;;151;1;111i1ii1i*lixi iii,.

liinliiliiiiilrlliiilii

ffieWW #ffiere
' ffi6, :dd'-

ffiffiw
16
MutnPLE TNTEGRATToN

ln this chapter we define the integral for functions of two

and three variables. We show that the computation of such
integrals can be reduced to successive integrals of functions
of one variable. These facts enable us to calculate areas,
volumes, and other teometric quantities for various regions
in the plane and solids in three-space.

DEFINITION OF THE DOUBLE INTEGNAT

L€t F be a reeion of area / situated in the xy plane.lVc shall always assume

that a regron iniludcs its boundary curve. Such regions are called clooed
rrgiom in analogy with closed intcrvals on thc real linc-that is, ones which
inilude their endpoints. We subdividc the xy plane into rectanglcs by drawing
lines parallel to the coordinate ax€s. These lines may or may not be equally
spaced (Fig. 1). Starting in some convenient place (such as the upper left

f2
r1 ?3
\
r4 r5
\
\\ F

fn-l lL

HGUNE 1

703
CHAPTEI 15 MUI.TIPIE INTCGRAIION

corner of F), we systematically number all the rectanlbs lytng entlrely within
F, Suppose there are n such and we label them t1,t2t...,r,,. We use the
symbols A(r), A(r2), .,., A(r,) for the areas of these rectangles. The collection
of n rectangles {r1 , 12, .. . , r"} is callcd a eubdivision A of f,'. The norm of the
cubdivieion, denoted as usual by llAll, is the length of the diagonal of the
largest rectangl€ in thc subdivision A.
Suppose that/(x, y) is a function dcfined for all (x, y) in the region F. The
definition of the double integral off oter the rcgion F is similar to the definition
of the integral for functions of one variable. (See page 187.) Select arbitrarily a
point in each of the rectanglcs of the subdivision A, denotfng the coordinates
of the point in the rectangle 11 by ((r, r?r). (See Fig. 2.) Now form thc sum
!{*, tttlA(r) +fGz, qzlA(rz) + '.. +I(*, 4,lA(r^l
or, more compactlj/,

(1)
,Lfle,,nr\A(rt),
This sum is an approximation to the double integral we shall deffne. Sums
such as (1) may be formed for subdivisions'with any positive norm and with
TIGURE.2 the ith point ({r, 4,) ehosen in any way whatsoever in the rectangle 11.

It oan be shown that if the number L exists, then it must be unique.

The double intcgral has a gcometric interpretation in terms of the volume

of a solid. We recall the methods of finding volumes of solids of revolution
developcd on page 228. Now we shall discuss the notion of volume in
somewhat more detail. The definition of volume depends on (i). the definition
of the volume of a cubc-namely, length times width times height, and (ii), a
limiting procass.
sEcTtoN 16.1 DEENmoN oF THE DouBtf lNrEGt L

I.et S be a solid in three-space. We divide all of space into cubes by

constructing planes parallel to the coordinate planeg at a distance apart of
Tvpe (2) lp units, with n some positive integer. In such a nctwork, the cubes are of
tfiree kinds: type ( l), those cubes eatirely within S; type (2), those cubes partly
in S and partly outside S; and type (3), those cubcs entirely outside S (Fig,3).
Wc define
t,r-a
t+-1
rL-J.I
|

V y,t*pe (t)
fl
Type (3)
r,-(s) = (*)' timcs the number of cubes of type (1),

tt
FrGt 3
rJ(S) = 1z" (S) + (+I times the number of cubes of typ€ (2).
Intuitively we e-xpect thd! however the volume of S'is define4 the numbcr
f,-(S) would be smaller than the volume, while the number I/"*(S) would be
larger. It can be shown that, as n increases, I/,-(S) gets larger or at least docs
not decrcass, while y"*(S) getstmaller or at least is nonincreasing. Clearly;
r;(s) < r.*(s),
for all n. Since bounded increasing sequenoes and bounded decreasing
seguenoes tend to limits, the following definitioins are appropriate.

DEFINITIONS Ttp lnaer volure of a soliil S denoteil I/-(S), it lim"-- ["-(S). Tlrc ogttfi
volume, ilenotedy*i^r, is lim,*. lzi* (S). A set of points S in tbee'space has a
volume wheneoer Iz-($) : I/*(S), This cowmn oakrc is denoteil by V{S) ann
is called tlw oolwu of S.

Renark. It is notdifrcult to congtruct point scto fior *Eich f -(S) < I/*(S).
For example, take ,S to be all points (x, y, z) cuch that x, y,and z arerational and
0<x< l,0Sy'3 1,032 < l. The readercan veri$ that [(S) - 0forevery
a while y"*( S) = I for every n. Howevsr, throughout this text we shall discuss
only regions for which I1'1S1-,y+(S), i.e., thosc which have volupe.
If 51 and S2 are two solids with.no polnts in common, it ca8 be shown" as
expoctd that (S1 u 52) = f(Sr) + (Sz). Also, 'the\-utdivisiirn of all of
space into cubes is not vital. Rectangular papllelepipq$.*ould do qually
well with the formula for'the volumc of a rccangular paralhlepipod taken as
length times width times height.
The volude of a solid is intimately coonect€d with thq doirble integral in
the same way that the area of a rcaioa is conoected with the single integral.
SIe now exhibit this connection.
Supposc that,f(a y) is a positive function defined for (x, y) in some region
F (Fig. 4). An item in the sum (1) apbroximating the double intcgrd is
fGt,rtt)A(r),
which we recognize as thc volume of the rectangular column of height/((,,4)
and area of basc {r1) (Fig.4). The sum of suctr s6[nmns is arapprorimation
to the volume of the cylindrical solid bouded by the surface z:f(x, y), the
Ectnt 4 plane figure 4 and lines parallel to the , '
,tr 's"gh tte boundary of F
cHA?Ttt 16 MUTTIru li{rEGRAnoN

(Fig. 5). It can be shown that, with appiopriate hl"porheses on the functionJ
the double integral
ft
,,,lft,
I Yt ae

measures the "volume under the surface" in the same way that a single
integral of a positive function /
lu f,t o*
Jo
measures the area under the curve.
FIGURT 5
The precise result is given in the next theorem which we state without
proof.

-ffi-*=ff*=ffi -
Iflrerev(g'."f:i1;ll.;::lrj;: ', "-,
;,
Methods for the evaluation of double integrals are discussed in Seetion 3.

Given/(x, y) : i + xy and the region F bounded by the lines y : O, y: x, and

x : 1 (Fig. 6), let A be the subdivision formed by the lines x:0,0.2"0.5,0.g, I
and y: 0, 0.2, 0.5, 0.7, 1. Find the value of the sum

i
i= 1
fG,, rilA(r)
if the points ((i,4) are selected at.the centers of the rectangles. Note that this
sum approximates the double integral

t'f
I lf1x. y) dA.
JoJ

Solution Referring to Fig. 6, we see that there are 6 rectangles in the subdivrsion which
we label r11t27 ...> ru, as shown. We compute:

0.7
.4(r,) :9.96, f(A35,0.1) : 1.935
0.5 '4(r') : 9'96' f(0'65,0'1): 1'965
'4(r') :6'64' /(0'9, 0'l) : 1'090
,{(ro) :9.69, /(0.65, 0.35) : 1.2275
.4(rr) :9.96, /(0.9, 0.35) i.315
=
.,{(rr) : g.gr4. /(0,9, 0.6) : 1.540
 7O.
sEcTtoN 15.1 DEflNmON Or THE DoUEtI INIEGTAI"

Multiplying and adding, we find that

n

L f{e,, q)A(r);0.420575 (Answer). tr

i= I

1 PROBIIMS

In each of Problems I through 12, calculate D= rI(8, tiA(ri), L*tf(x, y): fi - x!. For cach i, the point ((,,7,) is taken at
thc sum for the subdivision A of the region f'formed by the given the'lower right corner of thc rectangle r1.
lines and with the points ((i, r?i) selected as directed in each case. 7 Same as Problem 6 with ({r, 4,) taken at thc upper le&
corner of the rectangle r,.
I f(x, y)=x2+2y2; F is the rectangle 03xs l, 0<y< 1. 8 /(x, y) = *
:-O x - -
x2 2xy 3x 2y; F is the trapezoid bounded by
The subdivision A is: x - 0, 0.4, 0.E, 1; y: g, 0.3, 0.7, 1. For
each i the point ((,, 4i) is taken at the center ofthe rectangle
the lines x l, - *
y - Q, y x 1. The suMivision A is:
x:0, 0.4, l, 1.5,2; y:0, 0.6, 1, 1.4 1.8, 2, 3. For each i the
rl point ((,,4,) is taken at th€ ccnter ofthe rectangle r,.
2 Same as Problem I, with ((;, 41) taken at the point of r, 9 Same as Problem 8, with ((,, 4,) taken at the point of rr
which is closcst to the origin. farthest from the origin.
3 f{x, y): I + x2 y2i F is the triangular region formed by
the lines y: 0, y-: x, x : 2. The subdivision A is: x : 0, 0.5, l0 closest
Same as Problem 8, with ({,, 4,) takeo at the point of r;
to the origin.
l, 1.6, 2; ) : 0, 0.6, l, 1.5, 2. For each i, the point ((r, r?r) is
taken at the center of the rectangle r,.
tl Let
f(x,I): #,
r+x+y
4 Same as Problem 3, with ((i, rr,) taken at the point of r;
which is closest to the origin. F is the region bounded by the line y = Q a16 thc curvc y -
Same as Problem 3, with ((,,4,) selected on the lower edge 2x:x2. The suMivision A is: x-O 0.5, 1.0, 1.5, Z y:0,
of r,, midway between the vertical suMivision lines. 0.2,0.d 0.6 0.8, l. For each i, the point ((,,41) is taken at thc
l*t F :{(x, y): 0 s x s 6 0 < y < Ji:7\.The subdivi- cenier of the rectangle r,.
sion A is determined by vertical lines with x interccpts 0, 2, 12 Same as Problem 11 with the poiat ((r, Ti) takcn at the
3, 4,.6 and by horizontal lines with y intercepts 0,2, 4, 5, 6. point ofr, closest to the origin.

2
PROPERTIES OF THE DOUBLE INTECRAT

In analogy with the properties of the definite integral of functions of one

variable, we state several basic properties of the double integral. The simplest
properties are given in the two frrllowing theorems.

THEOREM 2 If c is any number and f is integrable oaer a closed region F, then cf is integrable
and
CHAPTER 16 MUTnru TNTEGMTION

THEORFJI4 3

The result holds for the sum of any finite number of integrable functions. The
proofs of Theorems 2 and 3 follow directly from the definition.

THEOREM 4 Suppose that f is integrable ouer a closed region F and

' m <f(x, y) 3M for all (x, y\ in F.
Then, if A(F) denotes tke area of F, we haue

The proof of Theorem 4 follows exactly the same pattern as does the proof in
the one-variable case. (Sce page 197, Theorem 7.)

THTOREM s I[f and g are integrable ooer F anil f(x, y) 3g(,x, y) for all (x, y) in F, then

The proof is established by the same argument used in the one-variable

case (page [98, Theorem 8).

THEOREM 6

The proof depends on the definition of double integral and on the basic
theorems on limits.
 709
stcTroN 15.2 Pnoprnrlrs oF THE DouBtt tNrtcRAt

2 PROBLEMS

In Problems l.through 7, use Theorem 4 to find in each case Curves /: 3x - x2 and Y: x2 - 3x.
estimates for the largest and smallest values the given double 8 Write a proof of Theorem 3.
integrals can possibly have. 9 Write a proof of Theorem 4.
, ilA where F is the region bounded by the lines x :0,
{r, l0 Write a proof of Theorem 5.
):0,x:2,y:x+3. ll Write a proof of Theorem 6.
2 [l\'1 + y21 dewhere F is the region bounded by the lines 12 Let Fb F2, ..., Fnbe n nonoverlapping regions. State and
x:-2, x:3,y:x*2. _r': -2. prove a generalization of Theorem 6.

3 fit t + 2x2 + y21 dA where F is the regior bo'":rded br lhe 13 Let [i.x. r') be continuous on a closed region F.

lines x : - 3, x : 3, y : q, y : -4.
O dAwhere F is the region bounded b;- the hee --,
: - a:":
$t- : it g" t" , s r"s; ::t:,:rir,:i!j :II i !,,:l:i{t lJ&:
the curve !:2x - x2.
5 JJ(x - y) dA where F is the region enclosed in the cirdc
tii'-l-..-I,rr-r{ < { :* i,,r * wr]: dirrif,
t [L 6|:.
x' + y2 :9. t
lr
6 jJ tI^ 1 + x2 + y2)) ilA where F is the region enclosed in the
14 Use the definition of double rntegial !o sho* dra: the
ellipse 4x2 +9y2:36. volume of a right circular cylinder with height i and radius
7 yr,Ii;4f dA where F is the region bounded by the of base r is nr2h.

EVATUATION OF DOUBTE INTEGRATS. ITERATED INTEGRALS

The definition of the double integral is almost useless as a tool for evaluation
in any particular case. Of course, it may happen that the function/(x, y) and
the region F are particularly simple, so that the limit of the sum

i
i= 1
ttC,, tr)A(ril

can be found directly. However, such limits cannot generally be found. As in

the case of ordinary integrals and line integrals, it is important to develop
simple and routine methods for determining the value of a given double
integral. In this section we show how the evaluation of a double integral may
be periormed by successivc evaluations of single integrals. In other words, we
reduce the problem to one we have already studied extensively.
Let F be the rectangJe with sides x:a, x:b, !:c, !:d, as shown in
Fig. 7. Suppose that/(.r, y) is continuous for (x, y) in F. We form the ordinary
integral u'ith respect to x.

lo fG, v) a*,
Jo
in u'hrch we keep .y f,red when performing the integration. Of course, the
fICURT 7 value of the above integral will depend on the value of y used, and so, using /4
CHAPTER 16 Munru NIEGTANON

for this value, we may write

fb
A(y): I f(x, y) dx.
Jo
The function ,a$) is defined for c < y < d and, in faet, it can be shown that if
I$, y) is continuous on F, then ,4(y) is continuous on [c, d]. The integral of
A(y) may be computed, and we write

Ii ^", *:l,lll,*, i a*fat (1)

We could start the other way around by fixing x and forming the integril

B(x): fo, r,, o,

['
Then

.[ ^,
*=
l:llin, r at)a, (z',)

Note that the integrals are computed successiuely: in (l) we first integrale
with respect to x (keeping y constant) and then with respect to y; in (2) we
first integrate with respect to y'(keeping r constant) and then with respect
to x.

Notation. The brackets in iterated integrals are unwieldy, and we will write

li tt,, y) itx tty to mean I" [l r,' t a.)ar,

t kr', y') dx to mean l)l["t*,tt ar)a*.
ity

Iterated integrals are computed in the usual way, as the next example shows.

ExAMru 1 Evaluate

y3) dx dy.
i, I_,,.'-2*y'+
Solution Keeping y fixed, we have

f'-G, -
2xy, + ytl dx: [*r, - xzyz * yrxfr-,
J-z
:9-9y2.*3yr-(-g- 4yr-2yrl
:i*_ Syz + Sy3.
 711
sEcItoN 16.3 EvALUAnoN oF DouBtt lvircnms. lrrrrrrD lxrrcrus

Therefore

I: i:,
(x2 - 2xv2 + v!) ttx d, : [' (*- 5y' + Syt) dy

- r35 v-*v3 +*vnlt

-L3 =?. tr
Iterated integrals may be defined over regions F which have curved
boundaries. This situation is more complicated than the one just discusscd.
-\ Consider a region F such as that shown in Fig. 8, in which the boundary
*
cqnsists of the lines x: a, x: b and the graphs of the functions p(x) and q(x)
RGURE 8 with lx) < q(x) for a3x < b. We may define
fb fq(x\
I I f(x, y) dy dx,
Jo J put

(r,u,l@,u)) in which we first integrate (for fixed x) from the'lower curve to the upper
curve, i.e., along a typical line as shown in Fig. 8; then we integrate with
respect to x over all such typical segments from a to b.
When we perform the first integration in (3),
c(r)
f(x, v) dv,
p(r)
(r;q(a),0)
we are actually finding the area of a region R,, asis represented schematically
(a,v,o) in Fig. 9. We then "continuously add up" the region R, as x ranges from c to D
(a, p (a), O) to get the volume. The curve C, that forms the top boundary of R, is siurply
FIGURE 9 the function g(y) =f(x, y), with x fixed.
In Fig. 8 two of the borders of F are irregular; that is, p(x), q(x) arc
continuous functions of x. We may also allow the other two borders to be
irregular: that.is, the "sides" bf F may be continuous functions of y. For
example, iterated integrals may be defined over a region F such as the one
shown in Fig. 10. Integrating first with respect to y, we have
Pb fq(xl
I I I@, y) dy dx.
Jo Jpt*l

On the other hand, the integral taken first with respect to x requires that we
represent F as shown in Fig. 11. Then wc have
ft, fs(yl

J
Ic J,ot
I .fG, yl dx dy,

a=d,
r(y)

r: il p(r) r=6 u:c

FIGURE 10 RCURE 11

p@):a ExAA,lPtE 2 Given the function/(x, y): xy and the triangular region F bounded by the
lines y = 0, y:
2x, x:2 (Fig. 12), find the value of both iterated integrals.
CHAPTER 16 Murnnr [trEcnATtoN

Solution Referring to Fig. 12" we see that for

lb fq(xl
xv itv itx'
.1" J*
we have p(x) :0, q(x) :2x, a:0, b:2. Therefore
12 f2x 12
xydyax:
J, J" )oL*xf)l'ax
: |fz 2x3 dx: $x41fr: g.
Jo
. Intqgrating with respect to x first (Fig. 13), we have

ff' 1t ib with 4il *y,s{y) 2, c o, d 4.

ttx' : : : :
J"
Therefore

:
t f ,xy dx rty
tn
t**,rx,, a,

FIGURE 13 :f rrr-ty.) ,ty=Ur-#ynl6:s. D

lt is not accidental that the two integrals in Example 2 have the sarne
value. The next theorem describes the general situation. trt is a special case of
a result known x Fubini's Theorem.

THEOREM 7 Suppose that F is a closeil region gioen by

F: {(x, y): a<x<b, p(x) <y<S(x)},
where p and q are continuous anil flx) < q{x)for a < x < b. Suppose thatf (x, y)
is continuous on F, Then

The conesponiling result holits if the closed region F has the rcpresentation

F : {(x, y,): c < y 3 il, di < x < s{y)}

where fly\ < s(y) for c < y < il. In such a cose,

, f<.t

J*
In other words, both iteraed integtals, wlen computable, are equal to
double integral and therefore equal to erch other.

Partial Proof we shall discuss the first result, the second bcing similar. Suppose first that
f(x, -v) is positive. A plane -r : corst intersects the surface , :flr, y) in a cunr
(Flg 1a). The area under this qrrve il th€ x:const plane is s-hown as e
 713
sEcTroN 15.3 EvAtuATroN or Dou$"E rrractAr.s. lrrurro rNrrculs

e *.f {*, yl

arc& * .'1 i.,r.i

Volume by cross sections parallel to yz plane

FIGURE 14

shaded region. Denoting the area of this region by A(x),we have the formula
f s((rl
1(x): I f(*, y) dy.
J P(x)

It can be shown that ,,{(x) is continuous. Furthermore, it can also be shown

that if ,4(x) is integrated between x: a and x : b, the volume 7 under the
surface/(x, y) is swept out. We recall that a similar argument was used in
obtaining volumes of revolution by the disk method. (See page 228.) The
double integral yields the volume under the surface, and so we write

v: Ifflf7, v) dA.
J,J

On the other hand, we obtain the volume'by integrating A(x); that is,

n: lo Ag\ dx: Jo
[u f''''r, x, y) tty dx.
Ja Jp@)

If f(x,y) is not positive but is bounded from bei,-''.,:'\y the plane z:c,
.
then subtraction of the volume of the cylinder of heiglr, ^-rd cross-section F
leads to the same result. '1€;*
n
Remqrks. We have considered two ways of expressing u ..!ion I in the xy
plane. They are

f: {(x, y): aSx<b, p(x) < y<q(x)} (4)

and

f: {(x, y): c<y<d, (y) <x<s(y)}. (s)

It frequently happeos that a region F is expressibie more simply in one of the

above forms than iu the other. In doubtful cases, a sketch of F rnay show
which is simpler and, therefore, which of the iterated integrals is evaluated
more easily.
A region F may not be expressible in either the form (4) or the form (5).
714
CHAPTEN 15 fuluLTtPl-[ iturEcf,AlioN

tn such cases. F n:a1' sometimes be subdivided into a riumber of regions, each

having one of the two forms. The integrations are then performed for each
s,ubregion and the results added. Figure 15 gives exarnples of how the
subdivision process might take place.

Subdividing a region
FIGURE 1'

qAMPtI3 Evaluate llF *'y' dA where F is the figure bounded by the lines y : l, y:\
.r : 0. and y: y (Fig. 16).

The closed region F is the set

r<p* 16 F : {(.x, y): I < y <2,0< x <y}.
We use Theorem 7 and evaluate the iterated integral, to find
ff f2 fv 12
an:J, x2y2 dxdy:
J).'t
F
J. ),1*x't'f|at
:+ [' Y' tl
Jt 'lY:i'
Note that in the above example the iterated integral in the other order is
more difrcult, since the curves p(x), q(x) are
(t for0<x<l)
P(x):tr;; i 'r(x):2' osx<2'
=.=rI'
The evaluation would have to take place in two parts, so that

o^: .,n d1, dx + *n ,ty itx.

.T
II,,n I: [i LL

EXAMPTT 4 Evaluate

Carrying out the integration first with respect to y is possible but difficult and
leads to a complicated integral lor x. Therefore we shall try to express the
integral as an iterated integral in the opposite order and use Theorem 7. We
 715
sEcTtoN't6"3 Ev[UAnoN OF TX)Itrt"E NTEGIAIS, hBAID INTECI^IS

construct the region F as shown in Fig 17. The region is expressed by

F:{(x,y): 0<x32 and 0<y<}r'}.

However, it is also expressed by
F:{(r, y):0<y<2 and .,fr<x<2}.
Thcrefore, integrating with respect to x first, we have

I:f,, j#vdyd.
:fiffi*:[:L,ffiedv
JG;r;;1'** : t]at
I,llt*- (r +

: [* ln ty + JF + sl + lyJf +
- y -ly,]8 s

:* h 5 +3-4-|-1".,fr: - I +f ln 5. tr
The next exampk shows how the volume of a solid may be found by
iterated integration

I,ct S be the solid bounded by thc surface z: x!, the cylinders / : x2 and
y2 : x, and the glane z:0. Find the volume Iz(S).
It *under" the surfae
The solid S is shown in Fig. 18. consists of all points
z: x!, bounded bf the cylinders, and'above" thc xy plane. The region F in
Hctnt 18
the xy plane is bounded by the curv€s y: :
x2, y2 I and is shown ia Fig 19.
Therefore

r(s):fJ,, ,^:[:l: ryity,it*=['|d#r-

s:E
g2:x :Ll ft (x'-xt) dx:*. tr
Jo
If a solid S is bounded by two surfaces of the form z:f(x, yl and z:
dx, y') with/(x, yl3g(x, y), then the volume between the surfaces may be
found as a double integral and that intcgal in turn may be evaluated by iter-
ated integrals. The closed region F over which the integration is performed
is found by the projection onto the xy p.lane of the curve of intersection of
the two surfaces. To find this projection we set

f{x, y) =g(.x, y)
and trace this curve ia the xy plane. The n€xt example shows the method.

qAMPtI6 Fid the volume bounded by the surfaces

z: x2
 716
CHAPTER t6 Munru NTEcRATTcN

and

z:4-x'-y'-
A portion of the solid s (the part corresponding to y < 0) is shown in Fig. 20.
We set

x2 :4- x'- y'

and find that theclosed region F in the xy plane is the elliprical disk (Fig. 2l)

r: {tr, ,r,* *t'.,}

Note that the surface z = 4 - x, - y2 g(x, y) is' above the surfae
z-x2 =I(x, i4 for (x, y) inside the above=ellipse. Therefore

Y(s) : l[,0 - v' - r' - x27 d,a

JJ
g F

: f' f'Zn_ rn - - 2x2r dx dy. y2

-2 -&-ftJz
J J

:\ We now use the fact that (4 - yltrz is an even function of y. Hence

2 _i o I 2
(s) =
I _,*n- f),t,
dy :+ I: {4 - ft
t, ity

=# cosa o *: r+ (, * * cos2
') f, l:''
2 oos 2o 2oy do
\
HGURI 21
-2
=*{. [r** NT' .* l:''(r
+ cos 4e)do : 4n,,fi.
n

3 PROBI*E^,|S

In Problems I through 18, evaluate the iterated integrals as 3

eff I1+1Glf
indicated. Sketch the region F in the xy plane over whiOh the
integration is taken. Describe F using set notation. 'f,i-.f
a, i'r!-tr
v,tv,t' _xdxdy
J -3 J -Jtt-2f
12
, * y, + xy-3) dx ,ty r0l I xfdydx t,
f2x2
x c,os y dy dx
l: !:(x2 J -3 Jt, J. J,
, ,, o ,,
f ,(x. + 2x2y - y. + xyt dx
I:
tty
J',J*',l+ I:l: dF,tv dx

, dx dy . [ [, de-, dx
i, i.
ez'*3v
"l"Jl ,-ar4 tty

t +2xy-3yz) dx ,t m y-y ws xt dx
[i I*r" tty
i" J:.,{x
rty

t
f [ @2-xy)ttydx , !:ff,(*2l+xy2)dxdy ,r l:[" rdrdy
 717
sEcTtoN 16.3 EVAI,UAnON Or IX)UBI"E INITGI^I"II. ITEnATED INItGrAtS

17
ra'4 rsccx
I I ty*sinx)rtu rtx x ity itx :ffi'3Ti::*"f'rfllH"H:l'i;1".:;i'ffiiJ"til,"tn'l;
Jtl6 Ju'r f::f,"',
" J*rc Jo
cos

describe it using set notation.

In Problems 19 through 25, evaluate the double integials as 36 S is bounded by the surfaces z:Q z:x,arrd y2:2- x..
indicated. Sketch the region F. 39 S is bounded by the planes z =0, !:0, !: x, x * y:),
and x*y *z=3.
u + y'l itt;r': {(x, yl; y' <x < +, o < y <2} CI S is bounded by the surfaces a:5, z=0,y2 :4- x,and z
JJtr' :y*2.
,, :4, y:0,
:^ cos y dA; r: {(x, y): o<x <fifi,o<y<,'} 41 S is bounded by the surfaces x2 + zz and x + y
I * z:3,
42 S is bounded by the surfaces y2 : z, y = z!, z: x, atrd .y2
z' ant r: {(x, v): I Sx </:, o<r<'} :2-x.
fi*fi 43 S is bounded by the coordinate planes and thc surface xrl2
* y'l' + zrl2 - at12.
n r:{(x, y):2<x<f,1<y<x-l} 4{ S is bounded by the surfaces : 4 - y. y= x2 and z2
[!ntae; 45 S is bounded by the surfaces y2:x, x*y= 2, x* z:0,
andz=x*1.
,, r:{(*,v):ocx <*,x<v<l} : ! * z, ! = O,z - O and x
[[60^t 46 S is bounded by the surfaces xz
-1
,o ll-,-:-=rn, r={(x,y): lsx <z,t<y<x} 47 S is bounded by the surfaces y2 + zz : 2x and y : x -,
J,Jlx'*Y" *48 Lct f: {(x, y): 0<x < 1, 0<yS l}. We definc
(t il Y is rational'
zs
fi)r,,o aa; r: {(x,y): I (x <t6,, *, <y<zl I$, y):
io othcrwisc.
Show that JJ/(r, y) d.,{ does not exist.
In each of Problems 26 through 37, (a) sketch the domaio F
over which the integration is performed; (b) writc the equivaleat '49 I*t F:{(x,y):0(x< 1,0<yS l}. We define
iterated integral in the reverse order; (c) evaluate the integral x= Un,n:
[O rI 1,2, 3, ...,
obtained in (b). Describe F using set notation (in both orders). J(x'/):11
(, otherwise.

,'ffii,'* 27
f2 f+R
IJ-zJ-64
I xydydx Show that llf$, i dzt exists and has the value 1.

1' 1uE7 50 Irt f: {(x, y): a3x3b, c<y <d} and suppose/(x, y)
o @'- y'l't' dy d* = g(x) . (y). Shos, that
J. J,
[!to' i
itA:(j. *,,
") ff. Hn
dv)
"[ ], {r+*,
ax ay n [' !' $;trat a,
i5l Suppose that/(x, y) is continuous
jjxsin(y3)dydx ,
otr a close4 bounded
,t
J" J'J$*r, region ^F. Assume that

v) dA:o
" j, l, ffi*at "l-,1-.*o'r,
v)d(x,
Jp,',
I

E) **
for all functioos flx, y) which are coatiauous on F. Show
l,inv"o,
"II that/(x,
where /
y)
=Oon F.(Htnr. Assune there is a point in F
is positive, chow f carefirlly, and then usc
n>0 thercby 636hing a contradiction.)
Theorem 4 with
" f f ,u#-o'o' 52 Find the volume of the elliprcid
(
" I. jj + 7, 6*i^x2 yzlzrz r: itx.,-. 4, ?x2 * Fv2 * 7z2- t =oI.)
CHAPTEI 16 i,IULTpLE INTEGRANON

4
AnEA, DENS|TY, AND MASS

The double integral of a nonnegative function z:f(x, y) taken over a region

F may be interpreted as a volume.The value of such an integral is the volume
ofthe cylinder having generators parallel to the z axis and situated between
the surface z:f(x, y) and the region F in the xy plane.
If we select for the surface/the particularly sirnple function 1, then z:
the volume Iz is given by the formula

v:[fi ae
F

On the other hand, the volume of a right cyliuder of cross section F and
height I is
V: A(F).1.
(See Fig. 22) Therefore

A(F)
IIr"
l I

EGIII
l By Theorem 7, we conclude that the iterated integral of the function 1 also
.yields the area of
12 F.

flAMPLE 1 Use iterated integration to find the area of the region F given by
F : {(.t, y): - 1 Sx S 1, x4 S yS4- 3x2}.

Solution The region F is shown in Fig. 23. Oneof the iterated integrals for the area is
fl l'4-tx2
A(F):J_, ity itx,
and its evaluation gives
J,.
rt fl
A(F)- I tyl:.-t'dx=l (4-tx2-xa)
J-r - -
itx
J-r'
:[4x-x3-]x511-r:?.
Note that the iterateil.integral in the other direction is more difficult to
evaluate. tr
If a flat object'is made of an extremely thin uniform material, then the
mass of the object is just a multiple of the area of the plape region on which
r the object rests. (The multiple depends on the units used.) If a thin object
resting on the xy plane is made of a nonunifonn material, then the mass of
the object may be expressed ia terms of the density dx, y) of the material at
sccrloH'lf.{ ArrA DEtE[rY, Atd, MAss

any point. We assume that the material is uniform in the z direction. Ittting
F dcnote the region occupied by the object, we decompose F into rectangles
t1tt2, ...sr, in the usual way. Then an approximation to the mass of the ith
ri.t*gf".i.'given by
'P(8t,4)A(rt\,

where .4(r,) is the area of ri and ((,, 4i) is a point in r,. The total mass of F is
approximated bY

L
i= I
PG,'tt)A(rt)'

and when we proceed to the limit in the customary manner, the mass M(r) is

M{F):ll^' y) dA.
F

1yg srrmmarize this discussion with a definition.

DEflMNON

: :::]::::'

In other words' tbe double integral is a useful device for tJrdiag the mass of a
thin flat object with variable dsnsrtv.

EXAMPII 2 A thin obiect occupies the region

F: {(x,il: y2 S r<4- )3, -r/A<y<,,,n}.
The densitv is given by d'' l) : 1 + 2r * y' Frnd the total mass'

Solution We have

M(F') =1J,, . zx + yt dA.

F

sketching rhe regioo I (Fig. 24),we obtain for M(F) the iterated intogral

M(F) : (1 + 2x + yt dx rty
I:"f-.
IJ'
:J_;O+x2+ xy]i,-v, ity

f,E
= ) _nQo
* 4v - tov2 - 2v3) ttv
: l2oy + zyz - *y, - *y1!:0: *&. tr
7n
cHA?Ttl 16 MULTIPTE INTEGRATION

4 PROBTEMS

la each of Problems 1 through 8, use iterated integration to find 13 F:{(x, y):.1 SxS4,f<y<5-x}; p:4y.
the areaof the given region F. Subdivide F and do each part
separately whenever necessary. Sketch the region F. 14 F: {(r,y):y'<x<y+2, -l<y<2h p:x2y2.
15 F is the interior of the triangle with vertices at (O 0), (c, 0),
t F- 0<xS 1, x3 <y<J;\.
{(x, y): (b, c), :2x.
a>b>0, c>0; p
2 F = l$, i: iy' < x 1 y, I < y <41. 16 F:{(x,y): -a<x<c,0< y<JV=V\; p:3y.
3F is determined by the inequalities 17 F is the interior ofthe rcctangle with vertices at (0, 0), (a, 0),
(4, b), (O b); P = 3xl0 * x2Y?'1.
xy34, !3x,27y>4x2.
18 F:{(x,y): -1(x(1, 0<y<4}; p:d+Y.
4 F consists of all (x, y) which satisfy the inequalities
19 F ={(x, y): 0<x< 1, 0<y<x}; p:y+sin(ax2).
y2 <6-x, y<x-2.
3x, y2
20 F :{(x, y): 0<y Sl, !2<x<y}; p : Ji.
5 F is the region lying between the curves Ji + ,,fi: l r"d, 21 F :{(x, y): 0<y< 1, 0<x<y2}; p : yd.
x+ Y:4. 22 F is the interior of the triangle formed by the lines 2x: !,
6 f is the bounded region determined by the curves y: y x: 1, and the x-axis; p: e-x2t2.
and x:4y- y2.
23 Suppose the density p(x, y) of a region F satisfies the
7 F is the bounded region determined by the curves x + y inequalities m1 3 p{x, y) 3mz.Show that M(F), the mass
: 5, xy:6. of F, is between the limits mt A < M(F) < mrA, where A is
8 F consists of all (x, y) which satisfy the inequalities the area of F.
+24 Suppose that p(x,y), the dernity, is continuous on a region
x2+y2<9, ySx*3, x*y30.
F which has positive area. Show that if JJ dx, y\ dA = 0,
F
In each of Problems 9 through 22, find the mass of the given
then p(x, y)
=0 on F.
r25 Suppdse that the density p(x, y) of a region F in the xy
region F. Draw a sketch of F.
plane is of the form I x, y) : pt!)pr(y). Let M( F) be the
9 F:t(x, y):x2+yr<il\; p:x2+y2. total mais of P. Show that
l0 I: {(x, y): 0<xS 1, *' <y <.,/i}; p=3y. M(F) <+(M1$) + M2G))
11 F: {(r, y): -1 <x< 2, x' <y <x +2}; p :xzy. M, is the mass of F with density pl
where and M, is the
12 F :{(x, y): 0<x( l, *' < y <,/i}; p:2x. rnass of F with density pl.

5
EvnIunnoN oF DoUBLE INTEGRAI.S BY PoLAR CooRDINATES

The polar coordinates (r, 0) of a point in the plane are related to the
rectangular coordinates (x, y) of the same point by the equations
x:rcos0, y=rsin0, r>0. (t)
We recall that certain problems in finding areas by integration are solved
more easily in polar coordinates than in rectangular coordinates. (See page
526). The same situation prevails in problems involving double integration.
Instead of considering (1) as a means of representing a point in two
different coordinate systems, we interpret the equations as a mapping
between the xy plane and the r0 plane. We draw the r0 plane as shown in
Fig.25, treating r:0 and 0:0 as perpendicular straight lines. A rectangle G
 721
sEcTloN 15.5 EvAtuAnoN or Dou8tr tN[crALs By prol t coo*DtMrts

in the rg plane bounded by the lines r : t 1, t : tz, altd A : 0 t, 0 : e z (0 in

radians) with 2z > 02> 0, > O tz) tr > 0 has an image F in the xy plane

x \,,,\.\,

bounded by two circular arcs and two rays. For the area of F, denoted /,.r,
we have
: Xrtr - rl'1P, - e
AX" '/ \rnt ''
A,, r( F ) r'1.

This area may be written as an iterated integral. A calculation shows that

A..tr,): , a,)ae
I;'U,:'
Because double integrals and iterated integrals are equivatent for evaluation
purpeses, we can also write

A,.tF): ae,,,, (2)

[[,
G

where ilA,.s is an element of area in the r0 plane, r aad 0 being treated as

rectangular coordinates. That is, dA,,r:47 49.
More generally, it can be shown that if G is any region in the r0 plane and
F is its image under the transformation ( 1), then the area of F may be found
by formula (2). Thus, areas of regions may be determined by expressing the
double integral in polar coordinates as in (2) and then evaluating the double
HG(nr 25 integral by iterated integrals in the usual way.

EXA ,lPlf 1 A region F above the x axis is bounded on the left by the line !: - x,and on
the right by the cune

c: {(a y): x2 + y2:3trpJf * 3xl,

as shown in Fig 26. Find its area

Solution We employ polar coordinates to describe the region. The curve C is the
cardioid r: 3(1 - oos 0), and the line y: - x is the tay 0 :3nl4.The region
F in Fig 25 is tbc image under the mapping (1) of the set G in the (r, 0) plane
{Ft*27), ErYen by
G: {(r, 9}: 0Sr<3(1-cos 0), 0<0<hl4l.
ftcre$orc for t}c arca,{F) we obtain
-"o"",
r{F} : oo.,: dA,.o:f"' d, de
RGI,,II 26
Ii?G iJ" i,'.'
3r\ ltila 1f -13(l-corO) o fX4a
,1"1, o':;J"
-{) :
Jo
(1 - cos ol2 ito'

To pc&rm tbc integration we multiply out and find that

l3/a

"":rr!r-,::;;;.-#:
RGWI 27 :Nfu-qJ1- rl. D
1an
CHAPTER 16 MUITIPTE INTEGRATION

The transformation of regions from the xy plane to'the r0 plane is useful

because general double integrals as well as areas may be evaluated by means
of polar coordinates. The theoretical basis for the method is the Fundamental
Lemma on Integration which we stete without proof.

THEOREM 8 (Fundamental Lemma on Integration) Assume that { and g are continuous on

some region F. Then for each e >O there is a 6 >0 such that

v)s(x, v) ool.,
l,Lt,r,oro,l- ffrr,,
for euery subdioision F1,'F 2, ..., Fo of F with norm less than 6 an{ any numbers
f ,,.fr, ...,fn, gt, Ez, ..., gowhere eachf, and each g, is between the minimunt and
maximum ualues of f and g, respectiuely, on F i.

The fundamental lemma is the basis for the next theorem, the proof of
which we sketch.

THEOREM 9 Suppose F and G are regions related according to the mapping x:r cos 0,
I: r sin 0, and f(x, y\ is cantinuous on F. Then the function g(r,0) :/(r cos 0,
r sin 0) is defned and contitruous on G and
,i.],].:::]..l:.iI];:.:,::],,::i,,:::|,:j,,,,:::,]]:i,,,:i::

1..1,,:1,,,,f{*;.,+},.d6_iii,,=,li::;l::g{,ri.,:fl}4,r,ffii,,0ir,,:,:,,,,r,::,,;ir ,r,:i

:J.r::r:::.::::.-::rr:.:.-.::r..::::-..... ..!!:. :!!..:!!!u!.!r:!: .i:! i:r::rrr! 1:r:!:1:rr !ii:t::i:rt:jii:::i:rt!:i:rl::lrt::]:ji:::!tn!:::ii:

The form (3) is the one most often used in actual computations of integrals in
polar coordinates.

Sketch of proof Consider a subdivision of G into "figures" G1, ..., G,. (See the discussion of
volume in Section 1.) Let (r,, 0,) be any point in G,, and let ((,, ry,) and F, be
the respective images of (r,,0,) and G;. Then (Fr, ...,F) is a subdivision of F.
From the expression for area in the xy plane as an integral, we obtain

A,.,(F,): 0o,,,
[!,
Gi,

Using Theorem 4 concerning bounds for integrals, we find

tt
dA,.o: riA,.{G;),
) Jr
Gi
 t
sEcTloN r5.5 Ev ru noN or Douru lNItGtAut By pc,l t cooi]x]tArEi

where r-, is between the minimum and maximum of r on G,. Thus

,i/t€,, t)A,,,(Fil:,f r(r,, )'tt' A,,Ac).

o,

The theorem follows by letting the norms of the subdivisions tend to zero,
using the Fundamental kmma to evaluate the limit of the sum on the right.

In terr-ns of iterated integrals, the result in Theorem 9 yields thc useful

formula

v\ (tA:il*" o)r dr do,

IJr,.'
rG
which is equivalent to (3). Of course, the integral on the right is equal to the
iterated integralin the reverse order. In any specific case the order of
integration will usually depend on the determination of the limits of
integration. One way may be easier than the other. We give an example

EXllMPlf 2 Use polar coordinates to evaluate

ii
where F is the region inside the circle x2 *y2:2x.
The region F (see Fig. 28) is the image of the set G (Fig. 29) given by

c: {tr, q, -; <o <;,0 < r < z co, e}.

t
Therefore, using Formula (3) we obtain

ff fl fil2 f2cuo
+ v2 dA'.v: , 'r dA,.s: dr do
)J* .|o 12
)
' :l
JJ

l*,,
J_",,

lcos30itl
J -rt2
tlz
: ,f If (t - sin2 g) cos 0 d0:*. tr
Jo

RGLff 28 EGt$ z)
724
CHAPTER 16 Muttnr TNTEGRATIoN

Although the construction of the region G in the"rg plane is helpful in

understanding the transformation (l), ir is not necessary jor determining the
limits in the iterated itegrals. The limits of integration in polar coordinates
may be found by using rectangular ind polar coordinates in the same plane
and using a sketch of the region F to read off the limits for r and 0.
Double integrals are useful for finding volumes bounded by surfaces.
cylindrical coordinates (r,0, z) are a natural extension to three-space of polar
coordinates in the plane. The z direction is selected as in rectangular
coordinates, as shown in Fig. 30. If a closed surface in space is expressed in
cylindrical coordinates, we may find the volume enclosed by this surface by
FIGURE 30
evaluating a double integral in polar coordinates. Here we have z:1r, 0l
giving the "height" of the object above the point (r,0), as shown in Fig. 31.
The next example illustrates how we can use cylindrical coordinates to find
the volume of in object.

HAMPTE 3 A region Sis bounded by the surface s x2 + yz

-2.x:0,42: *, + y.,22: x2
+ y2. Use cylindrical coordinates to find the volume Iz(S).

Solution In cylindrical coordinates, the paraboloid 4z: x2 *y2 has equation 4z: 12;
z : f (r,0) the cylinder x2+y2-2x:0 has equation r:2 cos 0; and, the cone
22:x2 +y2 haS equation 22:r2. The region is shown in Fig. 32, and we
note that the projection of s on the xy plane is precisely the plane region F of
Example 2; (See Fig. 28.) We obtain

,(s):
t v:[lttre)r dr do
[!lt*-t#)un.,
FlcunE 31 F.
:ff(' -!*\' ae"

:
I::,,1"'"-'(,, - i,,)
dr d0

: o-*,. o) d0
I:',:,,(;"",,
:'+-;l;" ('*2 cos ze+!lff!) ae
32
:;__
'3n
D
9 8'
Polar coordinates are also useful for calcurating definite integrals of
certain functions of one variable that might otherwise seem impossible. one
of the most important examples is the integral
t2*s-,,t2 dx, which is of
RGURE 32 fundamental.importance to probability and staiistics. {See section 10 of
Chapter 9.)

EXAtr,IPI,f 4 Use polar coordinates to evaluate the integral e-,'tz dx.

12*
Solution We begin by observing that since the function/( x) : e-xz1z is symmetric with
respect to the y axis, we have
fq |a
I e_,rtz dx :2 I e -x2t2 dx (see Fig. 33.)
J-' Jo
sEcTtoN 16.5 EvxuArroil or Dot st [rrrBGRALs By poL x cooruNArE

That is, the arca under the entire graph of e-"2t2 is simply rwice the area
under the graph for x positive. Next we set f : II e-*'r' dx. Then since the
variable choscn for integration is immaterial, we-can write I - l@ e-t'lz dy.
Therefore
g= 2-*lz ,,:(l; "-*',,7*)([l ,r,, a) (4)

aod it can bc shown that (4) is the same as the double integral
l'o /.o
Thc graph of e-x2t2 is symmetric about
Iz: I I e-'rt, s-t2tz 4* 4,
Jo Jo
the y axis fo Po
FGTJIT 33
' : e-t'z+Ya'n dx dY
J, .|"
:ffr-"'.r2)t2 itAx.y
,f

wherc the region F is the entire first quadrant in the xy plane. Changng to
polar coordinates we s6e that e'e'+r')/2 becomes e'"!2 Lnd, dr{r,, becomes
r ilr d0. The region F transforms into the region
G= {(a 0): 0<r ( @, 0 <0 <nl2l.
Thcrcfore
,,= P itA,,,
Il,
G
ltil2 fa
: r dr do
.|, )o "-"''
l*12
: I 1r-r'2tz)ff il|
Jo

=l:'' ue:;
We take the square we conclude that
1: Jvl2. Therefore

5 PROITEMS

h "r+
*-g
ofProblems I through
polar coordinates.
12, evaluate the given integral by 4f)or.t )f'_,ffiaya*
F-/G
-J1x-xz
r./t3
rz 2 rz+.Fx|
,J,J; ,ffiaxay J
t
sf-2 J2-J4-x2
I _n@aydx
f2 l"q=
I I I _ e-t 2+v2t
.
i -2 J -J1-x,
',7 ptr-j
dy dx
I i; g2 + itx y2) rty

3 I-,it_ft -ta-yz sin(x2+y')dxdy ,

,,) !,f ,-+y2)-,t2 dxdy
GHAPTER 16

paraboloid

t
I j,-
e"+Y' itY dx 3z:xz+y2.
28 S is bounded by the surfaccs z:0' 22: x2 4 y" alnd 2y

=x2*y2.
S is bounded by the cylinder x2+y2:4 and
the

*t;r#,,* 29
hyperboloid
xr+y2_zr:1.
,, x sin(v3) dv dr 30Sisboundedbytheconez2>x2}y2andthecylindcr
fi ij r:'l + cos o.

31 S is bounded by the surfaccs z = x rnd

2z: x2 + yz'
32 S is bounded by the surfaccs z=0'
z:x2+!2'andr= -
2(1 + cos 0).
insidc the cylin-
*33 S is inside the sphe te x2 + y2 + z' = c2z and
: t - d @s 73'
IneachofProblems14through20,usepolarcoordinatesto O* oo Lo" loop oi the curve O
-ii the area of the region
given' "tJ.a inside the
34 S is insiide the sphere x2 +y2+22:4
find and
t . region inside the circle x2 + yz '8y:0 and outside
cylinder etected o"-on" looi of thc curvc 2:[' 12:
the circle x2 + Y2 =9' 4 cos 20.
15 The region F: {(x, y)' *y' 3x 32y'0 s y < 8}' 35 S is bounded by the surfaces z2 : x2 I y2'y:0'y: r' and
y2\3 : l6x2'
16 The region interior to the curve (x2 + X=d.
1? The region inside the circle r:4 cos 0 but outside the *36 S is bounded by the surfaces 22 = x2 * y2 and x - 2z + 2

circle r: 2. :$.
lE The region inside the circle r:3 coq 0 but outside the from a spherical ballof radius C by two
-' A wedge isi"hcut
3?
on a diameter of thc ball' If the angle
circle r:cos 0. pt"n"rit interseci
wedge'
r: I+ cos 0 but outside the L"t the plancs is t/3, find the volume of the
19 The region inside the cardioid """n is
circlb r: 1.
What is the volume if the angle {?
a circular disk ol radius c
20 The region inside the circle r: I but outside the parabola *38 A torus is generated by revolving
;;; ixis outside the disk' If the distaoce of the axis
r(l + cos 0) = l'
fromttrectnterofthediskisb,usepolarcoordinatesto
the volume of S'
In each of Problems 21 through 36' find find the volume of the torus'
z: O 2z: x2 * y" : for all r' 0' Let F b a
21 S is the set bounded by the surfaces 39 a) Suppose that 3(r, 0) c(1,- - 0)
and x2 *Yz =4' symmeiric;ith respcct to thc x axis' Let Fs
-#the rrri"r,
itgion iq

22Sisthesetboundedbytheconez2:x2+y2andthe portion oi F abo"e the x axis' Show that

cylinder x2 + Y2 :4'
23 S is the set cut from a spherc of radius
4 by acylinder of o\r dr ito: r.fl*" o)r dr do'

radius 2 whose axis is a diameter of the spnere'

$*',
and inside the
24 S is the set above the cone z2:x2*y2 b) If 3(r, 0)= -glr, -0), show that
sPhere
dr do:o'
xr+y2+22=a2.
il*"
F
otr

25Sisthesetbsundedbytheconez2:x2+y2andthe
cylinder each of Problcms 4O through 42, use
polar coordinates to
In
x2 + yz -2Y:o' evaluate the given integral.

26 S is the set bounded by the sphere xz +

y2 + z2 :4 and the 0 !ff e-+" dx
cylinder 4l ff J; e-r' ilx (Hint l*t x=u2')'

l'+ Y' .='2'' , [' {rn fl}''

\x/)
itx. (Hint: t*t x = e-')'
Jo [
2?Sisth€s€tboundedbythecone22:x2+y2aridthe
 727
SECTION 15.5 MoMrMr OF INERT'A AND CENTTR OF MASS (OmoNAT)

6
MOMENT OF INERTIA AND CENTER OF MASS (OPTIONAI)

Consider the idealized situation in which an object of mass m occupies a

single point. lrut L be a line which we designate as an axis.

DEFINITION Ihe moment of inertia of a particle of mass ln about the axis L is mrz, where r is
the perpendiculqr ilistance of the object from the axis (Fig. 34). If a system of
particles nt1.tn2; ..., mnis at perpeniliculgr distances, respectiuely, ofr1;12, 2..,
r, from the axis L, then ,he moment of inertia of the system , I , is giuen by

rcLff 34

Let F be an object made of thin material occupying a region in the xy plane

(Fig. 35). We wish to define the moment of inertia of F about an axis L. The
axis L may be any line in three-dimensional space. We proceed as in the
definition of integration. First we make a subdivision of the plane into
rectangles or squares. We designate the rectangles either whdlly or partly in F
by Fr, F r, ...,F,. Since the object F may be of irregular shape and of variable
density, the mass of the subregions may not be calculable exactly. We select a
point in each subregion F; and denote its coordinates ({i,4r).We assume that
the entire mass of F,, denoted m(F,), is concentrated at the point (et 4).
Letting r, be the perpendicular distance of (il 4) from the line t we form the
sum n

I
i= 1
n(F;)r! .

FIGURE 35

DEflNITION If the aboue sums tend to a limit (called I) as the norms af the subdioisions tend
to zero, anil d this limit is independent of the manner in whiph the (et 4;\ are
selected within the F,, then we say that I is the rmoment of inertia of the mass
distribution about the axis L.

The above definition of rnoment of inertia leads in a natural way to the

next theorem.
CHAPTEN 16 A rJrrl?tt lMBGn noN

THEORE'I{ 10

Similarly, the moments of inertia about the x

toqnoethtolv

(see
The proof depends on the Fundamental Lemma on Integration.
Problem 29 at the end of this section')

COnOfUnY The moments tf inertia of F about the lines

L, ={(x, Y,z): x- s, z=0\, Lr={(x, Y, 2): Y:b, z:0\'
L3 = {(x' Y' z\: x : a' Y
:bl
(ne, respectiuery,

rl: -alzp(x,yl itA,

Jft,
f

c:Jfo- bt'p(x,v) dA,

r
,:'' : a)? + $ - b)'7p6, i dA.
IIn, -

EXAMru 1 Find thd moment of inertia about the x axis of the homogeneous plate (i.e.,
p=constant) bounded by the line y-0 and the curve y:4-x2 {Fig. 36).

According to Theorem 10, we have

u= oo: o![r2 dv dx
[ln
Ff

:o dv itx:'rl'-,n- x2)t dx
I-,li-'" n
:t,l _r(64-48x2 + tzxa - x61 dx:T# . n
sEcTroN 16.6 MoMENToI INERflA AND CENITT oF MAss (omoNAL)

Ul,r,fnf 2 Find the moment of inertia about the z axis of the homogeneous triangular
plate bounded by the lines y : O y : x, and x:4.

Solution 1 We have

r,: + y')p dA: el:j" ,"' - y21dy dx

JJt,'
F

: el'
[,'r* ir'].
o*:?
[i* o.:'+. tr

Solution 2 We may introduce porar coordinates as shown in Fig. 37. Then

,, =
II,.,
+ yz)p dAn.y: ,llu .r dr d0: r
l,'. J;-"'
rt ib d0
FG

-*rl;; w4 o do:o+n[t*e+]tan"):.- :ry

f nl+
tr
The moment of inertia about the z axis of twodimensional objects in
the xy plaoe is called the polar inoment of inprtia. Since the combination
xz + y2: r2 is always present in calculating polar moments, a change to
polar coordinates is frequently advantageous.
A quantity known as the radius of gyration is intimately connected with
the notion of moment of inertia.It is defined in terms of the total mass m of an
object and the moment of inertia I about a specific axis.

EGIJNE 37

DEFINITION ?he radius of gyration of an obiect about an axis I is the namber R such that

where I is tlrc moment of inertia about L, anil m is the total mass of the object.

If we imagine the total mass of a body as concentrated at one point which

is at distance R from the axis .L, then the moment of inertia of this idealized
"point mass" will be the same as the momint of inertia of the original body.

E[AIuru 3 Find the radius of gyration for the problem in Example 1. Do the same for
Example 2.
cHA?TEn 16 MUmPl.ENTEGTAnoN

Solution The mass m of the homogeneous plate in Example f is its area multiplied by
p. We have

^: o f_,1"-" o, o.-- o I:' ,(4- x,) a'= ol+*- +]_, :+

Therefore

^:{r: (1;r *;,,,:,rf;

The mass of the triangular plate in Example 2 is 84 and so

ft: :ey.#)'':4# -
The center of mass of an object was defined on page 329 fr. To calculate
the center of mass we make use of the monent of amass n with respect to one
of the coordinate axes. We rccall that if particles of masses-fii1;rn21 ...,mnate
situated at the points (xr, yr), (xz, yr),..., (xn, y,) respectively, then the
algebrait moment (sometimes called fint moment or simply monrent) of this
system about the y axis is

I
n
mrx, lmrx2*... *mnx,= r= tftixi.
I
Its algebraic moment about the x axis is

i ^"''
t=t
More generally, the algebraic moments about the line x: a and about the
line y: b are, respectively,
ni
.)'rmi(xi-a) and ,_r1
,=l Z *,0,-t).
we now define the moment of a thin object occupying a region F in the
xy plane.

DEHMNON Assume that a thinmass occupies aregion F inthe xy plane. Fr, F2, ..o F, l*t
be a subdioision of F as shown in Fig. 38. choose a point ((t,4ll in-each F, anil
a replace the mass in F, by a particle of mass tfiFr) locateit at ((y q). The n
idealized r&urses haoe morunt about the line a of x:

;= I
i te, - a)d.F).
If the abwe wms tmd to a limit M the nomx of the subdiuisions tend to zqo
-i.-+_L
tL(€; and for ony choices of the poin*
1as
(il ) in F 6 then the timit M , is defineil as the
'rJ 4
moment of the mess distribution about the line x: a. An analogous definition
for the limit Mz of swis of the form

u l= I
i tr,-b)nlF,)
yiekls the first fintnent about the line y = 6.
HGT.RT:}B
sEcTtoN't6.6 MoMENT or INERTIA AND CENTER or MAss (omoNAt)

The definition of first moment and the Fundamental Lemma on

Integration yield the next theorem.

THEOREI,I11

coRolLARy Giaen a distribution of mass ouer a region F in the ,y plon i, iihr*tl,

then there are_ unique aalues of a and b Qtenoted * and
!, respectioely) siuch that
Mr: Mz:O. Infact, the ualues ofi
and y are giuenby

*tn = y) itA.
IIr.,,

Proof If we set Mr:0, we get

,: - a)p(x, y) dA: 11 a,e - ,ll y) dA.

JI,, l[-rr., 'F
F P ^.,
Since n(F) :[t p8, y) dA,we find for the value of a:
F

I Ixo$'
il ae
o:l-
^@-
The value y is found similarly.
tr

DCTNMON

EXAt\{rc n Find the center of mass of the region

F: {(x, y):0<x< 1, x3 <y<J;)

if the density of F is given by p = 3x.
 AATJtTFtf hIIEGTATK,II.

(Soe Fig. 39.) For the first mmests, *t have

U:J; *,: dA:31"

l,:
x2 dy dx
$.u
I

[ ,'[rIr,
E, = ,.l, (xsrz -x5) dx : g,

*,: !!ro itA:3[ J,:- dv d.4l: {,'fr r.

t
,,
= iJ- (*' - x7) dx:7.

rt4:,[ '[r]:
n{.F) =
II,
F
o^:
'fi [:' rtv
"
: , [' (xtrz -x4) dx: ].
Jo
Therefore

. - Mr 5 5 25. - M, -5 25
x= ni.fr:T4'1:T2; Y: ,4F):-5G'5: 18'

EXAI{ru 5 Find the center of mass of a plate in the form of a circular sector of radius a
and central angle 2a if its thickness is proportional to its distance from the
center of the circle from which the sector is taken.

Select the sector so that the vertex is at the origin and the x axis bisects the
region (Fig. aO). Then the density is given by p:
kr, where k is a proportion-
ality constant. By symmetry we have !:0.
Using polar coordinates, we
obtain

u,: l!.*r dA*;,: ol-" J" r'"o, a r itr ito

'r
:+ o ito :
)kan
sin a'
f-,cos

m{F)=-1,[' dA,,y:-i:,i; r2 dr ito:!kat u,.

 733
SECTION 16.6 MoMTNT oF INERTI,A AND CENTEN OF MASS (OPTIONAT)

Therefore

- Mt 3asina tr
m(F) 4a

6 PROBLEMS

i: each of ProblemsI through 7, find the moment of inertia and In each of Problems 18 through ?8, find the center of mass of
radius of gyration about the given axis of the plate F whose the plate F described.
=e
is given
=:sit1
tr F is the square with vertices (0, 0), (a, O), (a, a), (0, a),
18 I:{(x,y): I <x<4,1<y<5-r}, p:ky.
-: : constant; -y axis.
19 F: 1$, y): y' <x <y+ 2, -l' < ! 32\, p = l1y.
I F is the triangle with vertices (0, 0), (a, 0), (b, c), with a > 0, 20 F is the interior ofthe triangle with vertices at (O 0), (c,0),
r > 0, P: constant; x axis. (b, c), with 0<b<a,0<c, p:kx.
3 r= {(x,y):0<x 11,x2 <y<J;},p:constant;
-yaxis.
2t P: {(x, y): 0<x Sl, x2 < I < aG\, p : 1rr.
a F: {(x, y): 1 <x <4,* <y<5- x}, p:ky; x axis. 22 f : {(.x, }): - I < x 12, x2 <} < x + 2}, p: constant.
5 F:{(x,y): -a<x<a, 0<y<Jr'-r'l, p:ky: 23 F is the square with vertices at (0,0), (a,0), (a, a), (O a),
r aris. p:k(xz+y21.
a6 i=i{x,y): 1<x<4, *<y<5-r}, p:constant; 24 : kr2.
F is the triangle with vertices at (0, 0), ( 1, 0), ( 1, l), p
: l-\ls. 25 .F is bounded by the cardioidr =2(l * cos 0), p: constant.
- ;: i1.x, y): 0<xS tt,O<y <sin x), p:constant; y axis. 26 i is bounded by one loop of the curve r=2 cos 24,
p : constant.
.*r tarh of Problems 8 through 17, find the moment of inertia
aorc*-: i:e given axis of the plate F whose density is given. 27 F is bounded by 3x2 *4y2=48 and (x-2)'*y2:1,
| p:*Jir+F; z axis.
:; = l{r, y): x2 + y, 1o,}, P = constant.
0 a: ir..y): -l<x<2, x2 1y1x+2), p:constant; 28 F is bounded by one loop of the curve 12 : a2 cos 20,
:.f,lj. P = constant.

lilllL "; s the interior of the square with vertices (0, 0), (a, 0), 29 The Mean Value Theorem for double integrals states that l/
is integrable ooer a region F of area ,4( F) and if m <f(x, y\
.r. rt. 10, a), p:constant; z axis.
:= < M for all (x, y) on F, then there is a number/between ,tt
x. y): o < x < 1, x2 1y <-Ji\, p: *y, y axis. and M such that

r:_-.:s line y:4.

: : = .x. _u): x2 + y, < ar]l, p: k!Gr+i; x axis.
'"* I s :ounded by the closed curve r:2a cos 0, p:k;
: &-::-<

5 : :s i:uoded by one loop of 12 : a2 cos 20, p: constant; Use the Fundamental Lemma on Integration and the Mean
: iJ-j.. Value Theorem to establish Theorem 10. Use the idea of the
proof of Theorem
by one loop of /2 : a2 cos 20, p = constant;
9.
" ; ;..I*oed
* : *-e region in the first quadrant inside the circle r: i, 30 Show that if a mass distribution F lies between the lines
= x: aandx: b and has centerofmass (i, )), then u < j < b.
rmrr :ounded by r-1,0:r, and 0:nl2; p:constant; Similarly, if F lies between the lines .i.: c and y : d, then
: &;!JS.. c<t<d.
cllAPTEl 16 ITJLNru NIEGIANON

be regions no t$'o of which have any 33 Find the moment bf inertia abiut thc z axis of a ring of
- Let F1,F2,..., F, and
31
let (ir, f , ), (*r, ! z),',', ( i', f') k uniform density in the xy planc bbundcd by the circles
points in
"o.rnon,
centers of mass. flenote the mass of F;, by x2 + y2 : rl and x2 + y2 : 13 with t r I t z. Find the radius
iheir resixctive
mr.ll F is the region consisting of all the points in every F1' of the solid disk with the same densrty having thc samc
show that the center of mass (i, y) of F is given by
moment of inertia about the z axis.

mr*r*m2i2+ "'+raoin v Let F' be a region in the xy plane with mass n(F). Sbow
that in the notation ofthis section
mr+m2+ "'+rnn
11,:It-ZaMr+a2m(Fl
m]r* mzlz* "' * m,!n
m1*m2* "'*m, where M, is thc first moment of .F about the y axis. Also,
show that
32 Show that if F is symmetric with respect to the x axis and Il'bt - I t - 2aM 1 - bM z+ (42 + bzln(F)
p(x, -y) p(x, y) for all (x, y) on F, then y:0' A similar
:
result holds for symmetry with respcct to the y axis' where M2 is the first moment of F with rcspect to the x axis'

SURFACE AREA

To defineiurface area we employ a procedure similar to that used for defining

area in the plane. First, we define surface area in the simplest case, ald
second, we employ a limiting process for the definition of surface area of a
generai curved surface.
Suppose two planes l,
and 12 intersect at an angle d (Fiq.41)' From
each p;int of Gr, r region in the plane I.r, we drop a perpendicular to ths
panJfr. The set of points of intersection of these perpendiculars with f2
iorms aiegion which we denote G2. The set G2 is called the proiection of Gt
on Fr. Weihall now determine the relationship between the area.'{(Gr) of G,
RGrff 41 and ahe area A(Gr) of G2. If G, is a rectangle-the simplest possible case-
the problem may be solved by elementary geometry. For conveoience, select
the iectangle in f , so that one side is parallel to the line of intersection of the
two planes (Fig.42). kt
the lengths of the sides of the rectangle be aand b, as
shown. The projection of the rectangle in I., onto I,
is a rectangle, as the
reader may easily verify. The lengths of the sides of the rectangle in are a f,
and b cos 4. Th" uteu A r :
abof the rectangle in and the area I,
z{ z ab cos 0 :
of the rebtangle in f,
satisfy the relation
Az: At cos d' (1)
$ lrrs rF

RGINE{2 Equation (1) is the basis of the next useful result.

ltM,lA l*t G r be a region in a plarc l r anil let Gt be the proiection of G l onto a plane
1.. Then

A{G):,{Gt) cos @, (2t

where $ is the angle between tlw planes l1 anil 12.

 735
SECTtON 16.7 SURFACE AREA
J

This lemma is proved by subdividing the plane I 1 rnto a network of

rectangles and observing that these rectangles project onto rectangles in 12
with areas related by Equation (1). Since the areas ofGr and Grarc obtained
as limits of sums of the areas of rectangles, Formula (2) holds in the limit
(Fig. a3). We observe that if the planes are parallel, then @ is zero, the region
and its projection are congruent, and the areas are equal. If the planes are
:
perpendicuiar, then 0 xl2, the projection of G, degenerates into a line
segment, and A(G) is zero. Thus Formula (2) is valid for all angles such {
that 0< 0<n12.
Suppose we have a surface S represented by an equation
fIGURE 43 z:I$, Y)

for (x, y) on some region F in the xy plane. We shall consider only functions/
which have continuous first partial derivatives for all (x, y) on F.
To define the area of the surface S we begin by subdividing the xy plane
into a rectangular mesh. Suppose 7,, a rectangle of the subdivision, is
completely contained in F. Select a point {1i,4) in T,. This selection may be
made in any manner whatsoever. The point P,((,,q,, (,), with Ci:f(€i,qi),is
on the surface S. Construct the plane tangent to the surface S at P, (Fig. a).
Planes parallel to the z axis and through the edges of T, cut out a portion
(denoted S,) of the surface, and they cut out a quadrilateral, denoted Q,, from
I the tangent plane. The projection of Q, on the xy plane is T,. If the definition
of surface area is to satisfy our intuition, then the area of Si must be close to
EGI.nI 44 the area of p; whenever the subdivision in the xy plane is sufficiently fine.
However, Q, is a plane region, and its area can be found exactly. We recall
from Chapter 15 (page 649) that we can determine the equation of a plane
tangent to a surface z:f(x, y) ai a given point on the surface. Such a
determination is possible because the quantities

f*(li,4),fr(t,,4), -l
form a set of attitude numbers for the tangent plane at the point (4i, b *)
where (; :f(Eo q).
On page 549 we showed that the formula for the angle between two
planes is

.
d: larb, * arb, * arbrl
cos
'1G71 oj;t, IETE+E'
where ar, a2, ca and b1, bs, b, are sets of attitude numbers of the fwo planes.
We now find the cosine"of the angle between the plan€ Jangent to the surfac
and the xy plane. I*tting @ denote the angle between the tangent plane and
the xy plane and recalling that the xy plane has attitude numbers O 0, - 1, we
get

cos $: l0.r+0.r,+l.ll :0 +11+fj1-rrz

,frTIT+V,
According to the above lemma, we have
A(Ti): A(Q) cos $
736
CHAPTER-I5 MUTT]PLE INTTGRATIoN

or

A(e): A(r)JTlfTC, N i7Tt" t1).

We add all expressions of the above type for rectangler I which are in F. We
obtain the sum

(3)

and we exp€ct that this sum is a good approximation to the (as yet undefined)
surface area if the norm of the rectangular subdivision in the _ry plane is
sufficiently small.

THEOREM 12 Suppose z:f(x, y) is defined on a region F in the (x, y) plane. If the first
deriuatiues f* and f, are continuous on F, then the sudace area A( S) of z:
f(x, y) exists and it is giuen by:

Theorem 12 is equivalcnt to stating that the sums in (3) tend to

rl
l.@ae
F

whenf, and{, are continuous. It is an immediate consequence of Theorem I

and the fact that ,frTTTTF, is continuous iff, andl, are. The integration
formula of the theorem may be used to calculate surface area, as the next
examples show.

Find the area of the surface z:!y3t2 +yttr) situated above the square
F: {(x, y):0Sx3 1,0<y< 1} (Fig. a5).
Solution Setting z = f(x, y), we have J, : xrt2 , f, - yt , and
12

1(s):|i,,. x*y)rrz41
F

: i' ['(1 *x* v)tt2 dv dx'

Jo Jo'
 7t7
sEcTtoN 15.7 Suxr cEAntA

Therefore
ft
/(s): I 3t(l + x + y)3/21[ ,r:
Jo
*}" (2+x1trz -(l + x)3t2f d.x

=*t(2 * x)stz- (l + x)t/'13 : *(1 + 9fi - srfrl. D

EXA tru 2 Find the area of the part of the cylinde,, :**' cut out by the planes y : 6,
Y:x, x:2'
Sotution See Figs.46 and 47, which show the surface S and the projection F. We have
Solution
lzf Ax: x, 0z/0Y = 0. Therefore

/(s)=
llF, oo=[:[',r;t or*
fz
: ,J, +x2 dx:*(l + x')'t'7lo
Jo
:3isr6- ry. D
The next example shows that it is sometimes useful to use polar
coordinates for the evaluation of the double integral.

EXAI{PT"E 3 Find the surface area ofthe part of the sphere x2 + y2 * 22 a2 cat out by :
the vertical cylinder erected on one loop of the curve whose equation in polar
coordinates is r=c eos 20.

Solution (See Fig. 48.) The surface consists of two parts, one above and one below the
xy plane, symmetrically placed. The area of the upper half will be found. We
have

2:16,:Tf 0z -.x oz -y
HGT.RE 47
,
0x Jo, -
-::- x, - yr' -::.
oy Jo'-x2-y2
Therefore (Fig. 48), we obtain

: l*'n lacos2e -Lar.ae.

J-,r+Jo Jo'-r'
This integral is an improper integral, but it can be shown io be convergent.
Taking this fact for granted, we get

/(s): ,ol*'o l- JV:Vl.co'2l do

Jo
rc]rf {8
:2a,
'f*11 (l _sin 20) d0:la21n_21.
| Jo
The total surface area is a21r*2'i. i,-.
L
7fi
CHAPTER 15 MLf,nrulNlfcr noN

If the given surface is of the form y :!(x, z\ or i:f(y, z), we get similar
formulas for the surface area. The three basic formulas are

7 PROBLEMS

sphere of radius a, center at the

In each of Problems 1 through 20, find the area of the surfacc f3 The part of the surface of a
origin, inside the cylinder erected on one loop of the curve
described.
1 The portion of the surface z:l{vttz 4y3/2) situated above
r: a cos 30.

the triangle t4 The part ofthe sphere x2 + y' * z2 = 4zinside the parabol-
oid x2 + y2:2.
f = {(x, y): 03x < y, 0 SyS l}.
15 ThE part of the cylinder yz + z2:22 cut off by the cone
2 The portion of the plane x/a* ylb+zlc= I in the ftrst v2-y2*22.
octant(a>0,b>0,c>0). l5 The part of the cylinder x2 + y2:2ax inside the sphere
3 The part of the cylinder x2+22 =42 inside the cylinder x2+y2*22:4a2.
x2 + Y2 :42.
17 Thc part of the cylinder y2 + 22 - 4a2 above the xy plane
4 The part of the cylinder x2+22 =42 above the square and bounded by the planes y:0, x : o, and, y: x'
lxl<14, lyl<*a.
It The part ofthe paraboloid y2+22:4cx cut offby the
5 The part of the colne zz = x2 * y2 inside the cYlinder cylinder y2 : ax and the plane x = 3a; outside the cylinder
x2 + Y2 =2a.
!2: ax'
6 The part of the cone z2 = xz + y2 above the figure bounded
19 The part of the parabolic cylinder z : y2 which lies overthe
by one loop of the curve 12:4 cos 20. (l' 1)'
triangle in the iy plane having vertices (0,0), (1,0)'
7 The part of the cone ,',: y' + z2 between the cylinder 20 The part of the surface 9(z - y)' = 4x3.which lies over the
y2 :i and the plane y : 2 - 2.
triangle in the xy pf"l* fi"riog vertices (0,0), (0, 2),(?-'21,
8 The part of the cone !2:x2+22 cut off by the plane
2l a) Use elanentary geometry (and trigonornctry) to es-
2y:\x +21,fr. tablish equation (2) for an arbitrary triangle.
9 The part of the cone x' : y? + z2 inside the sphere x2 + y'2
b) Use the result of (a) to establish equation (2) for an
+ z2 :22. arbitrary polygon.
10 The part of the surface z = xy inside the cylinder x2 + y2
t2 lf r : ,m7,0: &rctan (y/x), and z =f(x, y), establish
:az the formula for surfacc area in polar coordinates:
11 The part of the surface 4z:x2 -y2 above the region
bounded by the curve 12 =4 cos 0.
12 The part of the surface of a sphere of radius 2o inside a
cyliri-der 5f radius o if the center of the sphere is on the
surface ol the eylinder.
 SECTtoN 15.7 Sunmcr mrn

tr-re the polar coordinates formula (Problem 22) to find the u The portion of the cone x2+y2 =22 inside the cylinder
mrrface area in Problems 23 through 25. (x2 + y2'12 :22rr.
23 Th€ area ofthe surface ofthe paraboloid s: x2 + y2 which 25 The portion of the cone x2 + y': z2 inside the cylinder
b hside the cylinder x2 + y2 :4. xi + yz :1.

8.,
VOTUMES OF SOLIDS OF REVOIUTION (OPTIONAI.)

we previously developed methods for finding the volume of certain solids of

revolution. These techniques were applicable whenever the resulting process
reduced to a single integration. (See page 228 fr.) Now that areas may be
determined in a more general way by double integration, we can calculate the
volume of a greater variety of solids of revolution. The basic tool is the
Theorem of Pappus which we now state. (See also page 345.)

THEOREM 13 (Theorem of Pappus) If a plane figure F lies on one side of a line L in its prane ,
tlu oolume ofthe set s generatedby reuolaing F around Lis equalto the product
a
o.f A(F), the area of F, arul the length of the path desqibed by the eenter olmass
of F; in other words, if F is in the xy plane atd L is the x axis (see Fig. 49), then

(1)

rct,.lf at

The second equality in (l) above follows from the definition of y. In case F is a
rectangle of the form a < x < b, c s y < d where c-> Q then s is just a circular
ring of altitude b a, inner radius of base c (if c > 0) and outer radius d. Thus
-
V(S) : v14z - e\* - a) - *(+) ^(
F) :2neA(n :
IIr*
oo.

The theorem is proved in general by subdividing F and noting that

I Y(sr): ,\llr", ae: !!zw

n
ae,
i= 1
Fi&
.where Ff is the union of the F,; we then pass to the limit.

o(AIttPl"E 1 The region F: {(x,y):0<x( l, r3 .y<Ji} is revolved about the x axis

Find the volurne generated.
74tt
CHAPTTR 16 MUITTPLE INTEGMTIoN

Solution See Fig. 50. We have

I,(s):
^llt dA:2nf f
y dy ttx
F

:n |rr (x-x6) dx:*n.

Jo
The Theorem of Pappus in this more general form is especially useful
whenever the transformation to polar coordinates is appropriate. The next
example illustrates this Point.

EXAMPLE 2 Find the volume of the set S generated by revolving around the .x axis
the upper half F 6f the area bounded by the cardioid r:a(l*cos 0)'
r=ru(1*cos0) (See Fig. 51.)
-
Solution We have, since -v transforms to r sin 0 in polar coordinates,
T
v(tt: ae: nr sin o dA*,,
IJrnr
FF
[lz
FIGUN,E 5I
:2"r r2 srn o dr do
I;".*''
f*
:!na3 ,t * cos g)3 sin 0 d0
.|.

=Sr-(1+ cos ,ll-x:$.

The formula analogous to ( i) for revolving a region F in the xy plane
about the y axis is

8 PROBLEMS

In each of Problems I
through 18, find the volume of the set The upper halfofthe figure bounded by the right-hand loop
obtained by revolving the region described about the axis of the curve r2 = d2 cos 0; the x axis.
indicated. Sketch the region. Express in set notation each of the
The upper half of the right-hand loop of the curve r =
plane regions being revolved.
a cos2 6; the x axis.
1 The upper half of the ell\pse x2fa2 + y2lb': l; the x axis.
The region inside x2 * :64 and outside x2 + y2 :g*'
The region in.the first quadrant inside 3x2 * 4y2 :48 and
y2
2 the line x:8.
outside (, - 2)' * yz : l; the x axis.
3 The region satisfying the inequalities xy 3: 4, y S x, The region bounded by the curve y : x
3
and the lines y = 0,
27 -t > 4x1; the .t axis. x: 1; the x axis.
sEcTroN 16.8 VotuMts ot souDs or nEvoLUTtoN (oprloNAr)

E The region bounded by the curve ! : d andthe lines y : Q lE The upper half of the right-hand ioop tif the curve r?:
x: l; the x axis. a2 cos 20; the x axis.
9 The region bounded by the curve /: ln x and the lines
j/:0,x:e;theraxis. l9 A triangle T with base of length (r on ths.r axis and verte.x at
a point (b, c) in the first quadrant of the -xy plane is revolveri
l0 Same region as Problem 9; the y axis. about the x axis. tlse the formulas for the area of a triangle
ll The region in the first quadrant between the parabolas and the voiume of a right circular cone to find the cenrcr oi
x2 :4y, x2:8y - 4; the x axis. mass ol l.
12 Same region as Problem 1l; they axis. i, ith .liameter d on the x axis is revolved about
20 A semicircie
13 The loop of 13: Sin 20 in the first quadrant; the y axis. the x axis. Use the formuias fcrr ttre area of a semicircle ;rnt!
14 The upper half of the area outside the circle r = 4 and inside the volume of a sphere to find the center of mass of ttre'
the limagon r = 312 cos 0; the x axis. semicircle.
15 The upper half of the area to the right of ttre line I : ! and
inside the cardioid r :2(l + cos 0); the x axis.
2l a) A triangle in the first quadrant of the .xy plane witlr base
parallel to the x axis is revolved about the .x axis. F'incl the
16 The upper half of the area to the right of the parabola volume generated. (Hint: See Problem 19.)
r -9/(l * cos 0) and inside the cardioid r = 4{l + cos 0); b) A trapezoid in the first quadrant of the .t1, planc wittr
the x axis. bases parallel to the x axis is revolved about the x axis. Use
17 The right-hand horizontal loop of the curve r =2 cos 20; the results in (a) to find the volume generated and the
the y axis. Find the center of mass of that loop. center of mass oi the trapezoid.

9
Txr rRpu INTEGRAI

The definition of the triple integral parallels that of the double integral. In the
simplest case, we consider a rectangular box R bounded by the sii planes
x:ao, x:or, !:bo, !:br, z:co, z-cL (Fig.52).Letf(x, y, z) be a
function of three variables defined for ( x, y, z) in R. We subdivide the entire
three-dimensional space into rectangular boxes by constructing planes

ffi''
L- ---
--t- -

RGUNE 52
 MUTTIru INTECRATION
CHAPTER 16

parallel to the coordinate planes' FtIr' 82,.", B,'be those boxes of the
of the ith
subdivision which contain toints of R. Denote by Iz(Br) the volume
(,) may be made
L"-, f,. We now select a point P,16,, fi, in B;; this selection
in any manner whatsoever' The sum

L fG,,4t, (tlY(Bt)
i= I
is the
is an approximation to the triple integral. The norm of the subdivision
i"ngttt ri the longest diagonal of the boxes B.r, 8", "', B,'If the above sums

teni to a limit tt e nir-s of the subdivisions tend to zero and for any
"s triple integral of/over R. The
choices of the points P;; we call this limit the
expression

v, z\ ,tv
IIF,"R

is used to rePresent this limit'

triple integral has the same value as a threefold-iterated integral. In

the Qase

of the rectangular box R, we obtain

y,z) itv:.|:jtf t,a a,]atla*

ilf,' [J"]r,,,
Suppose a region S is bounded by the planes *=oo, !ar'!:bo'
y : ar, by thJ surfaees z : r(x, y), z: s(x, y), as shown in Fis' 53' The
iripieint.g*t'may be defined in the sarne way as for a rectangular box R, and
"nO
once again it is equal to the iterated integral' We have

y,zt dv:f {[[1",^' r,,t a,)atla*'

ilp,'
we state without proof the following theorem, which applies in the
general case.

HGUNE 53
 743
sEcTtoN 15.9 Tnr rnpti TNTEGRAI

THEORTM 14. Suppose that S is a region ilefined by the inequalities

S: {(x, y, z): a<x<b,p(x) < y Sq(x),r(x, y)<zSs(x, y)},

where the functions p, q, r, and s are continuous. Iff is a continuousfunction on
S, then

v, z) dv:i " n*,,,4 a')a\a'

iilr,,, {i:" tff
The iterated integrations are performed in turn by hording all variables
constant except the one being integrated. Brackets and braces in multiple
integrals will be omitted unless there is danger of confusion.

D(AMPII1 Evaluate the iterated integral

lr 16- 2z fa- ettly- lE,tjlz
Yz dx dY dz'
J" J. J.
Solution We have
,"
J;f In-,,,",-,n,,,.
y, d* dy d,

:f f "['r'],-'""'-'o'""0, d,
fl f6-22
:J, yz(4-ly-!zt dy dz
J,
: Ir3 l2zy2 -z*rt -ly2zz1o-2, 4,
Jo

: Ir3 lez - ?r\$ - 2zl2 - *z(6 - 2z)3) itz

Jo

:+ Ir3 z(6_22)t dz.

Jo
The integration may be performed by the substitution u: 6 - 22. The result
is 5415. tr
The determination of the limits of integration is th-e principal difficulty in
reducing a triple integral to an iterated integral. The ieadei who works a
large number of problems will develop good powers of visualization bf three-
dimensional figures. There is no simple mechinical technique for detdrmining
the limits of integration in the wide variety of problems we encoiinter. ThI
next examples illustrate the process. .

E(AMru 2 Evaluate

lli. *,
s
CHAPTET 16 AttJtmu m.{rt{n noN

where S is the region boundcd by the surfaes != x2,'y : x * 2,42 : xz + yz,

.-andz=x*3.

Solution To transform the triplc integral into an iterated integra[ we must.determine

the limits of integration. The region S is sketched in Fig. 54.

Ectnt lia

The projcction ofS on the xy plane is the region F bounded by the curvcc
!: x2 and-y = x + X as shown in Fig 55. From thib projection, the region
rises with vertical walls, bounded from bclow by the paraboloid z=,
ilx' + y2) and above by the plane z: x * 3. Sinco f is aacriUea by the
inequalities
F: {(x, y): -l 3 x 32, x2 S y tx * 2\;,

we have
S = {(-x, y, zri *l < x S 2, xz s y < x + 2, X*, + y2) < z< x + 3}.
Therefore
f f f f2 l'r+2 /'r+3
o': -,J,, x itz dv itx
JJ J ' J Jo,*nrn
s

=
I'_rf*.'
[x2 + 3x -l(r, + xy?lf (ty itx
=
f , {(r..,, - i,,) (2 + x -,,) - ;te+ x)3 -,r} *
:- 837
160'
 745
sEcrroN 16.9 fxr rnrru INTEGIAL

In the case of double integrals there are two possible orders of

integration, one ofthem often being easier to calculate than the other. In the
case of triple integrals there are six possible orders of integration. It becomes
a matter of practice and trial and error to find which order is the most
convenient.
The limits of integration may sometimes be found by projecting the
region on one ofthe coordinate planes and then finding the equations ofthe
'bottom" and "top" surfaces. This method was used in Example 2. If part of
the bounilary is a cylinder perpendicular to one of the coorilinate planes, that
fact can be used to help determine the limits of integtation.

EXAMPLE 3 Express the integral

,: rrt t, z) av
JJ
s
)f$,
as an iterated integral in six different ways if S is the region bounded by the
surfaces

z:0, z=x, and !2:4-2x.

The region S is shown in Fig. 56. The projection of S on the xy plane is the
two-dimensional region F,, bounded by x = 0 and y2 :4 - 2x, as shown in
Fig. 57. Therefore the integral * written
*11r
l'2 lr+-/4-2x l'x
r: l I I l(*, y,2) dz dy dx
Jo J_JT:T, Jo
f2 f2-(rl2ly2 fx
: z) dz dx dv'
.|_, J, )ot$,v,
The projection of S on the xz planeis the triangular region bounded by the
lines z :0,2: x, and x:2,as shown in Fig. 58. The iterated integral in this
case'oecomes
f 2 P, t'*-/i-'2.
: Y' z) itY dz dx
' J' J' J
: "=!(x'
P2 rz c+-fr=
Y' z) dY rtx itz'
J' J' )-'=''I(x'

HGUff 58
CHAPTEI 16 MULTIPLE INTEGRAfloN

The projection ofS on the yz plane is the plane region bounded by z:0 and
z:2 - iy' (Fig. 59). Then r takes the form
fZ f2'(ll2ly2 f2-(rl2'ty2
I:l I I f$,y,2)dxitzdy
J-rJo J,
: llz If+Jt-zz
_
f2-lrl2lYz
l ft*, y,2) dx dy dz.
Jo J -Jqa, J,

FrcUNE 59

9 PROBLEMS

In each of Problems I through I 1, find the value of the iterated In Problems 12 through 22, evaluate
integral. Express each region of integration in set notation. Itf
ft fr fr-u ''dzdvdx v, z) dv
JJJn.,
'J"J,J, "tt.
-" where S is bounded Uy giu"n surfaces and / is the given
, ,t"t.
i-, i" f dz dx tty function. In each case, express S in set notation.
ll z :0,y:0, y: x, x + y :2, x * ! * z - 3:f(x, y, z) = x
, xvdxdzdv 13 x : g, t = JV - y' - r.;f(x, y, z) : a
[],:f l{ z :0, xz + z : l, y2 + z = 1;f(x, !, z) = z2
pr (re-,,- ,.l4i:E-j
nJ"J. J, (x+y+z)dzdydx 15 x2 l- 22 : a2, y2 + z2 : a2:f(x, y, z): x2 + y2
f( x=0, t=0, z=O, (xla)+(ylb\+(zlc):1, (a, b, c>0);
flx,
y, zl = z

'ff fi',dxdyd, 17 y = 22, y2 = z,

/:0, z:0, xrlz
x= 0, x =
*
y.- z2;f{x,
y'l'+
y, z\ =' y * z2
rrtz: otlz'f(x, !, z):
.i.f !,ff*0,0. lE
19
x = 0,
x:0, y=0, z:0, !2=4-2, x:y+2;f(x, y, z\=x2;
z

y>0
fl f l-t f l-r-r
3dzdvdx lg z = i2 I y2, z : 27 - 2i2 - 2y2: flx, L z\ : I
'J"J" J, 2l z2 : 4ax, x2 + y2 : 2ax; f(x, y, z) = |
t l, [ '1"'-.-'xyttzttydx *22 y2 + z2 : 4ax, y2 : ax, x = 3a;f(x, y, z) : x2,. y2 < ax
In Problems 23 through 30, express each iterated integral
,l,l:
as a
triple integral by describing the set S over which the integration
lo,a,ara' is performed. Sketch the set S and then express the iterated
integral in two orders differing from the original. Do not
,, o' o, o.(',iar; Use svmmetrv to reduce evaluate the integrals.
I-, tZ[) ' fr fx l'r-e
.dzdvdz
this integral to the one of Problem 9.)
f2 ljT? f4-yz
" J,.l"J"
l1 fr-v'1,fr
rr I I azaydx
J -2 J -0Aj1-x2
-l )x2+3yz 'J-,J, ) n'f6dzttxdv
 747
sEcTtoN 16.9 THE TRTPLE INTIGRAI

fr fJy fy+z form/(r, y, z): g\x)h(y[(z), then

" j. J,, J,
xY dx dz dY
Q) dx ttv dz
JJjrt,rrtrD
,u (y, +,,) dz dx dy
f_,1;-'i;."
,',
f2 f4-yz lt+2 =[i",,,"][ ur dyfll' iat dz)
J _), J, xyz dz dx tty
In Problems 34 through 37, use the result of Problem 33 to
evaluate the integrals given.
,n, ttz rty dx
" iji" tl'-" 34filxz'z sinydx rlydz:S:{x,y, z:0<x<3, 1<y<5,
f-l Pl-x Pxtt .,'-vtz -3 < z <0).'
" j. J" J, dz dv dx
35 ffJ(x'?lz * yzl dx dy dz; S = lx, y, z:0<x54,0 3y36,
pr p$ pv-" 0<z<i0).
,o f(x, v, z) dz dv dx
J. J,, J" 36JJJez'+rr-a'ilxilyilz; S={x, y, z: -4<x34,
3l Express the integral of f(x, y,z) over the region S bounded -3 (y <3, -2<z<2\.
by the surface ,:rfri:;'--y'and the plane z=2in 37 jJJln(x?) dxdydz; S:{x, y, z: l<x<2,,1<y<2,
6 ways. l<z<2|.
32 Express the integral 38 tet S be the solid tetrahedron with vertices at (Q 0, 0),
(1, 0, 0); (0, 1, 0), and (0, 0, 1). Find the value of
(x'+ v'+ 22) dv dx dz jjj/t*, y, z) tlv where/(x, y, z):(x2 + 2xz + y2).
fi f i;
in 5 additional ways. 39 tet s be the solid pyramid with vertices (0,0,0), (1,0,0),
33 Suppose S is the rectangular box a< x3b, c<y<d, (lt I' O)' (0, 1, 0), and (0, 0, l). Find the value of
e3z3f. Show that when thc integrand has the special J[l/tr, v,zl dV where/(x, v,z):xyz*2y2.

10
MASS OF A SOTID. TRIPTE INTEGRALS IN CYLINDRICAL AND
SPHERICAL COORDINATES

From the definition of triple integral we see that if/(x, l, z) 1, then the triple
=
integral taken over a region S is precisely the volume Iz( S). In general, if a solid
object occupies a region S, and if the density at any point is given by 6{x, y, z),
then the total mass, m(S), is given by the triple integral

rT(s): v, z) itv'
iff^t
s

Notation. For the remainder of this chapter the symbol 6 will be used for
density. The quantity p, which we previously used for density, will denote one
of the variables in spheripal coordinates.
78
CHAPTTT.16 MUITIPTE INTTGRAT]ON

The solid S in the first octant is bounded by the surhces

. z:4-x2-y2, z:0, x*y:2, x:0, ,'l:0.

The density is given by 6( x, !, z) :22. Find the total mass.

We have (Fig. 6C)

t pt f2 f2-x f4-x2-y2
n{s)=JJJr,on:r),J" zdzdvdx
s
.|"
f2 f2-x
=lJo JoI t+-xz-y212dydx

: (16 + xt + yn - Bx2 - 8y2 + zx2y2) dy dx

I i;-"
FA

= I l$-
Jo
xr)ry -t(4-x2)y3 ++y5ll-, d*
t2
: I l2(4 - x2)2 - x(4 * x2)2 *N4- x\{2-x)3 +}(2-x)51 dx.
Jo
The above integral, a polynomial in x, can be evaluated. Answer: 7C/,145.
n
We found that certain double integrals are easy to evaluate if a polar
coordinate system is used. Similarly, thereare triple integrals which, although
difficult to evaluate in rectangular coordinates, are simple integrations when
transformed into other syster.ls. The most useful transformations are those to
cylindrical and spherical coordinates. (See Chapter 13, Section 6.)
Cylindrical coordinates consist of polar coordinates in the plane and a z
coordinate as in a rectangular system. The transformation from rectangular
to cylindrical coordinates is

,,'.,. .*,,r*..,'..i#...0i......,..,.......*..,U.'*,.+tul..*ll,...l'....ti;i,,ir".i, trl

A region S in (x, y, z) spacc corresponds to a region U in (r, 0, z) space.
The volume of S, I( S), may be found in.terms of a triple integral in (r, 0, z\
space by the formula (Fig. 6l)

/(s):
,fi,
**
a

This formula is a natural extension of thc formula relating area in rectangular

and in polar coordinates. (See page 526 ff.)
 sEcTroN 16.10 IvtAss or A soltD. TttPu TNTEGIALS lN cYuNDolcAt ANr] sPtlBlcAL coonuNAT[s

I
Az
1
1

FtGt f,E 51

A triple integral in cylindrical coordinates may be evaluated by iterated

integrals. We obtain

and, as before, there are five other orders of integration possible. Once again
the major problem is the determination of the limits of integration. For this
purpose it is helpful to superimpose cylindrical coordinates on a rectangular
system, sketch the surface, and read off the limits of integration. The next
example shows the method.

EXAMru 2 Find the mass of the solid bounded by the cylinder xz + y2: 4x ood the cone
22 : x2 + y2 if the density i, a: k$Tf (Fig. 62).

Solution We change to cylindrical coordinates. The cylinder is r:

c cos 0 and the cone
b z2: r2. The density is kr. The region S corresponds to the region U given
by

u: o, z}: o sr 3scos g, - i<o <1, -r=, = r}

I
{tr,
I Therefore
I

m(s):
! [SU[
raw dv,v, :-
iil'' r' dYn,

ftlz facos0 fr

l'*12 facmO t P*12

: rt dr ito: |kon t o'
l\
L-------J
// 'o ) -,,,J, J ---,,"'"n

mf 6t
:f,r,on
l-_::"(r*
2 cos ,, *uff!) or:r# !
CHAPTER 16 MUtnru NTEGNATON

The transformation from rectangular to spherical coordinates is given by

the equations
x:pcos0sin{, y=psin0sin{, s:pcos$. (2)

The region U given by

corresponds to a rectangular box in (p, 0, @) space. We wish to find the

volume of the solid S in (x, y, z) space which corresponds to U under the
transformation (2). Referring to Fig. 63, we see that S is a region such as
ABCDA'BC'D'between the spheres p= Pr and p= p2, between the planes
0 =,0, and 0:0zo and between the cones $ = OIOADD'A'I and
6: Or(OACC'B'). The region S is obtained by sweeping the plane region
F(ABCD shown in Fig. 64) through an angle L,0 : 0z- 0r. The Theorem of
Pappus (page 739) for determining volumes of solids of revolution applies
also for areas swept through any angle about an axis. We obtain

v(s): rcll* a,e,,.

F

t ---\
1,,
I

D/
l

u
FrcuRE63
Since p, { are polar coordinates in the zx plane (Fig. 64), we have

Therefore
fOz foz
r(s) =O'J,, p2 sin Q dp ilS
J,,
: 2

p2 sin 6 dp dd il|
I::!::r,
:Jfir"i, 0 tlV *0.
U
 751
stcrtoN 16.10 IVTASS OF A SOTID. TRIru NTEGRALS !N CYTINDRICAT AND SPHERICAT COORDINATES

More generally, if/(x, y, z) is continuous on a region S and if

g{p, 0, 0) :f(p cos 0 sin $, p sin0 sin @, p cos @),

then the triple integral of/may be transformed according to the formula

The triple integral is evaiuated by iterated integrations. The next 6xample

illustrates the process.

EXAMPLI 3 Find the volume above the cone z2 : x2 * y2 and inside the sphere
x2 + y2 * zz :2az (Fig. 65).

Solution In spherical coordinates the cone and sphere have the equations

O:; and p:2a cos f,

respectiveiy. Therefore

v(s) :
IJI
dv",, :
III
r sin g itvp66
SU
ftl4 f2acos6 f2r
:j. pz sin g do dp do
.}. .|,
f frt4 f zacos+
:rn p2 sin S ilp d0=
_l6atn |.",n
3 "or. Q sin 0 d0
Jo J. Jo

=*[-cosa $]!n:na'. x
 752
CHAPTE* 16 Munru tNTEGIAnoN

10 PROBLET\{S

In each of Problems I through 16, find the mass of the solid

having the given density d and bounded by the surfaces whose
equations are gi.ren.
I z2 = x2.* !2, x2 + y2 + 22 = a2, above the cone,,d: const.
2 The rectangular parallelepiped bounded by x a, : - x: a,
y: - !:
b, b, c, z: - z: d:
c, kl,x, + y2 + 21.
3 x2+y2+22:a2, x2+y2+22:b2, a1b, ,-
k!Er+7;7.
4 The rectangular parallelepiped bounded by x :0, x:2a,
:0, y : 2b, z :0, z : 2c, 6 : l4x, + y2 + zz).
y
5 x2+ y2:a2, x2+y2+z2=4a2;.6=kz2; outside the
cylinder.
6 T'he tetrahedron bounded by the coordinate planes and
x*y*z=li6-kxyz.
7 x2 + !2 :2ax, x2 + y2,+ z2 :4az: 6: lqx2 + y2).
8 z2:25(x2+y2), z:x2+y2*4; d=const; above the
paraboloid.
9 z2 : *, + yr, x2 + y2 + z2 : 2az; above the cone; d : fr2.
l0 Interior of x2+ y2+22=a2;6:\xz*y2+221,, n a
positive number.
ll y2: az, x2 + y2 + zz =2azi above the paraboloid;
x2 +
HGUTE 55
= const.
d
12 2z : xz + y2, z :2x; 6 = klGf;f.
n gl zrf;4Tr dx dz;S is bounded by z = 4, z :
= d2, 12 : a2 cos 29 (cylindrical. coordinates);
13 x2 + y2 + 22
ity xz + y2.
6 : const.
21
{J
1 dx ily dz;$ bounded by z: d'+t',2 = O aqd x2 + yz
14 x2 + y2 + z7:4a2, z:3a, above the plane; d = :16.
k,frrT7T7.
ls : *, + yr, (x, + yr), = a2(xi - y\; 6: kJGrTf .
z2
*16 z2:*'+y', x2+y2+22:2ax; Evaluate the integrals in problems 22 through 25 by using
d:const; bbove the spherical coordinates.
cone.
22 I dx dy dz where S is the intersection of the regions
Evaluate the integrals in problems 17 through zr oy using
{l
x2 + y2 + z? <2and x2 + y2 + z2 <22.
cylindrical coordinates.
l7 lll\Er+f dx dy dz; S is bounded Ay r:,/fiyi, z3 jff (x'? + yz+ z2)3 dx dy dz; S is determined by
z:0, and x2 + y2 : l. 0< x2 + y' <i, (x, * yzyttz<z<(l xz yzlrrz.
- -
t8 Z dx dy d4 S is determined by {O<xr+yr<t,
I{Jf ilx ilv dz; s is the hemisphere bounded by z: {Jr'
JT--;r:/ and z=0. $T/ <z< ll.(A truncated solid cone with vertcx at
(0,0,0)).
19 Jff 1 dx dy dz;S is in the first octant, inside the cylinder 25 IJJ sin([x, * y2 + zz11tt.z dx dy dz; S is the spherical shell
:
x2 + y2 ay and the paraboloid x2 + y2 + az : where
a is a constant. (Se Fig. 66.)
a2, between the spheres x, + y2 * z2
= I and x2 + y2 + z2
-9
 753
sEcTtoN 16.11 MoMENT or INERTTA. CrNtn oF MAss (ornoNAt)

11

MOMENT OF INERTIA. CTNTTN OF MASS (OPTIONAT)

The definition of moment of inertia of a solid body is similar to the definition

of moment of mass of a plane region (page 329). The following definition is
basic for this section.

DEFINIITION Suppose that a solid body occupies a region S and let L be any line in three-
space,Wemskeasubiliuisionof spaceintorectangularhoxesandlet Si,52,...,
Snbe those boxes which contain points of S. For each i, select any paint Pi(1;,4i.,
(;) in S;. If the sums

I r!m(S) (ri:
i= 1
distance of P,from L)

tend to a limit I as the norms of the subdioisions tend to zero, andfor any choices
of the P,, then I is called the moment of inertia of the solid S about L.

lt can be shown that if a solid S has continuous density 6(r, y, z), then the
moments of inertia 1,, I, and.I, about the x, y, and z axes, respectively, are
given by the triple integrals

DEFINITION If a solid S has a density 6(x, yrz\ and a mass m( S), the point (i, y, Z), defined
by the formulas

is called the center of mass of S.

In determining the center of mass, it is helpful to take into account all

available symmetries. The following rules are useful:
a) I/S is symmeffic in the xy plane and 5(x, y, -z):6(x, y, z),then Z:0.
A similar result holds for other coordinate planes.
 CHAPTER Mutnru tMEcRATtoN

b) I/S is symmetic in the x axis andd(x, -y, - z) : 6(x, y, z), then y : | : g.

A similar result holds for the other axes.

EXAMPLE r Find the moment of inertia of a homogeneous solid cone of base radius a and
altitude Ir about a line through the vertex and perpendicular to the axis.

Solution We take the vertex at the origin , the z axis as the axis of the cone, and thc
x axis as the line L about which the moment of inertia is to be computed
(Fig. 67).
+ "u Let a : arctan (alh) be half the angle opening of the cone. In spherical
coordinates we get (see Fig. 68)

': tlfo'* "t dV,,,

s

: p21sin2 { sin2 0 + cos2 d)p2 sin

i;J:-"'l;" 6 d0 dp d0.
Since

sinz 0 d0:n,
l:"
we obtain

TIGUR.E 58 ,: oJ,fa fhsccg

p4|o +cos2 4; sin 4 dp d0
J.
: nhs f"
TJo [(cos {)-5 + (cos {)-3] sin 6 dd
nhsfsecaa-l sec2a-11
:s[ 1 ----2 j: ff$*+o'),
since tan a: alh.

outMPLf 2 Find the center of mass of a solid hemisphere of radius a which has
proportional to the distance from the center.

solution we select the hemisphere so that the plane section is in the xy plane and
axis'is an axis of symmetry (Fig. 69). Then i: f :0. Changing to sph
coordinates, we have

:
!,xo' ,'
rr(s) sin s dp it$ d0 :lnkaa
I:" [:',',
and from this wE obtain

s'kP' P'
,:l:' f, tPT ds do
eos sin s dp

: .rn cos { sin Q dd :1"lSl1-,' 2a

#.!ro' !"'' 5'
FtGUtt 59 The center of mass is at (0, 0,2ali\.
 /55
sEcTtoN 16.11 MoMTNT oF INERTIA. CENTER oF MAss (opnoNAt)

11 PnoulrA{s

lneach or Problems I through 15, find the moment of inertia In each of Problems 16 through 32, find the center of mass of
about the given axis ofthe solid having the specified deosity 6 the solid having the given density and bounded by the surfaces
and bounded by the surfaces as described. as described. Describe each soiid in set notation.
I A cube of side a; 6 : consq about an edge. 16 x:0, ):0, z :A, xla+ ylb + zfc: l;6=const.
2 A cube ofiide a; , : const; about a line parallel to an edge, 17 x:0,.V:0, z :0', x + z : a, ! : z; 6 = kx.
at distance 2 from it, and in a plane of one of the faces. 18 x2 + yx : a2, x2 + zz : a', 6: const; the portion where
3 Bounded bY x:0, Y:0, z:0, x+z:a, Y:z;6:kx; x>0.
about the x axis. 19 z: x, z: - L, !2 :4 - 2x; , : const. i
4 x2 + y2 : a2; x2 + ?2 : a2; 6 -- const; about the z axis. 20 ;2:y'( I --x'),.y= l; 6:const.
5 z = x, 12 : 4 - 22, x = 0; 6 : const; about the z axis. 2l z :rJ, x2 + z: l, yz r z: l, 6:const.
6 x:0, !:0, z2: I - x - y; 6:const; about the z axis. 22 x:0, -l':0, z--Q, ytl? i y'i'+ zrl2 - ari2' 6:const.
7 z2 : y'(l - x2); / : 1; 6 : const; about the .x axis. 23 yz + z2 =4ax,.yz:ax, x=3a:6=const. (Inside y2:ax.l
t x2+ y2:a2, x2+y2:b2, z:O, z:h;6:ktE'+y'' 24 :4ax, x2 i- y2 :2,axi d = const.
z2
about the x axis (a < b).
25 z2 : *' + y', x2 + yx + 22 = d2,above the cone; d : const.
9 xz + y2 + 22 : a2, x2 + y2 + z2 = b2; 6 : k$+ fTZ: 26.22: *'+ y', x2 + y2 :2ax;6= k(xz + y2l.
about the z axis (c < b).
27 z2 : x2 + y2, x2 + y2 + z2 =2a2, above the cone; 6 : kz.
10 p: 4, p:5, z: l, z: 3 (1 < z <3); 6: const; about the
z axis. 28 .r2 + y2: az, x2 + yz + z2 :2a2, above the paraboloid;
: t, + y', (*' + y')' : 6 : const.
ll z2 a2(xz - y\; i: k\f;1-+f I

about the z axis. 29 x2 +- yz + z2 :4a2, z = 3a, above the plane; $ =

t2 z:O, z: 1E:7 -f 6: kz; about the x axis. ; kiG'+ ,4 rt.
13 x2la2 + yzlb' + z2fc2 - 1; 6 : const; about the z axis; p --4, p : J, :3
?t0 7 == 1, z (1 <z < 3); 6 :const.
(Divide the integral into two parts.)
14 (r-b\2+22:a2(0<a<b); 6:kz; about the e axis.
a3l z2 : r, + yr, {x, 4- yr), = a2{x2 - y2)i 6: krFr+f ;

Assume z > 0. the part for which x > 0.

15 p: : c,r : a(a <b < cj;outside r:

b, p c, betweeu p = b *32 z2:r.-l--r,,, x2+y2+22=2ax, above the cone;
and P : c; ,: const; about the z axis. 6 : const.

CHAPTER 16
RgvIrw PROBLEMS

ln I
through 11, evaluate the iterated integrals.
Problems
Caange order of integration, if necessary.
z
J'J''i, (i) r,* I I
f4 flnx

Js Jo
e-vdyax

f I f arcccy
, (x2+xy-y3\dxdy , jj dy dx l0 di" itx dY
J" i; I x2evo
J, J,
f2 fr
,,
,
J. J _,
(sin(xy) + xy\ dy dx
f fi [' *^, + x2'1 dz dx dy

f2 fxz f2 flnrr
,
J_, J,_,
(x2 -3y) dy dx o J, J, ' ;a* ar Iu Problems 12 through 17, evaluate the multiple integrals as
indicated.

t (xrrz -y2) tty dx . I L cos(lnxz) dx tty

n{|(r. 3)2y dx dy; F :{(x,y): o< x12, t < y <2}
ijI:""
7fi
CHA?TEI T6 MunnrNrtcr noN

(x2v2z2)
rsJJr,'-, ik tty F :{(x, y): I < x <z,o <y<2} 32 The reeion y, d,? *'o * I =rl. (volume)
{(x,
In each of Problems 33 through 38, find the moment of inertia
uJJ*<+"r) cos (|ry) dx tty. and the radius of gyration about the given axis of the plate F
whose density ls given.
F= {(x, y): -l Sx< 1, 0<y< 3} 33 Let F be the triangular region of Fig. 70 with density
rsJJ'i't, + y) ttx ar r <x P=3x)4yaxis.
=
{x,y):
0
=i, o =, =il v

,, + 3y2) ttx ity F :{(x, y): 0 < xz + y2 < t\

t[O.,
dx tllq F={(x, y):0< xl=!2,0<ys t}
" ff, P

In each of Problems 18 through 22, fiiid the area of the bounded FlGlrfE 70
region F detcrmined by the given curves.
lE xy:2,y:l,y:3x+1 34 Let F be the halfdisk of Fig. 7l with density p : 4x2 + t2);
x axis.
19 x2:4y,2y-x-4:O
fr y':5+x,y2:5-x
2l Y=Xz,!2:e'
D y-i:x',x2+y2:1
In each ofProblems 23 through 2T,evalaatethe given integral
by using polar coordinates.
23 lj cos( x'z + y2) dx dyF is the unit disk of radius I centered
at the origin. HGTNE 71

ulro-2 Iro _JitTfaya, 35 Let F be th€ region inside r:2 * sin 0 and outside r : l.
J J -J4-x' The density p:p(r,
g) is inversely proportional to the
distance from the pole, with proportionality constant fr;
x dx dyF as in problem 23. polar moment of inertia.
!lr-"'.r\
36 Let F be the region between two concentric circles of radii a
and b and center at the origin. The density is constant. Find
?5 | I yJ*'+
Jo Jzv
y2 dx dy the polar moment of inertia.
37 Let F be bounded by the curves y: sin2 x and y = - sin2 x,
x sin(y3) dy dx -lrSx S z. The density p is constant and equal to one;
" i, i, polar moment of inertia.
In each of Problems 2E through 32, use polar coordinates to 3E Let F be bounded by the curves !: d, !:0, 0<x < 14;
find the area, or the volume, as indicated, of the region given. P{x, y): xy; x axis.
23 The region inside the circle 7 = ! but to the right of the line In each of Problems 39 through 43, find the center ofmass ofthe
4r cos 0= 3. (Area) plate F described.
( _ -) 39 .F ii the bounded region described by y :
29 The resion 0): 0 < r < 2fos 0, 0 < d to.."l x 2 and x+y: Z.
{(r, = Z}. 40 F is the region bounded by x 2y + g : :0;
- 0; x * 3y + 5
( = -2; and x:4.
x
30 The ,.don a)' t <r <t + cos 0, o <0=
{{r, i}. ,o*", 41 Fis the region bounded by y:sin x, /=cos r,0 <x < |.
4
3l The region bounded below by the xy plane and above by 42 F is boundcd by y: ln x; y: ex: l; and x:2.
the paraboloid z: I - (x, + yrl. (Volume) 43
.F
is bounded by the curves.rE + 6: l; x = 0; and y : g.
 757
CHAPTER 16 REviEw PRoBIEMS

In each of Problems 44 through 46, find the area of the surface

described.
44 The part of the plane x + y + z- 4: 0 cut by the cylinder
" i, f i; e+z),tydzdx

x2 + y2 :16. y3 dz dy dx
45 That part of the surface z - !1x2 +y2) below the pl ane z : l.
" I i;l; zx2

rr rJTS ilt:17+7t
tt
45 That part of the cone x2 + y'
planesz:0and x-3=-22.
: z2 lyin| betwetn the two
J, J" j, dz dv dx

f rt2 f ft;4 f 2cNO

4? Find the volume of the solid generated by revolving the * J;-.J;--
J;'"
pz sin g dp dg do
triangle with vertices (2, 2),(5,2),(3" 3) about: (a) the x axis;
(b) the y axisl (c) the line y = x. Pt r-tr} rr
s7l l- l_idzdydx
48 Let F be the region in Fig. 72, with center of mass (i, y) J-t Jo JJ*'+y'
:(2,|),and areat. Find the volume of the solid generated
by revolving F about: (a) the x axis; (b) the y axis; and
(c) the line y: - x ++.
* J:['. J',n{,,*., dz dy dx
z2)3t2
59 JJJ et-'*r'* dx dy dz, where s is the solid bounded
, : ^ E';7 and above by
below by.the nappe of the co ns
the sphere x'+ y' *
- l. z2
.60 jfl
s
l(x-2)' +(y-3)'? +{z-2)21-rt2 dxdydz,wheresis
the unit sphere centered at the origin.
61 JJJdx ily dz,where S is the solid between two concentric
spheres of radii a and b,0 < a < b, centered at the origin.
62 Compute the mass of the solid described in Problem 6l ifthe
density y at each point is equal to the square ofthe distance
FC*.RE 72
from this point to the center of the two concentric spheres.

In each ofProblems 63 through 67, find the center ofmass ofthe

'49 Let/be a continuous function of one variable, and I el abe a
constant. Show that the following holds: solid having the grven density and bounded by the surfaces as
described.

dx dz dv-rli*-x)zr(x) dx 63 The solid is the tetrahedron with vertioes (0,0,0), (2,0,0),

I"J".['n,, (0, 3, 0), and (0, 0, 1). The density 6 - yz.

50 Show that the following formula is valid:

64 The solid is as in Problem 63; 6: xyz.

lt fy fl
dS fne solid is as in Problem 47; 6: y.

Jo
l' ,",.-',y1x) dx dy: JoI e - x){k-'f (x) dx,
I Jo 66 The solid is bounded above by xz + z:4, below by x + z
:2, arlrd on the sides by y: O and y: 3; d : I + y.
for constants a> 0, n, and a continuous functionl 67 The solid is bounded aboveby z:
y; below by the xy plane;
: : :
and on the sides by x 0, x 1, y 0, and y 1; 6 at a given :
In each of Problems 51 through 61, find the value ofthe iterated point is proportional to the square ofthe distance from the
:ntegral. point to the origin.
f1 fl-x fx+v *68 Let S be a sphere of radius a c€ntered at the origin. Show
tt
J. .|, j.',r+y-z)dzdydx that

t I i.'i, ,,. 4z) ttz dx rty {tJffidxdvdz:+