + C

10 . Find x 3e2 x dx . 2x Let u = x3, dv = e2x dx. Then du = 3x2 dx, v = 1 2 e , and

x e

3 2x

3 2x dx = 1 2 x e

3 2

x e

2 2x

dx

2x For the resulting integral, let u = x2 and dv = e2x dx. Then du = 2x dx, v = 1 2 e , and

x e

3 2x

3 2x dx = 1 2 x e

3 2

(
3 2

1 2

3 2 2x 2 2x 4 xe + x 2 e2 x xe2 x dx = 1 2 x e

3 2

xe

2x

dx

2x For the resulting integral, let u = x and dv = e2x dx. Then du = dx, v = 1 2 e , and

x e

3 2x

3 2 2x 3 2x dx = 1 xe + 4 2 x e

1 2

xe2 x

1 2

2x

3 2x 3 3 2 2x 3 2x xe +4 xe2 x 8 e +C dx = 1 4 2 x e

11. D erive the following reduction formula for sin m x dx.

sin

x dx = sin

m 1

x cos x + m 1 sin m 2 x dx m m

Let u = sinm1 x and dv = sin x dx.

u = sin m 1 x dv = sin x dx du = (m 1)sin m 2 x dx v = cos x

26 4

CHAPTER 31

Techniques of Integration I: Integration by P arts

Then

sin

x dx = cos x sin m 1 x + ( m 1) sin m 2 x cos2 x dx = cos x sin m 1 x + ( m 1) sin m 2 x(1 sin 2 x) dx = cos x sin m 1 x + (m 1) sin m 2 x dx (m 1) sin m x dx

H ence,

m sin m x dx = cos x sin m 1 x + (m 1) sin m 2 x dx

and division by m yields the req uired formula. 12 . Apply the reduction formula of P roblem 11 to find sin 2 x dx. W hen m = 2, we get

sin

x dx = = =

sin x cos x + 2 sin x cos x + 2

1 2

sin

x dx

1 2

1 dx

sin x cos x x x sin x cos x + +C= +C 2 2 2

13 . Apply the reduction formula of P roblem 11 to find sin 3 x dx. W hen m = 3, we get

sin

x dx = =

sin 2 x cos x + 3

2 3

sin x dx

sin 2 x cos x 2 3 cos x + C 3

= cos x (2 + sin 2 x) + C 3

S U P P LE M E N TA R Y P R O B LE M S

In P roblems 14 21, use integration by parts to verify the specified formulas.

14 .

x cos x dx = x sin x + cos x + C x sec cos

2 1 3x dx = 1 3 x tan 3 x 9 ln |sec x | + C

15 .

16 .

2 x dx = x cos 1 2 x

1 2

1 4x2 + C

17 .

x tan x e x
3

1 2 1 x dx = 1 2 ( x + 1) tan x 2 x + C

18 .

2 3 x

2 3 x (x2 + 2 dx = 1 3 x + 9) + C 3e

19 .

sin x dx = x 3 cos x + 3x 2 sin x + 6 x cos x 6 sin x + C

CHAPTER 31

Techniques of Integration I: Integration by P arts

26 5

20.

x sin

2 2 1 ( x 2 ) dx = 1 2 x sin ( x ) +

1 2

1 x4 + C

2 1.

ln x dx = ln x + 1 + C x x2
2

2 2 . Show that x sin nx dx = 2 for any positive integer n. 0 n

n2 2 3 . P rove the following reduction formula: sec n x dx = tan x sec x + n 2 sec n 2 x dx. n 1 n 1

2 4 . Apply P roblem 23 to find sec 4 x dx. Ans.

1 3

tan x(sec 2 x + 2) + C

2 5 . P rove the reduction formula:

(a
2 6 . Apply P roblem 25 to find Ans.

x2 dx = 1 2 x 2 n1 + 2 dx 2 n1 2n 2 (a + x ) + x 2 )n (a + x )

1 x + 1 tan 1 x + C 2 a2 + x 2 a a x n+1 [(n + 1)ln x 1)] + C for n 1. (n + 1)2

x2 dx. (a 2 + x 2 ) 2

2 7 . P rove x n ln x dx =

2 8 . P rove the reduction formula: x n eax dx = 1 x n eax n x n1eax dx. a a 2 9 . U se P roblem 28 and E xample 2 to show that: x 2 e x dx = e x ( x 2 2 x + 2) + C .

C H A P TE R 3 2

Techniques of Integration II: Trigonometric Integrands and Trigonometric Substitutions

T rig on ome tric In te g ran d s
1. Let us consider integrals of the form sin k x cosn x dx , where k and n are nonnegative integers. T yp e 1 . At least one of sin x and cos x occurs to an odd p ower: Then a substitution for the other function works.
E X A M P L E 3 2 .1 :

Let u = cos x. Then du = sin x dx. H ence,

sin

x cos 2 x dx .

sin

x cos 2 x dx = sin 2 x cos 2 x sin x dx = (1 cos 2 x)cos 2 x sin x dx = (1 u 2 )u 2 du = (u 4 u 2 ) du

1 1 3 1 3 5 5 =1 5 u 3 u + C = 5 cos x 3 cos x + C

E X A M P L E 3 2 .2 :

Let u = sin x. Then du = cos x dx, and

sin

x cos 7 x dx .

sin

x cos 7 x dx = sin 4 x cos6 x cos x dx = u 4 (1 u 2 )3 du = u 4 (1 3u 2 + 3u 4 u 6 ) du = (u 5 3u 6 + 3u 8 u10 ) du

3 7 1 9 6 =1 6u 7u + 3u 1 11

u11 + C
1 11

3 1 6 7 9 =1 6 sin x 7 sin x + 3 sin x

sin11 x + C

E X A M P L E 3 2 .3 :

Let u = cos x. Then du = sin x dx and

sin

x dx .

26 6 56789 : ;< =>? @@ABCAAABCAA@BCAD?E 8F<GHI!J9 KL%M N: O O(56P%7KQ: GR,BS Q IT5O : IU<G9 GV1692=1G9 PR36VWR,GT

CHAPTER 32

Techniques of Integration II

26 7

sin

x dx = sin 4 x sin x dx = (1 cos 2 x)2 sin x dx = (1 u 2 )2 du = (1 2u 2 + u 4 ) du

1 5 3 = (u 2 3u + 5u )+C 2 3 5 =1 5 cos x + 3 cos x cos x + C

T yp e 2. Both p owers of sin x and cos x are even: This always involves a more tedious computation, using the identities
cos 2 x =
E X A M P L E 3 2 .4 :

1 + cos 2 x 2

and sin 2 x =

1 cos 2 x 2

cos

x sin 4 x dx = (cos 2 x)(sin 2 x)2 dx = 1 + cos 2 x 1 cos 2 x 2 2

( )( ) dx = ( 1 + cos 2 x ) ( 1 2 cos 2 x + cos 2 x ) dx 2 4

2 2

= 1 (1(1 2 cos 2 x + cos 2 2 x) + (cos 2 x)(1 2 cos 2 x + cos 2 2 x)) dx 8 = 1 (1 2 cos 2 x + cos 2 2 x + cos 2 x 2 cos 2 2 x + cos3 2 x) dx 8 = 1 (1 cos 2 x cos 2 2 x + cos3 2 x) dx 8 =1 8

( 1 dx cos 2x dx cos 2x dx + cos 2x dx )

2 3

= 1 x sin 2 x 1 + cos 4 x dx + (cos 2 x)(1 sin 2 2 x) dx 8 2 2 = 1 x sin 2 x 1 x + sin 4 x + cos 2 x dx 1 u 2 du 2 8 2 2 4

3

)
[ letting u = sin 2 x]

( ) ( ) = 1 ( x sin 2 x x sin 4 x + sin 2 x 1 sin 2 x ) + C 2 2 3 8 2 2 8 = 1 ( x sin 4 x sin 2 x ) + C 8 2 8 6

3 3 = x sin 4 x sin 2 x + C 16 64 48

2 . Let us consider integrals of the form tan k x sec n x dx, where k and n are nonnegative integers. R ecall that sec2 x = 1 + tan2 x. T yp e 1. n is even: Substitute u = tan x.
E X A M P L E 3 2 .5 :

tan

x sec 4 x dx

Let u = tan x, du = sec2 x dx. So,

tan

x sec 4 x dx = tan 2 x(1 + tan 2 x)sec 2 x dx = u 2 (1 + u 2 ) du

1 1 1 3 5 5 3 = (u 4 + u 2 ) du = 1 5 u + 3 u + C = 5 tan x + 3 tan x + C

26 8

CHAPTER 32

Techniques of Integration II

T yp e 2. n is odd and k is odd:

E X A M P L E 3 2 .6 :

Substitute u = sec x.

tan

x sec x dx

Let u = sec x, du = sec x tan x dx. So,

tan

x sec x dx = tan 2 x sec x tan x dx = (sec 2 x 1)sec x tan x dx

1 3 3 = (u 2 1) du = 1 3 u u + C = 3 sec x sec x + C

T yp e 3. n is odd and k is even:

E X A M P L E 3 2 .7 :

This case usually req uires a tedious calculation.

tan

x sec x dx = (sec 2 x 1)sec x dx = (sec3 x sec x) dx = 1 (sec x tan x + ln |sec x + tan x |) ln |sec x + tan x | + C 2 (by P roblem 8 of Chapter 31)

=1 2 (sec x tan x ln |sec x + tan x |) + C

3 . Let us consider integrals of the form sin Ax cos Bx dx , sin Ax sin Bx dx, and cos Ax cos Bx dx. W e shall need the identities
sin Ax cos Bx = 1 2 (sin( A + B) x + sin( A B) x ) sin Ax sin Bx = 1 2 (cos( A B) x cos( A + B) x ) cos Ax cos Bx = 1 ( A B) x + cos( A + B) x) 2 (cos(
E X A M P L E 3 2 .8 :

sin 7 x cos 3x dx =

1 2

(sin(7 + 3) x + sin(7 3) x) dx = 1 2 (sin 10 x + sin 4 x ) dx

1 1 1 =1 2 ( 10 cos 10 x 4 cos 4 x ) + C = 40 (2 cos 10 x + 5 cos 4 x ) + C

E X A M P L E 3 2 .9 :

sin 7 x sin 3x dx =

1 2

(cos(7 3) x cos(7 + 3) x) dx = 1 2 (cos 4 x cos 10 x ) dx

1 10

1 =1 2 ( 4 sin 4 x

sin 10 x) + C =

1 40

(5 sin 4 x 2 sin 10 x) + C

E X A M P L E 3 2 .1 0 :

cos 7 x cos 3x dx =

1 2

(cos(7 3) x + cos(7 + 3) x) dx = 1 2 (cos 4 x + cos 10 x ) dx

1 10

1 =1 2 ( 4 sin 4 x +

sin 10 x) + C =

1 40

(5 sin 4 x + 2 sin 10 x) + C

T rig on ome tric S ub stitution s

There are three principal kinds of trigonometric substitutions. W e shall introduce each one by means of a typical example.
E X A M P L E 3 2 .1 1 :

Find

dx . 4 + x2 4 + x 2 = 4 + 4 tan 2 = 2 1 + tan 2 = 2 sec 2 = 2 |sec |

Let x = 2 tan q, that is, q = tan1 (x/2). Then

dx = 2 sec 2 d

and

CHAPTER 32

Techniques of Integration II

26 9

B y definition of the inverse tangent, /2 < q < /2. So, cos q > 0 and, therefore, sec q > 0. Thus,
sec = |sec | = 4 + x 2 / 2 . H ence,

2 sec 2 d dx = 4 tan 2 (2 sec ) 4 + x2 sec d 1 cos d 1 = = (sin )2 cos d =1 4 sin 2 4 4 tan 2

1 =1 4 ( (sin ) ) + C =

1 +C 4 sin

N ow we must evaluate sin q. Analytic method:

sin = tan = sec x/2 = 4 + x2 /2 x . 4 + x2

G eometric method: D raw the right triangle shown in Fig. 32-1. From this triangle we see that sin = x / 4 + x 2 . (N ote that it follows also for q < 0.) H ence,

4 + x2 dx = +C 4x 4 + x2

Fig. 32 -1

This example illustrates the following general rule:

S tr a te g y I . If

a 2 + x 2 occurs in an integrand, try the substitution x = a tan q.

E X A M P L E 3 2 .1 2 :

dx . 9 x2 Let x = 3 sin q, that is, q = sin1 (x/3). Then dx = 3 cos q dq and

Find

9 x 2 = 9 9 sin 2 = 3 sin 2 = 3 cos 2 = 3 |cos |

B y definition of the inverse sine, /2 < q < /2 and, therefore, cos q > 0. Thus, cos = |cos | = 9 x 2 / 3 . N ow,

3 cos d dx = =1 2 9 x 2 9 sin (3 cos ) 9

csc

d
9 x2 /3 +C= 1 x /3 9 9 x2 +C x

= 1 cot + C = 1 cos + C = 1 9 9 sin 9

This example illustrates the following general method:

S tr a te g y I I .
E X A M P L E 3 2 .1 3 :

If

a 2 x 2 occurs in an integrand, try the substitution x = a sin q.

Find

x2 dx . x2 4

27 0

CHAPTER 32

Techniques of Integration II

Let x = 2 sec q, that is, q = sec1 (x/2). Then dx = 2 sec q dq and

x 2 4 = 4 sec 2 4 = 2 sec 2 1 = 2 tan 2 = 2 |tan |

B y definition of the inverse secant, q is in the first or third q uadrant and, therefore, tan q > 0. So, tan = |tan | = x 2 4 / 2 . N ow,

4 sec 2 (2 sec tan ) x2 dx = d 2 2 tan x 4 f Chapter 31) = 4 sec3 d = 2(sec tan + ln |sec + tan | + C ) ( by P roblem 8 of = 2 x 2 = = x2 4 + ln x + 2 2 x2 4 +C 2

x + x2 4 x x2 4 + 2 ln +C 2 2 x x2 4 + 2 ln x + x 2 4 + K 2 where K = C 2 ln 2

This example illustrates the following general method:

S tr a te g y I I I .

If

x 2 a 2 occurs in an integrand, try the substitution x = a sec q.

S O LV E D P R O B L E M S
In P roblems 1 23, verify the given solutions. R ecall the identities
sin 2 u = 1 cos 2 u 1 sin 2 x = 2 sin x cos x 2 (1 + cos 2u ) 2 (1 cos 2u )

1.

sin

1 1 1 x dx = 1 2 (1 cos 2 x ) dx = 2 ( x 2 sin 2 x ) + C = 2 ( x sin x cos x ) + C .

2.

cos (3x) dx =
2

1 2

1 (1 + cos 6 x) dx = 1 2 ( x + 6 sin 6 x ) + C .

3.

sin

x dx = sin 2 x sin x dx = (1 cos 2 x)sin x dx = sin x dx + cos 2 x( sin x) dx

3 = cos x + 1 (by Quick k Formula I) 3 cos x + C

4.

sin

x cos3 x dx = sin 2 x cos 2 x cos x dx = sin 2 x(1 sin 2 x)cos x dx = sin 2 x cos x dx sin 4 x cos x dx
1 5 3 =1 (by Quick Formula I) 3 sin x 5 sin x + C

CHAPTER 32

Techniques of Integration II

27 1

5.

sin (3x)cos (3x) dx = (1 cos

3 5

(3x))cos5 (3x)sin (3x) dx

= cos5 (3x)sin(3x) dx cos 7 (3 x)sin 3x dx = 1 cos5 (3x)(3 sin(3x)) dx + 1 cos 7 (3x)(3 sin(3x)) dx 3 3
1 6 =1 3 6 cos (3 x ) + 1 1 3 8

cos8 (3x) + C
6

(by Quick Formula I)

1 72

(3 cos (3x) 4 cos (3x)) + C

8

6.

cos

x dx = 1 sin x cos x dx (3 ) ( ( 3 )) 3 = (1 sin ( x )) cos x dx = cos x dx sin ( x ) cos x dx 3 3 3 3 3 = 3 sin x 3 sin ( x ) ( 1 cos x ) dx 3 3 3 3 = 3 sin x 3 sin ( x ) + C (by Quick Formula I) 3 3 = 3 sin x sin ( x ) + C 3 3
2 2 2 2 1 3 3 3

7.

sin

x dx = (sin 2 x)2 dx = 1 (1 cos(2 x))2 dx 4 = 1 1 dx 1 cos(2 x) dx + 1 cos 2 (2 x)dx 4 4 2 = 1 x 1 sin(2 x) + 1 (1 + cos 4 x)) dx 4 4 8
1 1 1 =1 4 x 4 sin (2 x ) + 8 ( x + 4 sin (4 x )) + C 3 =8 x 1 4 sin (2 x ) + 1 32

sin (4 x) + C

8.

sin x cos
2

x dx = 1 sin 2 (2 x) dx = 1 (1 cos(4 x)) dx 8 4

1 1 =8 (x 1 4 sin(4 x )) + C = 8 x 1 32

sin(4 x) + C

9.

sin

(3x)cos 2 (3x) dx = (sin 2 (3x)cos 2 (3x))sin 2 (3x) dx = 1 sin 2 (6 x)(1 cos(6 x)) dx 8 = 1 sin 2 (6 x) dx 1 sin 2 (6 x)cos(6 x) dx 8 8 = 1 (1 cos(12 x)) dx 1 sin 2 (6 x)(6 cos(6 x)) dx 48 16 = =
1 16 1 16

(x x

1 12

sin(12 x)) sin(12 x))

1 144 1 144

sin 3 (6 x) + C sin 3 (6 x) + C

( by Quick Formula I)

1 192

10 .

sin 3x sin 2 x dx =
=
1 2

1 2

(cos(3x 2 x) cos(3x + 2 x)) dx

1 2

(cos x cos 5x)) dx =

1 10

(sin x 1 5 sin 5 x ) + C

=1 2 sin x

sin 5x + C

27 2

CHAPTER 32

Techniques of Integration II

11.

sin 3x cos 5x dx =

1 2

(sin(3x 5x) + sin(3x + 5x)) dx

= 1 (sin(2 x) + sin(8 x)) dx = 1 ( sin(2 x) + sin(8 x)) dx 2 2

1 1 1 =1 2 ( 2 cos(2 x ) 8 cos(8 x )) + C = 4 cos(2 x ) 1 16

cos(8 x) + C

12 .

cos 4 x cos 2 x dx = 2 (cos(2 x) + cos(6 x)) dx

1 1 1 =1 2 ( 2 sin(2 x ) + 6 sin(6 x )) + C = 4 sin(2 x ) + 1 12

sin(6 x) + C

13 .

1 cos x dx = 2 sin x dx 2 =

x = 1 cos x ( ) ( by sin ( 2 ) 2 ) 2 ( 2 cos ( x )) + C = 2 2 cos ( x ) + C 2 2

2

14 .

(1 + cos 3x)

3/ 2

dx = 2 2 cos3

3x) ( 32x ) dx since cos ( 32x ) = 1 + cos( 2 3x 3x = 2 2 1 sin ( 2 ) cos ( 2 ) dx 3x 3x 3x = 2 2 cos ( ) dx sin ( ) cos ( ) dx 2 2 2 2 3x 2 3x 3 3x = 2 2 sin ( ) sin ( ) cos ( ) dx 2 3 2 2 2 3 3x 21 3x 2 sin ( ) + C = 2 2 sin ( ) 2 33 2 3 3x 3x 4 2 3 sin ( ) sin ( ) + C = 2 2 9
2 2 2 2 3 3

15 .

dx = 1 sin 2 x

dx 1 cos 2 x 2

)
1 cos 2 x 2 sin 2 nce sin x = 4 2

2 = 2

( ) ( ) 2 2 csc ( x ) dx = ln csc ( x ) cot ( x ) + C 2 4 2 4 4

dx x sin 4

16 .

tan

x dx = tan 2 x tan 2 x dx = tan 2 x(sec 2 x 1) dx = tan 2 x sec 2 x dx tan 2 x dx

3 2 =1 ( by Quick Formula I) 3 tan x (sec x 1)dx 3 =1 3 tan x (tan x x ) + C 3 =1 3 tan x tan x + x + C

CHAPTER 32

Techniques of Integration II

27 3

17 .

tan

x dx = tan 3 x tan 2 x dx = tan 3 x(sec 2 x 1) dx = tan 3 x sec 2 x dx tan 3 dx

4 2 ( by Quick Formula I) =1 4 tan x tan x (sec x 1) dx 4 2 =1 4 tan x tan x sec x dx + tan x dx 1 2 4 =1 4 tan x 2 tan x + ln |sec x | + C

( by Quick Fo ormula I)

18 .

sec

(2 x) dx = sec 2 (2 x)sec 2 (2 x) dx = sec 2 (2 x)(1 + tan 2 (2 x)) dx = sec 2 (2 x) dx + sec 2 (2 x) tan 2 (2 x)) dx =1 2 tan(2 x ) + =1 an(2 x) + 2 ta
1 2

tan

(2 x)(2 sec 2 (2 x)) dx (by Quick Formula I )

1 1 2 3

tan 3 (2 x) + C

1 3 =1 2 tan(2 x ) + 6 tan (2 x ) + C

19 .

tan (3x)sec
3

(3x) dx = tan 3 (3x)(1 + tan 2 (3x))sec 2 (3x) dx = tan 3 (3x)sec 2 (3x) dx + tan 5 (3x)sec 2 (3x) dx = =
11 34 1 12

tan 4 (3x) + tan 4 (3x) +

2

11 36 1 18

tan 6 (3x) + C tan 6 (3x) + C

20.

cot (2 x) dx = cot(2 x)(csc (2 x) 1) dx

3 1 2 =1 4 cot (2 x ) + 2 ln |csc(2 x )| + C

2 1.

cot

(3x) dx = cot 3 (3x)(csc 2 (3x) 1) dx = cot 2 (3x) csc 2 (3x) dx cot 2 (3x) dx
3 2 =1 9 cot (3 x ) (csc (3 x ) 1) dx 1 3 =1 9 cot (3 x ) + 3 cot(3 x ) + x + C

22.

csc

x dx = csc 2 x(1 + cot 2 x)2 dx = csc 2 x dx + 2 cot 2 x csc 2 x dx + cot 4 x csc 2 x dx

1 3 5 = cot x 2 3 cot x 5 cot x + C

23.

cot

x csc5 x dx = cot 2 x csc 4 csc x cot x dx = (csc 2 x 1) csc 4 x csc x cot x dx = csc6 x csc x cot x dx csc 4 x csc x cot x dx
1 5 7 =1 7 csc x + 5 csc x + C

27 4

CHAPTER 32

Techniques of Integration II

2 4 . Find

9 4x2 dx . x
9 4

9 4x2 = 2 dx = 3 2 cos d

x 2 . So, let x = 3 2 sin . Then

and

9 4 x 2 = 9 9 sin 2 = 3 cos 2 = 3 |cos | = 3 cos

H ence,

2 2 3 cos ( 3 9 4x2 2 cos ) d dx = = 3 cos d = 3 1 sin d 3 sin sin x sin 2

= 3 (csc sin ) d = 3 ln|csc c cot | +3 cos + C

B ut
csc = 1 = 3 sin 2 x

and

cot = cos = sin

9 4 x 2 /3 = 2x /3

9 4x2 2x

So,

2 5 . Find

9 4x2 3 9 4x2 dx = 3 ln + 9 4x2 + K x x

where

K = C 3 ln 2

dx . 9 4x2

3 2 Let x = 3 9 4 x 2 = 3 sec . H ence, 2 tan . (See Fig. 32-2.) Then dx = 2 sec and 3 sec 2 d dx = 3 2 2 ( 2 tan )(3 sec ) 9 + 4x

= 1 csc d = 1 3 ln|csc cot | + C 3 =1 3 ln 9 + 4x2 3 +K x

Fig. 32 -2

2 6 . Find

(16 9 x 2 )3/ 2 dx . x6

4 16 9 x 2 = 4 cos . H ence, Let x = 4 3 sin . (See Fig. 32-3.) Then dx = 3 cos d and

(64 cos3 )( 4 (16 9 x 2 )3/ 2 3 cos d ) dx = 6 4096 sin 6 x 729 = 243 cot 4 csc 2 d = 243 cot 5 + C 16 80 = 243 80 (16 9 x 2 )5/ 2 (16 9 x 2 )5/ 2 +C= 1 +C 80 243x 5 x5

CHAPTER 32

Techniques of Integration II

27 5

Fig. 32 -3

2 7 . Find

x 2 dx x 2 dx = . 2 2x x 1 ( x 1)2

Let x 1 = sin q. (See Fig. 32-4.) Then dx = cos q dq and 2 x x 2 = cos . H ence,

x 2 dx = (1 + sin )2 cos d 2 x x 2 cos

1 = (1 + sin )2 d = ( 3 2 + 2 sin 2 cos 2 ) d 1 =3 2 2 cos 4 sin 2 + C 1 2 2 1 =3 2 sin ( x 1) 2 2 x x 2 ( x 1) 2 x x + C 1 2 1 =3 2 sin ( x 1) 2 ( x + 3) 2 x x + C

Fig. 32 -4

2 8 . Find

dx dx = . (4 x 2 24 x + 27)3/ 2 (4( x 3)2 9)3/ 2

3 4 x 2 24 x + 27 = 3 tan . So, Let x 3 = 3 2 sec . (See Fig. 32-5.) Then dx = 2 sec tan d and

(4 x

3 dx 2 sec tan d = 1 csc cot d 3/ 2 = 18 24 x + 27) 27 tan 3

x3 = 1 csc + C = 1 +C 18 9 4 x 2 24 x + 27

(f from Fig. 32-5)

Fig. 32 -5

S U P P LE M E N TA R Y P R O B LE M S
29.

cos sin

1 x dx = 1 2 x + 4 sin 2 x + C

30.

1 3 2 x dx = 1 6 cos 2 x 2 cos 2 x + C

27 6

CHAPTER 32

Techniques of Integration II

3 1.

sin

3 1 sin 4 x + 2 x dx = 8 x 8

1 64

sin 8 x + C

32.

cos sin

4 1 2

3 x dx = 8 x+ 1 2 sin x +

1 16

sin 2 x + C

33.

3 5 3 7 x dx = 1 7 cos x 5 cos x + cos x cos x + C

34.

cos sin sin sin sin

6 1 2

x dx =

5 16

x+ 1 2 sin x +

3 32

sin 2 x

1 24

sin 3 x + C

35.

1 2 5 7 3 x cos5 x dx = 1 3 sin x 5 sin x + 7 sin x + C

36.

1 3 5 x cos 2 x dx = 1 5 cos x 3 cos x + C

37.

x cos3 x dx =

1 48

cos3 2 x

1 16

cos 2 x + C

38.

x cos 4 x dx =

1 128

1 (3x sin 4 x + 8 sin 8 x + C

39.

sin 2 x cos 4 x dx = cos 3x cos 2 x dx = sin 5x sin x dx = 1 sin x sin sin

cos3 x dx
1 8

1 4

cos 2 x

1 12

cos 6 x + C

40.

1 2

sin x +

1 10

sin 5x + C

4 1.

sin 4 x

1 12

sin 6 x + C

42.

2 = sin x + 1 2 sin x + C

43.

cos 2 / 3 x dx = 3 cot 5/ 3 x + C 5 8/3 x cos3 x dx = csc x 1 csc3 x + C 3 4 x

3

44.

45.

x(cos tan tan tan

3

x 2 sin 3 x 2 ) dx =

1 12

(sin x 2 + cos x 2 )(4 + sin 2 x 2 ) + C

46.

2 x dx = 1 2 tan x + ln|cos x | + C

47.

1 3 3x sec 3x dx = 1 9 sec 3 x 3 sec 3 x + C

48.

3/ 2

2 5/ 2 tan 9 / 2 x + C x sec 4 x dx = 2 x+ 9 5 tan

49.

tan

1 5 7 x sec 4 x dx = 1 7 tan x + 5 tan x + C

CHAPTER 32

Techniques of Integration II

27 7

50.

cot cot cot csc

2 x dx = 1 2 cot x ln|sin x | + C

5 1.

1 6 4 x csc 4 x dx = 1 4 cot x 6 cot x + C

52.

1 3 5 x csc3 x dx = 1 5 csc x + 3 csc x + C

53.

1 3 2 x dx = 1 2 cot 2 x 6 cot 2 x + C

54.

1 1 sec x ( tan x ) dx = 3 tan x tan x + C

4 3

55.

cot 3 x dx = sin x csc x + C csc x sec x dx = 2 sec x + C x +C 4 4 x2

56.

tan x
dx

57.

(4 x ) x

2 3/ 2

58.

25 x 2 5 25 x 2 dx = 5 ln + 25 x 2 + C x x
2 2 dx = a 2 x + C 2 2 a x a x

59.

60.

2 x 2 + 4 dx = 1 x2 + 4 ) + C 2 x x + 4 + 2 ln( x +

6 1.

(a

x 2 dx = x 2 )3 / 2

x sin 1 x + C a a2 x2

()

62.

2 x 2 4 dx = 1 x2 4 + C 2 x x 4 2 ln x +

63.

a2 + x2 a x2 + a2 +C dx = x 2 + a 2 + a ln 2 2 x a + x2 + a x 2 dx = x3 +C 12(4 x 2 )3/ 2

64.

(4 x ) (a x
2 2

2 5/ 2

65.

dx x = +C + x 2 )3 / 2 a 2 a 2 + x 2 9 x2 dx = +C 2 9x 9 x

66.

67.

x 2 dx = 1 x x 2 16 + 8 ln x + x 2 16 + C x 2 16 2

27 8

CHAPTER 32

Techniques of Integration II

68.

2 2 2 5/ 2 a (a 2 x 2 )3 / 2 + C a 2 x 2 dx = 1 5 (a x ) 3

69.

dx = ln( x 2 + x 2 4 x + 13 ) + C x 2 4 x + 13 dx = x2 +C 4 4x x2 x tan 1 x + +C 3 18(9 + x 2 )

70.

(4 x x ) (9 + x )
dx
2 2

2 3/ 2

7 1.

1 54

()

In P roblems 72 and 73, first apply integration by parts.

72.

x sin

2 2 1 x dx = 1 x+ 1 4 x 1 x + C 4 (2 x 1)sin

73.

x cos

1 2 2 1 x dx = 1 4 (2 x 1)cos x 4 x 1 x + C

C H A P TE R 3 3

Techniques of Integration III: Integration by P artial Fractions

N ( x) W e shall give a general method for finding antiderivatives of the form dx, where N(x) and D(x) are D( x) N ( x) is called a rational function. (N(x) is the numerator and D(x) is polynomials. A function of the form D( x) the denominator.) As examples, consider

x 1 dx 3 +8

and

x3 x dx x+2

Two restrictions will be assumed, neither of which limits the applicability of our method: (i) the leading coefficient (the coefficient of the highest power of x) in D(x) is +1; (ii) N(x) is of lower degree than D(x). A q uotient N(x)/D(x) that satisfies (ii) is called a p rop er rational function. Let us see that the restrictions (i)(ii) are not essential.
E X A M P L E 3 3 .1 : note that

Consider the case where

N ( x) 2x3 is . H ere, our first restriction is not satisfied. H owever, D( x) 5 x 8 + 3 x 4 x3 2x3 dx = 1 8 23 5 x + 5x + 3x 4 dx

5x

4 5

The integral on the right side satisfies restrictions (i) and (ii).
5 N ( x) E X A M P L E 3 3 .2 : Consider the case where is 2 x2 + 7 . H ere, our second restriction is not satisfied. B ut we can D( x) x +3 divide N(x) by D(x):

2 x 5 + 7 = 2 x 3 6 x + 18 x + 7 x2 + 3 x2 + 3

H ence,

2 x 5 + 7 dx = 1 x 4 3x 2 + 18 x + 7 dx x2 + 3 2 x2 + 3

and the problem is reduced to evaluating 182x + 7 dx, which satisfies our restrictions. x +3

A polynomial is said to be irreducible if it is not the product of two polynomials of lower degree. Any linear polynomial f (x) = ax + b is automatically irreducible, since polynomials of lower degree than f (x) are constants and f (x) is not the product of two constants. N ow consider any q uadratic polynomial g(x) = ax2 + bx + c. Then g(x) is irreducible if and only if b2 4ac < 0

XYZ[\ ] ^_ `ab ccdefdddefddcefdgbh [i_jkl!m\ no%p q] r r(XYs%Znt] ju,ev t lwXr ] lx_j\ jy1Y\2`1j\ su3Yyzu,jw

27 9

28 0

CHAPTER 33 Techniques of Integration III

To see why this is so, assume that g(x) is reducible. Then g(x) = (Ax + B)(Cx + D). H ence, x = B/A and x = D/C are roots of g(x). The q uadratic formula
x = b b 2 4 ac 2a

should yield these roots. Therefore, b2 4ac cannot be negative. Conversely, assume b2 4ac 0. Then the q uadratic formula yields two roots of g(x). B ut, if r is a root of g(x), then g(x) is divisible by x r. H ence, g(x) is reducible.
E X A M P L E 3 3 .3 : (a) x2 + 4 is irreducible, since b2 4ac = 0 4(1)(4) = 16 < 0. (b) x2 + x 4 is reducible, since b2 4ac = 1 4(l)(4) = 17 0.

W e will assume without proof the following fairly deep property of polynomials with real coefficients.
T H E O R E M 3 3 .1 : Any polynomial D(x) with leading coefficient 1 can be expressed as a product of linear factors of the form x a and of irreducible q uadratic factors of the form x2 + bx + c. (R epetition of factors is permitted.) EXA (a) (b) (c) (d) M P L E 3 3 .4 : x3 4x = x(x2 4) = x(x 2)(x + 2) x3 + 4x = x(x2 + 4) (x2 + 4 is irreducible.) 4 2 2 x 9 = (x 3)(x + 3) = (x 3 )(x + 3 )(x2 + 3) x3 3x2 x + 3 = (x + l)(x 2)2

(x2 + 3 is irreducible.)

M e th od of P artial Frac tion s

Assume that we wish to evaluate
N ( x) is a proper rational function and D(x) has leading D( x) coefficient 1. First, write D(x) as a product of linear and irreducible q uadratic factors. O ur method will depend on this factoriz ation. W e will consider various cases and, in each case, we will first explain the method by means of an example and then state the general procedure.

D( x) dx , where

N ( x)

C ase I
D(x) is a product of distinct linear factors.
E X A M P L E 3 3 .5 :

Find 2dx . x 4

In this case, D(x) = x2 4 = (x 2)(x + 2). W rite

1 = A + B ( x 2)( x + 2) x 2 x + 2

It is assumed that A and B are certain constants, that we must now evaluate. Clear the denominators by multiplying both sides by (x 2)(x + 2): 1 = A(x + 2) + B(x 2) First, substitute 2 for x in (1): 1 = A(0) + B(4) = 4B. Thus, B = 1 4. Second, substitute 2 for x in (1): 1 = A(4) + B(0) = 4A. Thus, A = 1 4 . H ence,
1 =1 1 1 1 ( x 2)( x + 2) 4 x 2 4 x + 2

(1)

In general, if a polynomial h(x) has r as a root, then h(x) must be divisible by x r.

CHAPTER 33 Techniques of Integration III

28 1

So,

dx = 1 1 1 1 dx = 1 ln | x 2 | 1 ln | x + 2 | + C 4 4 4 4 x2 4 x+2 =1 4 (ln | x 2 | ln | x + 2 |) + C x2 +C =1 4 ln x+2

( x + 1) dx . x3 + x 2 6x x +1 Factoring the denominator yields x(x2 + x 6) = x(x 2)(x + 3). The integrand is . x ( x 2)( x + 3) R epresent it in the following form:
E X A M P L E 3 3 .6 :

Find

x +1 =A+ B + C x( x 2)( x + 3) x x 2 x 3

Clear the denominators by multiplying by x(x 2)(x + 3):

x + 1 = A(x 2)(x + 3) + Bx(x + 3) + Cx(x 2) Let x be 0 in (2): 1 = A(2)(3) +B(0)(3) + C(0)(2) = 6A. So, A = 1 6. Let x be 2 in (2): 3 = A(0)(5) + B(2)(5) + C(2)(0) = 10B. So, B =
3 10

(2)

2 . Let x be 3 in (2): 2 = A(5)(0) + B(3)(0) + C(3)(5) = 15C. So, C = 15

H ence,

( x + 1) dx = 1 1 + 3 1 2 1 dx + x 2 6 x 6 x 10 x + 2 15 x + 3
3 2 =1 6 ln | x | + 10 ln | x + 2 | 15 ln | x + 3 | + C

G e n e ral R ule for C ase I A R epresent the integrand as a sum of terms of the form for each linear factor x a of the denominaxa tor, where A is an unknown constant. Solve for the constants. Integrating yields a sum of terms of the form A ln |x a|.

Remark: W e assume without proof that the integrand always has a representation of the req uired kind. For every particular problem, this can be verified at the end of the calculation.

C ase II
D(x) is a product of linear factors, some of which occur more than once.
E X A M P L E 3 3 .7 :

x2 x + 1 First factor the denominator:

3

Find

(3x + 5) dx

x3 x2 x + l = (x + l)(x l)2 Then represent the integrand

3x + 5 as a sum of the following form: x 3 x2 x + 1 3x + 5 = A + B + C x 3 x2 x + 1 x + 1 x 1 ( x 1)2

In trying to find linear factors of a denominator that is a polynomial with integral coefficients, test each of the divisors r of the constant term to see whether it is a root of the polynomial. If it is, then x r is a factor of the polynomial. In the given example, the constant term is 1. B oth of its divisors, 1 and 1, turn out to be roots.

28 2

CHAPTER 33 Techniques of Integration III

N ote that, for the factor (x 1) that occurs twice, there are terms with both (x 1) and (x 1)2 in the denominator. N ow clear the denominators by multiplying both sides by (x + l)(x 1)2: 3x + 5 = A(x l)2 + B(x + l)(x 1) + C(x + 1) Let x = 1. Then 8 = (0)A + (2)(0)B + (2)C = 2C. Thus, C = 4. Let x = 1. Then 2 = (4)A + (0)(2)B + (0)C = 4A. Thus, A = 1 2. To find B, compare the coefficients of x2 on both sides of (1). O n the left it is 0, and on the right it is A + B. 1 H ence, A + B = 0. Since A = 1 2 , B = 2 . Thus,
3x + 5 = 1 1 1 1 +4 1 2 x3 x 2 x + 1 2 x + 1 2 x 1 ( x 2)

(1)

Therefore,

x
B y Quick Formula I,

(3x + 5) dx dx = 1 ln | x + 1 | 1 2 ln | x 1| + 4 ( x 1)2 x2 x + 1 2

( x 1)
So,

dx

= ( x 1)2 dx = ( x 1)1 = 1 x 1

(3x + 5) dx 1 +C = 1 ln | x + 1| 1 2 ln | x 1| 4 x 1 x2 x + 1 2 | x + 1| 4 + C =1 2 ln | x 1| x 1

E X A M P L E 3 3 .8 :

Find

( x + 1) dx . x 3 ( x 2)2 ( x + 1) in the following form: x 3 ( x 2)2 ( x + 1) E =A+ B + C + D + x 3 ( x 2)2 x x 2 x 3 x 2 ( x 2)2

R epresent the integrand

Clear denominators by multiplying by x3(x 2)2: x + 1 = Ax2(x 2)2 + Bx(x 2)2 + C(x 2)2 + Dx3(x 2) + Ex3 Let x = 0. Then 1 = 4C. So, C = 1 4.
3 Let x = 2. Then 3 = 8E. So, E = 8 . 1 Compare coefficients of x. Then 1 = 4B 4C. Since C = 1 4,B= 2. 1 1 2 Compare coefficients of x . Then 0 = 4A 4B + 4C. Since B = 2 and C = 1 4, A= 4. 1 Compare coefficients of x4. 0 = A + D. Since A = 1 4, D = 4.

So,

( x + 1) 1 =11+1 1 +1 1 1 1 +3 x 3 ( x 2)2 4 x 2 x 2 4 x 3 4 x 2 8 ( x 2)2

and

x ( x 2)
3

( x + 1) dx
2

= 1 ln | x | 1 1 1 12 1 ln | x 2 | 4 2x 8x 4 1 3 1 +C = 1 ln x 4 x + 4 x2 8 x2 8x2

3 8

1 +C x2

CHAPTER 33 Techniques of Integration III

28 3

G e n e ral R ule for C ase II A A2 Ak For each repeated linear factor (x r) that occurs k times in the denominator, use 1 + +. . .+ x r ( x r )2 ( x r)k as part of the representation of the integrand. E very linear factor that occurs only once is handled as in Case I.

C ase III
D(x) is a product of one or more distinct irreducible q uadratic factors and possibly also some linear factors (that may occur more than once).
G e n e ral R ule for C ase III Linear factors are handled as in Cases I II. For each irreducible q uadratic factor x2 + bx + c, place a term Ax + B in the representation of the integrand. 2 x + bx + c
E X A M P L E 3 3 .9 :

( x 1) dx . x( x 2 + 1)( x 2 + 2) R epresent the integrand as follows:

Find

( x 1) + C + Dx + E = A + Bx x( x 2 + 1)( x 2 + 2) x x2 + 1 x2 + 2
Clear the denominators by multiplying by x(x2 + l)(x2 + 2). x 1 = A(x2 + l)(x2 + 2) + (Bx + C )x(x2 + 2) + (Dx + E )x(x2 + 1) M ultiply out on the right: x 1 = (A + B + D)x4 + (B+ E )x3 + (3A + C + D)x2 + (2C + E )x + 2A Comparing coefficients, we get: 2A = 1, 2C + E= 1, 3A + C + D = 0, B + E = 0, A+B+D=0

3 1 H ence, A = 1 2 and, therefore, C + D = 2 , B + D = 2 . From the latter two eq uations, C B = 1. From 2C + E = 1 and B + E = 0, we get 2C B = 1. N ow, from C B = 1 and 2C B = 1, we get C = 0. H ence, from C B= 1, B = 1. 3 Then, from B + D = 1 2 , it follows that D = 2 . Finally, from B + E = 0, E = 1.

Thus,
( x 1) +2 = 1 1 2 x + 1 3x 2 x x + 1 2 x2 + 2 x( x 2 + 1)( x 2 + 2)

H ence,

x( x

( x 1) dx x dx + 1 3x + 2 dx =1 2 ln | x | 2 2 x2 + 2 x +1 + 1)( x 2 + 2)

N ow, Also,

x dx = 1 2 x dx = 1 ln( x 2 + 1) 2 2 x2 + 1 +1

(by Quick Formula II)

3x + 2 dx = 3x 2 2 x 2 + 2 dx + x 2 + 2 dx +2 = 3 22 x dx + 2 tan 1 x = 3 ln( x 2 + 2) + 2 tan 1 x 2 x +2 2 2 2

Therefore,

x( x

( x 1) dx 2 1 2 tan 1 x + C =1 2 ln | x | + 4 ln( x + 1) + 2 + 1)( x 2 + 2) 2

28 4

CHAPTER 33 Techniques of Integration III

C ase IV
D(x) is a product of z ero or more linear factors and one or more irreducible q uadratic factors.
G e n e ral R ule for C ase IV Linear factors are handled as in Cases I II. For each irreducible q uadratic factor x2 + bx + c that occurs to the k th power, insert as part of the representation of the integrand.

A1 x + B1 A x + B2 . . . + 2Ak x + Bk k + 2 2 2 2 + x + bx + c ( x + bx + c) ( x + bx + c)
E X A M P L E 3 3 .1 0 :
2 + 3 dx. Find 2 x ( x 2 + 1)2

2 + 3 = Ax + B + Cx + D . Then Let 2 x 2 ( x + 1)2 x 2 + 1 ( x 2 + 1)2

2x2 + 3 = (Ax + B)(x2 + 1) + Cx + D = Ax3 + Bx2 + (A + C )x + (B + D)

Compare coefficients: A = 0, B = 2, A + C = 0, B + D = 3. H ence, C = 0, D = 1. Thus,

(x

2 x 2 + 3 dx = 2 1 2 x 2 + 1 dx + ( x 2 + 1)2 dx + 1)2 = 2 tan 1 x + 1 dx ( x 2 + 1)2

In the second integral, let x = tan q. Then

sec 2 d 1 = cos 2 d = 1 2 ( + sin cos ) 2 dx = + 1) sec 4

(x

= 1 + tan = 1 tan 1 x + 2 x 2 x +1 tan 2 + 1 2

2 x 2 + 3 dx = 5 tan 1 x + 1 x + C 2 2 2 x2 + 1 + 1)2

) (

Thus,

(x

S O LV E D P R O B L E M S
1.
4 x 3 x 1 dx. Find x 3 x x2 The integrand is an improper fraction. B y division,

x 4 x3 x 1 = x x + 1 = x x + 1 x3 x 2 x3 x 2 x 2 ( x 1)

W e write

x + 1 = A + B + C and obtain x 2 ( x 1) x x 2 x 1

x + 1 = Ax(x 1) + B(x 1) + Cx2

CHAPTER 33 Techniques of Integration III

28 5

For x = 0, 1 = B and B = 1. For x = 1, 2 = C. For x = 2, 3 = 2A + B + 4C and A = 2. Thus,

x 4 x 3 x 1 dx = x dx + 2 dx + dx 2 dx x x2 x 1 x3 x 2 1 2 ln | x 1| + C = 1 x 2 1 + 2 ln x + C 2 =1 2 2 x + 2 ln | x | x x x 1

2.

Find

( x + 2)( x + 3) .
( x + 2 ) ( x + 3)
x = A + B . Clear the denominators: x+2 x+3

x dx

Let

x = A(x + 3) + B(x + 2) Let x = 2. Then 2 = A. Let x = 3. Then 3 = B. So, B = 3.

( x + 2)( x + 3) = 2 x + 2 dx + 3 x + 3 dx
= 2 ln | x + 2 | + 3 ln | x + 3 | + C = ln(( x + 2)2 ) + ln(| x + 3 |)3 + C = ln ( x + 3)3 +C ( x + 2)2

x dx

3.

Find Let

x( x + 2)( x 1) dx.
x2 + 2 = A + B + C . Clear the denominators: x( x + 2)( x 1) x x + 2 x 1

x2 + 2

x2 + 2 = A(x + 2)(x 1) + Bx (x 1) + Cx (x + 2) Let x = 0. Then 2 = 2A. So, A = 1. Let x = 2. Then 6 = 6B. So, B = 1. Let x = 1. Then 3 = 3C. So, C = 1. H ence,

x( x + 2)( x 1) dx = x dx + x + 2 dx + x 1 dx
= ln | x | + ln | x + 2 | + ln | x 1| + C = ln ( x + 2)( x 1) +C x

x2 + 2

4.

Find

x3 + 1 dx. ( x + 2)( x 1)3 x3 + 1 = A + B + C + D . Clear the denominators: ( x + 2)( x 1)3 x + 2 x 1 ( x 1)2 ( x 1)3

Let

x3 + l = A(x 1)3 + B(x + 2)(x l)2 + C(x + 2)(x 1) + D(x + 2)

7 3 Let x = 2. Then 7 = 27A. So, A = 27 . Let x = 1. Then 2 = 3D. So, D = 2 3 . Compare coefficients of x . Then 1 = A + B. 7 7 7 2 Since A = 27 , B = 20 27 . Compare coefficients of x . 0 = 3A + C. Since A = 27 , C = 9 .

Thus,

( x + 2)( x 1)

x3 + 1

dx = 7 1 dx + 20 1 dx + 7 1 2 dx + 2 1 3 dx 27 x + 2 27 x 1 9 ( x 1) 3 ( x 1) 1 +C = 7 ln | x + 2 | + 20 ln | x 1| 7 1 1 27 9 x 1 3 ( x 1)2 27

28 6

CHAPTER 33 Techniques of Integration III

5.

Find

x 3 + x 2 + x + 2 dx. x 4 + 3x 2 + 2
3 + x 2 + x + 2 = Ax + B + Cx + D and obtain x4 + 3x2 + 2 = (x2 + l)(x2 + 2). W e write x 4 x + 3x 2 + 2 x2 + 1 x2 + 2

x3 + x2 + x + 2 = (Ax + B)(x2 + 2) + (Cx + D)(x2 + 1) = (A + C )x3 + (B + D)x2 + (2A + C )x + (2B + D) H ence A + C = 1, B + D = 1, 2A + C = 1, and 2B + D = 2. Solving simultaneously yields A = 0, B = 1, C= l, D = 0. Thus,

x 1 x 3 + x 2 + x + 2 dx = x 2 + 1 dx + x 2 + 2 dx x 4 + 3x 2 + 2
2 = tan 1 x + 1 2 ln( x + 2) + C

6.

Find

x 5 x 4 + 4 x 3 4 x 2 + 8 x 4 dx. ( x 2 + 2)3

W e write x 5 x 4 + 4 x 3 4 x 2 + 8 x 4 = Ax + B + Cx + D + Ex + F . Then x 2 + 2 ( x 2 + 2)2 ( x 2 + 2)3 ( x 2 + 2)3 x5 x4 + 4x3 4x2 + 8x 4 = (Ax + B)(x2 + 2)2 + (Cx + D)(x2 + 2) + Ex + F = Ax5 + Bx4 + (4A + C)x3 + (4B + D)x2 + (4A + 2C + E)x + (4B + 2D + F) from which A = 1, B = 1, C = 0, D = 0, E = 4, F = 0. Thus the given integral is eq ual to

B y Quick Formula II,

( x 1) dx x dx x dx x dx dx + 4 + 4 2 = ( x 2 + 2)3 x2 + 2 ( x + 2)3 x 2 + 2 x 2 + 2

x
and by Quick Formula I,

2 x dx 1 x dx =1 = ln( x 2 + 2) 2 + 2 2 x2 + 2 2

(x
So,

x dx 1 2 2 = 1 ( x 2 + 2)3 (2 x) dx = 1 =1 21 2 2 ( 2 )( x + 2) 4 ( x + 2) + 2)3 2

x 5 x 4 + 4 x 3 4 x 2 + 8 x 4 dx = 1 ln( x 2 + 2) 2 ta an 1 x 2 1 2 + C 2 2 ( x 2 + 2)3 2 ( x + 2)

S U P P LE M E N TA R Y P R O B LE M S
In P roblems 7 25, evaluate the given integrals. 7.

x x x

dx = 1 ln x 3 + C 9 6 x+3 x dx = 1 ln ( x + 1)( x 4)4 + C 3x 4 5

8.

9.

x 2 3x 1 dx = ln x1 2 ( x + 2)3 2 + C 3 x 1 + x2 2x

CHAPTER 33 Techniques of Integration III

28 7

10 .

x x

dx = 1 ln x + 1 + C + 7x + 6 5 x + 6

11.

x 2 + 3x 4 dx = x + ln ( x + 2)( x 4)4 + C 2 2x 8 x dx = ln | x 2 | 2 + C x2
2 6 dx = 1 2 x 3 x ln(1 x )

12 .

( x 2) (1 x) x
3

13 .

x4

4 + 1 +C 1 x 2(1 x)2

14 .

dx = ln +x

x +C x2 + 1

15 .

(x

x 3 + x 2 + x + 3 dx = ln x 2 + 3 + tan 1 x + C 2 + 1)( x 2 + 3) x +C x2 2x + 3

16 .

x 4 2 x 3 + 3x 2 x + 3 dx = 1 x 2 + ln 2 x 3 2 x 2 + 3x 2 x 3 dx = ln( x 2 + 1) + 2 1 + C 2 + 1)2 x +1

17 .

(x

18 .

2 x 3 + x 2 + 4 dx = ln( x 2 + 4) + 1 tan 1 x + 4 + C 2 2 x2 + 4 ( x 2 + 4) 2 x 3 + x 1 dx = ln x 2 + 1 1 tan 1 x 1 x + C 2 2 x2 + 1 ( x 2 + 1)2 x 4 + 8 x 3 x 2 + 2 x + 1 dx = ln x 3 x 2 + x 3 + 2 tan 1 2 x 1 + C x +1 ( x 2 + 3)( x 3 + 1) ( x + 1)2 3 3 x 3 + x 2 5x + 15 dx = ln x 2 + 2 x + 3 + 5 ta an 1 x + 1 5 tan 1 x + C 2 + 5)( x 2 + 2 x + 3) 2 5 2

()

19 .

20.

2 1.

(x

22.

2 x 6 + 7 x 5 + 15x 4 + 23x 2 + 25x 3 dx = 1 3 + ln 2 x + 1 + C ( x 2 + x + 2)2 ( x 2 + 1)2 x2 + x + 2 x2 + 1 x +x+2 x dx 3 +C = 1 + 1 ln e ex 3e x 3e x 9

23.

2x

(H int: Let ex = u.) (H int: Let cos x = u.)

24.

cos x(1 + cos

sin x dx

x)

= ln

1 + cos 2 x +C cos x

25.

(2 + tan 2 )sec 2 d = ln |1 + tan | + 2 tan 1 2 tan 1 + C 1 + tan 3 3 3

C H A P TE R 3 6

Applications of Integration III: Area of a Surface of R evolution

If an arc of a curve is revolved about a line that does not intersect the arc, then the resulting surface is called a surface of revolution. B y the surface area of that surface, we mean the area of its outer surface. Let f be a continuous function on [ a, b] that is differentiable in (a, b) and such that f (x) 0 for a x b. Then the surface area S of the surface of revolution generated by revolving the graph of f on [ a, b] about the x axis is given by the formula
b b dy S = 2 y 1 + dx = 2 f ( x) 1 + ( f ( x))2 dx a a dx 2

(36.1)

For a justification of this formula, see P roblem 11. There is another formula like (36.1) that is obtained when we exchange the roles of x and y. Let g be a continuous function on [ c, d ] that is differentiable on (c, d ) and such that g( y) 0 for c y d. Then the surface area S of the surface of revolution generated by revolving the graph of g on [ c,d ] about the y axis is given by the formula:
d d dx S = 2 x 1 + dy = 2 g( y) 1 + (g ( y))2 dy c c dy 2

(36.2)

Similarly, if a curve is given by parametric eq uations x = f (u), y = g(u) (see Chapter 37), and, if the arc from u = u1 to u = u2 is revolved about the x axis, then the surface area of the resulting surface of revolution is given by the formula
u2 dx dy S = 2 y + du u1 du du 2 2

(36.3)

H ere, we have assumed that f and g are continuous on [ u1, u2] and differentiable on (u1, u2), and that y = g(u) 0 on [ u1, u2]. Another such formula holds in the case of a revolution around the y axis.

S O LV E D P R O B L E M S
1. Find the area S of the surface of revolution generated by revolving about the x axis the arc of the parabola y2 = 12x from x = 0 to x = 3. B y implicit differentiation,
dy 6 = dx y dy y 2 + 36 1+ = dx y2
2

and

XYZ[\ ] ^_ `ab ccdefdddefddcefdgbh [i_jkl!m\ no%p q] r r(XYs%Znt] ju,ev t lwXr ] lx_j\ jy1Y\2`1j\ su3Yyzu,jw

301

302

CHAPTER 36

Applications of Integration III

B y (36.1),
S = 2 y
0 3 3 y 2 + 36 dx = 2 12 x + 36 dx 0 y
3

= 2 (8(12 x + 36) 2 )]3 0 = 24(2 2 1)

2.

Find the area S of the surface of revolution generated by revolving about the y axis the arc of x = y3 from y = 0 to y = 1.
dx = 3y 2 and 1 + dx = 1 + 9 y 4. So, by (36.2), dy dy S = 2 x 1 + 9 y 4 dy = 2 y3 1 + 9 y 4 dy
0 0 1 1 2

= (1 + 9 y 4 ) (36 y3 ) dy 18 0
1
1 2 3 = 2 (1 + 9 y 4 ) 2 ]1 0 18 3 = (10 10 1) 27

3.

Find the area of the surface of revolution generated by revolving about the x axis the arc of y2 + 4x = 2 ln y from y = 1 to y = 3.
d 3 3 1 + y2 S = 2 y 1 + dx dy = 2 y dy = (1 + y 2 ) dy = 32 e 1 1 3 2y dy 2

4.

Find the area of the surface of revolution generated by revolving a loop of the curve 8a2y2 = a2x2 x4 about the x axis. (See Fig. 36-1.)

Fig. 36 -1

H ere

dy a 2 x 2 x 3 = dx 8a 2 y

and

(a 2 2 x 2 ) 2 (3a 2 2 x 2 )2 1 + dx = 1 + 2 2 2 = dy 8a (a x ) 8a 2 (a 2 x 2 )

H ence

a a x a2 x2 dy 3a 2 2 x 2 dx S = 2 2 1 + dx = 2 0 0 dx 2a 2 2a 2 a 2 x 2

= 2 4a

(3a 2 2 x 2 ) x dx = 1 a 2 4

5.

2 y2 = 1. Find the area of the surface of revolution generated by revolving about the x axis the ellipse x + 16 4

S = 2 y
4

16 y 2 + x 2 dx = 4y 2

64 3x 2 dx
4

x 3 x 3 4 3 64 3x 2 + 32 sin 1 = = 8 1 + 9 8 2 3 2 4

CHAPTER 36

Applications of Integration III

303

6.

Find the area of the surface of revolution generated by revolving about the x axis the hypocycloid x = a cos3 q, y = a sin3 q. The req uired surface is generated by revolving the arc from q = 0 to q = . W e have 2 2 dx = 3a cos 2 sin , dy = 3a sin 2 cos , and dx + dy = 9a 2 cos 2 sin 2 . Then d d d d

( )
2

S = 2(2 ) y
2

dx + dy d = 2(2 ) (d ) d
2

(a sin 3 )3a cos sin d

= 12a 5
2

(sq uare units)

7.

Find the area of the surface of revolution generated by revolving about the x axis the cardioid x = cos 3q cos 2q, y = 2 sin q sin 2q. The req uired surface is generated by revolving the arc from q = 0 to q = . (See Fig. 36-2.) W e have
dx = 2 sin + 2 sin 2 , d dy = 2 cos 2 cos 2 , d

Fig. 36 -2

and
dx + dy (d ) d
2 2

= 8 (1 sin sin 2 cos cos 2 ) = 8 (1 cos )

Then
S = 2 (2 sin sin 2 )(2 2 1 cos ) d
0

16 2 3 (1 cos )5 2 = 8 2 sin (1 cos ) 2 d = 0 5 0

= 128 5

(sq uare units)

8.

Show that the surface area of a cylinder of radius r and height h is 2rh. dy = 0, The surface is generated by revolving about the x axis the curve y = r from x = 0 to x = h. Since dx 2 dy 1 + = 1. Then, by (36.1), dx
S = 2 r dx = 2 ( rx ) = 2 rh
0 0 h h

9.

Show that the surface area of a sphere of radius r is 4r2. The surface area is generated by revolving about the x axis the semicircle y = r 2 x 2 from x = r to x = r. B y symmetry, this is double the surface area from x = 0 to x = r. Since y2 = r2 x2,
2y dy = 2 x dx

and therefore

dy =x dx y

and

2 x 2 + y2 r 2 dy 1 + = 1 + x2 = = 2 dx y y2 y

304

CHAPTER 36

Applications of Integration III

H ence, by (36.1),
2 r r 2 S = 2 2 y r 2 dx = 4 r 1 dx = 4 r x 0 = 4 r 0 0 y r

10 . (a) Show that the surface area of a cone with base of radius r and with slant height s (see Fig. 36-3) is rs. (b) Show that the surface area of a frustum of a cone having bases of radius r1 and r2 and slant height u (see Fig. 36-4) is (r1 + r2)u. (N ote that the frustum is obtained by revolving the right-hand segment of the slant height around the base of the triangle.)

Fig. 36 -3

Fig. 36 -4

(a) Cut open the cone along a slant height and open it up as part of a circle of radius s (as shown in Fig. 36-5). N ote that the portion of the circumference cut off by this region is 2p r (the circumference of the base of the cone.) N ow the desired area S is the difference between p s2 (the area of the circle in Fig. 36-5) and the area A1 of the circular sector with central angle q. This area A1 is ( s 2 ) = 1 s 2. Since the arc cut off by q is 2 2 2p s 2p r, we get = 2 s 2 r . Thus, A1 = p (s r)s. H ence, S = p s2 p ( s r )s = p rs sq uare units. s

Fig. 36 -5

(b) From the similar triangles in Fig. 36-4, we get sq uare units.

ru u1 u1 + u . Then r2 u1 = r1 ul + r1u. So, u1 = r 1 . N ow, by = r1 r2 2 r 1 part (a), the surface area of the frustum is p r2(u1 + u) p rl ul = p (r2 - rl)ul + p r2 u = p r1 u + p r2 u = p (r1 + r2)u

CHAPTER 36

Applications of Integration III

305

11. Sketch a derivation of formula (36.1). Assume that [ a, b] is divided into n eq ual subintervals, [ xk-1, xk], each of length x = b a . The total surface n area S is the sum of the surface areas Sk generated by the arcs between the points (xk-1, f (xk-1)) and (xk, f (xk)), each of which is approximated by the surface area generated by the line segment between (xk-1, f (xk-1)) and (xk, f (xk)). The latter is the area of a frustum of a cone. In the notation of Fig. 36-6, this is, by virtue of P roblem 10(b):

f ( xk 1 ) + f ( xk )

( x )2 + ( y )2 = 2

N ow, f ( xk 1 ) + f ( xk ), being the average of f (xk-1) and f (xk), is between those two values and, by the 2 2 2 2 y x * for some x * in (x 1 x + y = + ( ) ( ) intermediate value theorem, is eq ual to f ( xk x ) . Also, .B y k1, k ) k x y # # = f x ( k ) for some xk in (xk-1, xk). Thus, S is approximated by the sum the mean value, theorem x

f ( xk 1 ) + f ( xk ) 2

( x )2 + ( y )2

2 f ( x )
* k k =1

# 1 + f ( xk ) x 2

and it can be shown that this sum can be made arbitrarily close to 2 f ( x ) 1 + ( f ( x )) dx. H ence, the latter is a eq ual to S.
b 2

Fig. 36 -6

S U P P LE M E N TA R Y P R O B LE M S
In P roblems 12 20, find the area of the surface of revolution generated by revolving the given arc about the given axis:

12 . y = mx from x = 0 to x = 2; x axis 13 . y = 1 x 3 from x = 0 to x = 3; x axis 3

Ans.

4 m 1 + m 2

Ans. p (8282 - 1)/9

In general, the following result can be proved: B lis s s T h e o r e m : Assume f and g are continuous on [ a, b]. D ivide [ a, b] into subintervals [ xk-1, xk] with a = x0 < x1 < < xn < b,

* # and let Dk x = xk xk-l. In each [ xk-1, xk], choose xk and xk . Then the approximating sum

to

f ( x ) g ( x ) dx by letting n + and making the maximum lengths of the subintervals approach 0.

f ( x ) g ( x ) xcan be made arbitrarily close

* k # k k k =1

306

CHAPTER 36

Applications of Integration III

3 14 . y = 1 3 x from x = 0 to x = 3; y axis

Ans. Ans. Ans. Ans.

1 9 82 + ln 9 + 82 2 1 4 + ln 2 ) ( 15 4 34 2 + ln 3 + 2 2

15 . O ne loop of 8y2 = x2(l x2); x axis 16 . y = x3/6 + l/2x from x = 1 to x = 2; y axis 17 . y = ln x from x = 1 to x = 7; y axis 18 . O ne loop of 9y2 = x(3 x)2; y axis 19. An arch of x = a(q sin q ), y = a(1 cos q ); x axis 20. x = e t cos t, y = et sin t from t = 0 to t = 1 ; x axis 2

Ans. 28p 3/5 Ans. 64p a2/3 Ans. 2p 2(2ep + l)/5

21. F ind the su rface are a of a z one cu t from a sp he re of rad iu s r b y tw o p aralle l p lane s, e ach at a d istance 1 a from 2 the ce nte r. Ans. 2par

22. F ind the su rface are a of a toru s (d ou g hnu t) g e ne rate d b y re v olv ing the circle x2 + (y b)2 = a2 ab ou t the x axis. Assu me 0 < a < b. Ans. 42ab

C H A P TE R 3 7

Parametric Representation of Curves

Parametric Equations
If the coordinates (x, y) of a point P on a curve are given as functions x = f (u), y = g(u) of a third variable or p aram eter, u, the equations x = f (u) and y = g(u) are called p aram etric eq uatio ns of the curve.
EXAMPLE 37.1:

(a) x = cos , y = 4 sin2 are parametric equations, with parameter , of the parabola 4x2 + y = 4, since 4x2 + y = 4 cos2 + 4 sin2 = 4. (b)
2 x= 1 2 t , y = 4 t is another parametric representation, with parameter t, of the same curve.

It should be noted that the first set of parametric equations represents only a portion of the parabola (F ig. 3 7 -1(a)), whereas the second represents the entire curve (F ig. 3 7 -1(b)).

y q = 1p 2 q = 3p 4 q= p O (a) q= 0 q = 1p 4 x

Fig. 37-1 EXAMPLE 37.2 :

(a) The equations x = r cos , y = r sin represent the circle of radius r with center at the origin, since x 2 + y 2 = r 2 cos2 + r 2 sin2 = r 2(cos2 + sin2 ) = r 2. The parameter can be thought of as the angle from the positive x axis to the segment from the origin to the point P on the circle (Fig. 37-2). (b) The equations x = a + r cos , y = b + r sin represents the circle of radius r with center at (a, b), since (x a)2 + (y b)2 = r 2 cos2 + r 2 sin2 = r 2 (cos2 + sin2 ) = r 2.

{|}~   ~! % ({|%} , {  1|21 3|,

307

308

CHAPTER 37 Parametric Representation of Curves

Fig. 37-2

A ssume that a curve is specified by means of a pair of parametric equations x = f (u) and y = g(u). Then dy d2y the first and second derivatives and 2 are given by the following formulas.
dx dx

(37.1) F irst D eriv ativ e

dy dy dx = dx du du

This follows from the C hain R ule formula

dy dy dx = . du dx du

(37.2 ) S econd D eriv ativ e

d 2 y d dy = dx 2 du dx dx du

This follows from the C hain R ule formula

d dy d 2 y dx = . du dx dx 2 du

Arc Leng th for a Parametric C urv e

If a curve is given by parametric equations x = f (t), y = g(t), then the length of the arc of the curve between the points corresponding to parameter values t1 and t2 is
L=
t2

t1

dx dy + dt dt dt

This formula can be derived by an argument similar to that for the arc length formula (29.2).

S O LV ED PR O B LEMS
1. Find
dy d2y and 2 if x = t sin t, y = 1 cos t. dx dx

dx = 1 cos t dt

and

dy dy sin t . Then = sin t. B y (37.1), = dt dx 1 cos t

d dy = (1 cos t ) (cos t ) (sin t ) (sin t ) dt dx (1 cos t )2 = cos t (cos 2 t + sin 2 t ) cos t 1 1 = = (1 cos t )2 (1 cos t )2 cos t 1

CHAPTER 37 Parametric Representation of Curves

309

H ence, by (37.2),
d2y 1 1 (1 cos t ) = = (1 cos t )2 dx 2 cos t 1

2.

Find

dy d2y and if x = et cos t, y = et sin t. dx dx 2 dy cos t + sin t dx = et (cos t sin t ) and dy = et (cos t + sin t ). B y (37.1), dx = cos t sin t . Then, dt dt
2 d dy = (cos t sin t ) (cos t + sin t )( sin t cos t ) dt dx (cos t sin t )2

= =

(cos t sin t )2 + (cos t + sin t )2 2(cos 2 t + sin 2 t ) = (cos t sin t )2 (cos t sin t )2 2 (cos t sin t )2

S o, by (37.2),
d2y 2 2 = et (cos t sin t ) = t dx 2 (cos t sin t )2 e (cos t sin t )

3.

Find an equation of the tangent line to the curve x = t , y = t 1 at the point where t = 4. t dy dy 1 1 dx = 1 = 1 + 3/ 2 . B y (37.1), = 2 t + . S o, the slope of the tangent line when t = 4 is and dt 2 t dt dx t 2t 17 7 17 1 2 4 + 4 = 4 . W hen t = 4, x = 2 and y = 2 . A n equation of the tangent line is y 7 2 = 4 ( x 2). The position of a particle that is moving along a curve is given at time t by the parametric equations x = 2 3 cos t, y = 3 + 2 sin t, where x and y are measured in feet and t in seconds. (S ee Fig. 37-3.) N ote that 1 1 2 2 9 ( x 2) + 4 ( y 3) = 1, so that the curve is an ellipse. Find: (a) the time rate of change of x when t = /3; (b) the time rate of change of y when t = 5/3; (c) the time rate of change of the angle of inclination of the tangent line when t = 2/3. dy 2 dy dy = cot t . = 2cos t . Then tan = = 3sin t and dx 3 dt dt
3 3 ft/sec (a) W hen t = , dx = 2 3 dt dy 1 (b) W hen t = 5 , 3 dt = 2( 2 ) = 1 ft/sec 2 6 csc 2 t d = 2 3 csc t = (c) = tan 1 ( 2 . 3 cot t ). S o, 4 2 dt 1 + 9 cot t t 9 + 4 cot 2 t

4.

y q t = 1p 3 t= 0 t = 2p 3 (2, 3)

t = 5p 3 x O

Fig. 37-3

31 0

CHAPTER 37 Parametric Representation of Curves

2 24. Thus, the angle of inclination of the tangent line is decreasing at the W hen t = 2 , d = 6(2 / 3 ) 3 dt 9 + 4(1/ 3 )2 = 31 rate of 24 31 radians per second.

5.

Find the arc length of the curve x = t2, y = t3 from t = 0 to t = 4. 2 2 dx = 2t , dy = 3t 2 and dx + dy = 4t 2 + 9t 4 = 4t 2 (1 + 9 t 2 ). 4 dt dt dt dt Then

( )

2 L = 2t 1 + 9 4 t dt = 0

4 9

2 1/ 2 9 (1 + 9 ( 2 t ) dt 4t )

4 2 9 3

2 3/ 2 (1 + 9 ]0 = 4t ) 4

8 27

(37 37 1)

6.

Find the length of an arch of the cy cloid x = sin , y = 1 cos between = 0 and = 2. 2 2 dx = 1 cos , dy = sin and dx + dy = (1 cos )2 + sin 2 = 2(1 cos ) = 4 sin 2 . Then d d d 2 d

( )

()

4 L = 2 sin d = 4 cos = 4(cos cos 0) = 8 0 2 2 0

()

()

S U PPLEMEN T AR Y PR O B LEMS
In P roblems 711, find: (a) 7. 8. 9. x = 2 + t, y = 1 + t2 x = t + 1/t, y = t + 1 x = 2 sin t, y = cos 2t
dy d2y ; (b) . dx dx 2

Ans. Ans. Ans. Ans. Ans.

(a) 2t; (b) 2 (a) t2/(t2 1 ); (b) 2t3/(t 2 1 )3 (a) 2 sin t; (b) 1 (a) tan ; (b) 1/(3 cos4 sin ) (a) tan ; (b) 1/(a cos3 )

10 . x = cos3 , y = sin3 11. x = a(cos + sin ), y = a(sin cos )

12 . Find the slope of the curve x = et cos 2t, y = e2t sin 2t at the point t = 0. Ans. 2

13 . Find the rectangular coordinates of the highest point of the curve x = 96t, y = 96t 16t2. (H int: Find t for maximum y.) Ans. (28 8 , 1 44)

14 . Find equations of the tangent line and normal line to the following curves at the points determined by the given value of the parameter: (a) x = 3et, y = 5e t at t = 0 (b) x = a cos4 , y = a sin4 at = 4 Ans. (a) 3y + 5 x = 30 , 5 y 3x = 1 6 ; (b) 2x + 2y = a, y = x

CHAPTER 37 Parametric Representation of Curves

31 1

15 . Find an equation of the tangent line at any point P(x, y) of the curve x = a cos3 t, y = a sin3 t. S how that the length of the segment of the tangent line intercepted by the coordinate axes is a. Ans.
x sin t + y cos t = a sin 2t 2

16 . For the curve x = t2 1, y = t3 t, locate the points where the tangent line is (a) horiz ontal, and (b) vertical. S how that, at the point where the curve crosses itself, the two tangent lines are mutually perpendicular. Ans. (a) t =
3 ; (b) t = 0 3

In P roblems 1720, find the length of the specified arc of the given curve. 17 . The circle x = a cos , y = a sin from = 0 to = 2. Ans. 2a

18 . x = et cos t, y = et sin t from t = 0 to t = 4. Ans.

2 (e4 1)

19 . x = ln 1 + t 2 , y = tan1 t from t = 0 to t = 1. Ans.

ln(1 + 2 )

2 0 . x = 2 cos + cos 2 + 1, y = 2 sin + sin 2. Ans. 16

1 2 3/ 2 2 1. The position of a point at time t is given as x = 1 2 t , y = 9 (6t + 9) . Find the distance the point travels from t = 0 to t = 4.

Ans.

20

2 2 . Identify the curves given by the following parametric equations and write equations for the curves in terms of x and y: (a) (b) (c) (d) x = 3t + 5, y = 4t 1 x = t + 2, y = t2 x = t 2, y = t t2 x = 5 cos t, y = 5 sin t Ans. Ans. Ans. Ans. S traight line: 4x 3y = 23 P arabola: y = (x 2)2 H y perbola: y = 2 + 1 x C ircle: x2 + y2 = 25

2 3 . (G C ) U se a graphing calculator to find the graphs of the following parametric curves: (a) x = + sin , y = 1 cos (b) x = 3 cos3 , y = 3 sin3 (c) x = 2 cot , y = 2 sin2 2 (d) x = 3 3 , y = 3 3 (1 + ) (1 + ) (cy cloid) (hy pocy cloid) (witch of A gnesi) (folium of D escartes)

C H A P TE R 3 8

Curvature
D eriv ativ e of Arc Leng th
L et y = f (x) have a continuous first derivative. L et A(x0, y0) be a fixed point on its graph (see Fig. 38-1) and denote by s the arc length measured from A to any other point P(x, y) on the curve. W e k now that, by formula (29.2),
s=
x x0

1+

dy dx dx

if s is chosen so as to increase with x. L et Q(x + x, y + y) be a point on the curve near P. L et s denote the arc length from P to Q. Then
ds s dy = lim = 1+ dx x0 x dx
2

and, similarly ,
ds s dx = lim = 1+ dy y0 y dy
2

The plus or minus sign is to be tak en in the first formula according as s increases or decreases as x increases, and in the second formula according as s increases or decreases as y increases.

Fig. 38 -1

W hen a curve is given by parametric equations x = f (u), y = g(u),

ds s = lim = du u0 u dx dy + du du
2 2

H ere the plus or minus sign is to be tak en according as s increases or decreases as u increases.

31 2 {|}~   ~! % ({|%} , {  1|21 3|,

CHAPTER 38

Curvature

31 3

To avoid the repetition of ambiguous signs, we shall assume hereafter that the direction on each arc has been established so that the derivative of arc length will be positive.

C urv ature
The curvature K of a curve y = f (x) at any point P on it is defined to be the rate of change of the direction of the curve at P, that is, of the angle of inclination of the tangent line at P, with respect to the arc length s. (S ee Fig. 38-2.) Intuitively , the curvature tells us how fast the tangent line is turning. Thus, the curvature is large when the curve bends sharply .

Fig. 38 -2

A s formulas for the curvature, we get:

d2y d dx 2 K= = lim = 2 3/ 2 ds s0 s dy 1 + dx
(38.1)

or, in terms of y,
K= d 2x dy 2
3/ 2

dx 2 1+ dy

(38.2)

For a derivation, see P roblem 13. K is sometimes defined so as to be positive. If this is assumed, then the sign of K should be ignored in what follows.

T h e R ad ius of C urv ature

The radius of curvature R at a point P on a curve is defined by R =
1 , provided that K 0. K

T h e C ircle of C urv ature

The circle of curvature, or o sculating circle of a curve at a point P on it, is the circle of radius R ly ing on the concave side of the curve and tangent to it at P. (S ee Fig. 38-3.)

Fig. 38 -3

31 4

CHAPTER 38

Curvature

To construct the circle of curvature, on: the concave side of the curve, construct the normal line at P and on it lay off a segment PC of length R. The point C is the center of the required circle.

T h e C enter of C urv ature

The center of curvature for a point P(x, y) of a curve is the center C of the circle of curvature at P. The coordinates ( , ) of the center of curvature are given by
2 dy dy 1 + dx dx = x 2 2 d y / dx

dy dx = y+ 2 d y / dx 2 1+

or by
dx 1+ dy = x+ 2 d x / dy 2
2 2 dx dx 1 + dy dy = y 2 2 d x / dy

S ee P roblem 9 for details.

T h e Ev olute
The evolute of a curve is the locus of the centers of curvature of the given curve. (S ee P roblems 1112.)

S O LV ED PR O B LEMS
1. Find ds at P(x, y) on the parabola y = 3x2. dx
ds = 1 + dy = 1 + (6 x)2 = 1 + 36 x 2 dx dx
2

2.

Find ds and ds at P(x, y) on the ellipse x2 + 4y2 = 8. dx dy S ince 2 x + 8 y

4y dy dy = 0, = x and dx = . Then dx dy x dx 4y
2 x 2 + 16 y 2 32 3x 2 dy 1+ = 1+ x 2 = = and ds = dx dx 16 y 16 y 2 32 4 x 2 2

32 3x 2 32 4 x 2

16 y 2 x 2 + 16 y 2 2 + 3y 2 1 + dx = 1 + 2 = = and ds = dy dy 2 y2 x x2

2 + 3y 2 2 y2

3.

Find ds at P( ) on the curve x = sec , y = tan . d

ds = d dx + dy (d ) d
2 2

= sec 2 tan 2 + sec 4 = |sec c | tan 2 + sec 2

CHAPTER 38

Curvature

31 5

4.

The coordinates (x, y) in feet of a moving particle P are given by x = cos t 1, y = 2 sin t + 1, where t is the time in seconds. A t what rate is P moving along the curve when (a) t = 5/6, (b) t = 5/3, and (c) P is moving at its fastest and slowest?
ds = dt

( )
dx dt

dy + = sin 2 t + 4 cos 2 t = 1 + 3 cos 2 t dt

3 (a) W hen t = 5/6, ds/dt = 1 + 3( 4 ) = 13 / 2 ft/sec.

(b) W hen t = 5/3, ds/dt = 1 + 3( 1 7 /2 ft/sec. 4) = ds dS 2 = 3cos t sin t . S olving dS/dt = 0 gives the critical numbers t = 0, (c) L et S = = 1 + 3 cos t . Then dt dt S /2, , 3/2. W hen t = 0 and , the rate ds/dt = 1 + 3(1) = 2 ft/sec is fastest. W hen t = /2 and 3/2, the rate ds/dt = 1 + 3(0) = 1 ft/sec is slowest. The curve is shown in Fig. 38-4.

Fig. 38 -4
3 , 3); (c) (0, 0). Find the curvature of the parabola y2 = 12x at the points: (a) (3, 6); (b) ( 4

5.

dy dy 6 d2y dy 36 = 62 = 36 = ; so 1 + dx = 1 + 2 and y dx y y3 dx 2 y dx 2 dy d2y 1/ 6 = 1, so K = 23/ 2 = 24 . (a) A t (3, 6): 1 + = 2 and dx 6 dx 2 2 d2y dy 4 5 3 (b) A t ( 4 , 3): 1 + = 5 and 2 = 4 , so K = 43//3 = . dx 3 75 5 2 dx 2 2 y dy = 1 , and K = 1 . (c) A t (0, 0), is undefined. B ut dx = = 0, 1 + dx = 1, d x 6 dy 6 dx dy dy 2 6
2

6.

Find the curvature of the cy cloid x =

sin , y = 1 cos

at the highest point of an arch. (S ee Fig. 38-5.)

Fig. 38 -5

31 6

CHAPTER 38

Curvature

To find the highest point on the interval 0 < x < 2; dy/d = sin , so that the critical number on the interval is x = . S ince d2y/d 2 = cos < 0 when = , the point = is a relative maximum point and is the highest point of the curve on the interval. To find the curvature,
dx = 1 cos , d dy = sin , d dy = sin , dx 1 cos d2y d sin d = 1 = dx 2 d 1 cos dx (1 cos )2

A t 7.

1 = , dy/dx = 0, d 2 y / dx 2 = 1 4 , and K = 4 .

Find the curvature of the cissoid y2(2 x) = x3 at the point (1, 1). (S ee Fig. 38-6.)

Fig. 38 -6

D ifferentiating the given equation implicitly with respect to x, we obtain

y 2 + (2 x)2 yy = 3x 2

(1)

and
2 yy + (2 x)2 yy + (2 x)2( y )2 2 yy = 6 x

(2)

From (1), for x = y = 1, 1 + 2y = 3 and y = 2. S imilarly , from (2), for x = y = 1 and y = 2, we find y = 3. Then K = 3 /(1 + 4)3/ 2 = 3 5 / 25. 8. Find the point of greatest curvature on the curve y = ln x.
dy 1 = dx x

and

d2y = 12 . dx 2 x

S o,

K=

x (1 + x 2 )3/ 2

and

dK = 2 x 2 1 dx (1 + x 2 )5/ 2

The critical number is, therefore, x = 1 . The required point is 1 , ln 2 . 2 2 2

9.

Find the coordinates of the center of curvature C of the curve y = f (x) at a point P(x, y) at which y 0. (S ee Fig. 38-3.) The center of curvature C( , ) lies: (1) on the normal line at P and (2) at a distance R from P measured toward the concave side of the curve. These conditions give, respectively ,
[1 + ( y ) 2 ] 3 y = 1 ( x) and ( x)2 + ( y)2 = R 2 = y ( y )2

From the first,

x = y( y). S ubstitution in the second y ields

( y) 2 [ 1 + ( y ) 2 ] = [1 + ( y ) 2 ] 3 ( y )2

and, therefore,

y=

1 + ( y ) 2 y

CHAPTER 38

Curvature

31 7

P,

To determine the correct sign, note that, when the curve is concave upward, y > 0 and, since C then lies above y > 0. Thus, the proper sign in this case is +. (Y ou should show that the sign is also + when y < 0.) Thus,

= y+

1 + ( y ) 2 y

and = x

y [1 + ( y )2 ] y

10 . Find the equation of the circle of curvature of 2xy + x + y = 4 at the point (1, 1). D ifferentiating y ields 2y + 2xy + 1 + y = 0. A t (1, 1), y = 1 and 1 + (y)2 = 2. D ifferentiating again y ields 4y + 2xy + y = 0. A t (1, 1), y = 4 3. Then
K = 4 /3 , 2 2 R= 3 2 , 2

= 1

1(2) 5 = , 4 /3 2

= 1+ 2 = 5 4 /3 2

9 5 2 2 The required equation is (x )2 + (y )2 = R2 or ( x 5 2 ) + ( y 2 ) = 2.

11. Find the equation of the evolute of the parabola y2 = 12x. A t P(x, y):
dy 6 = = dx y 3 , x dy 1 + = 1 + 36 = 1+ 3 , dx x y2
2

d2y 3 = 3/ 2 = 36 2x y3 dx 2

Then

= x
and

3 / x (1 + 3/x) 2 3 ( x + 3) = x+ = 3x + 6 3 3/2 x 3/ 2

= y+

1 + 36/y 2 y3 + 36 y y3 = y = 36 36 36/y3

The equations = 3x + 6, = y3/36 may be regarded as parametric equations of the evolute with x and y, connected by the equation of the parabola, as parameters. H owever, it is relatively simple in this problem to eliminate the parameters. Thus, x = ( 6)/3, y = 3 36 , and substituting in the equation of the parabola, we have
(36 )2 / 3 = 4( 6) or 81 2 = 4( 6)3

The parabola and its evolute are shown in Fig. 38-7.

Fig. 38 -7

31 8

CHAPTER 38

Curvature

12 . Find the equation of the evolute of the curve x = cos A t P(x, y):
dx = cos , d dy = sin , d

sin , y = sin

cos .

dy = tan , dx

d 2 y sec 2 sec3 = = dx 2 cos

Then
2 = x tan 3sec = x sin = cos (sec )/

and

= y+
and = cos , = sin

sec 2 = y + cos = sin (sec3 )/

are parametric equations of the evolute (see Fig. 38-8).

Fig. 38 -8

13 . D erive formula (38.1). tan is the slope of the tangent line and, therefore,
dy = tan . dx d dy = d dy d ds dx d dx ds

S o,

H ence
d dy dx = sec 2 d dx dx ds ds

This y ields
d2y dx 2 dy 2 d = 1+ 2 dx ds dy 1+ dx 1

from which
d2y d = dx 2 2 3/ 2 ds dy 1+ dx

CHAPTER 38

Curvature

31 9

S U PPLEMEN T AR Y PR O B LEMS
In P roblems 1416, find ds and ds . dx dy 14 . x2 + y2 = 25 Ans.
ds = dx ds = dx
1 2

5 , ds = 25 x 2 dy 4 + 9 x , ds = dy

5 25 y 2 4 + 9 y2/3 3y1/ 3
1/ 3

15 . y2 = x3

Ans.

16 . x2/3 + y2/3 = a2/3 In P roblems 17 and 18, find ds . dx 17 . 6xy = x4 + 3

Ans.

ds = (a / x)1/ 3, ds = a dx dy y

Ans.

ds = x 4 + 1 dx 2x2 ds = ( x + 2a) /3a dx

18 . 27ay2 = 4(x a)3 In P roblems 1922, find ds. dt 19 . x = t2, y = t3

Ans.

Ans.

t 4 + 9t 2

2 0 . x = 2 cos t, y = 3 sin t

Ans.

4 + 5 cos 2 t

2 1. x = cos t, y = sin t

Ans. 1

2 2 . x = cos3 t, y = sin3 t

Ans.

3 2

sin 2t

2 3 . Find the curvature of each curve at the given points: (a) y = x3/3 at x = 0, x = 1, x = 2 (c) y = sin x at x = 0, x = 1 2 Ans. (a) 0 , (b) x2 = 4ay at x = 0, x = 2a 2 (d) y = e x at x = 0
2 / 8a; (c) 0 , 1 ; (d) 2

2 / 2, 4 17 / 289; (b) 1 /2a,

2 4 . S how (a) the curvature of a straight line is 0; (b) the curvature of a circle is numerically the reciprocal of its radius.
3 2 5 . Find the points of maximum curvature of (a) y = ex; (b) y = 1 3 x .

Ans.

1 (a) x = 1 2 ln 2; (b) x = 1/ 4 5

2 6 . Find the radius of curvature of (a) x3 + xy2 6y2 = 0 at (3, 3). (b) x = 2a tan , y = a tan2 at (x, y). (c) x = a cos4 , y = a sin4 at (x, y). Ans. (a) 5 5 ; (b) 2a | sec3 |; (c) 2a(sin4 + cos4 )3/2

32 0

CHAPTER 38

Curvature

2 7 . Find the center of curvature of (a) P roblem 26(a); (b) y = sin x at a maximum point. Ans. (a) C(7 , 8 ); (b) C , 0 2

( )

2 8 . Find the equation of the circle of curvature of the parabola y2 = 12x at the points (0, 0) and (3, 6). Ans. (x 6 )2 + y2 = 36 ; (x 1 5 )2 + ( y + 6 )2 = 28 8

2 9 . Find the equation of the evolute of (a) b2x2 + a2y2 = a2b2; (b) x2/3 + y2/3 + a2/3; (c) x = 2 cos t + cos 2t, y = 2 sin t + sin 2t. Ans. (a) (a )2/3 + (b )2/3 = (a2 b2)2/3; (b) ( + )2/3 + ( )2/3 = 2a2/3; (c) = 1 3 (2 cos t cos 2t ), =1 ( 2 sin t sin 2 t ) 3

C H A P TE R 4 1

Pol ar Coordinates
The position of a point in a plane may be described by its coordinates (x, y) with respect to a given rectangular coordinate system. Its position may also be described by choosing a fixed point O and specifying the directed distance = O and the angle that O makes with a fixed half-line OX . (See Fig. 41-1.) This is the p o lar co o rdinate system . The point O is called the p o le, and OX is called the p o lar axis.

Fig. 4 1-1

To each number pair (, ) there corresponds one and only one point. The converse is not true. For example, (1, 0 ) and (1, 2) describe the same point, on the polar axis and at a distance 1 from the pole. That same point also corresponds to (1, ). (W hen is negative, the point corresponding to (, ) is obtained as follows: R otate the polar axis OX through radians (counterclockwise if is positive and clockwise if is negative) to a new position OX and then move | | units on the half-line opposite to OX .) In general, a point with polar coordinates (, ) also can be described by (, 2n) and (, (2n + 1)), where n is any nonnegative integer. In addition, the pole itself corresponds to (0 , ), with arbitrary .
EXAMPLE 4 1.1: ordinates 1, 3 . 2

A p o lar eq uatio n of the form = f ( ) or F(, ) = 0 determines a curve, consisting of those points corresponding to pairs (, ) that satisfy the eq uation. For example, the eq uation = 2 determines the circle with center at the pole and radius 2. The eq uation = 2 determines the same set of points. In general, an eq uation = c, where c is a constant, determines the circle with center at the pole and radius |c|. A n eq uation = c determines the line through the pole obtained by rotating the polar axis through c radians. For example, = /2 is the line through the pole and perpendicular to the polar axis.
(2, p ) 2

( )

In Fig. 41-2, several points and their polar coordinates are shown. N ote that point C has polar co-

3p ) ( 1 , 4

(1 , p) 4

x (1 ,0 )

3p ) ( 1 , 2

Fig. 41-2

  ! % (% ,  121 3,

339

340

CHAPTER 41 Polar Coordinates

Polar and Rectangular Coordinates

Given a pole and polar axis, set up a rectangular coordinate system by letting the polar axis be the positive x axis and letting the y axis be perpendicular to the x axis at the pole. (See Fig. 41-3.) Then the pole is the origin of the rectangular system. If a point has rectangular coordinates (x, y) and polar coordinates (, ), then
x = cos
and

y = sin

(41.1)

These equations entail

2 = x 2 + y2

and

tan =

y x

(41.2)

Fig. 41-3 E X A M PL E 4 1 .2 : C onsider the polar curve = cos . M ultiplying by , w e get 2 = cos . H ence, x2 + y2 = x holds for the rectangular coordinates of points on the 1 2 2 curve. That is equivalent to x2 x + y2 = 0 and completion of the square w ith respect to x yields ( x 1 2) + y = 4. 1 1 H ence, the curve is the circle w ith center at ( 2 , 0) and radius 2 . N ote that, as varies from 0 to /2, the upper semicircle is traced out from (1, 0) to (0, 0), and then, as varies from to , the low er semicircle is traced out from (0, 0) 2 back to (1, 0). This w hole path is retraced once more as varies from to 2. Since cos has a period of 2, w e have completely described the curve. E X A M PL E 4 1 .3 : C onsider the parabola y = x2. In polar coordinates, w e get sin = 2 cos2 , and, therefore, = tan sec , w hich is a polar equation of the parabola.

S om e T y p ical Polar Curv es

(a) (b) (c) (d) C L R L ardioid: = 1 + sin . See Fig. 41-4(a). ima on: = 1 + 2 cos . See Fig. 41-4(b). ose w ith three petals: = cos 3. See Fig. 41-4(c). emniscate: 2 = cos 2. See Fig. 41-4(d).

A t a point on a polar curve, the angle from the radius vector O to the tangent T to the curve (see Fig. 41-5) is given by
tan = d = , d
w here

d d

(41.3)

For a proof of this equation, see P roblem 1. Tan plays a role in polar coordinates similar to that of the slope of the tangent line in rectangular coordinates.

CHAPTER 41 Polar Coordinates

341

Fig. 41-4

A ngle of Inclination
The angle of inclination of the tangent line to a curve at a point (, ) on it (see Fig. 41-5) is given by
tan =

cos + sin sin + cos

(41.4)

For a proof of this equation, see P roblem 4.

Points of Intersection
Some or all of the points of intersection of tw o polar curves = f1( ) and = f2() (or equivalent equations) may be found by solving
f1 ( ) = f2 ( )
(41.5)

342

CHAPTER 41 Polar Coordinates

Fig. 41-5 E X A M PL E 4 1 .4 : Find the points of intersection of = 1 + sin and = 5 3 sin . Setting 1 + sin = 5 3 sin , w e obtain sin = 1. Then = 2 and = /2. The only point of intersection is (2, /2). N ote that w e need not indicate the infinite number of other pairs that designate the same point.

Since a point may be represented by more than one pair of polar coordinates, the intersection of tw o curves may contain points that no single pair of polar coordinates satisfies (41.5).
E X A M PL E 4 1 .5 : Find the points of intersection of = 2 sin 2 and = 1. Solution of the equation 2 sin 2 = 1 yields sin 2 = 1 2 , and, therefore, w ithin [0, 2), = /12, 5 /12, 13 /12, 17 /12. W e have found four points of intersection: (1, /12), (1, 5 /12), (1, 13 /12), and (1, 17 /12). B ut the circle = 1 also can be represented as = 1. N ow solving 2 sin 2 = 1, w e get sin 2 = 1 2 and, therefore, = 7 /12, 11 /12, 19 /12, and 23 /12. H ence w e get four more points of intersection (1, 7 /12), (1, 11 /12), (1, 19 /12), and (1, 23 /12).

W hen the pole is a point of intersection, it may not appear among the solutions of (41.5). The pole is a point of intersection w hen there exist 1 and 2 such that f1(1) = 0 = f2(2).
E X A M PL E 4 1 .6 : Find the points of intersection of = sin and = cos . From the equation sin = cos , w e obtain the points of intersection ( 2 / 2, /4) and ( 2 / 2, 5 /4). H ow ever, both curves contain the pole. O n = sin , the pole has coordinates (0, 0), w hereas, on = cos , the pole has coordinates (0, /2). E X A M PL E 4 1 .7 : Find the points of intersection of = cos 2 and = cos . Setting cos 2 = cos and noting that cos 2 = 2 cos2 1, w e get 2 cos2 cos 1 = 0 and, therefore, (cos 1) 1 (2 cos + 1) = 0. So, cos = 1 or cos = 1 2 . Then = 0, 2 /3, 4 /3, yielding points of intersection (1, 0), ( 2 , 2 /3), 1 and ( 2 , 4 /3). B ut the pole is also an intersection point, appearing as (0, /4) on = cos 2 and as (0, /2) on = cos .

A ngle of Intersection
The angle of intersection, , of tw o curves at a common point (, ), not the pole, is given by
tan = tan 1 tan 2 1 + tan 1 tan 2

(41.6)

w here 1 and 2 are the angles from the radius vector O to the respective tangent lines to the curves at . (See Fig. 41-6.) This formula follow s from the trigonometric identity for tan(1 2), since = 1 2 .

Fig. 41-6

CHAPTER 41 Polar Coordinates

343

E X A M PL E 4 1 .8 : Find the (acute) angles of intersection of = cos 2 and = cos . The points of intersection w ere found in E xample 7. W e also need tan 1 and tan 2. For = cos , formula (41.3) yields tan 1 = cot . For = cos 2, formula (41.3) yields tan 2 = 1 2 cot 2 . A t the point (1, 0), tan 1 = cot 0 = and, lik ew ise, tan 2 = . Then 1 = 2 = /2 and, therefore, = 0. A t the point 1 , 2 , tan 1 = 3 / 3 and tan 2 = 3 / 6 . So, by (41.6), 2 3

tan =

( 3 / 3) + ( 3 / 6) 3 3 = 1 (1/ 6) 5

and, therefore, the acute angle of intersection 46 6 . B y symmetry, this is also the acute angle of intersection at the point ( 1 2 , 4 /3). A t the pole, on = cos , the pole is given by = /2. O n = cos 2, the pole is given by = /4 and = 3 /4. Thus, at the pole there are tw o intersections, the acute angle being /4 for each.

T h e D eriv ativ e of th e A rc L ength

The derivative of the arc length is given by
ds = 2 + ( ) 2 d d w here = and it is understood that s increases w ith . d For a proof, see P roblem 20.
(41.7)

Curv ature
The curvature of a polar curve is given by
K=

2 + 2( )2 [ 2 + ( ) 2 ] 3 / 2

(41.8 )

For a proof, see P roblem 17.

S O L V E D PRO B L E M S
1.

d D erive formula (41.3): tan = d = , w here . d d In Fig. 41-7, Q( + , + ) is a point on the curve near P. From the right triangle PS Q, sin sin SP = tan = S P = = = S Q OQ OS + cos (1 cos ) + sin + 1 cos

N ow as Q P along the curve, 0, OQ OP, PQ PT, and .

Fig. 41-7

344
A s 0, sin 1 and 1 cos 0. Thus,
tan = lim tan =
0

CHAPTER 41 Polar Coordinates

= d dp /d d

In P roblems 2 and 3, use formula (41.3) to find tan for the given curve at the given point. 2.

= 2 + cos at = . (See Fig. 41-8 .) 3 3 5 , and tan = = 5 . A t = , = 2+ 1 2 = 2 , = sin = 3 2 3

Fig. 41-8

3.

= 2 sin 3 at = . (See Fig. 41-9.) 4 A t = , = 2 1 = 2 , = 6 cos 3 = 6 1 = 3 2 and tan = = 1 3. 4 2 2

Fig. 41-9

4.

D erive formula (41.4): tan =

cos + sin . sin + cos From Fig. 41-7, = + and d + sin d cos tan + tan tan = tan( + ) = = 1 tan tan 1 d sin d cos
d cos + sin cos + sin d = = d sin + cos cos sin d

5.

Show that, if = f () passes through the pole and 1 is such that f (1) = 0, then the direction of the tangent line to the curve at the pole (0, 1) is 1. (See Fig. 41-10.)

Fig. 41-10

CHAPTER 41 Polar Coordinates

345

A t (0, 1), = 0, and = f (1). If 0

tan = 0 + f (1 )sin 1 cos + sin = tan 1 = sin + cos 0 + f (1 )cos1

If = 0,
tan = lim f ( )sin = tan 1 f ( )cos

In P roblems 6 8 , find the slope of the given curve at the given point. 6. = 1 cos at = . (See Fig. 41-11.) 2

Fig. 41-11

A t = , 2
sin = 1, cos = 0, = 1, = sin = 1 tan =

and

cos + sin = 1 0 + 1 1 = 1 sin + cos 1 1 + 1 0

7.

= cos 3 at the pole. (See Fig. 41-12.) W hen = 0, cos 3 = 0. Then 3 = /2, 3 /2, 5 /2, and = /6, /2, 5 /6. B y P roblem 5, tan = 1/ 3 , , and 1 3 .

Fig. 41-12

8.

= a at = . 3 2 2 A t = /3: sin = 3 / 2 , cos = 1 2 , = 3a/, and = a/ = 9a/ . Then

tan =

cos + sin 3 3 = sin + cos 3 + 3

346

CHAPTER 41 Polar Coordinates

9.

Investigate = 1 + sin for horiz ontal and vertical tangents. (See Fig. 41-13.)

Fig. 41-13

A t P(, ):
tan = cos (1 + 2 sin ) (1 + sin )cos + cos sin = (sin + 1)(2 sin 1) (1 + sin )sin + cos 2

W e set cos (1 + 2 sin ) = 0 and solve, obtaining = /2, 3 /2, 7 /6, and 11 /6. W e also set (sin + 1) (2 sin 1) = 0 and solve, obtaining = 3 /2, /6, and 5 /6. For = /2: There is a horiz ontal tangent at (2, /2). 1 For = 7 /6 and 11 /6: There are horiz ontal tangents at ( 1 2 , 7 /6) and ( 2 , 11 /6). 3 3 For = /6 and 5 /6: There are vertical tangents at ( 2 , /6) and ( 2 , 5 /6). For = 3 /2: B y P roblem 5, there is a vertical tangent at the pole. 1 0 . Show that the angle that the radius vector to any point of the cardioid = a(1 cos ) mak es w ith the curve is one-half that w hich the radius vector mak es w ith the polar axis. A t any point P( , ) on the cardioid,

= a sin
So = 1 2 .

and

tan =

1 cos = = tan . sin 2

In P roblems 11 13, find the angles of intersection of the given pair of curves. 1 1 . = 3 cos , = 1 + cos . (See Fig. 41-14.)

Fig. 41-14

Solve 3 cos = 1 + cos for the points of intersection, obtaining (3/2, /3) and (3/2, 5 /3). The curves also intersect at the pole. For = 3 cos : For = 1 + cos :

= 3 sin = sin

and and

tan 1 = cot tan 2 = 1 + cos sin

CHAPTER 41 Polar Coordinates

347

A t = /3, tan 1 = 1 3 , tan 2 = 3 , and tan = 1/ 3 . The acute angle of intersection at ( 3 2 , /3) and, by symmetry, at ( 3 , 5 /3) is /6. 2 A t the pole, either a diagram or the result of P roblem 5 show s that the curves are orthogonal.
2 1 1 2 . = sec 2 1 2 , = 3 csc 2 . 2 1 Solve sec 2 = 3 csc 2 1 2 for the points of intersection, obtaining (4, 2 /3) and (4, 4 /3).

For = sec 2 1 2 : For = 3 csc 2 1 2 :

1 = sec 2 1 2 tan 2 2 1 = 3 csc 2 cot 1 2

and and

tan 1 = cot 1 2 tan 2 = tan 1 2

A t = 2 /3, tan 1 = 1/ 3 , and tan 2 = 3 , and = 1 2 ; the curves are orthogonal. L ik ew ise, the curves are orthogonal at = 4 /3. 1 3 . = sin 2, = cos . (See Fig. 41-15.) The curves intersect at the points ( 3 / 2 , /6) and ( 3 /2 , 5 /6) and the pole. For = sin 2 : For = cos :

= 2 cos 2 = sin

and and

tan 1 = 1 2 tan 2 tan 2 = cot

A t = /6, tan 1 = 3 / 2 , tan 2 = 3 , and tan 1 = 3 3 . The acute angle of intersection at the point ( 3 / 2 , /6) is = tan 1 3 3 = 79 6. Similarly, at = 5 /6, tan 1 = 3 / 2 , tan 2 = 3 , and the angle of intersection is tan 1 3 3 . A t the pole, the angles of intersection are 0 and /2.

Fig. 41-15

In P roblems 14 16, find 1 4 . = cos 2.

ds at the point P (, ). d

= 2 sin 2 and ds = 2 + ( )2 = cos 2 2 + 4 sin 2 2 = 1 + 3 sin 2 2 d

1 5 . (1 + cos ) = 4. D ifferentiation yields sin + (1 + cos ) = 0. Then

sin = 4 sin 1 + cos (1 + cos )2

and

4 2 ds = 2 + ( )2 = d (1 + cos )3/ 2

348

CHAPTER 41 Polar Coordinates

1 6 . = sin 3 . (A lso evaluate ds at = .) 3 d 2

1 = sin 2 1 3 cos 3

()

and

ds = sin 6 1 + sin 4 1 cos 2 1 = sin 2 1 3 3 3 3 d

1 2 1 A t = 1 2 , ds /d = sin 6 = 4 .

1 7 . D erive formula (41.8 ): K =

2 + 2( )2 . [ 2 + ( ) 2 ] 3 / 2 B y definition, K = d . N ow , = + and, therefore, ds

d = d + d = d + d d = d 1 + d ds ds ds ds d ds ds d

w here = tan 1 . A lso, d [( )2 ] / ( )2 ( )2 = ; = 2 d 1 + ( / )2 + ( ) 2

so 1+

( )2 2 + 2( )2 d =1+ 2 = d + ( ) 2 2 + ( ) 2

Thus,

2 + 2( )2 d 1 + d /d 1 + d /d = = = K = d 1 + 2 2 ds d ds /d [ 2 + ( ) 2 ] 3 / 2 + ( )

1 8 . L et = 2 + sin . Find the curvature at the point P(, )

K=

2 + 2( )2 (2 + sin )2 + 2 cos 2 + (sin )(2 + sin ) 6(1 + sin ) = = [ 2 + ( ) 2 ] 3 / 2 [(2 + sin )2 + cos 2 ] 3/ 2 (5 + 4 sin )3/ 2

1 9 . L et (1 cos ) = 1. Find the curvature at = and = 4 . 3 2

sin (1 cos )2

and

cos + 2 sin 2 ; (1 cos )2 (1 cos )3

so

K = sin 3 2

A t = /2, K = (1/ 2 )3 = 2 / 4 ; at = 4 /3, K = ( 3 / 2)3 = 3 3 /8 . 2 0 . D erive formula (41.7): ds = 2 + ( )2 . d C onsider as a function of . From x = cos and y = sin , w e get dx/d = sin + (cos ) and dy/d = cos + (sin ). H ence,
dx = [ sin + ( ) cos 2 sin cos ] (d )
2 2 2 2 2

and Thus,

dy = [ 2 cos 2 + ( )2 sin 2 + 2 sin cos ] d dy = dx + ( dds ) ( d ) d

2 2 2

= 2 + ( ) 2

Since s increases w ith , ds > 0 and w e obtain formula (41.7). d 2 1 . For = cos 2, find ds at = . (A ssume as usual that s increases w ith .) d 4 d . B y Formula (41.7), = 2 sin 2 = d ds = cos 2 (2 ) + 4 sin 2 (2 ) = 1 + 3 sin 2 (2 ) d
= 1 + 3 sin 2 ( /2) = 2

CHAPTER 41 Polar Coordinates

349

S U PPL E M E N T A RY PRO B L E M S
In P roblems 22 25, find tan for the given curve at the given points.

2 2 . = 3 sin at = 0, = 3 /4

Ans.

3; 3 2 1

2 3 . = a(1 cos ) at = /4, = 3 /2

Ans.

2 1; 1

2 4 . (1 cos ) = a at = /3, = 5 /4

Ans.

3 / 3; 1 + 2

2 5 . 2 = 4 sin 2 at = 5 /12, = 2 /3

Ans.

1 3 ; 3

In P roblems 26 29, find tan for the given curve at the given point. 2 6 . = 2 + sin at = /6 Ans.
3 3

2 7 . 2 = 9 cos 2 ; at = /6

Ans.

2 8 . = sin3 (/3) at = /2

Ans.

2 9 . 2(1 sin ) = 3 at = /4

Ans.

1+ 2

3 0 . Investigate = sin 2 for horiz ontal and vertical tangents. Ans. horiz ontal tangents at = 0, , 54 44, 125 16, 234 44, 305 16; vertical tangents at = /2, 3 /2, 35 16, 144 44, 215 16, 324 44

In P roblems 31 33, find the acute angles of intersection of each pair of curves. 3 1 . = sin , = sin 2 Ans.

= 79 6 at = /3 and 5 /3; = 0 at the pole

3 2 . = 2 sin , 2 = cos 2

Ans.

= /3 at = /6, 5 /6; = /4 at the pole

3 3 . 2 = 16 sin 2, 2 = 4 csc 2

Ans.

= /3 at each intersection

3 4 . Show that each pair of curves intersects at right angles at all points of intersection. (a) = 4 cos , p = 4 sin (c) 2 cos 2 = 4, 2 sin 2 = 9 (b) = e, = e (d) = 1 + cos , = 1 cos

3 5 . Find the angle of intersection of the tangents to = 2 4 sin at the pole. Ans. 2 /3

35 0

CHAPTER 41 Polar Coordinates

3 6 . Find the curvature of each of these curves at P(, ): (a) = e; (b) = sin ; (c) 2 = 4 cos 2; (d) = 3 sin + 4 cos . Ans. (a) 1/( 2 e ) ; (b) 2; (c)
3 2

cos 2 ; (d)

2 5

3 7 . Find ds for the curve = a cos . d Ans. a

3 8 . Find ds for the curve = a(1 + cos ). d Ans.

a 2 + 2 cos

3 9 . Suppose a particle moves along a curve = f () w ith its position at any time t given by = g(t), = h(t). 2 2 (a) M ultiply the equation ds = 2 + ( )2 obtained in P roblem 20 by d to obtain d dt 2 2 2 d ds d v2 = = 2 + . dt dt dt d d d (b) From tan = = /dt , obtain sin = d and cos = 1 . v dt d d /dt v dt

( )

( )

( )

( )

In P roblems 40 43, find all points of intersection of the given equations. 4 0 . = 3 cos , = 3 sin Ans. (0, 0), ( 3 2 / 2, /4)

4 1 . = cos , = 1 cos

Ans.

1 (0, 0), ( 1 2 , /3), ( 2 , /3)

4 2 . = , =

Ans.

(, ), (, )
(2n + 1) (0, 0), 2 , for n = 0, 1, 2, 3, 4, 5 6 2

4 3 . = sin 2, = cos 2

Ans.

4 4 . (GC ) Sk etch the curves in P roblems 40 43, find their graphs on a graphing calculator, and check your answ ers to P roblems 40 43.

4 5 . (GC ) Sk etch the graphs of the follow ing equations and then check your answ ers on a graphing calculator: (a) = 2 cos 4 (b) = 2 sin 5 2 (d) = 2(1 cos ) (e) = 1 + cos (g) = 2 sec (h) = 2 (In parts (g) and (h), look for asymptotes.) (c) 2 = 4 sin 2 (f ) 2 = 1

4 6 . C hange the follow ing rectangular equations to polar equations and sk etch the graphs: (a) x2 4x + y2 = 0 (d) x = a Ans. (b) 4x = y2 (e) y = b (c) xy = 1 (f ) y = mx + b

(a) = 4 cos ; (b) = 4 cot csc ; (c) 2 = sec csc ; (d) = a sec ; (e) = b csc ; b (f ) = sin m cos

CHAPTER 41 Polar Coordinates

35 1

4 7 . (GC ) C hange the follow ing polar equations to rectangular coordinates and then sk etch the graph. (V erify on a graphing calculator.) (a) = 2c sin ; (b) = ; (c) = 7 sec Ans. (a) x2 + ( y c)2 = c2; (b) y = x tan( x 2 + y 2 ) ; (c) x = 7

4 8 . (a) Show that the distance betw een tw o points w ith polar coordinates (1, 1) and (2, 2) is
2 2 12 cos(1 2 ) 12 + 2

(b) W hen 1 = 2, w hat does the distance simplify to? E xplain w hy this is so. Ans. |1 2|

(c) W hen 1 2 = , w hat does the formula yield? E xplain the signficance of the result. 2 Ans.
2 12 + 2

(d) Find the distance betw een the points w ith the polar coordinates (1, 0) and 1, . 4 Ans.
2 2

( )

4 9 . (a) L et f be a continuous function such that f () 0 for < < . L et A be the area of the region bounded by the lines = and = , and the polar curve = f (). D erive the formula A = 1 ( f ( ))2 d = 1 2 d . 2 2 (H int: D ivide [ , ] into n equal parts, each equal to . E ach resulting subregion has area approximately * * 2 equal to 1 2 ( f ( i )) , w here i is in the ith subinterval.) (b) Find the area inside the cardioid = 1 + sin . (c) Find the area of one petal of the rose w ith three petals, = cos 3. ( H int: Integrate from 6 to 6 .)

C H A P TE R 4 2

Infinite Sequences
Infinite S eq uences
An infinite sequence sn is a function whose domain is the set of positive integers; sn is the value of this function for a given positive integer n. Sometimes we indicate sn just by writing the first few terms of the sequence s1, s2, s3, . . . , sn, . . . . We shall consider only sequences where the values sn are real numbers.
E X A M PL E 4 2 .1 : (a) 1 is the sequence 1, 1 , 1 , . . . , 1 , . . .. 2 3 n n

(b)

is the sequence 1 , 1 , 1 , . . . , 1n , . . . . 2 4 8 2 (c) n 2 is the sequence of squares 1, 4, 9 , 16 , . . ., n2, . . . . (d) 2n is the sequence of positive even integers 2, 4, 6 , 8, . . . , 2n, . . .. (e) 2n 1 is the sequence of positive odd integers 1, 3, 5 , 7 , . . ..

(1 2)

L im it of a S eq uence
If sn is an infinite sequence and L is a number, then we say that lim sn = L if sn gets arbitrarily close to L n+ as n increases without bound. F rom a more precise standpoint, lim sn = L means that, for any positive real number > 0 , there ex ists n+ a positive integer n0 such that, whenever n n0, we have |sn L| < . T o illustrate what this means, place the points L, L , and L + on a coordinate line (see F ig. 42-1), where is some positive real number. N ow, if we place the points s1, s2, s3, . . . on the coordinate line, there will eventually be an index n0 such that sn0, sn0 +1, sn0 + 2 , sn0 +3 , . . . and all subsequent terms of the sequence will lie inside the interval ( L , L + ).
s1 s2 sm sm + 1 L L L+

Fig. 42-1

If lim sn = L , then we say that the sequence sn co nv erg es to L. If there is a number L such that sn n+ converges to L, then we say that sn is co nv erg ent. When sn is not convergent, then we say that sn is div erg ent.
1 is convergent, since lim 1 = 0. T o see this, observe that 1/n can be made arbitrarily close to E X A M PL E 4 2 .2 : n n+ n 0 by mak ing n large enough. T o get an idea of why this is so, note that 1/10 = 0.1, 1/100 = 0.01, 1/1000 = 0.001, and so on. T o check that the precise definition is satisfied, let be any positive number. T ak e n0 to be the smallest positive integer greater than 1/. So, 1/ < n0. H ence, if n n0, then n > 1/ and, therefore, 1/n < . T hus, if n n0, |1/n 0| < . T his proves lim 1 = 0 . n+ n
E X A M PL E 4 2 .3 : 2n is a divergent sequence, since lim 2n L for each real number L. In fact, 2n gets arbitrarily n+ large as n increases.

35 2   ! % (% ,  121 3,

CHAPTER 42

Infinite S equences

35 3

We write lim sn = + if sn gets arbitrarily large as n increases. In such a case, we say that sn diverges to n+ +. M ore precisely, lim sn = + if and only if, for any number c, no matter how large, there ex ists a positive n+ integer n0 such that, whenever n n0, we have sn > c. L ik ewise, we write lim sn = if sn gets arbitrarily small as n increases. In such a case, we say that sn n+ diverges to . M ore precisely, lim sn = if and only if, for any number c, no matter how small, there n+ ex ists a positive integer n0 such that, whenever n n0, we have sn < c. We shall write lim sn = if lim | sn| = + , that is, the magnitude of sn gets arbitrarily large as n increases.
n+ n+

E X A M PL E 4 2 .4 : (a) lim 2n = + ; (b) lim (1 n)3 = ; (c) lim (1)n (n 2 ) = . N ote that, in case (c), the sequence n+ n+ n+ converges neither to + nor to . E X A M PL E 4 2 .5 : T he sequence (1)n is divergent, but it diverges neither to +, nor to , nor to . Its values oscillate between 1 and 1.

A sequence sn is said to be bo unded abo v e if there is a number c such that sn c for all n, and sn is said to be bo unded belo w if there is a number b such that b sn for all n. A sequence sn is said to be bo unded if it is bounded both above and below. It is clear that a sequence sn is bounded if and only if there is a number d such that |sn | d for all n.
E X A M PL E 4 2 .6 : (a) T he sequence 2n is bounded below (for ex ample, by 0) but is not bounded above. (b) T he sequence (1)n is bounded. N ote that (1)n is 1, 1, 1, . . . So, |(1)n| 1 for all n. T h eorem 4 2 .1 :

E very convergent sequence is bounded.

F or a proof, see P roblem 5 . T he converse of T heorem 42.1 is false. F or ex ample, the sequence (1)n is bounded but not convergent. Standard arithmetic operations on convergent sequences yield convergent sequences, as the following intuitively obvious results show.
T h eorem 4 2 .2 :

Assume lim sn = c and lim tn = d . T hen:

n+ n+

(a) (b) (c) (d) (e) (f )

n+

lim k = k , where k is a constant. lim ksn = kc , where k is a constant. lim (sn + tn ) = c + d . lim (sn tn ) = c d . lim (sn tn ) = cd . lim (sn /tn ) = c /d provided that d 0 and tn 0 for all n.

n+ n+ n+

n+ n+

F or proofs of parts (c) and (e), see P roblem 10. T he following facts about sequences are intuitively clear.
T h eorem 4 2 .3 :

If lim sn = and sn 0 for all n, then lim 1 = 0 . n+ n+ sn

F or a proof, see P roblem 7 .

T h eorem 4 2 .4 :

(a) If |a| > 1, then lim a n = .

n+

In particular, if a > 1, then lim a n = + . n+ (b) If |r | < 1, then lim r n = 0 .

n+

F or proofs, see P roblem 8.

T h eorem 4 2 .5 (S q ueez e T h eorem ): then lim tn = L .
n+

If lim sn = L = lim un , and there is an integer m such that sn tn un for all n m,

n+ n+

F or a proof, see P roblem 11.

35 4

CHAPTER 42

Infinite S equences

Corollary 4 2 .6 :

If lim un = 0 and there is an integer m such that |tn| |un| for all n m, then lim t n = 0.
n+ n+ n+ n+

T his is a consequence of T heorem 42.5 and the fact that lim an = 0 is equivalent to lim |an| = 0 .
E X A M PL E 4 2 .7 :
n+

lim (1)n 12 = 0. T o see this, use C orollary 42.6 , noting that (1)n 12 1 and lim 1 = 0 . n n+ n n n

T h eorem 4 2 .7 : Assume that f is a function that is continuous at c, and assume that lim sn = c , where all the terms n+ sn are in the domain of f. T hen lim f (sn ) = f (c) .
n+

See P roblem 33. It is clear that whether or not a sequence converges would not be affected by deleting, adding, or altering a finite number of terms at the beginning of the sequence. C onvergence depends on what happens in the long run. We shall ex tend the notion of infinite sequence to the case where the domain of a sequence is allowed to be the set of nonnegative integers or any set consisting of all integers greater than or equal to a fix ed integer. F or ex ample, if we tak e the domain to be the set of nonnegative integers, then 2n + 1 would denote the 1 1 sequence of positive odd integers, and 1/ 2n would denote the sequence 1, 1 2 , 4 , 8 , . . ..

M onotonic S eq uences
(a) (b) (c) (d) (e) A sequence sn is said to be no ndecreasing if sn sn+ 1 for all n. A sequence sn is said to be increasing if sn < sn+ 1 for all n. A sequence sn is said to be no nincreasing if sn sn+ 1 for all n. A sequence sn is said to be decreasing if sn > sn+ 1 for all n. A sequence is said to be mo no to nic if it is either nondecreasing or nonincreasing.

C learly, every increasing sequence is nondecreasing (but not conversely), and every decreasing sequence is nonincreasing (but not conversely).
E X A M PL E 4 2 .8 : (a) T he sequence 1, 1, 2, 2, 3, 3, 4, 4, . . . is nondecreasing, but not increasing. (b) 1, 1, 2, 2, 3, 3, 4, 4, . . . is nonincreasing, but not decreasing.

An important basic property of the real number system is given by the following result. Its proof is beyond the scope of this book .
T h eorem 4 2 .8 :

E very bounded monotonic sequence is convergent.

T here are several methods for showing that a given sequence sn is nondecreasing, increasing, nonincreasing, or decreasing. L et us concentrate on the property that sn is increasing. M eth o d 1: S h o w th at sn+ 1 sn > 0.
E X A M PL E 4 2 .9 : C onsider sn =

3n . T hen s = 3(n + 1) = 3n + 3 . So, n +1 4n + 1 4(n + 1) + 1 4 n + 5

(12n 2 + 15 n + 3) (12n 2 + 15 n) sn+1 sn = 3n + 3 3n = 4n + 5 4n + 1 (4 n + 5 )(4 n + 1) = 3 >0 (4 n + 5 )(4 n + 1)

since 4n + 5 > 0 and 4n + 1 > 0

Method 2: When all sn > 0, show that sn+1/sn > 1.

EXAMPLE 42.10:

Using the same example sn =

sn+1 = 3n + 3 4n + 5 sn

3n as ab o v e, 4n + 1
2 2

3 4 n + 1 = 12n + 15n + 3 > 1, ) ( 4n3n+ 1 ) = 3n3+ n 4n + 5 12n + 15n

since 12n2 + 15n + 3 > 12n2 + 15n > 0.

CHAPTER 42 Infinite Sequences

355

Method 3: F ind a differentiable function f (x) such that f (n) = sn for all n, and show that f (x) > 0 for all x 1 (and, hence, that f is an increasing function for x 1).
EXAMPLE 42.11:

C o nsider sn =

3 3n again. L et f ( x) = 3x . T hen f ( x) = > 0 fo r all x. 4x + 1 4n + 1 (4 x + 1)2

S O LV ED PR O B LEMS
1. F o r each o f the fo llo w ing seq u ences, w rite a fo rmu la fo r the nth term and determine the limit (if it exists). It is assu med that n = 1, 2, 3, . . .. (a) 1 , 1 , 1 , 1 , . . .. (b ) 1 , 2 , 3 , 4 , . . 2 4 6 8 2 3 4 5 (c) 1, 1 , 1 , 1 , 1 , 1 , . . . (d) 0.9, 0.99, 0.999, 0.9999, . . . 2 3 4 5 6 2 3 4 (e) sin , sin , sin 3 , sin 2, sin 5 , . . . (f ) 2 , 3 , 4 , 5 , . . . 2 2 2 1 2 3 4

()()()

(a) sn = 1 ; lim 1 = 0 . 2n n + 2n

(b ) sn = n ; lim n = lim 1 1 = 1 lim 1 = 1 0 = 1. n + 1 n + n + 1 n + n +1 n + n + 1 (1)n+1 (1)n+1 = 0 . T his is intu itiv ely clear, b u t o ne can also apply T heo rem 42.3 to the seq u ence ; lim (c) sn = n n n + (1)n+1 n , since lim (1)n+1 n = . n + (d) sn = 1 1 n ; lim 1 1 n = 1 lim 1 n = 1 0 = 1. 10 n + 10 n + 10

N o te that lim

n +

1 = 0 b y v irtu e o f T heo rem 42.4(b). 10 n

(e) sn = sin n . N o te that the seq u ence co nsists o f repetitio ns o f the cy cle 1, 0, 1, 0 and has no limit. 2 (f ) sn = n + 1 ; lim n + 1 n n n + 2.

( )

1 = e b y (26.17). ( ) = lim (1 + n )
n n n +

E v alu ate lim sn in the fo llo w ing cases: n + 2 2 (a) sn = 5n2 4 n + 13 (b ) sn = 8n 3 2n + 5 3n 95n 7

(c)

3n + 7 n 3 2n 9

2 2 (a) R ecall that lim 5x 2 4 x + 13 = 5 b y C hapter 7, P ro b lem 13. T herefo re, lim 5n2 4 n + 13 = 5 . A similar 3 3 x + 3 x 95 x 7 n + 3n 95n 7 resu lt ho lds w henev er sn is a q u o tient o f po ly no mials o f the same degree. 2 2 (b ) R ecall that lim 8 x 3 = + b y C hapter 7, P ro b lem 13. T herefo re, lim 8n 3 = +. A similar resu lt ho lds 2 2 5 + +5 x n x + n + w henev er sn is a ratio nal fu nctio n w ho se nu merato r has greater degree than the deno minato r (and w ho se leading co efficients hav e the same sign). (c) R ecall that lim 3 3x + 7 = 0 b y C hapter 7, P ro b lem 13. T herefo re, lim 3 3n + 7 = 0 . T he same resu lt x + x 2 x 9 n + n 2n 9 ho lds w henev er sn is a ratio nal fu nctio n w ho se deno minato r has greater degree than the nu merato r.

3.

F o r each o f the fo llo w ing seq u encies, determine w hether it is no ndecreasing, increasing, no nincreasing, decreasing, o r no ne o f these. T hen determine its limit, if it exists. (a) sn = 5n 2 (b ) sn = nn (c) 1n 7n + 3 3 2
(7 x + 3)(5) (5x 2)(7) 29 (a) L et f ( x) = 5x 2 . T hen f ( x) = = > 0. 7x + 3 (7 x + 3)2 (7 x + 3)2 H ence, f(x) is an increasing fu nctio n and, therefo re, sn is an increasing seq u ence. 2 x x(ln x)2 x 1 x(ln 2) = (b ) L et f ( x) = xx . T hen f ( x) = . 22 x 2x 2

356

CHAPTER 42 Infinite Sequences

S ince ln 2 > 1 2 (b y (2 5 .1 2 )), x(ln2) > x/2 1, w hen x 2. T hu s, 1 x(ln2) < 0 w hen x 2 and, therefo re, f (x) < 0 w hen x 2. S o , f(x) is decreasing fo r x 2 and this implies that sn is decreasing fo r n 2. N o te that s1 = 1 2 = s2 . H ence, sn is no nincreasing. N o w let u s find the limit. B y L H pitals R u le,
n =0 1 lim xx = lim n x = 0 and, therefo re, lim n + 2 x + (ln 2)2 2

x +

(c)

< 1. H ence, s is decreasing. ( ) ( 31 ) = 1 3 T heo rem 42.4(b) tells u s that lim 1 = lim ( 1 ) = 0 . 3 3 sn+1 = 1 sn 3n+1
n n n n + n n +

4.

S ho w that the seq u ence sn =

1 3 5 7 . . . (2n 1) is co nv ergent. 2 4 6 8 . . . (2n) L et u s u se T heo rem 42.8. sn is b o u nded, since 0 < sn < 1. L et u s sho w that sn is decreasing. N o te that sn+1 = 1 3 5 7 (2n + 1) = s 2n + 1 < s 2 4 6 8 (2n + 2) n 2n + 2 n

5.

P ro v e T heo rem 42.1: E v ery co nv ergent seq u ence sn is b o u nded. L et lim sn = L . T ak e = 1. T hen there exists a po sitiv e integer n0 su ch that, w henev er n n0, w e hav e n + |sn L| < 1 . H ence, fo r n n0, u sing the triangle ineq u ality , w e get
|sn| = |(sn L ) + L | |sn L| + |L | < 1 + | L |

S o , if w e tak e M to b e the maximu m o f 1 + |L | and |s1 |, |s2 |, |s3 |, . . . , |sn0| , then |sn| M fo r all n. T hu s, sn is b o u nded. 6.
n! is div ergent. 2n ! = 1 2 3 . . . n = 1 3 4 . . . n > n fo r n > 4, the seq u ence is no t b o u nded. S o , b y T heo rem 42.1, S ince nn 222 . . . 2 2 2 2 2 2 2 the seq u ence canno t b e co nv ergent.

S ho w that the seq u ence

7.

P ro v e T heo rem 42.3: If lim sn = and sn 0 fo r all n, then lim 1 n = 0. n + n + s C o nsider any > 0 . S ince lim sn = , there exists so me po sitiv e integer m su ch that, w henev er n m,
n +

|sn | > 1

and, therefo re,

1 0 = 1 < . sn sn

So,

n +

lim 1 = 0. sn

8.

P ro v e T heo rem 42.4: (a) if |a| > 1, then lim a n = ; (b ) If |r | < 1 , then lim r n = 0 . n +
n +

(a) L et M > 0, and let |a | = 1 + b . S o , b > 0. N o w , |a|n = (1 + b)n = 1 + nb + . . . > 1 + nb > M w hen n M . b (b ) L et a = 1/r. S ince |r | < 1, |a | > 1 . B y part (a), lim a n = . H ence, lim (1/r n ) = . S o , b y T heo rem 42.3, n + n + lim r n = 0 .
n +

9.

P ro v e: lim 1n = 0 . n + 2 lim 2n = b y T heo rem 42.4(a). H ence, lim 1n = 0 b y T heo rem 42.3. n + n + 2

10 . P ro v e T heo rem 42.2(c) and (e). A ssu me lim sn = c and lim tn = d .

n + n +

CHAPTER 42 Infinite Sequences

357

(c) lim (sn + tn ) = c + d . L et > 0 . T hen there exist integers m1 and m2 su ch that |sn c| < / 2 fo r n m1 and n + |tn d | < / 2 fo r n m2. L et m b e the maximu m o f m1 and m2. S o , fo r n m, |sn c| < / 2 and |tn d | < / 2. H ence, fo r n m,
|(sn + tn ) (c + d )| = |(sn c) + (tn d )| |sn c| + |tn d | < + = 2 2

(e)

lim (sn tn ) = cd . S ince sn is co nv ergent, it is b o u nded, b y T heo rem 42.1 and, therefo re, there is a po sitiv e n + nu mb er M su ch that |sn | M fo r all n. L et > 0 . If d 0, there exists an integer m1 su ch that |sn c| < / 2 |d | fo r n m1 and, therefo re, |d ||sn c| < / 2 fo r n m1. If d = 0, then w e can cho o se m1 = 1 and w e w o u ld still hav e |d ||sn c| < / 2 fo r n m1. T here also exists m2 su ch that |tn d | < / 2 M fo r n m2. L et m b e the maximu m o f m1 and m2. If n m, |sn tn cd | = |sn (tn d ) + d (sn c)| |sn (tn d )| + |d (sn c)| = |sn ||tn d | + |d ||sn c| M = ( 2M ) + 2

11. P ro v e the S q u eez e T heo rem: If lim sn = L = lim un , and there is an integer m su ch that sn tn un fo r all n m, n + n + then lim tn = L . n + L et > 0. T here is an integer m1 m su ch that |sn L| < / 4 and |un L| < / 4 fo r n m1. N o w assu me n m1. S ince sn tn un, |tn sn| |un sn|. B u t
|un sn| = |(un L ) + ( L sn )| |un L| + |L sn| < + = 4 4 2 T hu s, |tn sn | < / 2. H ence, |tn L| = |tn sn | + (sn L )| |tn sn | + |sn L| < + < 2 4

S U PPLEMEN T AR Y PR O B LEMS
In each o f P ro b lems 12 29, determine fo r each giv en seq u ence sn w hether it is b o u nded and w hether it is no ndecreasing, increasing, no nincreasing, o r decreasing. A lso determine w hether it is co nv ergent and, if po ssib le, find its limit. (N ote: If the seq u ence has a finite limit, it mu st b e b o u nded. If it has an infinite limit, it mu st b e u nb o u nded.) 12 .
n+ 2 n sin n 4

Ans.

no ndecreasing; increasing fo r n 2; limit +

13 .

Ans.

b o u nded; no limit

14 . 3 n 2
n! 10 n ln n n

Ans.

increasing; limit +

15 .

Ans.

increasing fo r n 10; limit +

16 .

Ans.

decreasing fo r n 3; limit 0

n +1 ) 17 . 1 2 (1 + (1)

Ans.

b o u nded; no limit

18 .

ln n + 1 n

Ans. decreasing; limit 0

358

CHAPTER 42 Infinite Sequences

19 .

2n n!

Ans.

no nincreasing; decreasing fo r n 2; limit 0

2 0 . n n
3n n+2 co s n 4n + 5 n 3 2n + 3 sin n n

Ans.

decreasing fo r n 3; limit 1

2 1.

Ans.

increasing; limit 3

22.

Ans.

increasing, limit 1

23.

Ans.

decreasing; limit 0

24.

Ans.

limit 0

2 5 . n +1 n

Ans.

decreasing; limit 0

26.

2n 3n 4 n sin n 1 n2 + 1 n nn n!

Ans.

decreasing; limit 0

27.

Ans.

increasing, limit

28.

Ans.

increasing; limit +

29.

Ans.

increasing; limit +

In each o f P ro b lems 30 32, find a plau sib le fo rmu la fo r a seq u ence w ho se first few terms are giv en. F ind the limit (if it exists) o f y o u r seq u ence.

3 0 . 1, 3 , 9 , 27 , 81 , . . . 2 4 6 8 3 1. 1, 1, 1, 1, 1, 1, . . . 3 2 . 3 , 7 , 11 , 3 , 19 , . . . 1 4 7 2 11

Ans.

sn =

3n1 ; limit is + 2(n 1)

Ans.

sn = (1)n; no limit
sn = 4 n 1 ; decreasing, limit is 3n 2

Ans.

4 3

3 3 . P ro v e T heo rem 42.7. (H int: L et > 0 . C ho o se > 0 su ch that, fo r x in the do main o f f fo r w hich |x c| < , w e hav e | f ( x ) f (c)| < . C ho o se m so that n m implies |sn c| < .)

3 4 . S ho w that lim

n +

1/n p = 1 fo r p > 0. (H int: np/n = e(p ln n)/n.)

CHAPTER 42 Infinite Sequences

359

3 5 . (G C ) Use a graphing calcu lato r to inv estigate sn = b ehav io r o f the seq u ence. Ans. decreasing; limit is
1 2

n 2 + 5 fo r n = 1 to n = 5. T hen determine analy tically the 4n 4 + n

5 3 6 . (G C ) Use a graphing calcu lato r to inv estigate sn = nn fo r n = 1 to n = 10. T hen determine analy tically the 2 b ehav io r o f the seq u ence.

Ans.

decreasing fo r n 7 ; limit is 0

3 7 . P ro v e that lim an = 0 is eq u iv alent to lim |an | = 0 .

n + n +

2 3 8 . If sn > 0 fo r all n and lim sn = c , pro v e that lim sn = c . n + n +

3 9 . (G C ) D efine sn b y recu rsio n as fo llo w s: s1 = 2 and sn+1 = 1 sn + 2 fo r n 1. 2 sn (a) Use a graphing calcu lato r to estimate sn fo r n = 2, . . . , 5. (b ) S ho w that, if lim sn exists, then lim sn = 2 . n + n + (c) P ro v e that lim sn exists.
n +

4 0 . D efine sn b y recu rsio n as fo llo w s: s1 = 3, and sn+1 = 1 2 ( sn + 6) fo r n 1. (a) (b ) (c) (d) Ans. P S P E ro v e sn < 6 fo r all n. ho w that < sn > is increasing. ro v e that lim sn exists. n + v alu ate lim sn .
n +

(d) 6

4 1. P ro v e T heo rem 42.2, parts (a), (b ), (d ), (f ).

C H A P TE R 4 3

Infinite Series
Let sn be an infinite sequence. We can form the infinite sequence of partial sums Sn as follows:
S1 = s1 S2 = s1 + s2 S3 = s1 + s2 + s3 Sn = s1 + s2 + + sn

We usually will designate the sequence Sn by the notation

s
the series. We usually designate S by sn .
n =1 n + n + +

= s1 + s2 + + sn +

The numbers s1, s2, . . ., sn, . . . will be called the terms of the series. If S is a number such that lim Sn = S , then the series sn is said to conv erg e and S is called the sum of the series sn is said to div erg e. If lim Sn = + , then If there is no number S such that lim Sn = S , then +
n + n =1 n +

to div erge to and we write sn = .

n =1

the series is said to div erge to + and we write sn = + . Similarly, if lim Sn = , then the series is said +

EXAMPLE 43 .1: C onsider the sequence ( 1) n+1. The terms are s1 = 1, s2 = 1, s3 = 1, s4 = 1, and so on. Hence, the p artial sums begin with S1 = 1, S2 = 1 + ( 1) = 0, S3 = 1 + (1) + 1 = 1, S4 = 1 + (1) + (1) + (1) = 0, and continue with alternating 1s and 0s. So, lim Sn does not ex ist and the series div erges (but not to + or ).
n +

G e o m e tric S e rie s
C onsider the sequence ar n1, which consists of the terms a, ar, ar2, ar3, . . . . The series ar n1 is called a g eometric series with ratio r and first term a. Its nth p artial sum Sn is giv en by
Sn = a + ar + ar 2 + + ar n1 Multip ly by r: Subtract: Hence, , rSn = ar + ar 2 + + ar n1 + ar n Sn rSn = a ar n (1 r ) Sn = a(1 r n ) Sn = a(1 r n ) 1 r

36 0   ! % (% ,  121 3

,

CHAPTER 43

Infinite Series

36 1

E v erything now dep ends on the ratio r. If |r| < 1, then lim r n = 0 (by Theorem 42.4(b)) and, therefore, n + lim Sn = a / (1 r ) . If |r| > 1, then lim r n = (by Theorem 42.4(a)) and, therefore, lim Sn = . (A triv ial n + n + n + ex cep tion occurs when a = 0. In that case, all terms are 0, the series conv erges, and its sum is 0.) These results are summariz ed as follows:
: a (a) If |r| < 1, the series conv erges and has sum 1 r . (b) If |r| > 1 and a 0, the series div erges to .
T h e o re m 43 .1: EXAMPLE 43 .2:

G iv en a geometric series

ar

n 1

Tak e the geometric series

1 n 1 2

with ratio r =

1 2

and first term a = 1:

1 1 1+ 1 2 + 4 + 8 +
+

( ) = 2. We can multip ly a series s by a constant c to obtain a new series cs , and we can add two series s and t to obtain a new series (s + t ) .
B y Theorem 43.1(a), the series conv erges and has sum
n 1 n 1 2 n =1 n n n n

1 = 1 =2 . Thus, 1 1 (1 2) 2

T h e o re m 43 .2:

If c 0, then csn conv erges if and only if sn conv erges. Moreov er, in the case of conv ergence,

cs
n =1

= c sn
n =1

To obtain this result, denote by Tn = cs1 + cs2 + + csn the nth p artial sum of the series csn . Then Tn = cSn, where Sn is the nth p artial sum of sn . So, lim Tn ex ists if and only if lim Sn ex ists, and when the limits ex ist, lim Tn = c lim Sn . This yields Theorem 43.2.
n + n + n + n +

T h e o re m 43 .3 :

A ssume that two series sn and tn both conv erge. Then their sum (sn + tn ) also conv erges and

(s
n =1

+ t n ) = sn +
n =1

t
n =1

To see this, let Sn and Tn be the nth p artial sums of sn and t n, resp ectiv ely. Then the nth p artial sum Un of (sn + t n ) is easily seen to be Sn + Tn. So, lim U n = lim Sn + lim Tn . This yields Theorem 43.3.
n + n + n +

C o ro lla ry 43 .4: conv erges and

A ssume that two series

s
+

and

both conv erge. Then their difference

(s

tn ) also

(s
n =1

t n ) = sn t n
n =1 n =1

This follows directly from Theorems 43.2 and 43.3. J ust note that series (1) t n .
T h e o re m 43 .5 :

(s

t n ) is the sum of sn and the

If

conv erges, then lim sn = 0 .

+ n + n =1 n +

To see this, assume that sn = S . This means that lim Sn = S, where, as usual, Sn is the nth p artial sum of the series. We also hav e lim Sn1 = S. B ut, sn = Sn Sn1. So, lim sn = lim Sn lim Sn1 = S S = 0 .
n + n + n + n +

36 2

CHAPTER 43

Infinite Series

C o ro lla ry 43 .6 (T h e D iv e rg e n c e T h e o re m ):

If lim sn does not ex ist or lim sn 0, then sn div erges.

n + n +

This is an immediate logical consequence of Theorem 43.5.

EXAMPLE 43 .3 :
3 2 4 The series 1 3 + 5 + 7 + 9 + div erges. Here, sn = n . Since lim n = 1 0 , the D iv ergence Theorem imp lies that the series div erges. 2n + 1 2 n + 2n + 1

The conv erse of Theorem 43.5 is not v alid: lim sn = 0 does not imp ly that sn conv erges. This is shown n + by the following ex amp le.
1 1 1 1 1 EXAMPLE 43 .4: C onsider the so-called harmonic series = 1 + + + + + . Let us look at the following n 2 3 4 5 p artial sums of this series: S2 = 1 + 1 2 S4 = 1 + 1 + 1 + 1 > 1 + 1 + 1 + 1 = 1 + 1 + 1 = 1 + 2 2 2 2 4 4 2 3 4 2 S8 = S4 + 1 + 1 + 1 + 1 > S4 + 1 + 1 + 1 + 1 = S4 + 4 = S4 + 1 5 6 7 8 8 8 8 8 8 2 > 1+ 3 2 S16 = S8 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 9 10 11 12 13 14 15 16 > S8 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 = S8 + 1 16 16 16 16 16 16 16 16 2 > 1+ 4 2
6 C ontinuing in this manner, we would obtain S32 > 1 + 5 2 , S64 > 1 + 2 , and, in general, S2k > 1 + k / 2 when k > 1. This imp lies that lim Sn = + and, therefore, the harmonic series div erges. B ut notice that lim sn = lim 1/n = 0. n + n + n +

R emark: C onv ergence or div ergence is not affected by the addition or deletion of a finite number of terms at the beginning of a series. F or ex amp le, if we delete the first k terms of a series and the sum of the deleted terms is c, then each new p artial sum Tn has the form Sn+k c. (F or ex amp le, T1 is Sk+1 c.) B ut lim (Sn+ k c) ex ists if and only if lim Sn+ k ex ists, and lim Sn+ k ex ists if and only if lim Sn ex ists.
n + n + n + n +

N otation: It will often be useful to deal with series in which the terms of sn are index ed by the nonnegativ e integers: s0, s1, s2, s3, . . . . Then the p artial sums Sn would also begin with S0 = s0, and the sum of a conv ergent series would be written as sn .
n= 0 +

S O LV ED PR O B LEMS
1. E x amine the series 1 + 12 + 13 + for conv ergence. 5 5 5 1 This is a geometric series with ratio r = 1 5 and the first term a = 5 . Since |r | = / / a 1 5 1 5 = = = 1. that the series conv erges and that its sum is 1 r 1 (1/ 5) 4 / 5 4

1 5

< 1, Theorem 43.1 (a) tells us

2.

E x amine the series 1 + 1 + 1 + 1 + for conv ergence. 1 2 2 3 3 4 4 5 1 The nth term is . This is equal to 1 1 . Hence, the nth p artial sum n (n + 1) n n +1

CHAPTER 43

Infinite Series

36 3

1 Sn = 1 + 1 + 1 + 1 + + 1 2 2 3 3 4 4 5 n (n + 1) = 1 1 + 1 1 + 1 1 + 1 1 + + 1 1 1 2 n n +1 2 3 3 4 4 5

)(

)(

)(

= 1 1 n +1 1 Thus, lim Sn = lim 1 = 1 0 = 1. Hence, the series conv erges and its sum is 1. n +1 n + n +

3.

1 1 1 We k now that the geometric series 1 + 1 2 + 4 + 8 + 16 + conv erges to S = 2. E x amine the series that results when: (a) its first four terms are drop p ed; (b) the terms 3, 2, and 5 are added to the beginning of the series. 1/16 = 1/16 = 1 . N ote 1 1 (a) The resulting series is a geometric series 16 + 32 + with ratio 1 2 . It conv erges to 1 (1/ 2) 1/ 2 8 15 1 1 1 1 that this is the same as S (1 + 2 + 4 + 8 ) = 2 ( 8 ) = 8 . 1 1 1 (b) The new series is 3 + 2 + 5 + 1 + 1 2 + 4 + 8 + 16 + . The new p artial sums are the old ones p lus (3 + 2 + 5). Since the old p artial sums conv erge to 2, the new ones conv erge to 2 + 10 = 12. Thus, the new series is conv ergent and its sum is 12.

4.

3 7 15 Show that the series 1 2 + 4 + 8 + 16 + div erges. n 1 = 1 1 . Since lim 1 = 0 , it follows that lim s = 1 0 = 1 0 . So, by the D iv ergence Here, sn = 2 n n n + 2 n + 2n 2n Theorem, the series div erges.

5.

E x amine the series 9 12 + 16

64 3

256 9

for conv ergence.

4 This is a geometric series with ratio r = 4 3 . Since |r | = 3 > 1, Theorem 43.1(b) tells us that the series div erges.

6.

E v aluate

(1) n = 1 1 + 1 1 1 . 2 4 8 16 2n n=0 1 This is a geometric series with ratio r = 1 2 and first term a = 1. Since |r | = 2 < 1, the series conv erges and its 1 = 1 = 2. sum is a = 1 r 1 (1/ 2) 3/ 2 3

7.

Show that the infinite decimal 0.999 . . . is equal to 1. 0.999 = 9 + 9 + 9 + . This is a geometric series with first term a = 10 100 1000 Hence, it conv erges to the sum a = 9 /10 = 9 /10 = 1. 1 r 1 (1/10) 9 /10

9 10

and ratio r =

1 10

8.

E x amine the series 1 + 1 + 1 + 1 + . 1 3 3 5 5 7 7 9 1 1 1 1 =1 Here, sn = . N ote that . Hence, the nth p artial sum Sn is (2n 1)(2n + 1) 2 2n 1 2n + 1 (2n 1)(2n + 1)

1 1 1 + 1 1 1 + 1 1 1 + + 1 1 1 = 1 1 1 2n + 1 2 1 3 2 3 5 2 5 7 2 2n 1 2n + 1 2

) (

) (

) (

So, lim Sn = 1 . Thus, the series conv erges to 1 2. 2

n +

9.

E x amine the series 3 + 3 + 3 3 + 4 3 + .

sn = n 3 = 31/ n = e(ln 3) / n . Then lim sn = e 0 = 1 0 . B y the D iv ergence Theorem, the series div erges.
n +

10 . E x amine the series

1 10

1 1 1 + 11 + 12 + 13 + .

This series is obtained from the harmonics series by deleting the first nine terms. Since the harmonic series div erges, so does this series.

36 4

CHAPTER 43

Infinite Series

11. (Z e n o s P a r a d o x ) A chilles (A) and a tortoise (T ) hav e a race. T gets a 1000 ft head start, but A runs at 10 ft/sec, whereas T only does 0.01 ft/sec. When A reaches T s starting p oint, T has mov ed a short distance ahead. When A reaches that p oint, T again has mov ed a short distance ahead, etc. Z eno claimed that A would nev er catch T. Show that this is not so. When A reaches T s starting p oint, 100 seconds hav e p assed and T has mov ed 0.01 (100) = 1 ft. A cov ers that additional 1 ft in 0.1 seconds, but T has mov ed 0.01(0.1) = 0.001 ft further. A needs 0.0001 seconds to cov er that distance, but T meanwhile has mov ed 0.01(0.0001) = 0.000001 ft, etc. The limit of the distance between A and T ap p roaches 0. The time inv olv ed is 100 + 0.1 + 0.0001 + 0.0000001 + , which is a geometric series with first term a = 100 and ratio r = 1/1000. Its sum is
a = 100 = 100 = 100000 1 r 1 (1/1000) 999 /1000 999

which is a little more than 100 seconds. The seeming p aradox arises from the artificial div ision of the ev ent into infinitely many shorter and shorter step s.

S U PPLEMEN T AR Y PR O B LEMS
12 . E x amine each of the following geometric series. If the series conv erges, find its sum.
1 (a) 4 1 + 1 4 16 + 3 9 27 (b) 1 + 2 + 4 + 8 + 1 1 (c) 1 1 3 + 9 27 + 1 2 (d) 1 + e + e + e 3 +

Ans. Ans. Ans. Ans.

S = 16 5 D iv erges 3 S= 4 S= e e 1

13 . A rubber ball is drop p ed from a height of 10 ft. Whenev er it hits the ground, it bounces straight up three-fourths of the p rev ious height. What is the total distance trav eled by the ball before it stop s? Ans. 70 ft
+ 1 + 1 + . n(n1+ 4) = 11 5 26 37

14 . E x amine the series Ans.

S=
25 48

15 . E x amine the series Ans.

S=
1 4

= 1 + 1 + 1 + n(n + 11 )(n + 2) 1 2 3 2 3 4 3 4 5

16 . E v aluate sn when sn is the following:

n =1

(a) 3 Ans.

(b)
1 2

1 n(n + 2)
3 4

(c)
11 18

1 n(n + 3)

(d)

n (n + 1)!

(a)

; (b)

; (c)

; (d) 1

17 . Show that each of the following series div erges: (a) (d)
7 9 3+ 5 2 + 3 + 4 +

e+

e2 8

e3 27

e4 64

(b) 2 + 2 + 3 2 + 4 2 + 1 (e) n + n 1

(c) 1 + 1 + 31 + 41 + 2 2 2 2

18 . E v aluate the following + (a) 1n + 1n 2 7 n=0

(b)

41n
n =1

(c)

n +1 n2 (n + 1)
2 n =1

CHAPTER 43

Infinite Series

36 5

(d) (g) ( j)

+3 2 5
n n n= 0 + 2 2

(e) (h)

23
n =1

n 1 n

(f ) (i)

5 3
n =1

n n

3 n n+ n+2
n =1 +

(1) n 52 n n =1

2 3
n =1

3n

2n

1
n =1

Ans.

(a)

19 6

; (b) +; (c) 1; (d)

25 6

1 ; (e) 1; (f ) +; (g) +; (h) 26 ; (i) 8; ( j) +

19 . (G C ) In P roblems 1 and 6, use a calculator to comp ute the first 10 p artial sums and determine to how many decimal p laces the 10th p artial sum is a correct estimate of the sum of the series. 2 0 . (G C ) (a) If |x| < 1, what function is rep resented by x n = 1 + x + x 2 + x 3 + ? (b) U se a grap hing calculator to grap h 1 + x + x 2 + x 3 + + x 9 on the interv al (1, 1) and comp are the grap h with that of the function in (a). Ans. (a)
1 1 x
n=0 +

2 1. In each of the following, find those v alues of x for which the giv en series conv erges, and then find the function rep resented by the sum of the series for those v alues of x. n + + + + n (a) (3x) n (b) ( x 2) n (c) x (d) x 1 2 2 n= 0 n= 0 n= 0 n=0

()

( )

Ans.

(a) |x| < 1 , 1 ; (b) 1 < x < 3, 1 ; (c) |x| < 2 , 2 ; (d) 1 < x < 3, 2 3 1 3x 3 x 2 x 3 x

C H A P TE R 4 4

Series w ith P ositive Term s.The IntegralTest.C om parison Tests

S e rie s o f Po s itiv e T e rm s
If all the terms of a series

F or a p ositiv e n Sn + 1 = Sn + sn + 1 > Sn . T his y ield s the follow in g u sefu l resu lt.

T h e o re m 44.1:

s are p ositiv e, then the series is c alled a positiv e series. series s , the seq u en c e of p artial su ms S is an in c reasin g
n n

seq u en c e, sin c e

A p ositiv e series

c on v erg es if an d on ly if the seq u en c e of p artial su ms sn is b ou n d ed .

T o see this, n ote first that, if sn c on v erg es, then , b y d efin ition , Sn c on v erg es an d , therefore, b y T heorem 42.1, Sn is b ou n d ed . C on v ersely , if Sn is b ou n d ed , then , sin c e Sn is in c reasin g , T heorem 42.8 imp lies that Sn c on v erg es, that is, sn c on v erg es.
sn b e a p ositiv e series an d let f (x) b e a c on tin u ou s, p ositiv e d ec reasin g fu n c T h e o re m 44.2 (In te g ra l T e s t): L et tion on [1, + ) su c h that f (n) = sn for all p ositiv e in teg ers n. T hen :

s
F rom F ig . 44-1 w e see that
u 1

c on v erg es if an d on ly if

f ( x) dx c on v erg es

so, f ( x) dx w ill b e b ou n d ed for all u 1 an d , therefore, w e hav e s2 + s3 + + sn <

Sn <

f ( x) dx < s1 + s2 + + sn 1 = Sn 1 . If sn c on v erg es, then Sn is b ou n d ed ;

f ( x) dx an d , therefore, Sn

<
1

1 n

f ( x) dx c on v erg es. C on v ersely , from F ig . 44-1

f ( x) dx + s1. T hu s, if

f ( x) dx + s, an d so Sn w ill b e b ou n d ed . H en c e, b y

T heorem 44.1, s

1 n

f ( x) dx c on v erg es, then

c on v erg es. T his p rov es

T heorem 44.2.

Sn 1 2 3 4 5 n 2 n 1 n

Fig. 44-1

36 6   ! % (% ,  121 3

,

CHAPTER 44 Series w ith P ositive Term s

36 7

EXAMPLE 44.1:

ln n d iv erg es. n

L et f ( x) =

ln x . N ow , x

H en c e, b y the in teg ral test,

EXAMPLE 44.2:

ln x dx = lim u ln x dx = lim 1 (ln x) 2 = lim 1 ((ln u ) 2 0) = + 2 x x u + 1 u + 1 u + 2

lnnn d iv erg es.

n1

c on v erg es.

1 L et f ( x) = 2 . N ow , x

H en c e, b y the in teg ral test,

1 dx = lim u + x2

1 dx = lim 1 = li im 1 1 = 1 u x u + x2 1 u +

( )

1
2

c on v erg es.

R emark: T he in teg ral test c an b e easily b e ex ten d ed to the c ase w here the low er limit of the in teg ral is c han g ed from 1 to an y p ositiv e in teg er.
T h e o re m 44.3 (C o m p a ris o n T e s t): L et an an d for w hic h ak bk for all in teg ers k m. T hen :

b e tw o p ositiv e series su c h that there is a p ositiv e in teg er m

(1) (2)

If

b If a

c on v erg es, so d oes an ; d iv erg es, so d oes bn .

W e may assu me in the d eriv ation of T heorem 44.3 that m = 1, sin c e c on v erg en c e is n ot affec ted b y d eletion of a fin ite n u mb er of terms at the b eg in n in g of a series. N ote also that (2) is a log ic al c on seq u en c e of (1). T o p rov e (1), assu me that bn c on v erg es. L et Bn = b1 + b2 + + bn b e the nth p artial su m for bn an d let An = a1 + a2 + + an b e the nth p artial su m an . T hen An Bn , sin c e ak bk for all k. F rom the fac t that bn c on v erg es, it follow s, b y T heorem 44.1, that the seq u en c e Bn is b ou n d ed . S in c e An Bn for all n, it follow s that the seq u en c e An is b ou n d ed . S o, b y T heorem 44.1, an c on v erg es. T his p rov es T heorem 44.3 .
EXAMPLE 44.3 :

1 c on v erg es. +5 1 1 an d bn = 12 . T hen an < bn for all n. B y E x amp le 2, 2 c on v erg es. S o, b y the c omp arison test, L et an = 2 n n n +5 n 2 1+ 5 c on v erg es.
2

EXAMPLE 44.4:

3 n1+ 5 d iv erg es.

1 L et an = 4 n an d bn = 1 . N ow , an bn for n 5. (T o see this, ob serv e that 1 1 is eq u iv alen t to 3 n + 3n+5 4n 3 n + 5 5 4n, w hic h is eq u iv alen t to 5 n.) R ec all that the harmon ic series 1 d iv erg es (b y C hap ter 43 , E x amp le 4). n H en c e, 1 d iv erg es b y T heorem 43 .2. T he c omp arison test imp lies that 1 d iv erg es. 4n 3n+5

S ometimes, as in E x amp le 4, c omp lic ated man eu v ers are n eed ed in ord er to ap p ly the c omp arison test. T he follow in g resu lt offers a mu c h more flex ib le tool.
a lim n ex ists T h e o re m 44.4 (Lim it C o m p a ris o n T e s t): L et an an d bn b e tw o p ositiv e series su c h that L = n + bn an d 0 < L < +. T hen an c on v erg es if an d on ly if bn c on v erg es.

36 8

CHAPTER 44 Series w ith P ositive Term s

Assume that bn converges. Let c be a positive number such that L < c. Then there exists a positive integer m such that an /bn < c for all n m. Hence, an < cbn for all n m. But, since bn converges, so does cbn . Therefore, by the comparison test, an converges. Conversely, if an converges, then bn converges. b 1 (In fact, lim n = > 0 and w e can use the same k ind of argument that w as just given.) L n + an
EXAMPLE 44.5:

5n + 4 diverges. 3n7n +2
2 3

W hen dealing w ith q uotients of polynomials, a good rule of thumb is to ignore everything except the leading 2 terms. In this case, w e have 3n 3 = 3 1 . Let us try a limit comparison w ith 1 . N ow 7n 7n n
3n 2 5n + 4 lim 7n 3 + 2

n +

1 = lim 3n 3 5n 2 + 4 n = 3 . n + 7 n 7n 3 + 2 5n + 4 diverges. 3n7n +2

2 3

S ince 1 diverges, the limit comparison test tells us that n

EXAMPLE 44.6 :

5n 2 converges. n6 4n 2 + 7 5n = 5n = 5 . 3 n 2 S o, let us try a limit comparison n6 n

U sing the rule of thumb given in E xample 5, w e should look at 1 w ith 2 : n

lim 6 5n 2 2 n n + 4 7

n +

1 = lim 5n 3 2n 2 n + n2 n6 4n 2 + 7

Let us divide the numerator and denominator by n3. N ote that, in the denominator, w e w ould get
1 n6 4n 2 + 7 = 1 n3 n6 n 6 4 n 2 + 7 = 1 44 + 66 n n

S o, the result w ould be

lim 5 2 n = 5 =5 1 4 7 1 4 + 6 n x

n +

Hence, since w e k now , by E xample 2, that converges.

n1

converges, the limit comparison test implies that

5n 2 n6 4n 2 + 7

S O LV ED PR O B LEMS
1. Consider the series 1p , w here p is constant. This is called a p-series. Then: n (a) If p > 1, the series 1p converges. n (b) If p 1, the series 1p diverges. n W e may assume that p 1, since w e already k now that the harmonic series 1 diverges. W e may also assume n that p > 0; if p 0, lim 1p 0 and the D ivergence Theorem implies that the series diverges. Let us apply the n + n integral test w ith f ( x) = 1/ x p . (f (x) is positive and decreasing in [1, +).) N ow ,

1 dx = lim u + xp

1 dx = lim x1 p u + 1 p xp 1

1 p = lim u 1 . 1 p u + 1 p

CHAPTER 44 Series with Positive Terms

369

u1 p 1 = 1 . By the integral test, (a) p > 1. Then p 1 > 0 and lim u1 p = lim 1 p 1 = 0. S o, lim 1 p p 1 u + u + u u + 1 p 1 converges. np 1 p (b) p < 1. Then 1 p > 0 and lim u1 p = +. S o, lim u 1 = + and, by the integral test, 1p 1 p n u + u + 1 p diverges.

In P roblems 2 7, examine the given series for convergence. 2.

1 + 1 + 1 + 1 + . 3 5 7 sn = 1 . Let f ( x) = 2n 1 1 . O n [1, +), f ( x) > 0 and f is decreasing. 2x 1

u dx = lim 1 u (2 x 1)1/2 1 2 dx = lim (2) dx 1 + u 2x 1 2 x 1 u + 2 1

= lim

u +

1 2

(2)(2 x 1)1/ 2 = lim ((2u 1)1/ 2 1) = + 1 u +

Hence, the series diverges by the integral test. 3.

1 + 1 + 1 + + 1 + . 3 10 29 n3 + 2 1 < 1. n3 + 2 n3 is convergent. 1 n
3

is convergent, since it is a p-series w ith p = 3 > 1. Thus, by the comparison test,

n 1 +2
3

4.

1 + 1 + 1 + 1 + . 2! 3! 4! 1 sn = 1 . N ote that 1 = is a convergent geometric series (w ith 1 for n 2. S ince 1 n! n(n 1) 3 2 2n1 n! 2n1 1 1 ratio 2 ), is convergent by the comparison test. n!

5.

2 + 33 + 4 + 5 + . 2 33 4 3 1 . U se limit comparison w ith n = 1 . sn = n + n3 n3 n 2

n +

1 1 = lim n 3 + n 2 = 1 lim n + n3 n 2 n + n 3

1 converges. W e k now that 12 converges. S o, by the limit comparison test, n + n3 n

6.

1 + 12 + 1 + 1 + 2 33 4 4 1 sn = 1n . N ow , 1n = is a convergent geometric series (r = 1 1 and 1 2 ). S o, by the n n n 2n1 2n1 n n 1 comparison test, n converges. n

7.

2 2 2 1 + 23 + 1 + 33 + 1 + 4 3 + 1 + . 2 +1 3 +1 4 +1 2 2 sn = n 3 + 1 . U se limit comparison w ith n 3 = 1 : n n +1 n

n +

n2 + 1 lim 3 n +1

1 = lim n 3 + n = 1 3 n n + n + 1

37 0

CHAPTER 44 Series with Positive Terms

W e k now that the harmonic series 8.

1 + 1 + 1 + . 2 ln 2 3 ln 3 4 ln 4 sn = 1 is defined for n 2. n ln n
+

1 diverges. S o, by the limit comparison test, n n n

2 3

+ 1 diverges. +1

dx = lim x ln x u +

dx = lim ln(ln u ) = lim (ln (ln u ) ln (ln 2)) = + . 2 u + x ln x u +

Hence, the series diverges by the integral test. 9. How many terms of 12 suffice to obtain tw o-decimal place accuracy (that is, an error < 5/103)? n If w e use k terms, then w e req uire that the error
u

1 1 = 1 + 1 dx = lim 2 2 u + n n n 2 k x 2 n =1 n =1 n = k +1 = 1 < 53 = 1 200 k 10

1 dx = lim 1 = lim 1 1 x u + x2 k u + u k

Hence, 200 < k. Thus, it suffices to use 201 terms of the series. (The graphing calculator can be used to find 201 1 1.64.) 2 n n =1 10 . A ssume sn converges by virtue of the integral test applied to f (x) and, for each n, the error (or remainder) Rk after k terms is defined to be

s s .
n n n =1 n =1 +

Then

Rk =

n = k +1

s <
n

f ( x) dx.

F ind a bound on the error w hen 12 is approximated by the first five terms: 1 + 1 + 1 + 1 + 1 = 5269 1.4636. 4 9 16 25 3600 n n =1 + The error R5 < 12 dx = 1 . . = 0 2 5 5 x 11. A ssume sn and

are positive series, cn converges, and sn cn for all n. Then the error Rk after k terms is
+ k + +

s s
n n =1 n =1

n = k +1

n = k +1

c.
n

A t least how many terms w ill suffice to estimate

n 1+ 1 w
5 n =1 +

ith an error < 0.00001?

+

+ In this case, sn = 51 and cn = 15 . It suffices to have 15 < 0.00001. N ow , 15 < 15 dx = 1 4 . k x n n +1 n 4k n n = k +1 n = k +1 1 . E q uivalently, 100,000 < 4k4, 25,000 < k4, k 13. S o, w e need 1 4 < 0.00001 = 100, 000 4k

CHAPTER 44 Series with Positive Terms

37 1

S U PPLEMEN T AR Y PR O B LEMS
F or P roblems 12 43, determine w hether the series converges.

12 .

n(n3+ 1)
n (n + 1)( n + 2)

Ans.

converges; comparison w ith 32 n diverges; limit comparison w ith 1 n diverges; limit comparison w ith
1 n

13 .

Ans.

14 .

n n+ 1
2

Ans.

15 .

en

Ans.

converges; integral test converges; limit comparison w ith 12 n converges; limit comparison w ith

16 .

n (n + 1)(n2 + 2)(n + 3)

Ans.

17 .

(2n 1 + 1) n 1 1
3

Ans.

n1

18 .

Ans.

converges; limit comparison w ith 13 n

19 .

2 nn
3

Ans.

converges; limit comparison w ith

n1

20.

nln+n2
2

Ans.

converges; limit comparison w ith

n1

3/ 2

2 1.

1 n sin ( n )

Ans.

diverges; D ivergence Theorem

22.

1 n

Ans.

diverges; p-series, p = 1 3 <1 converges; comparison w ith 1 ,n2 2n1 diverges; comparison w ith ln n n diverges; comparison w ith 1 n diverges; limit comparison w ith 1 n diverges; integral test

23.

n1

n 1

Ans.

24.

n lnn

Ans.

25.

1 +1ln n n
n +1 3n 2 (for n 3)

Ans.

26.

Ans.

27.

1 n ln n ln (ln n) 1 n ln n (ln (ln n))

Ans.

28.

(for n 3)

Ans.

converges, integral test

37 2

CHAPTER 44 Series with Positive Terms

29.

1 + 1 + 1 + 1 + . 4 2 72 10 2 132

Ans. sn =

1 ; converges; limit comparison w ith (3n + 1)2

n1

+ 3 + 3 + . 3 0 . 3+ 3 21/ 3 31/ 3 41/ 3

Ans.

sn = 3 ; diverges; p-series, p = 1 3 <1 n1/ 3

3 1. 1 + 1 + 1 + 1 + . 5 9 13 Ans.
sn = 1 ; diverges; limit comparison w ith 4n 3 1 n

1 + . 32. 1+ 1 + 1 + 2 3 4 4 56 56 7 8

Ans.

sn =

1 ; converges; limit comparison w ith (n + 1)(n + 2) (2n)

n1

3 3 . 2 + 3 2 + 4 3 + 5 4 + . 3 23 33 4 3 Ans.
; converges; limit comparison w ith sn = n + 1 n 3n

31

3 4 . 1 + 1 2 + 1 3 + 1 4 + . 2 22 32 42 Ans. 35.
sn = 1 n ; converges; comparison w ith n2

21

2 + 3 + 4 + 5 + . 1 3 2 4 3 5 4 6

Ans.

sn =

n + 1 ; diverges; limit comparison w ith 1 n n(n + 2)

+ 3 + 4 + . 36. 1+ 2 2 3 2 4 3 54

Ans.

sn =

n ; converges; comparison w ith (n + 1)n

21

n 1

+ 1 + . 3 7 . 1 + 12 + 1 2 35/ 2 4 3

Ans.

sn =

1 ; converges; comparison w ith n ( n+ 2)/ 2

n1

3 8 . 1 + 3 + 4 + 5 + . 5 10 17 Ans.
sn = n2 + 1 ; diverges; limit comparison w ith n +1 1 n

3 9 . 2 + 2 4 + 2 4 6 + 2 4 6 8 + . 5 5 8 5 8 11 5 8 11 14 Ans.
sn = 2 4 (2n) n ; converges; comparison w ith ( 2 3) 5 8 (2 + 3n)

CHAPTER 44 Series with Positive Terms

37 3

40. 3 + 5 + 7 + 9 + . 2 10 30 68 Ans.
n + 1 ; c o n v e rg e s , lim it c o m p a ris o n w ith sn = 2 3 n +n

n1

41 . 3 + 10 + 29 + 66 + . 2 24 108 320 Ans. 42 .

3 sn = n4 + 23 ; d iv e rg e s ; lim it c o m p a ris o n w ith n +n

1 n

1 + 2 + 3 + 4 + . 2 2 1 32 2 4 2 3 52 4

Ans. 43 .

sn =

n ; d iv e rg e s ; lim it c o m p a ris o n w ith ( n + 1) 2 n

1 n

1 + 1 + 1 + 1 + . 23 12 33 22 4 3 32 53 4 2

Ans.

sn =

1 ; c o n v e rg e s ; lim it c o m p a ris o n w ith ( n + 1) 3 n 2

1 n

44. (G C ) E s tim a te th e e rro r w h e n : + (a ) n 1 is a p p ro x im a te d b y th e s u m o f its firs t s ix te rm s . 3 +1 n =1

+

(b ) Ans.

4
n =1

1 is a p p ro x im a te d b y th e s u m o f its firs t s ix te rm s . +3

(a ) 0.0007; (b ) 0.00009

45 . (G C ) (a ) E s tim a te th e e rro r w h e n th e g e o m e tric s e rie s

23

is a p p ro x im a te d b y th e s u m o f its firs t s ix te rm s .

(b ) H o w m a n y te rm s s u ffic e to c o m p u te th e s u m if th e a llo w a b le e rro r is 0.00005? Ans. (a ) 0 .0 4 7 ; (b ) 1 6

46 . (G C ) (a ) H o w m a n y te rm s s u ffic e to a p p ro x im a te

n1
n =1

w ith a n e rro r < 0.001?

(b ) F in d a b o u n d o n th e e rro r if w e a p p ro x im a te 14 b y th e s ix th p a rtia l s u m . n + n =1 (c ) W h a t is y o u r a p p ro x im a tio n to 14 b y th e s ix th p a rtia l s u m , c o rre c t to fo u r d e c im a l p la c e s ? n n =1 Ans. (a ) 7 ; (b ) 0 .0 0 1 5 ; (c ) 1 .0 8 1 1

47 . (G C ) L e t Sn b e th e nth p a rtia l s u m 1 + 1 + + 1 o f th e d iv e rg e n t h a rm o n ic s e rie s . 2 n (a ) P ro v e ln ( n + 1) Sn 1 + ln n. (b ) L e t En = Sn ln n . P ro v e th a t En is b o u n d e d a n d d e c re a s in g . (c ) P ro v e th a t En c o n v e rg e s . Its lim it is d e n o te d a n d is c a lle d Eulers co nstant. (d ) U s e a g ra p h in g c a lc u la to r to a p p ro x im a te E999 to e ig h t d e c im a l p la c e s . Ans. (d ) 0.57771608 (in fa c t, g ~ 0.57721566.)

48 . (E x te n s io n o f th e lim it c o m p a ris o n te s t.) A s s u m e sn a n d sn (a ) If lim = 0 a n d t n c o n v e rg e , s o d o e s sn . n + t n sn (b ) If lim = + a n d t n d iv e rg e s , s o d o e s sn . n + t n

a re p o s itiv e s e rie s . P ro v e :

37 4

CHAPTER 44 Series with Positive Terms

4 9 . U se the extension of the limit comparison test to determine w hether Ans. converges; use

(ln n)4 converges. n3

n1

and P roblem 48 (a)

5 0 . A ssume sn is a positive series and lim nsn exists and is positive. P rove that n + comparison w ith (1/ n).)

s diverges. (Hint: Limit

n

5 1. A ssume

and

are convergent positive series. P rove that

s t

n n

converges.

C H A P TE R 4 6

Power Series
Po w e r S e rie s
An infinite series
+

a ( x c)
n n= 0 +

= a0 + a1 ( x c) + a2 ( x c)2 +

(46.1)

is called a po w er series in x about c with coefficients an . An important special case

a x
n n= 0

= a0 + a1 x + a2 x 2 +

(46.2)

is a power series about 0. For a given value of x, the series (46.1) either converges or diverges. Hence, (46.1) determines a function f whose domain is the set of all x for which (46.1) converges and whose corresponding value f (x) is the sum of the series. Note that (46.1) converges when x = c.
EXAMPLE 46 .1 :

The power series about 0

x
n=0

= 1 + x + x2 +

is a geometric series with ratio x. Thus, it converges for |x| < 1, and its sum is 1 . S o, the domain of the 1 x corresponding function is an interval around 0.

T h e o re m 46 .1 :

Assume that the power series

a ( x c)
n n=0

converges for x0 c. Then it converges absolutely for all

x such that |x - c| < |x0 - c| (that is, for all x that are closer to c than x0).

For a proof, see P roblem 4.

T h e o re m 46 .2 : For a power series an ( x c)n, one of the following three cases holds: n=0 (a) it converges for all x; or (b) it converges for all x in an open interval (c - R1, c + R1) around c, but not outside the closed interval [c - R1, c + R1]; or (c) it converges only for x = c.
+

B y the interv al o f co nv erg ence of In case (a): (- , +) In case (b): (c - R1, c + R1) In case (c): { c}

a ( x c)
n n=0

we mean:

38 3

  
!#"$%&('# )
*,+ - . ./ 0,)1 $234 1&5 .  &6#$ $7898$ 02:
7;23$5

38 4
+

CHAPTER 46

Power Series

B y the radius o f co nv erg ence of In case (a): In case (b): R1 In case (c): 0

a ( x c)
n n=0

we mean:

N o te: In case (b), whether the power series converges at neither endpoint of its interval of convergence or at one or both of those endpoints depends upon the given series. For a sk etch of a proof of Theorem 46.2, see P roblem 5.
EXAMPLE 46 .2 :

The power series

( x 2)n ( x 2)2 ( x 2)3 = ( x 2) + + + 2 3 n n =1

is a power series about 2. L et us use the ratio test to find the interval of convergence.
n +1 sn+1 = |x 2| sn n +1

|x 2|n = n |x 2|. Thus, n n +1

n+

lim

sn+1 = |x 2|. sn

S o, by the ratio test, the series converges absolutely for |x 2| < 1. The latter ineq uality is eq uivalent to 1 < x 2 < 1, which, in turn, is eq uivalent to 1 < x < 3. Hence, the interval of convergence is (1, 3) and the radius of convergence is 1. At the endpoint x = 1, the series becomes point x = 3, the series becomes
EXAMPLE 46 .3 :
+ n =1

(1/ n), the divergent harmonic series. Thus, the power series converges for 1 x < 3.

[(1) / n], which converges by

n n =1

the Alternating S eries Theorem. At the end-

The power series

=1+ x + x + x + x 2! 3! n!
n= 0 + n 2 3

is a power series about 0. (R ecall that 0! = 1.) L et us use the ratio test:
n +1 sn+1 = | x| sn (n + 1)!

|x|n = |x| . n! n +1

S o,

n+

lim

sn+1 = 0. sn

Hence, by the ratio test, the series converges (absolutely ) for all x. Its interval of convergence is (- , +) and its radius of convergence is .
EXAMPLE 46 .4:

The power series

n! x
n= 0

= 1 + x + 2! x 2 + 3! x 3 +

is a power series about 0. L et us use the ratio test again:

sn+1 (n + 1)! |x|n+1 = = (n + 1) |x|. S o, sn n! |x|n sn+1 = + . sn

n+

lim

ex cept when x = 0. Thus, the series converges only for x = 0. Its (degenerate) interval of convergence is { 0} and its radius of convergence is 0.

CHAPTER 46

Power Series

38 5

U n ifo rm C o n v e rg e n c e
L et fn be a seq uence of functions, all defined on a set A. L et f be a function defined on A. Then fn is said to co nv erg e unifo rmly to f on A if, for every > 0, there ex ists a positive integer m such that, for each x in A and every n m, |fn(x) f(x)| < .
T h e o re m 46 .3 :

If a power series
k n=0

Sk ( x), where Sk ( x) = an ( x c)n, converges uniformly to

a ( x c)
n n=0

converges for x0 c and d < |x0 c|, then the seq uence of partial sums

a ( x c)
n n=0

on the interval consisting of all x such that

|x c| < d. Hence, the convergence is uniform on any interval strictly inside the interval of convergence.

The reader is referred to more advanced book s on analy sis for a proof of this result.
T h e o re m 46 .4:

If fn converges uniformly to f on a set A and each fn is continuous on A, then f is continuous on A.

For a proof, see P roblem 6.

C o ro lla ry 46 .5: convergence.

The function defined by a power series

a ( x c)
n n=0

is continuous at all points within its interval of

This follows from Theorems 46.3 and 46.4.

T h e o re m 46 .6 (In te g ra tio n o f Po w e r S e rie s ): L et f be the function defined by a power series its interval of convergence (with radius of convergence R1). Then:

a ( x c)
n n=0

on

(a)

f ( x) dx = a
n= 0

( x c)n+1 +K n +1

for | x c| < R1

(46.3)

where the interval of convergence of the power series on the right side of formula (46.3) is the same as that of the original series. K is an arbitrary constant of integration. Note that the antiderivative of f is obtained by term-by term integration of the given power series. (b) If a and b are in the interval of convergence, then:

Thus,

f ( x)dx = an n= 0

( x c)n+1 n +1

(46.4)
a

f ( x)dx is obtained by term-by -term integration.

For a proof of Theorem 46.6, the reader should consult a more advanced book on analy sis.
T h e o re m 46 .7 (D iffe re n tia tio n o f Po w e r S e rie s ):

L et f be the function defined by a power series an ( x c)n

n=0

on its interval of convergence (with radius of convergence R1). Then f is differentiable in that interval and
f ( x) = nan ( x c)n1
n=0 +

for

| x c | < R1

(46.5)

Thus, the derivative f is obtained by term-by -term differentiation of the power series. The interval of convergence of the power series on the right side of formula (46.5) will be the same as for the original power series.

For a proof, the reader is referred to more advanced tex ts in analy sis.

38 6

CHAPTER 46

Power Series

EXAMPLE 46 .5:

W e already k now by E x ample 1 that, for |x| < 1,

1 = x n = 1 + x + x 2 + x3 + + x n + 1 x n= 0
+

(46.6)

Now, Dx

( 1 1 x ) = (1 1x) . S o, by Theorem 46.7,

2

1 = 1 + 2 x + 3x 2 + + nx n1 + for |x| < 1 (1 x)2 = nx n1 = (n + 1) x n

n =1 n= 0 + +

EXAMPLE 46 .6 :

W e k now already that

1 = x n = 1 + x + x 2 + x3 + + x n + 1 x n=0
+

for |x| < 1

R eplace x by x. (This is permissible, since |x| = |x| < 1.) The result is
1 = ( x)n = (1)n x n = 1 x + x 2 x 3 + 1+ x n=0 n=0
+ +

(46.7)

B y Theorem 46.6(a), we can integrate term by term:

1 + x = (1)
dx
n= 0 +

x n+1 + K = (1)n1 x n + K n +1 n n =1 for |x| < 1

for |x| < 1

n ln 1 + x = (1)n1 x + K n n =1

L etting x = 0 and noting that ln 1 = 0, we find that K = 0. Note also that, for |x| < 1, we have 1 < x < 1, 0 < 1 + x < 2, and, therefore, |1 + x| = 1 + x. Hence,
n ln (1 + x) = (1)n1 x n n =1 +

for |x| < 1

= x 1 x2 + 1 x3 1 x 4 + 2 3 4

(46.8 )

The ratio test shows that this series converges. If we replace x by x 1, we obtain:
ln x = (1)n1
n =1 +

( x 1)n n

for |x 1| < 1

(46.9 )

Note that |x 1| < 1 is eq uivalent to 0 < x < 2. Thus, ln x is definable by a power series within (0, 2).
T h e o re m 46 .8 (Ab e ls T h e o re m ):

Assume that the power series an ( x c)n has a finite interval of convergence
n=0

|x c| < R1 and let f be a function whose values in that interval are given by that power series. If the power series also converges at the right-hand endpoint b = c + R1 of the interval of convergence, then lim f ( x) ex ists and is eq ual to the sum of the series at b. The analogous result holds at the left-hand endpoint a = c R1.
x b

The reader is referred to advanced book s on analy sis for a proof.

CHAPTER 46

Power Series

38 7

EXAMPLE 46 .7 :

This is a continuation of E x ample 6. B y formula (46.8 ),

n ln (1 + x) = (1)n1 x n n =1 +

for | x | < 1

At the right-hand endpoint x = 1 of the interval of convergence, the power series becomes the convergent alternating harmonic series

(1)
n =1 x 1

n 1

1 =1 1 + 1 + 1 + 2 3 4 n

B y Abels Theorem, this series is eq ual to lim ln (1 + x) = ln 2 . S o,

1 1 ln 2 = 1 1 2 + 3 4 +

(46.10)

EXAMPLE 46 .8 :

S tart again with

1 = x n = 1 + x + x 2 + x 3 + + x n + for |x | < 1 1 x n=0
+

R eplace x by x2, obtaining

1 = (1)n x 2 n = 1 x 2 + x 4 x 6 + 1 + x2 n=0
+

(46.11)

S ince |x2| < 1 is eq uivalent to |x| < 1, (46.11) holds for |x| < 1. Now, by Theorem 46.6(a), the antiderivative tan1 x of 1 2 can be obtained by term-by -term integration: 1+ x
2 n +1 tan 1 x = (1)n x +K +1 n 2 n= 0 +

for |x| < 1

1 7 1 5 3 =K+x 1 3 x + 5 x 7 x +

Here K is the constant of integration. If we let x = 0 and note that tan1 0 = 0, it follows that K = 0. Hence,
2 n +1 1 7 1 5 3 =x 1 tan 1 x = (1)n x 3 x + 5 x 7 x + n 2 +1 n= 0 +

(46.12)

At the right-hand endpoint x = 1 of the interval of convergence, the series in (46.12) becomes

(1)
n= 0

1 =1 1 + 1 1 + 3 5 7 2n + 1

which converges by virtue of the Alternating S eries Theorem. S o, by Abels Theorem,

1 1 1 1 tan 1 ( x) = tan 1 1 = 3 + 5 7 + = lim 4 x 1

(46.13)

38 8
+

CHAPTER 46

Power Series

EXAMPLE 46 .9 : W e k now already , by E x ample 3, that B y term-by -term differentiation (Theorem 46.7),
+

x n!
n= 0 +

converges for all x. L et f ( x) = x for all x. n! n= 0

n 1 n = x = f ( x) f ( x ) = x n n 1 ( )! ! n =1 n= 0

Note that f (0) = 1. Therefore, by formula (28 .2), f (x) = ex. Thus,
ex = x n! n= 0
+ n

for all x

(46.14)

S O LV ED PR O B LEMS
1. Find the interval of convergence of the power series

( x 2)n ( x 2)2 ( x 2)3 = ( x 2) + + + n 2 3 n =1

and identify the function represented by this power series. U se the ratio test:
n +1 sn+1 = |x 2| sn n +1

|x 2|n = n |x 2| . n n +1

S o,

n+

lim

sn+1 = | x 2| sn

Hence, the interval of convergence is |x 2| < 1. (This is eq uivalent to 1 < x 2 < 1. which, in turn, is eq uivalent to 1 < x < 3.) At the right endpoint x = 3, the series is the divergent harmonic series, and, at the left endpoint x = 1, the series is the negative of the convergent alternating harmonic series. S o, the series converges for 1 x < 3. + + ( x 2)n L et h ( x) = . By Theorem 46.7, h ( x) = ( x 2)n1 . This series is a geometric series with n n =1 n =1 1 = 1 . Thus, h ( x) = 1 . H ence, first term 1 and ratio (x 2); so its sum is 3 x 1 ( x 2) 3 x h( x) = dx = ln |3 x| + C . N ow, 3 x
h(2) = (2 2)n =0 n n =1
+

and

ln |3 2| + C = 0. S o,

C= 0

M oreov er, since x < 3 in the interv al of conv ergence, 3 x > 0 and, therefore, |3 x| = 3 x. Thus, h(x) = ln (3 x). In P roblems 2 and 3, find the interv al of conv ergence of the giv en series and the behav ior at the endp oints (if any). 2.
x n
n =1 + n 2 2 3 = x + x + x + . 4 9

U se the ratio test:

n +1 sn+1 = |x| 2 sn (n + 1)

|x|n = n |x| . n +1 n2

( )

H ence,

n+

lim

sn+1 = |x|. sn

H ence, the interv al of conv ergence is | x | < 1. The radius of conv ergence is 1. A t x = 1, we obtain the conv ergent p-series with p = 2. A t x = 1, the series conv erges by the alternating series test. Thus, the series conv erges for 1 x 1.

CHAPTER 46 Power Series

389

3.

( x + 1)n ( x + 1)2 ( x + 1)3 = ( x + 1) + + + . 2 3 n n =1

U se the ratio test:

n +1 sn+1 = |x + 1| sn n +1

|x + 1|n = n

n |x + 1|. n +1

H ence,

n+

lim

sn+1 = |x + 1| . sn

H ence, the interv al of conv ergence is |x + 1| < 1. This is eq uiv alent to 1 < x + l < 1, which, in turn, is eq uiv alent + to 2 < x < 0. The radius of conv ergence is 1. A t the right endp oint x = 0, we get the div ergent p-series 1 + n n =1 (1)n (with p = 1 , which conv erges by the 2 ). A t the left endp oint x = 2, we get the alternating series n n =1 A lternating S eries Theorem. Thus, the series conv erges for 2 x < 0.

4.

P rov e Theorem 46.1. S ince an ( x0 c)n conv erges, lim an ( x0 c)n = 0 by Theorem 43.5. H ence, there is a p ositiv e number M such that |an| |x0 c|n < M for all n, by Theorem 42.1. A ssume |x c| < |x0 c|. L et
r = |x c| < 1. |x0 c|
n+

Then,

|an ||x c|n = |an ||x0 c|n r n < Mr n.

Therefore, |an ( x c)n | is conv ergent by comp arison with the conv ergent geometric series Mr n . Thus, an ( x c)n is absolutely conv ergent.

5.

P rov e Theorem 46.2. O nly a v ery intuitiv e argument is p ossible here. A ssume that neither case (a) nor case (c) holds. S ince case (a) does not hold, the p ower series does not conv erge for some x c. S ince case (c) does not hold, the series does conv erge for some x c. Theorem 46.1 imp lies that there is an interv al (c K, c + K) around c in which the series conv erges. The interv al of conv ergence is the max imal such interv al. (U sing Theorem 46.1, one tak es the least up p er bound R1 of all K such that the series conv erges in (c K, c + K). Then, (c R1, c + R1) is the desired interv al.)

6.

P rov e Theorem 46.4. A ssume x is in A and > 0. S ince fn conv erges uniformly to f on A, there is a p ositiv e integer m such that, if n m, then | fn ( y) f ( y)| < /3 for all y in A. S ince fm is continuous at x, there ex ists d > 0 such that, for any x* in A, if |x* x| < d, then | fm(x*) fm(x)| < /3. H ence if |x* x| < d,
| f ( x * ) f ( x)| = |( f ( x * ) fm ( x * )) + ( fm ( x * ) fm ( x)) + ( fm ( x) f ( x))| | f ( x * ) fm ( x * )| + | fm ( x * ) fm ( x) | + | fm ( x) f ( x)| < + + = 3 3 3

This p rov es the continuity of f at x.

f ( x) dx = lim fn ( x) dx . n+ a A ssume > 0. There is a p ositiv e integer m such that, if n m, then | fn ( x) f ( x)| < for all x in [a, b]. ba b Therefore, a | fn ( x) f ( x)| dx < . Then
a

7.

If fn conv erges uniformly to f on [a, b] and each fn is continuous on [a, b], then

f ( x) dx fn ( x) dx =
a

( fn ( x) f ( x)) dx | fn ( x) f ( x)| dx <

a

for n m

390

CHAPTER 46 Power Series

8.

P rov e that the function f defined by a p ower series is continuous within its interv al of conv ergence (C orollary 46.5).
f ( x) = lim Sn ( x) and the conv ergence is uniform by Theorem 46.3. E ach Sn(x), being a p olynomial, is
n+

continuous. H ence, f is continuous by Theorem 46.4. F ind a p ower series about 0 that rep resents the function x 2 . In what interv al is the rep resentation v alid? 1+ x + By formula (46.11), 1 2 = (1)n x 2 n for |x| < 1. H ence 1+ x n= 0
x = (1)n x 2 n+1 1 + x2 n=0
+

9.

for |x| < 1

The series div erges at both endp oints x = 1 and x = 1. In P roblems 10 and 11, use the ratio test to find the interv al of conv ergence and indicate what hap p ens at the endp oints (if any).
n 10 xn . sn+1 (n + 1) |x|n+1 = sn 10 n+1 n |x|n = n + 1 |x| . n 10 10 n

10.

( )

H ence,

n+

lim

sn+1 = |x| . sn 10

W e get conv ergence when |x|/10 < 1, that is, when |x| < 10. That is the interv al of conv ergence. The series div erges at both endp oints 10.
n (x ) . 3
n n

11.

sn+1 (n + 1) |x |n+1 = sn 3n+1

n | x |n = n + 1 | x | n 3 3n

H ence,

n+

lim

sn+1 = |x | . sn 3

S o, the interv al of conv ergence is |x | < 3. The series div erges at both endp oints. 1 2. F ind the interv al of conv ergence of A p p ly the ratio test:
sn+1 ((n + 1)!)2 |x|n+1 = sn (2n + 2)!

(2n)!x .
n

(n!)2

(n + 1)2 (n!)2 |x|n = |x| (2n + 2)(2n + 1) (2n)!

H ence,

n+

lim

sn+1 = |x| . 4 sn

S o, the interv al of conv ergence is |x| < 4. 1 3 . F ind a p ower series about 0 that rep resents x 3 . + 1 x S tart with 1 = x n for |x| < 1. R ep lace x by x3: 1 x n=0
1 = x 3n 1 x3 n= 0
+

for

|x| < 1

(since |x3| < 1 is eq uiv alent to |x| < 1). M ultip ly by x:

x = x 3n+1 1 x3 n= 0
+

for

|x| < 1

In P roblems 14 16, find simp le formulas for the function f (x) rep resented by the giv en p ower series.

CHAPTER 46 Power Series

391

14.

x + x 2 + x 3 + . 2! 3! 4! + n L et f ( x) = x . ( n + 1)! n =1
n +1 n x xf ( x) = x =x n! 1 x = e 1 x n 1 ( + )! n= 0 n =1 x H ence, f ( x) = e 1 x . x 1 9 6 x3 + 1 6 x + 9 x + . + + 3 n L et f ( x) = x . Then f ( x) = x 3n1 = x 2 + x 5 + x 8 + . 3n n =1 n =1 This is a geometric series with ratio x3. S o, it conv erges for |x3| < 1, which is eq uiv alent to |x| < 1. H ence, 2 2 3 f ( x) = x 3 for |x| < 1. Therefore, f ( x) = x 3 dx = 1 3 ln |1 x | + C . But f (0) = 0. H ence, C = 0. A lso, 1 x 1 x 1 x3 > 0 for |x| < 1. Therefore, 1 3 3 f ( x) = 1 3 ln (1 x ) + +

15.

for

|x| < 1.

1 6 . x + 2 x 3 + 3 x 5 + 4 x 7 + . The ratio test shows that the series conv erges for |x| < 1. L et
g( x) = x + 2 x 3 + 3x 5 + 4 x 7 + = nx 2 n1
+

Then 2 g( x) = 2nx
n =1

n =1

2 n 1

H ence, tak ing antideriv ativ es,

+

2 g( x)dx = K + x 2 n = K +
n =1

x2 1 x2

(since x 2 n is a geometric series with ratio x 2 )

n =1

N ow differentiate:
2 g( x ) = D
x

( 1 x x ) = (1 2xx ) ,
2 2 2 2

g( x ) =

x (1 x 2 )2

for

|x| < 1

1 7 . (G C ) A p p rox imate

ln (1 + x) dx to two-decimal-p lace accuracy (that is, with an error < 5/103). x 1 4 1 3 2 By formula (46.8), ln (1 + x) = x 1 2 x + 3 x 4 x + for |x| < 1, S o

1/ 2

+ ln (1 + x) (1)n x n 1 3 1 2 =1 1 2 x + 3 x 4 x + = x n +1 n= 0

By Theorem 46.6(b),

1/ 2

+ + ln (1 + x) (1)n x n+1 (1)n 1 dx = = x n n 1 1 + + n ( + 1)2 2n+1 n= 0 n= 0 0

1/ 2

which is a conv ergent alternating series. In order to get an ap p rox imation with an error less than 5/103, we must find n such that the first omitted term 1 1 is 5 = 1 . S o, we must hav e 200 (n + 1)2 2n+ 1. Trial and error shows that n 3. H ence, we 10 3 200 (n + 1)2 2n+1 can use the terms corresp onding to n = 0, 1, 2:
1 1 + 1 = 65 ~ 0.45 2 16 72 144

This answer can be confirmed by a grap hing calculator (which yields 0.44841421 as an ap p rox imation).

392
+

CHAPTER 46 Power Series

1 8 . F ind the function defined by

2
n=0

xn .

This is a geometric series with ratio 2x and first term 1. H ence, it conv erges for |2x| < 1, that is, for |x| < 1 2 , and 1 . its sum is 1 2x 1 9 . F ind the interv al of conv ergence of A p p ly the ratio test: . ln (x n + 1)
n =1 + n

n +1 sn+1 = |x| sn ln (n + 2)

ln (n + 1) |x|n |x| = ln (n + 1) ln (n + 2)

s By L H p itals rule, lim n+1 = |x|. H ence, the interv al of conv ergence is giv en by |x| < 1. (F or x = 1, we get n+ sn + + (1)n 1 , which we k now is div ergent. F or x = 1, we get the conv ergent alternating series .) 1 ln ( n + ) ln (n + 1) n =1 n =1

20 . A p p rox imate 1 with an error less than 0.0001. e By formula (46.14),

ex = x n! n= 0
+ n

for all x. H ence,

(1)n 1 = e1 = e n! n=0

By the A lternating S eries Theorem, we seek the least n such that 1/n! 0.0001 = 1/10,000, that is, 10,000 n!. Trial and error shows that n 8. S o, we must use the terms corresp onding to n = 0, 1, . . . ,7:
1 1 + 1 1 + 1 1 + 1 1 = 103 ~ 0.3679 2 6 24 120 720 5040 280

(A grap hing calculator yields the answer 0.367 8794412, correct to 10 decimal p laces.) 21 . A p p rox imate

e
0

x2

dx to two-decimal-p lace accuracy, that is, with an error less than 5/103 = 0.005.

By formula (46.14),
ex = x n! n= 0
+ n

for all x.

H ence,

e x =
2

(1)n 2 n x n! n=0

for all x

By Theorem 46.6(b),

e x dx =
2

+ (1)n x 2 n+1 (1)n 1 = + 2 1 ! n n n= 0 0 n = 0 n! 2n + 1 +

1 should be (2n + 1)n! 0.005 = 1/200. S o, 200 (2n + 1)n!. Trial and error shows that n 4. H ence, we should use the first four terms, that is, those corresp onding to n = 0, 1, 2, 3:

W e can ap p ly the A lternating S eries Theorem. The magnitude of the first term omitted

1 1 + 1 1 = 26 ~ 0.743 3 10 42 35

(A grap hing calculator yields the ap p rox imation 0.74682413, correct to eight decimal p laces.) 22. F ind a p ower series ex p ansion for 1 about 0. x+3 + 1 =1 1 . By formula (46.7), 1 = (1)n x n = 1 x + x 2 x 3 + x + 3 3 ( x /3) + 1 1 + x n= 0

for |x| < 1.

CHAPTER 46 Power Series

393

H ence,
1 = (1)n x 3 ( x /3) + 1 n= 0
+ +

x ( ) = (1) 3
n + n n=0

n n

for r

x <1 3

Thus, The series div erges at x = 3.

1 = (1)n x n x+3 3n+1 n= 0

for |x| < 3

23 . F ind a p ower series ex p ansion for 1 about 1. x + 1= 1 . By formula (46.7), 1 = (1)n x n for |x| < 1. H ence, x 1 + ( x 1) 1 + x n= 0
1= 1 = (1)n ( x 1)n x 1 + ( x 1) n= 0
+

for |x 1| < 1

SUPPLEMENTARY PROBLEMS
In P roblems 24 31, find the interv al of conv ergence of the giv en p ower series. 24 .

nx

Ans.
n

1 < x < 1 1 x 1 5 x < 5 1 x 1

25 .

n(nx+ 1) nx5
n n

Ans.

26 .

Ans.
2n

27 .

n(n + x 1)(n + 2) (ln (xn + 1))

x 1+ n
n 3 n +1 2

Ans.

28 .

Ans. 1 x < 1 Ans. 1 x 1 3x5

1 < x <
7 3

29 .

30.

( x 4) n n2 (3x 2)n 5n

Ans.

31.

Ans.

3 2. E x p ress e2x as a p ower series about 0. Ans.

(1)n 2n n x n! n=0

3 3 . R ep resent e x/2 as a p ower series about 2. Ans.

2 (en!) ( x 2)
n n=0

394

CHAPTER 46 Power Series

3 4 . R ep resent ln x as a p ower series about 2. Ans.

ln 2 + (1)n+1 ( x 2)n n2n n =1
+

3 5 . (G C ) F ind ln (0.97) with sev en-decimal-p lace accuracy. (Hint: U se the p ower series for ln (1 x) about 0.) Ans. 0.0304592

3 6 . H ow many terms in the p ower series for ln (1 + x) about 0 must be used to find ln 1.02 with an error 0.00000005? Ans. Three

3 7 . (G C ) U se a p ower series to comp ute e2 to four-decimal-p lace accuracy. Ans. 0.1353

3 8 . (G C ) E v aluate Ans. 0.4940

1/ 2

dx to four-decimal-p lace accuracy. 1 + x4

In P roblems 39 and 40, find the interv al of conv ergence of the giv en series.
x n
n =1 + + n n

39.

Ans.

(, + )

40.

n! x 10
n n= 0

Ans.

x=0

x x 4 1 . R ep resent cosh x = e + e as a p ower series about 0. 2

Ans.

x (2 n)!
n= 0

2n

4 2. F ind a p ower series about 0 for the normal distribution function Ans.

e t / 2 dt .
2

n!(2 ) 2xn + 1
n n= 0

(1)n

2 n +1

4 3 . F ind a p ower series ex p ansion about 0 for ln 1 + x . 1 x Ans.

2 n +1 2 x n +1 2 n=0 +

4 4 . (G C ) A p p rox imate tan 1 1 2 to two-decimal-p lace accuracy. Ans. 0.46

CHAPTER 46 Power Series

395

4 5 . S how that the conv erse of A bels Theorem is not v alid, that is, if f ( x) = an x n for |x| < r, where r is the radius of conv ergence of the p ower series, and lim f ( x) ex ists, then xr 1 f ( x) = .) 1+ x
+

a r
n n=0

n= 0

need not conv erge. (Hint: L ook at

4 6 . F ind a simp le formula for the function f (x) rep resented by n 2 x n . Ans.
x( x + 1) (1 x)3
n =1

4 7 . F ind a simp le formula for the function f (x) rep resented by Ans. x + (1 x) ln (1 x)

xn . n 1)n ( n=2

4 8 . (a) S how that (b) S how that

x = nx n for |x| < 1. (Hint: U se E x amp le 5.) (1 x)2 n =1

+

2 x 2 = n(n 1) x n for |x| < 1. (Hint: F irst div ide the series by x, integrate, factor out x, use (1 x)3 n=2 p art (a), and differentiate.) x( x + 1) = n 2 x n for |x| < 1. (1 x)3 n =1
+ n +

(c) S how that (d) E v aluate Ans.

2n
n =1

and

n =1

n2 2n

(d) 2 and 6

C H A P TE R 4 7

Tayl or and M acl aurin Series. Tayl ors Form ul a with R em ainder
Ta y lo r a n d Ma c la u rin Se rie s
L et f be a function that is infinitely differentiable at x = c, that is, the deriv ativ es f (n)(c) ex ist for all p ositiv e integers n. The T aylo r series fo r f abo ut c is the p ower series

a ( x c)
n n=0

= a0 + a1 ( x c) + a2 ( x c)2 +

f ( n ) ( c) where an = for all n. N ote that f (0) is tak en to mean the function f itself, so that a0 = f (c). n! The Maclaurin series fo r f is the Taylor series for f about 0, that is, the p ower series

a x
n n= 0

= a0 + a1x + a2 x 2 +

where an =

f ( n ) (0) for all n. n!

EX AMPLE 4 7 .1 : T h e M a c la u r in s e r ie s fo r s in x L et f (x) = sin x. Then

f ( x) = cos x, f ( x) = sin x, f ( x) = cos x,

S ince f (x) = sin x, further deriv ativ es rep eat this cycle of four functions. S ince sin 0 = 0 and cos 0 = 1, f (2k)(0) = 0 and (1) k f (2k+1)(0) = (1)k. H ence, a2k = 0 and a2 k +1 = . S o, the M aclaurin series for sin x is (2 k + 1)!
(4)

(2k + 1)! x
k =0

(1) k

2 k +1

3 5 7 = x x + x x + 3! 5! 7!

A n ap p lication of the ratio test shows that this series conv erges for all x. W e do not k now that sin x is eq ual to its M aclaurin series. This will be p rov ed later.
EX AMPLE 4 7 .2 :

L et us find the M aclaurin series for f ( x) = 1 . 1 x

f ( x ) = 1 , (1 x)2 f ( x) = 2 , (1 x)3 f ( x) = 32 , (1 x)4

f 4 ( x) = 4 3 2 , (1 x)5

f 5 ( x) = 5 4 3 62 (1 x)

396
<=>
?@ A BCDEFGGHIJHHHIJHH
GIJHKFL?#MCNOP(Q#@ R
S,T UA V V/<=W,>RXA NY3IZ XP[<V A P\#CN@ N]8=@9D8N@ WY:=
]^Y3N[

CHAPTER 47

Tayl or and M acl aurin Series

397

n! . H ence, a = f ( n ) (0) = 1 for all n, and the M aclaurin series for 1 is n 1 x n! (1 x)n+1 + 1 n x . In this case, we already k now that is eq ual to its M aclaurin series for |x| < 1. 1 x n= 0

W e can see the p attern: f ( n ) ( x) =

Th e o re m 4 7 .1 :

If f ( x ) = bn ( x c)n for some x c, then this series is the Taylor series for f, that is, bn =
n=0 + n= 0

f ( n ) ( c) for n!

all n. In p articular, if f ( x ) = bn x n for some x 0, then this series is the M aclaurin series for f.

A ssume f ( x ) = bn ( x c)n for some x c. Then f (c) = b0. By term-by-term differentiation (Theorem 46.7),
f ( x ) = nbn ( x c)n1 in the interv al of conv ergence of bn ( x c)n . H ence f (c) = b1. D ifferentiating again, we
+ n=0 +

get f ( x ) = n(n 1)bn ( x c)n 2. S o, f (c) = 2b2 and, therefore, b2 =

n=0 +

n=0

n= 0

f (c) . 2! f (c) . 3!

D ifferentiating again, we get f ( x ) = n(n 1)(n 2)bn ( x c)n3. S o, f (c) = 3! b3 and, therefore, b3 = n=0 Iterating this p rocedure, we obtain
bn = f ( n ) ( c) n! for all n 0

Thus, the series is the Taylor series for f.

EX AMPLE 4 7 .3 :

W e already k now by formula (46.8) that

n ln(1 + x) = (1)n1 x n n =1 +

for | x | < 1

n H ence, by Theorem 47.1, the series (1)n1 x must be the M aclaurin series for ln (1 + x). It is not necessary to n n =1 go through the laborious p rocess of comp uting the M aclaurin series for In (1 + x) directly from the definition of M aclaurin series.

EX AMPLE 4 7 .4 :

If f ( x) = 1 , find f (47) (0). 1 x + f ( n ) (0) 1 n . W e k now that = x for |x| < 1. H ence, by Theorem 47.1, the coefficient of xn, namely 1, is eq ual to 1 x n! n=0 ( 47 ) f (0) and, therefore, f (47)(0) = (47)!. S o, for n = 47, 1 = (47)!

Th e o re m 4 7 .2 (Ta y lo rs F o rm u la w ith Re m a in d e r): L et f by a function such that its (n + 1)st deriv ativ e f (n+1) ex ists in (a, b ). A ssume also that c and x are in (a, b). Then there is some x* between c and x such that

f ( x) = f (c) + f (c)( x c) + =
k =0 n

f (c) f ( n ) ( c) f ( n+1) ( x* ) ( x c) 2 + + ( x c) n + ( x c)n+1 2! (n + 1)! n!

f ( k ) ( c) ( x c) k + Rn ( x) k!

(47.1)

H ere, Rn ( x) =

f ( n+1) ( x * ) ( x c)n+1 is called the remainder term or the erro r. (n + 1)!

Theorem 47.2 can be deriv ed from Theorem 13.6 (the H igher-O rder L aw of the M ean).

398

CHAPTER 47

Tayl or and M acl aurin Series

Ap p lic a tio n s o f Ta y lo rs F o rm u la w ith Re m a in d e r

(I ) S h o w in g t h a t c e r t a in f u n c t io n s a r e r e p r e s e n t e d b y t h e ir T a y lo r s e r ie s b y p r o v in g t h a t lim R n (x ) = 0
n+

F rom Taylors formula (47.1),

Rn ( x) = f ( x)
k =0 n

f ( k ) ( c) ( x c) k k!

If lim Rn ( x) = 0 then
n+

f ( x) = lim

n+

k =0

+ f ( k ) ( c) f ( k ) ( c) ( x c) k ( x c) k = k! k! k =0

that is, f (x) is eq ual to its Taylor series.

+ dn xn Remark: lim conv erges for all x. H ence, by Theorem = 0 for any d. To see this, recall that n! n= 0 x n n+ n ! 43.5, lim = 0 for any x. n+ n !

EX AMPLE 4 7 .5 : sin x is eq ual to its M aclaurin series. W hen f (x) = sin x, then ev ery deriv ativ e f (n)(x) is either sin x, cos x, sin x, or cos x, and, therefore, | f (n)(x)| 1. S o,

|Rn ( x)| =

f ( n+1) ( x* ) |( x c)n+1| ( x c)n+1 (n + 1)! (n + 1)!

x c)n+1 | By the R emark abov e, lim | ( lim Rn ( x) = 0. Therefore, sin x is eq ual to its M aclaurin series: = 0 . H ence, n + (n + 1)! n+ sin x =
3 5 7 (1) k x 2 k +1 = x x + x x + (2 k + 1)! 3! 5! 7! k =0 +

(47.2)

(I I ) A p p r o x im a tin g v a lu e s o f fu n c tio n s o r in te g r a ls U se a bound on Rn(x) to get a bound on the error when we ap p rox imate the sum of an infinite series by a p artial sum.
EX AMPLE 4 7 .6

L et us ap p rox imate e to four decimal p laces, that is, with an error less than 0.00005. + n P reliminary result: e < 3. To see this, note that, since e x = x , n! n=0
e = e1 = 1 = 1 + 1 + 1 + 1 + 1 + 1 + 2! 3! 4! 5! n! n=0 < 1 + 1 + 1 + 12 + 13 + 14 + 2 2 2 2 1 = 1 + 1n = 1 + =1+ 2 = 3 1 (1/ 2) 2 n= 0
+ +

N ow, for the function f (x) = ex, we wish to mak e the magnitude of the error Rn(1) < 0.00005. By Taylors formula with remainder, with x = 1,
f ( n+1) ( x* ) , (n + 1)!

|Rn (1)| =

where 0 < x* < 1

CHAPTER 47

Tayl or and M acl aurin Series

399

S ince D x (e x ) = e x, f ( n+1) ( x) = e x for all x. S o, f ( n+1) ( x * ) = e x . S ince ex is an increasing function, e x < e1 = e < 3. Thus, 3 . S ince we wish to mak e the error <0.00005, it suffices to hav e | Rn (1)| < (n + 1)!
* *

3 0.00005, (n + 1)!

that is,

3 1 , (n + 1)! 20, 000

60, 000 (n + 1)!.

8

Trial and error shows that this holds for n 8. S o, we can use the p artial sum 1 ~ 1.7183. n! n=0
Th e o re m 4 7 .3 (Th e Bin o m ia l Se rie s ):

A ssume r 0. Then
+

(1 + x)r = 1 +
n =1

r (r 1)(r 2) (r n + 1) n x n!

for | x |< 1

= 1 + rx +

r (r 1) 2 r (r 1)(r 2) 3 x + x + 2! 3!

(47.3)

A p p ly the ratio test to the giv en series:

sn+1 r (r 1)(r 2) (r n) x n+1 = (n + 1)! sn r (r 1)(r 2) (r n + 1) x n n!

S o,
lim sn+1 (r n) x = lim = | x| sn n +1 n+

n+

H ence, the series conv erges for |x| < 1. F or a sk etch of the p roof that this series is eq ual to (1 + x)r, see P roblem 31. N ote that, if r is a p ositiv e integer k, then the coefficients of xn for n > k are 0 and we get the binomial formula
(1 + x) k =
n= 0 k

k! xn n!( k n)!

EX AMPLE 4 7 .7 :

L et us ex p and 1 + x as a p ower series about 0. This is the binomial series for r = 1 2.

1 + x = 1 + 1/ 2 x + 1! + (1/ 2)(1/ 2) 2 (1/ 2)(1/ 2)(3/ 2) 3 x + x 2! 3!

(1/ 2)(1/ 2)(3/ 2)(5/ 2) 4 x + 4!

= 1 + 1 x 1 x 2 + 1 x3 5 x 4 + 2 8 16 128
EX AMPLE 4 7 .8 :

(47.4)
1 1 x

L et us find a p ower series ex p ansion about 0 for

1 2

Tak e the binomial series for r = , and then rep lace x by x:

1 = 1 + 1/ 2 ( x) + (1/ 2)(3/ 2) ( x)2 + (1/ 2)(3/ 2)(5/ 2) ( x)3 + 1! 2! 3! 1 x + 1 3 5 (2n 1) n x + n! 2 n 1 3 5 (2n 1) n x 2 4 6 (2n) n =1
+

=1+

(47.5)

400
+ +

CHAPTER 47

Tayl or and M acl aurin Series

+

Th e o re m 4 7 .4 :

If f ( x) = an x n for |x| < R1 and g( x) = bn x n for |x| < R2, then f ( x) g( x) = cn x n for |x| < minimum
n n= 0 n= 0 n=0 k =0

(R1, R2), where cn = ak bn k.

The reader is referred to more adv anced treatments of analysis for a p roof. Theorem 47.4 guarantees that, if f and g hav e p ower series ex p ansions, then so does their p roduct.

SOLV ED PROBLEMS
1. F ind a p ower series ex p ansion about 0 for cos x. W e k now by E x amp le 5 that
sin x =
k =0 + 3 5 7 (1) k x 2 k +1 = x x + x x + for all x 3! 5! 7! (2 k + 1)!

Then, by Theorem 46.7, we can differentiate term by term:

cos x =
k =0 + 2 4 6 (1) k 2 k x = 1 x + x x + for all x (2 k )! 2! 4! 6!

2.

F ind a p ower series about for sin x. 2 U se the identity sin x = cos x . Then, by P roblem 1, 2

sin x =
k =0

(1) k x (2 k )! 2

2k

=1 1 x 2! 2

) + 41! ( x 2 )
2 4

3.

If f (x) = tan1 x, ev aluate f (38)(0). W e k now by formula (46.12) that

2 n +1 1 5 1 7 3 for | x | < 1 =x 1 tan 1 x = (1)n x 3 x + 5 x 7 x + n 2 +1 n= 0 +

f (38) (0) . But the coefficient of x38 is 0. H ence, by Theorem 47.1, the coefficient of x38 in this p ower series is eq ual to (38)! (38) S o, f (0) = 0.

4.

F ind p ower series ex p ansions about 0 for the following functions: (a) cos (x2) (a) cos x =
+

(b) xe2x

(c) 1/ 3 1 + x

+ (1) k 2 k (1) k 4 k x by P roblem 1. Therefore, cos( x 2 ) = x . (2 k )! (2 k )! k =0 k =0 + + k (1) k 2 k k x . H ence, (b) W e k now that e x = x . S o, e2 x = k ! k! k =0 k =0

xe2 x =
k =0

(1) k 2 k k +1 + (1)n1 2n1 n x x = (n 1)! k! n =1

(c) This is the binomial series for r = 1 3.

(1/ 3)(4 / 3) 2 (1/ 3)(4 / 3)(7 / 3) 3 x + 1/ 3 1 + x = 1 1 x + x 3 2! 3! + (1/ 3)(4 / 3)(7 / 3)(10 / 3) 4 x + 4! (1)n (1 4 7 (3n 2)) n x 3n n ! n =1
+

=1+

CHAPTER 47

Tayl or and M acl aurin Series

401

5.

F ind the first fiv e terms of the M aclaurin series for ex(sin x). Metho d 1: L et f (x) = ex(sin x). Then
f ( x) = e x (sin x + cos x), f ( x) = 2e x (cos x), and f ( x) = 2e x (cos x sin x)

f ( 4 ) ( x) = 4 e x (sin x),

f (5) ( x) = 4 e x (sin x + cos x)

H ence, since an =

f ( n ) (0) 1 , we get a0 = 0, a1 = 1, a2 = 1, a3 = 1 3 , a4 = 0 , and a5 = 30 . Thus n!

3 5 e x (sin x) = x + x 2 + x x + 3 30

2 3 3 5 Metho d 2: e x (sin x) = 1 + x + x + x + x x + x . If we multip ly out according to the rule in 2! 3! 3! 5! 1 1 1 1 12 + 120 = 30 . Theorem 47.4, we get the same result as abov e. F or ex amp le, c5 = 24

)(

6.

3 5 W e k now that sin x = x x + x . F or what v alues of x will ap p rox imating sin x by x p roduce an error < 0.005? 3! 5! f (3) ( x* ) 3 | x |3 | R2 ( x)| = x | f (3) ( x)| 1 since f (3) is cos x.) S o, we req uire | x |3/6 < 0.005, which is 3! 6 . (H ere, eq uiv alent to |x|3 < 0.03. S o, we want | x | < 3 0.03 ~ 0.31.

7.

3 If we ap p rox imate sin x by x x for |x| < 0.5, what is a bound on the error? 3! S ince sin x is eq ual to an alternating series for any x, the error will be less than the magnitude of the first term omitted, in this case |x|5/5!. W hen |x| < 0.5, the error will be less than 1 (0.5)5 ~ 0.00026. 120

8.

A p p rox imate

sin x dx with an error less than 0.005. x sin x =

3 5 7 (1) k x 2 k +1 = x x + x x + 7 ! ( 2 k + 1 )! 3 ! 5 ! k =0 +

H ence,

+ 2 4 6 (1) k sin x = x2k = 1 x + x x + 2 1 3 5 7 + x ( k )! ! ! ! k =0

Therefore,

+ + 1 (1) k (1) k x 2 k +1 sin x dx = 2k = x dx 0 x 0 + 2 1 2 ( k )! ( k + 1)! 2 k + 1 k =0 k =0 1

=
k =0

(1) k 1 (2 k + 1)! 2 k + 1

This is an alternating series. W e must find k so that

1 It is true for k 2. S o, we need 1 18 = 17 18

1 1 0.005, or, eq uiv alently, 200 (2k + 1)!(2k + 1). (2 k + 1)! 2 k + 1

~ 0.944.

9.

F ind a p ower series about 0 for sin1 x. By formula (47.5),

1 = 1 + 1 3 5 (2n 1) x n 2 4 6 (2n) 1 x n =1
+

for r | x| < 1

R ep lace x by t2.
1 3 5 (2n 1) 2 n 1 t =1+ 2 2 4 6 (2n) 1 t n =1
+

for | t | < 1

402

CHAPTER 47

Tayl or and M acl aurin Series

S o, for |x| < 1,

sin 1 x =
x

1 dt = x + 1 3 5 (2n 1) x 2 n+1 2 4 6 (2n) 2n + 1 1 t2 n =1

1 0 . F ind M aclaurin series for the following functions: (a) sin(x3); (b) sin2 x. R ecall that, if a function has a p ower series ex p ansion in an interv al about 0, then that p ower series is the M aclaurin series of the function.
+ (1) k (1) k x 2 k +1 for all x. H ence, sin( x 3 ) = x 6 k +3 and this series is the M aclaurin series for 2 1 2 + ( k )! ( k + 1)! k =0 k =0 sin (x3). + 1 cos(2 x) 1 (1) k 22 k 2 k + (1) k +1 22 k 1 2 k x by P roblem 1. S o, the M aclaurin series for x = = 1 (b) sin 2 x = (2 k )! 2 2 (2 k )! k =1 k =0 + (1) k +1 22 k 1 2 k x . sin2 x is (2 k )! k =1

(a) sin x =

1 1 . F ind the first four nonz ero terms of the M aclaurin series for f (x) = sec x. It would be v ery tedious to comp ute the successiv e deriv ativ es. Instead, since sec x cos x = 1, we can p roceed differently. W e assume sec x = an x n . Then
n= 0 +

+ + (1) k 2 k an x n x =1 k = 0 (2 k )! n=0
2 4 6 (a0 + a1 x + a2 x 2 + a3 x 3 + ) 1 x + x x + = 1 2 24 720

W e then multip ly out, comp are coefficients on the two sides of the eq uation, and solv e for the an.
a0 = 1, a1 = 0, a2 = 1 2 ; a3 = 0; a4 =
5 24

; a5 = 0; a6 =

61 720

Thus,
sec x = 1 + 1 x 2 + 5 x 4 + 61 x 6 + 2 24 270
2 4 6 A n alternativ e method would be to carry out a long div ision of 1 by 1 x + x x + 2 24 720

SUPPLEMENTARY PROBLEMS
1 2. F ind the M aclaurin series for the following functions: (a) sin (x5); (b) 1 5 ; (c) cos2 x. 1+ x Ans. (a)

(2k + 1)! x
k =0 +

(1) k

10 k + 5

; (b)

(1)
k =0

x 5n ; (c) 1 +

(1) k 22 k 1 2 k x (2 k )! k =1

1 3 . F ind the Taylor series for ln x about 2. Ans.

ln 2 + (1)n1
n =1

( x 2)n n2n

1 4 . F ind the first three nonz ero terms of the M aclaurin series for (a) sinx x ; (b) ex cos x. e Ans.
1 3 3 (a) x x 2 + 1 3 x + ; (b) 1 + x 3 x +

CHAPTER 47

Tayl or and M acl aurin Series

403

1 5 . C omp ute the first three nonz ero terms of the M aclaurin series for tan x. Ans.
3 x+ 1 3 x + 2 15

x5 +

1 6 . C omp ute the first three nonz ero terms of the M aclaurin series for sin1 x. Ans.
3 x+ 1 6 x + 3 40

x5 +

1 7 . F ind the Taylor series for cos x about . [ Hint : U se an identity for cos + x .] 3 3 3 + + 2k 2 k +1 k k (1) (1) 3 Ans. 1 x x 2 k = 0 (2 k )! 3 2 3 (2 k + 1)! k =0

( ( ))

1 8 . (G C ) U se p ower series to ap p rox imate Ans. 0.4872

1/ 2

tan 1 x dx x

1 9 . (G C ) U se p ower series to ap p rox imate Ans. 0.4484

1/ 2

ln(1 + x) dx correctly to four decimal p laces. x

20 . (G C ) U se p ower series to ap p rox imate Ans. 1.0948

1 + x 2 dx correctly to four decimal p laces.

2 0.05 < 1.06.) 21 . (G C ) W hat is a bound on the error if we ap p rox imate ex by 1 + x + 1 2 x for |x| 0.05? (Y ou may use e

Ans.

0.0000221

22. (G C ) W hat is a bound on the error if we ap p rox imate ln (1 + x) by x for |x| 0.05? Ans. 0.00125

23 . (G C ) U se the Taylor series for sin x about to ap p rox imate sin 62 correctly to fiv e decimal p laces. 3 Ans. 0.88295

24 . (G C ) In what interv al can you choose the angle if the v alues of cos x are to be comp uted using three terms of its Taylor series about and the error must not ex ceed 0.000 05? 3 Ans.
x 0.0669 3

25 . (G C ) U se p ower series to comp ute to four-decimal-p lace accuracy: (a) e2; (b) sin 32 ; (c) cos 36 . Ans. (a) 0.1353; (b) 0.5299; (c) 0.8090

26 . (G C ) F or what range of x can:

2 (a) ex be rep laced by 1 + x + 1 2 x if the allowable error is 0.0005? 1 1 3 (b) sin x be rep laced by x 6 x + 120 x 5 if the allowable error is 0.00005?

Ans.

(a) |x| < 0.1; (b) |x| < 47

404
27 . U se p ower series to ev aluate: (a) lim e e x3 x0 Ans. (a) 1 ; (b) e 6 2
sin x

CHAPTER 47
; (b) lim e e x2 x0
cos x

Tayl or and M acl aurin Series

28 . (G C ) U se p ower series to ev aluate: (a) (b) (c)

/2

0 1

2 1/ 2 (1 1 dx (to three-decimal-p lace accuracy). 2 sin x )

cos
0 1/ 2 0

x dx (to fiv e-decimal-p lace accuracy).

dx (to four-decimal-p lace accuracy). 1 + x4

Ans. (a) 1.854; (b) 0.76355; (c) 0.4940

3 29 . (G C ) U se p ower series to ap p rox imate the length of the curv e y = 1 3 x from x = 0 to x = .5, with four-decimalp lace accuracy.

Ans.

0.5031

3 0 . (G C ) U se p ower series to ap p rox imate the area between the curv e y = sin (x2) and the x ax is from x = 0 to x = 1, with four-decimal-p lace accuracy. Ans. 0.3103

3 1 . P rov e that the binomial series ex p ansion in Theorem 47.3 is correct.

+ dy Hint: L et y = 1 + r (r 1)(r 2) (r n + 1) x n . U se term-by-term differentiation to find the series for and dx n ! n =1 dy ry dy r dx . Then deriv e y = (1 + x)r. [U se sep aration of v ariables ; .] show that = = dx 1 + x y 1+ x

3 2. E x p and the p olynomial f (x) = x4 11x3 + 43x2 60x + 14 as a p ower series about 3, and find Ans. 1.185

3.2

f ( x) dx.

C H A P TE R 4 8

PartialD erivatives
F u n c tio n s o f Se v e ra l V a ria b le s
If a real number z is assigned to each p oint (x, y) of a p art of the xy p lane, then z is said to be giv en as a function, z = f (x, y), of the indep endent v ariables x and y. The set of all p oints (x, y, z) satisfying z = f (x, y) is a surface in three-dimensional sp ace. In a similar manner, functions w = f (x, y, z, . . .) of three or more indep endent v ariables may be defined, but no geometric p icture is av ailable. There are a number of differences between the calculus of one and two v ariables. H owev er, the calculus of functions of three or more v ariables differs only slightly from that of functions of two v ariables. The study here will be limited largely to functions of two v ariables.

Lim its
By an o pen disk with center (a, b) we mean the set of p oints (x, y) within some fix ed distance d from (a, b), that is, such that ( x a)2 + ( y b)2 < . By a deleted disk around (a, b) we mean an op en disk without its center (a, b). L et f be a function of two v ariables and assume that there are p oints in the domain of f arbitrarily close to (a, b). To say that f (x, y) has the limit L as (x, y) ap p roaches (a, b) means intuitiv ely that f (x, y) can be made arbitrarily close to L when (x, y) is sufficiently close to (a, b). M ore p recisely,
( x , y )( a ,b )

lim

f ( x , y) = L

if, for any > 0, there ex ists d > 0 such that, for any (x, y) in the domain of f and in the deleted disk of radius d around (a, b), | f (x, y) L| < . This is eq uiv alent to saying that, for any > 0, there ex ists d > 0 such that 0 < ( x a)2 + ( y b)2 < imp lies | f (x, y) L| < for any (x, y) in the domain of f. N ote that it is not assumed that f (a, b) is defined. L aws for limits analogous to those for functions of one v ariable (Theorems 7.1 7.6) also hold here and with similar p roofs.
EX AMPLE 4 8 .1 :

U sing these standard laws for limits, we see that

3xy 2 + 1 xy = 3(3)(1) + 1 (3)(1) = 9 + 3 = 8 2 2 7 +1 2 ( x , y )( 3,1) 7 + y lim
21 8

In some cases, these standard laws do not suffice. 3xy 2 0 , which is indeterminate. S o, we need a L et us show that lim 2 2 = 0. O ur usual limit rules would yield 0 ( x , y ) ( 0 ,0 ) x + y more inv olv ed argument. A ssume > 0. N ow,
EX AMPLE 4 8 .2 :

3xy 2 3xy 2 y2 3 | x | 3 x 2 + y 2 < 3 = 2 2 0 = 2 2 = 3 | x| 2 x +y x +y x + y2

if we choose d = /3 and we assume that 0 < x 2 + y 2 < .

405
<=>
?@ A BCDEFGGHIJHHHIJHH
GIJHKFL?#MCNOP(Q#@ R
S,T UA V V/<=W,>RXA NY3IZ XP[<V A P\#CN@ N]8=@9D8N@ WY:=
]^Y3N[

406

CHAPTER 48

Partial D erivatives

x 2 y2 2 2 does not ex ist. ( x , y ) ( 0 ,0 ) x + y x 2 y2 x 2 = = 1. S o, the limit along the x ax is is 1. N ow L et ( x, y) (0, 0) along the x ax is, where y = 0. Then 2 x + y2 x 2 2 2 2 x y y = 2 = 1. S o, the limit along the y ax is is 1. H ence, let ( x, y) (0, 0) along the y ax is, where x = 0. Then 2 x + y2 y there can be no common limit as one ap p roaches (0, 0), and the limit does not ex ist.
EX AMPLE 4 8 .3 :

L et us show that

lim

EX AMPLE 4 8 .4 :

x 2 y2 2 2 does not ex ist. ( x , y ) ( 0 , 0 ) x + y 2 x 2 y2 H ere, we cannot use the same argument as in E x amp le 3, since 2 2 ap p roaches 1 as (x, y) ap p roaches x +y (0, 0) along both the x ax is and the y ax is. H owev er, we can let (x, y) ap p roach (0, 0) along the line y = x. Then 2 2 2 x 2 y 2 = x 2 x 2 = 0. S o, x 2 y 2 0 along y = x. S ince this is different from the limit 1 ap p roached along x2 + x2 x 2 + y2 x 2 + y2 the x ax is, there is no limit as ( x, y) (0, 0).

L et us show that

lim

C o n tin u ity
L et f be a function of two v ariables and assume that there are p oints in the domain of f arbitrarily close to (a, b). Then f is co ntinuo us at (a, b) if and only if f is defined at (a, b), lim f ( x, y) ex ists, and ( x , y ) ( a , b ) lim f ( x, y) = f (a, b)

( x , y ) ( a , b )

W e say that f is continuous on a set A if f is continuous at each p oint of A. This is a generaliz ation to two v ariables of the definition of continuity for functions of one v ariable. The basic p rop erties of continuous functions of one v ariable (Theorem 8.1) carry ov er easily to two v ariables. In addition, ev ery p olynomial in two v ariables, such as 7 x 5 3xy 3 y 4 + 2 xy 2 + 5, is continuous at all p oints. E v ery continuous function of one v ariable is also continuous as a function of two v ariables. The notions of limit and continuity hav e obv ious generaliz ations to functions of three or more v ariables.

Pa rtia l D e riv a tiv e s

L et z = f (x, y) be a function of two v ariables. If x v aries while y is held fix ed, z becomes a function of x. Then its deriv ativ e with resp ect to x
lim f ( x + x , y) f ( x , y) x

x0

is called the (first) partial deriv ativ e of f with resp ect to x and is denoted fx (x, y) or

z f or . x x S imilarly, if y v aries while x is held fix ed, the (first) partial deriv ativ e of f with resp ect to y is f y ( x , y) = z f f ( x, y + y) f ( x, y) = = lim y y y y0

EX AMPLE 4 8 .5 : L et f (x, y) = x2 sin y. Then fx(x, y) = 2x sin y and fy(x, y) = x2 cos y. N ote that, when fx is comp uted, y is temp orarily treated lik e a constant, and, when fy is comp uted, x is temp orarily treated lik e a constant. The p artial deriv ativ es hav e simp le geometric interp retations. C onsider the surface z = f (x, y) in F ig. 48-1. Through the p oint P(x, y, z), there is a curv e APB that is the intersection with the surface of the p lane through P p arallel to the xz p lane (the p lane determined by the x ax is and the z ax is). S imilarly, CPD is the curv e through P that is the intersection with the surface z = f (x, y) of the p lane through P that is p arallel to the yz p lane. A s x v aries while y is held fix ed, P mov es along the curv e APB , and the v alue of z at (x, y) is the slop e of the tangent line to the curv e APB at P. x S imilarly, as y v aries while x is held fix ed, P mov es along the curv e CPD , and the v alue of z at (x, y) is the slop e of y the tangent line to the curv e CPD at P.

CHAPTER 48

Partial D erivatives

407

Fig. 48-1

Pa rtia l D e riv a tiv e s o f H ig h e r Ord e r

W e can tak e the p artial deriv ativ es with resp ect to x and y of
z , yielding x z z 2 z 2 z = fyx ( x, y) = and 2 = f xx ( x , y) = x x y x y x x

S imilarly, from

z we obtain y 2 z z 2 z z = fxy ( x, y) = and 2 = f yy ( x , y) = y y y x y x y

A ssume that fxy and fyx ex ist and are continuous in an op en disk . Then fxy = fyx at ev ery p oint of the disk .

Th e o re m 4 8 .1 :

F or a p roof, see P roblem 30.

EX AMPLE 4 8 .6 :

L et us v erify Theorem 48.1 for f (x, y) = x2(sin yx).

fx ( x, y) = x 2 (cos yx)( y) + 2 x(sin yx) = x[ xy(cos yx) + 2 sin yx] fy ( x, y) = x 2 (cos yx) x + x 3 (cos yx) fyx ( x, y) = x[ x( y( sin yx)( x) + cos yx) + 2(cos yx)( x)] = x 2 [ xy sin yx + 3 cos yx] fxy ( x, y) = x 3 ( sin yx)( y) + 3x 2 cos yx = x 2 [ xy sin yx + 3 cos yx]

P artial deriv ativ es also can be defined for functions of three or more v ariables. A n analogue of Theorem 48.1 holds for any two orderings of giv en subscrip ts. N ote that p artial deriv ativ es may fail to ex ist when the req uired limits do not ex ist.

SOLV ED PROBLEMS
1. E v aluate: (a)
x y x cos . 4 S ince the standard limit laws ap p ly, the limits are: (a) 2(3)(2)4 7(3)2(2)2 = 96 252 = 156; (b) cos = 2 4 2
( x , y ) ( 3 ,2 )

lim (2 xy 4 7 x 2 y 2 ); (b)

( x , y ) ( ,0 )

lim

2.

E v aluate

x2 . 2 ( x , y ) ( 0 ,0 ) x + y A s ( x, y) (0, 0) along the y ax is, x = 0 and lim

2

x 2 = 0 0. x 2 + y2

408

CHAPTER 48

Partial D erivatives

2 2 A s ( x, y) (0, 0) along the x ax is, y = 0 and 2 x 2 = x 2 = 1 1. x +y x H ence, the limit does not ex ist.

3.

E v aluate

( x , y ) ( 0 ,0 )

lim

xy . x 2 + y2 xy | y| 0 as ( x, y) (0, 0). S o, x 2 + y2
( x , y ) ( 0 , 0 )

S ince |x | = x 2 x 2 + y 2 ,

lim

xy = 0. x 2 + y2

4.

The function f ( x, y) =

sin( x + y) is continuous ev erywhere ex cep t at (0, 0) and on the line y = x, where it is not x+y defined. C an f (0, 0) be defined so that the new function is continuous? sin( x + y) 1, since lim sin u = 1. S o, if we let f (0, 0) = 1, the A s ( x, y) (0, 0), x + y 0 and, therefore, x+y u u0 new function will be continuous at (0, 0). Thus, the original discontinuity was remov able.

In P roblems 5 9, find the first p artial deriv ativ es. 5. z = 2x2 3xy + 4y2. Treating y as a constant and differentiating with resp ect to x yields z = 4 x 3y . x Treating x as a constant and differentiating with resp ect to y yields z = 3x + 8 y . y
2 y2 z= x + . y x y2 Treating y as a constant and differentiating with resp ect to x yields z = 2 x 2 . y x x 2 2y Treating x as a constant and differentiating with resp ect to y yields z = x 2 + . x y y

6.

7.

z = sin (2x + 3y).

z = 2 cos(2 x + 3y) x z = tan 1 ( x 2 y) + tan 1 ( xy 2 ). z = 2 xy + y 2 x 1 + x 4 y 2 1 + x 2 y 4 z = ex
2 + xy

and

z = 3 cos(2 x + 3y) y

8.

and

z = x 2 + 2 xy y 1 + x 4 y 2 1 + x 2 y 4

9.

z = e x 2 + xy (2 x + y) x

and

z = xe x 2 + xy y

1 0 . The area of a triangle is giv en by K = 1 2 ab sin C . W hen a = 20, b = 30, and C = 30 , find: (a) The rate of change of K with resp ect to a, when b and C are constant. (b) The rate of change of K with resp ect to C, when a and b are constant. (c) The rate of change of b with resp ect to a, when K and C are constant.
15 (a) K = 1 b sin C = 1 2 (30)(sin 30 ) = 2 a 2 (b) K = 1 ab cos C = 1 2 (20)(30)(cos 30 ) = 150 3 C 2 2( 1 ab sin C ) (c) b = 2 K and b = 22 K = 2 2 =b =3 a sin C a 2 a sin C a sin C a

CHAPTER 48

Partial D erivatives

409

In P roblems 11 13, find the first p artial deriv ativ es of z with resp ect to the indep endent v ariables x and y. 1 1 . x2 + y2 + z2 = 25. [This is the eq uation of a sp here of radius 5 and center (0, 0, 0).] D ifferentiate imp licitly with resp ect to x, treating y as a constant, to obtain:
2 x + 2 z z = 0. x

H ence,

z = x z x

D ifferentiate imp licitly with resp ect to y, treating x as a constant:

2 y + 2 z z = 0. y

H ence,

z = y z y

1 2. x2(2y + 3z) + y2(3x 4z) + z2(x 2y) = xyz. D ifferentiate imp licitly with resp ect to x:
2 x(2 y + 3z) + 3x 2 z + 3y 2 4 y 2 z + 2 z( x 2 y) z + z 2 = yz + xy z x x x x 4 xy + 6 xz + 3y 2 + z 2 yz S olv ing for z yields: z = 2 . 3x 4 y 2 + 2 xz 4 yz xy x x

D ifferentiate imp licitly with resp ect to y:

2 x 2 + 3x 2 z + 2 y(3x 4 z) 4 y 2 z + 2 z( x 2 y) z 2 z 2 = xz + xy z y y y y 2 x 2 + 6 xy 8 yz 2 z 2 xz S olv ing for z yields: z = 2 . 3x 4 y 2 + 2 xz 4 yz xy y y

1 3 . xy + yz + zx = 1. D ifferentiating with resp ect to x yields y + y z + x z + z = 0 , whence x x z D ifferentiating with resp ect to y yields x + y + z + x z = 0 , whence y y

z = y + z . x+y x z = x + z . x+y y

1 4 . C onsidering x and y as indep endent v ariables, find r , r , , when x = e2r cos q, y = e3r sin q. x y x y F irst differentiate the giv en relations with resp ect to x:
1 = 2e2 r cos r e2 r sin x x

and

0 = 3e3r sin r + e3r cos x x

Then solv e simultaneously to obtain r = 2 r cos 2 and = 2 r 3 sin 2 . e (2 + sin ) x e (2 + sin ) x N ow differentiate the giv en relations with resp ect to y:
0 = 2e2 r cos r e2 r sin y y

and

1 = 3e3r sin r + e3r cos y y

Then solv e simultaneously to obtain r = 3r sin 2 and = 3r 2 cos 2 . y e (2 + sin ) y e (2 + sin ) 1 5 . F ind the slop es of the tangent lines to the curv es cut from the surface z = 3x2 + 4y2 6 by p lanes through the p oint (1, 1, 1) and p arallel to the xz and yz p lanes. The p lane x = 1, p arallel to the yz p lane, intersects the surface in the curv e z = 4y2 3, x = 1. Then z = 8 y = 8(1) = 8 is the req uired slop e. y The p lane y = 1, p arallel to the xz p lane, intersects the surface in the curv e z = 3x2 + 2, y = 1. Then z = 6 x = 6 x is the req uired slop e.

410

CHAPTER 48

Partial D erivatives

In P roblems 16 and 17, find all second p artial deriv ativ es of z and v erify Theorem 48.1. 1 6 . z = x2 + 3xy + y2.
z = 2 x + 3y, x z = 3x + 2 y, y
2 2 N ote that z = z . y x x y

2 z = z = 2, x 2 x x 2 z = z = 2, y 2 y y

( )

2 z = z = 3 y x y x 2 z = z = 3 x y x y

( )

1 7 . z = x cos y y cos x.
z = cos y + y sin x, x 2 z = z = y cos x x 2 x x

( )

2 z = z = sin y + sin x y x y x z = x sin y cos x, y 2 z = z = x cos y y 2 y y

( )

2 z = z = sin y + sin x x y x y
2 2 N ote that z = z . y x x y

1 8 . L et f (x, y, z) = x cos (y z). F ind all p artial deriv ativ es of the first, second, and third order. fx = cos (yz), fy = xz sin (yz), fz = xy sin (yz), fxx = 0, fyx = z sin (yz),
2

fzx = y sin (yz)

fyy = xz cos (yz), fzz = xy2 cos (yz),

fxy = z sin (yz) fxz = y sin (yz)

fzy = x(zy cos (yz) + sin( yz)) fyz = x(zy cos (yz) + sin (yz)) N ote that fxy = fyx and fxz = fzx and fyz = fzy. fxxx = 0, fxxy = fxyx = 0, fxxz = fxzx = 0 fxyy = z2 cos (yz), fxzz = y2 cos (yz) fyyy = xz3 sin (yz), fyxx = 0, fyxy = fyyx = z2 cos (yz) fyxz = fyzx = (yz cos (yz) + sin (yz)) fyyz = fyzy = x(z2y sin (yz) + z cos (yz) + z cos (yz)) = xz(zy sin (yz) 2 cos (yz)) fyzz = x(y2z sin (yz) + 2y cos (yz)) = xy(z sin (yz) 2 cos (yz)) fzzz = xy3 sin (yz),
2

fxyz = fxzy = (zy cos (yz) + sin (yz))

fzxx = 0,

fzxy = fzyx = (zy cos (yz) + sin (yz))

fzxz = fzzx = y cos (yz) fzyy = x(z2y sin (yz) + 2z cos (yz)) = xz(zy sin (yz) 2 cos (yz)) fzyz = fzzy = x(zy2 sin (yz) + y cos (yz) + y cos (yz)) = xy(zy sin (yz) 2 cos (yz)) N ote that, in the third order, any two rearrangements of subscrip ts will be eq ual. F or ex amp le, fxyz = fxzy = fyxz = fyzx = fzxy = fzyx = (zy cos (yz) + sin (yz)).

CHAPTER 48

Partial D erivatives

411

2 2 1 9 . D etermine whether the following functions are solutions of L ap laces eq uation z + z = 0: x 2 y 2

(a) z = ex cos y (a) z = e x cos y, x

z = e x sin y, y

x+ y ) (b) z = 1 2 (e

(c) z = x2 y2

2 z = e x cos y x 2 2 z = e x cos y y 2

2 2 + z = 0. Then z x 2 y 2

(b) z = 1 (e x + y ), x 2
z = 1 (e x + y ), y 2

2 z = 1 (e x + y ) x 2 2 2 z = 1 (e x + y ) y 2 2

2 2 + z = ex+ y 0 . S o, z x 2 y 2

z (c) x = 2 x, z = 2 y, y

2 z = 2 x 2 2 z = 2 y 2

2 2 S o, z + z = 0. x 2 y 2

SUPPLEMENTARY PROBLEMS

In P roblems 20 24, ev aluate the giv en limit.

20 .
lim x 2y x2 + y xy x 2 + y2 3xy 2 x 2 + y2 xy 2 x 2 + y4 x 2 + y2 x + y2 + 4 2
2

( x , y )( 1, 2 )

Ans.

5 3

21 .

( x , y ) ( 0 , 0 )

lim

Ans.

no limit

22.

( x , y ) ( 0 , 0 )

lim

Ans.

no limit

23 .

( x , y ) ( 0 , 0 )

lim

Ans.

no limit

24 .

( x , y ) ( 0 , 0 )

lim

Ans.

25 . D etermine whether each of the following functions can be defined at (0, 0) so as to be continuous: (a) Ans.
y2 x 2 + y2

(b)

xy x+y

(c)

x 3 + y3 x 2 + y2

(d)

x+y x 2 + y2

(a) no; (b) no; (c) yes; (d) no

412

CHAPTER 48

Partial D erivatives

26 . F or each of the following functions z, find z and z . x y (a) z = x2 + 3xy + y2

y (b) z = x2 2 y x

Ans. Ans. Ans. Ans. Ans. Ans. Ans.

(c) z = sin 3x cos 4y

y (d) z = tan 1 x

(e) x2 4y2 + 9z2 = 36 (f ) z3 3x2y + 6xyz = 0 (g) yz + xz + xy = 0

z = 2 x + 3y; z = 3x + 2 y x y z = 1 + 2 y ; z = 2 x 1 y3 x 2 x y 2 x 3 y z = 3 cos 3x cos 4 y ; z = 4 sin 3x sin 4 y x y z = y ; z = x x x 2 + y 2 y x 2 + y 2 z = x ; z = 4 y 9 z y 9 z x z = 2 y( x z) ; z = x( x 2 z) x z 2 + 2 xy y z 2 + 2 xy z = y + z ; z = x + z . x+y x + y y x

2 2 z , 2 z , and 2 z . 27 . F or each of the following functions z, find z 2, x y x x y y 2

(a) z = 2x2 5xy + y2

y (b) z = x2 2 y x

Ans. Ans. Ans. Ans.

2 z = 4; 2 z = 2 z = 5; 2 z = 2 x 2 x y y x y 2 2 2 2 z = 6 y ; z = z = 2 1 1 ; 2 z = 6 x x 4 x y y x x 2 x 3 y3 y 2 y 4 2 z = 9 z ; 2 z = 2 z = 12 cos 3x sin 4 y ; 2 z = 16 z x 2 y 2 x y y x 2 2 2 2 2 2 xy z = z = z = z = y x2 2 2 2; 2 2 ( x + y ) x y y x ( x 2 + y 2 )2 x y

(c) z = sin 3x cos 4y

y (d) z = tan 1 x

28 . (a) If z =

2 xy 2 z + y2 2 z = 0. , show that x 2 z 2 + 2 xy xy x x y y 2 2 2 (b) If z = e x cos y and = , show that z + z = 0. x 2 y 2 2 2 + z = z . (c) If z = et(sin x + cos y), show that z x 2 y 2 t

2 2 2 = k2 z +z . (d) If z = sin ax sin by sin kt a 2 + b 2 , show that z t 2 x 2 y 2

29 . F or the gas formula p + a2 (v b) = ct , where a, b, and c are constants, show that v

p 2a(v b) ( p + a /v 2 )v 3 = , v 3 (v b) v t = v b , c p cv 3 v = t ( p + a /v 2 )v 3 2a(v b) p v t = 1 v t p

3 0 . F ill in the following sk etch of a p roof of Theorem 48.1. A ssume that fxy and fyx ex ist and are continuous in an op en disk . W e must p rov e that fxy(a, b) = fyx(a, b) at ev ery p oint (a, b) of the disk . L et h = ( f (a + h, b + h) f (a + h, b)) ( f (a, b + h) f (a, b)) for h sufficiently small and 0. L et F(x) = f (x, b + h) f (x, b). Then h = F (a + h) F (a). A p p ly the M ean-V alue Theorem to get a* between a and a + h so that F(a + h) F(a) = F(a*)h = [fx(a*, b + h) fx(a*, b)] h, and ap p ly the M ean-V alue Theorem to get b* between b and b + h so that fx(a*, b + h) fx(a*, b) = fxy(a*, b*)h. Then,
h = h 2 fxy (a* , b* ) and lim
h 0

h = lim f (a*, b* ) = fxy (a, b) h 2 ( a* , b* )( a , b ) xy

CHAPTER 48

Partial D erivatives

413

By a similar argument using h = ( f (a + h, b + h) f (a, b + h)) ( f (a + h, b) f (a, b)) and the M ean-V alue Theorem, we get
lim
h 0

h = fyx (a, b) h2

3 1 . S how that Theorem 48.1 no longer holds if the continuity assump tion for fxy and fyx is drop p ed. U se the following function:
xy( x 2 y 2 ) f ( x , y) = x 2 + y 2 0 if (x, y) (0, 0) if ( x, y) = (0, 0)

[F ind formulas for. fx(x, y) and fy(x, y) for (x, y) (0, 0); ev aluate fx(0, 0) and fy(0, 0), and then fxy(0, 0) and fyx(0, 0).]