Calculus III (part 3): Multiple Integration (by Evan Dummit, 2021, v. 4.

00)

Contents
3 Multiple Integration 1

3.1 Double Integrals in Rectangular Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

3.1.1 Double Integrals via Riemann Sums . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

3.1.2 Iterated Integrals and Fubini's Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

3.1.3 Changing the Order of Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

3.2 Triple Integrals in Rectangular Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

3.2.1 Triple Integrals via Riemann Sums . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3.2.2 Fubini's Theorem and Methods of Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3.3 Alternative Coordinate Systems and Changes of Variable . . . . . . . . . . . . . . . . . . . . . . . . 12

3.3.1 General Changes of Variable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

3.3.2 Double Integrals in Polar Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

3.3.3 Triple Integrals in Cylindrical Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

3.3.4 Triple Integrals in Spherical Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

3.3.5 Additional Examples of Coordinate Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

3.4 Applications of Multiple Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

3.4.1 Areas, Volumes, and Average Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

3.4.2 Masses, Centroids, and Moments of a Plate and Solid . . . . . . . . . . . . . . . . . . . . . . 23

3 Multiple Integration
In this chapter we develop the theory of integration in multiple variables. Our focus is on double and triple integrals,
as they are the ones that show up in most applications of multivariable calculus in a 3-dimensional universe.

We start by dening double integrals over regions in the plane (and triple integrals over a region in 3-space) in
terms of Riemann sums, and then discuss how to evaluate double and triple integrals as iterated integrals. We
then discuss how to set up double and triple integrals in alternative coordinate systems, focusing in particular on
polar coordinates and their 3-dimensional analogues of cylindrical and spherical coordinates. We nish with some
applications of multiple integration for nding areas, volumes, masses, and moments of solid objects.

3.1 Double Integrals in Rectangular Coordinates

• Our motivating problem for integration of one variable was to nd the area below the curve y = f (x) above
an interval on the x-axis. The motivating problem for double integrals is to nd the volume below the surface
z = f (x, y) above a region R in the xy -plane.
• Integrals in one variable are initially dened using Riemann sums, and we will do the same for double integrals.

◦ The idea is the following: rst, we approximate the region R by many small rectangular pieces, and in
each piece we draw a rectangular prism with base in the xy -plane and upper face intersecting z = f (x, y).

1
 ◦ Then we take the limit over better and better approximations of the region R, and (so we hope) the
collective volume of the rectangular prisms will ll the volume under the graph of z = f (x, y).
◦ Here is a series of such approximations for the volume under f (x, y) = 2 − x − y 2 on the
2
region
R = {(x, y) : −1 ≤ x ≤ 1, −1 ≤ y ≤ 1}:

3.1.1 Double Integrals via Riemann Sums

• Here are the details of the formal denition of a double integral using Riemann sums:

• Denition: A region is a closed, bounded subset of the plane. If R is a region in the plane, a (tagged) partition
ofR into n pieces is a list of n disjoint rectangles inside R, where the k th rectangle contains the point (xk , yk ),
has width ∆xk , height ∆yk , and area ∆Ak = ∆yk · ∆xk . The norm of the partition P is the largest number
among the widths and heights of all of the rectangles in P .

• Denition: If P is a partition of the region R, and f (x, y) is a continuous function, we dene the Riemann sum
Xn
of f (x, y) on R corresponding to P to be RSP (f ) = f (xk , yk ) ∆Ak .
k=1
n
X
◦ Reminder: If g(x) is a function, then the notation g(k) means g(1) + g(2) + g(3) + · · · + g(n).
k=1

• Denition: For a continuous function

˜ f (x, y) dened on a region R, we dene the (double) integral of f on R,
denoted f (x, y) dA, to be the value of L such that, for every  > 0, there exists a δ > 0 (depending on )
R
such that for every partition P with norm(P ) < δ , we have |RSP (f ) − L| < .

◦ Remark: It can be proven (with signicant eort) that, if f (x, y) is continuous, then a value of L satisfying
the hypotheses actually does exist.

◦ Essentially, what this denition means is: the value of the denite integral is the limit of the Riemann
sums of f, as the size of the subregions in the partition becomes small.

◦ Note that our geometric motivation for integration involved nding the area under the graph of a function
z = f (x, y). But as with a function of one variable, the denition via Riemann sums does not require
thatf be nonnegative, and we interpret the integral of a negative function as giving a negative volume.

• Like with integrals of a single variable, double integrals have a number of formal properties that can be
deduced from the Riemann sum denition. Specically, for C an arbitrary constant and f (x, y) and g(x, y)
continuous functions, the following properties hold:
˜
◦ Integral of constant: C dA = C · Area(R).
R ˜ ˜
◦ Constant multiple of a function: C f (x, y) dA = C · R f (x, y) dA.
˜ R ˜ ˜
◦ Addition of functions:
R˜
f (x, y) dA + R g(x, y) dA = R [f (x, y) + g(x, y)] dA.
˜ ˜
◦ Subtraction of functions: f (x, y) dA − R g(x, y) dA = R [f (x, y) − g(x, y)] dA.
R ˜
◦ Nonnegativity: if f (x, y) ≥ 0, then
R
f (x, y) dA ≥ 0.
∗ As a corollary, if f (x, y) ≥ g(x, y), f (x, y) − g(x, y) ≥ 0 and using
then by applying this property to
˜ ˜
the subtraction property, we see that f (x, y) dA ≥ g(x, y) dA.
R ˜ R ˜ ˜
◦ Union: If R1 and R2 don't overlap and have union R, then
R1
f (x, y) dA+ R2 f (x, y) dA = R f (x, y) dA.

2
3.1.2 Iterated Integrals and Fubini's Theorem

• Evaluating double integrals via Riemann sums is generally quite hard, even for very simple functions. However,
we can give an alternative way, using the idea that volume is given by integrating cross-sectional area:

◦ Suppose for simplicity that we want to nd the volume underneath the portion of z = f (x, y) lying above
the rectangular region a ≤ x ≤ b, c ≤ y ≤ d.
◦ If we imagine taking the solid volume and slicing it into thin pieces perpendicular to the x-axis from
´b
x=a to x = b, then the volume is given by the integral
a
A(x) dx, where A(x) is the cross-sectional
area at a given x-coordinate.
◦ If we then look at each cross-section, we see that the area A(x0 ) is simply the area under the curve
´d
z = f (x0 , y) between y =c and y = d, which is
c
f (x0 , y) dy , where here we are thinking of x0 as a
constant independent of y .

◦ By putting all of this together, we see that the volume under z = f (x, y) on the region R should also be
´ b h´ d i
given by the iterated integral
a c
f (x, y) dy dx, where we integrate rst (on the inside) with respect

to the variable y, and then second (on the outside) with respect to the variable x.
´b´d
◦ We will usually write iterated integrals without the brackets:
a c
f (x, y) dy dx.
◦ Note that there are two limits of integration, and they are paired with the two dierential variables
inside out: the inner limits [c, d] are paired with the inner dierential dy , and the outer limits [a, b] are
paired with the outer dierential dx.

• Example: Find the volume under the surface z = 6 − x2 − y 2

0 ≤ x ≤ 1, 0 ≤ y ≤ 2.
that lies above the region

´1´2

◦ From the discussion above, we see that the volume is given by the iterated integral 0 0 6 − x2 − y 2 dy dx.
´2

◦ To evaluate the inner integral 0 6 − x2 − y 2 dy , we view x as a constant and take the antiderivative
(with respect to y ):

ˆ 2    
2 2 2 1 3 2 2 8 28
− 2x2 .


6−x −y dy = 6y − x y − y  = 12 − 2x − − (0 − 0 − 0) =
0 3 y=0 3 3

´1
   
28 28 2 3 1 26
◦ Now we can evaluate the outer integral
0
− 2x2 dx = x− x  = .
3 3 3 x=0 3

• We can generalize this idea to compute the volume of a solid lying above a non-rectangular region.

◦ Explicitly, suppose that the region R is dened by the inequalities a ≤ x ≤ b, c(x) ≤ y ≤ d(x): this
represents the region above y = c(x) and below y = d(x), between x = a and x = b.
◦ Then by the same logic as before, the volume of the solid below z = f (x, y) above the region R in the
´b
xy -plane is given by the integral a
A(x) dx, where A(x) is the cross-sectional area at a given x-coordinate.
◦ The areaA(x0 ) of each cross section will be the area under the curve z = f (x0 , y) between y = c(x0 ) and
´ d(x )
y = d(x0 ), which is c(x00) f (x0 , y) dy : here, again, we are thinking of x0 as a constant independent of y .
´ b h´ d(x) i
◦ So, we see that the volume is given by the iterated integral a c(x) f (x, y) dy dx, where now the inner
limits depend on x.

• Example: Find the volume under the surface z = 6 − x2 − y 2 that lies above the region 0 ≤ x ≤ 1, x ≤ y ≤ 2x.
´ 1 ´ 2x

◦ From the discussion above, we see that the volume is given by the iterated integral
0 x
6 − x2 − y 2 dy dx.
´ 2x

◦ To evaluate
x
6 − x2 − y 2 dy , we take the antiderivative with respect to y:
ˆ 2x      
2 2 2 1 3 2x 8 3 1 3 10
= 12x − 2x − x − 6x − x − x = 6x − x3 .
3 3


6−x −y dy = 6y − x y − y 
x 3 y=x 3 3 3

3
 ´1
   
10 5 4 1 13
◦ Now we can evaluate the outer integral
0
6x − x3 2
dx = 3x − x  = .
3 6 x=0 6

• Notice also that instead of slicing perpendicular to the x-axis, we could have tried slicing perpendicular to
they -axis. By going through the same logic as before, the volume should be given by the iterated integral
´d´b
c a
f (x, y) dx dy in the other order.
• Since we have three ways to interpret the integral of a function on a region, namely by Riemann sums and
by iterated integration (in two possible orders), we would hope that these denitions all agree. It turns out
that as long as the function is continuous on the entire region, they do:

• Theorem (Fubini): If f (x, y) is continuous on a region R = {(x, y) : a ≤ x ≤ b, g1 (x) ≤ y ≤ g2 (x)} and R =

¨ ˆ b ˆ g2 (x) ˆ d ˆ h2 (y)
{(x, y) : c ≤ y ≤ d, h1 (y) ≤ x ≤ h2 (y)}, then f (x, y) dA = f (x, y) dy dx = f (x, y) dx dy .
R a g1 (x) c h1 (y)

◦ Fubini's Theorem says that in order to evaluate a double integral, we need only choose an order of
integration, slice up the region accordingly, and then write down and evaluate the resulting iterated
integral.

• Using Fubini's Theorem, we can give the following process for evaluating double integrals:

◦ Step 0: If necessary, determine the region of integration, and sketch it (if it is complicated).

◦ Step 1: Decide on an order of integration and slice up the region according to the chosen order.

◦ Step 2: Determine the limits of integration one at a time, starting with the outer variable. Note that
the region may need to be split into several pieces, if the boundary of the region is complicated.
´ b ´ g2 (x)
∗ For R = {(x, y) : a ≤ x ≤ b, g1 (x) ≤ y ≤ g2 (x)}, the integral is
a g1 (x)
f (x, y) dy dx.
´ d ´ h2 (y)
∗ For R = {(x, y) : c ≤ y ≤ d, h1 (y) ≤ x ≤ h2 (y)}, the integral is
c h1 (y)
f (x, y) dx dy .
◦ Step 3: Evaluate each iterated integral as a single-variable integral in the appropriate variable.

∗ Remember that the outer variable is a constant, for the purposes of the inner integral. Think of the
inner integration as taking an anti-partial-derivative and then evaluating at the endpoints, just like
with a regular integral in one variable.

˜
• Example: Evaluate I= R
x2 y dA, where R is the region {(x, y) : 0 ≤ x ≤ 2, 0 ≤ y ≤ 3}.

◦ We've been given the inequalities for the region, so we just need to write down the iterated integral. We
can use either integration order dy dx or dx dy , so let's do both.
´2´3
◦ For dy dx, we have I= 00
x2 y dy dx. To evaluate the inner integral, we take the antiderivative of x2 y
1 2 2
with respect to y , yielding x y , and then plug in to evaluate the outer integral, as follows:
2
ˆ 2 ˆ 3  ˆ 2  ˆ 2
1 2 2 3 9 2 3
I= x2 y dy dx = x y  dx = x dx = x3 |2x=0 = 12 .
0 0 0 2 y=0 0 2 2
´3´2
◦ For dx dy we have I= 0 0
x2 y dx dy . To evaluate the inner integral, we take the antiderivative of x2 y
1 3
with respect to x, x y , and then plug in like above:
yielding
3
ˆ 3 ˆ 2  ˆ 3  ˆ 3
1 3 2 8 4
I= x2 y dx dy = x y  dy = y dy = y 2 |3y=0 = 12 .
0 0 0 3 x=0 0 3 3
˜
• Example: Evaluate
R
ex+y dA, where R is the region {(x, y) : 0 ≤ x ≤ ln(2), 0 ≤ y ≤ x}.
´ ln(2) ´ x
◦ Since the y -inequalities depend on x, the order dy dx is easiest: the resulting iterated integral is
0 0
ex+y dA.

4
 ◦ Now we compute the inner integral, then plug in and evaluate the outer one:

ˆ ln(2) ˆ x  ˆ ln(2) ˆ ln(2)  

 x+y  x 1 2x ln(2) 1
ex+y dy dx
 2x
e − ex dx = e − ex 

= e dx = = .
2 2

0 0 0 y=0 0 x=0

• Example: Integrate the function f (x, y) = xy 2 over the nite region R between the curves y = 2x and y = x2 .

◦ We can integrate in either order, so we will set up both integrals to illustrate the ideas in each case.

◦ We can see that the two curves y = 2x and y = x2 will intersect at (0, 0) and (2, 4), so the region looks
like this:

◦ We have two choices for setting up the integral: one with the order dx dy having horizontal slices, and
the other with the order dy dx having vertical slices.

◦ For the order dy dx, we see that the allowable values for x range from 0 ≤ x ≤ 2. Then for any given
value of x, the allowable values of y range from the lower curve y = x2 to the upper curve y = 2x.
ˆ 2 ˆ 2x
Therefore, the bounds on y are x2 ≤ y ≤ 2x, and so the integral is xy 2 dy dx .
0 x2

◦ We then compute

ˆ 2 ˆ 2x ˆ 2 ˆ 2
8 · 25 28
    
1 3 2x 8 4 1 7 8 5 1 8 2 32
xy 2 dy dx = xy y=x 2 dx = x − x dx = x − x x=0 = − = .
0 x2 0 3 0 3 3 15 24 15 24 5

◦ For the order dx dy , we see that the allowable values for y range from 0 ≤ y ≤ 4. Then for any given
value of y , the allowable values of x range from the left curve y = 2x to the right curve y = x2 . Therefore,
ˆ 4 ˆ √y
1 √
the bounds on x are y ≤ x ≤ y , and so the integral is xy 2 dx dy .
2 0 y/2

◦ We then compute

ˆ 4 ˆ √
y ˆ 4 ˆ 4
1 2 2 √y 44 45
    
1 3 1 4 1 4 1 32
xy 2 dx dy = y − y 5 4y=0 =

x y x=y/2 dy = y − y dy = − = .
0 y/2 0 2 0 2 8 8 40 8 40 5

• Example: Integrate the function f (x, y) = y on the region R given by the nite area between the curves
x = y2 − 1 and y = 1 − x.

◦ Here is a sketch of the region:

◦ To nd the intersection points of the boundary curves y = 1−x and x = y 2 −1, plugging the rst equation
2 2
into the second yields x = (1 − x) − 1 = x − 2x, so that x = 0, 3. Therefore, the two intersection points
are (0, 1) and (3, −2). Also, the vertex of the parabola is (−1, 0).

5
 ◦ We can set up the integral with either integration order, but to use vertical slices we need to divide the
region into two pieces, since the top curve changes from the upper half of the parabola to the line at
x = 0. If we use horizontal slices then it is not necessary to divide the region of integration.

◦ For horizontal slices, with integration order dx dy :

∗ We see that the range for y is −2 ≤ y ≤ 1, and then the corresponding range for x is y 2 −1 ≤ x ≤ 1−y .
´ 1 ´ 1−y
∗ The desired integral is therefore
−2 y 2 −1
y dx dy . Now we can evaluate it:
ˆ 1 ˆ 1−y ˆ 1 1−y ˆ 1
y(1 − y) − y(y 2 − 1) dy
 
y dx dy = [xy]  dy =

−2 y 2 −1 −2 x=y 2 −1 −2
ˆ 1  
 2 3
 2 1 3 1 4 1 9
= 2y − y − y dy = y − y − y  = −
−2 3 4 y=−2 4

◦ For vertical slices, with integration order dy dx:

∗ The range for x is −1 ≤ x ≤ 3. We need to divide it into two pieces, since the upper curve changes
x = 0.
at
√ √
∗ Solving x = y 2 −1 for y in terms of x yields y = x + 1 (the upper half of the parabola), y = − x + 1
(the lower half of the parabola. The line segment is y = 1 − x.
√
∗ For −1 ≤ x ≤ 0 the lower curve is the bottom √ half of the parabola y = − x + 1 and the upper
√ √
curve is the top half of the parabola y = x + 1, so the range for y is − x + 1 ≤ y ≤ x + 1, and
the integral is

ˆ 0 ˆ √
x+1 ˆ 0   √ ˆ 0
1 2  x+1
√
y dy dx = y  √ dx = 0 dx = 0.
−1 − x+1 −1 2 y=− x+1 −1
√
∗ For 0≤x≤3 the lower curve is still the bottom half of the parabola
√ y = − x + 1 but the upper
curve is now the line segment y = 1 − x, so the range for y is − x + 1 ≤ y ≤ 1 − x and the integral
is
ˆ 3 ˆ 1−x ˆ 3   ˆ 3
1 2 1−x 1
(1 − x)2 − (x + 1) dx

√
y dy dx = y  √ dx =
0 − x+1 0 2 y=− x+1 0 2
ˆ 3  
1 2  1 1 3 3 2 3 9
= x − 3x dx x − x  =− .
0 2 2 3 2 x=0 4
ˆ 0 ˆ √
x+1 ˆ 3 ˆ 1−x
9
∗ The answer is then the sum
√
y dy dx + √
y dy dx = − .
−1 − x+1 0 − x+1 4

3.1.3 Changing the Order of Integration

• Although Fubini's Theorem guarantees that integrating in either order will yield the same result, it can happen
that one order of integration is easier to evaluate than the other. In such situations, reversing the order of
integration can be useful.

• To change the order of integration in a multiple integral, follow these steps:

◦ Step 1: Sketch the region of integration.

◦ Step 2: Slice up the region according to the new order, and determine the new limits of integration one
at a time, starting with the outer variable. Note that the region may need to be split into several pieces,
if the boundary of the region is complicated.

◦ Step 3: Evaluate the new integral.

ˆ 1 ˆ y
• Example: Reverse the order of integration for xy dx dy .
0 y2

6
 ◦ The region is dened by 0≤y≤1 and y2 ≤ x ≤ y, and the current order of integration has horizontal
slices:

◦ To reverse the order of integration, we want to slice up the region perpendicular to the x-axis so that it
has vertical slices, as in the second diagram above.
√
◦ We see that the range for x is0 ≤ x ≤ 1, and then the new limits for y are x≤y≤ x.
ˆ 1 ˆ √x
◦ Hence the integral becomes xy dy dx .
0 x

ˆ π ˆ π
sin(y)
• Example: Evaluate the integral dy dx by reversing the order of integration.
0 x y

sin(y)
◦ As written it is not possible to evaluate the inner integral, as does not have an elementary
y
antiderivative. So we try reversing the order of integration.

◦ The region is dened by 0≤x≤π and x ≤ y ≤ π, and the current order of integration has vertical
slices:

◦ To reverse the order of integration, we want to slice up the region perpendicular to the y -axis so that it
has horizontal slices, as in the second diagram above.

◦ We see that the range for y is 0 ≤ y ≤ π , and then the new limits for x are 0 ≤ x ≤ y.
ˆ ˆ
π y
sin(y) ´π
◦ Then the integral is dx dy = 0 sin(y) dy = 2 .
0 0 y

3.2 Triple Integrals in Rectangular Coordinates

• Now that we have discussed double integrals and integrating a function over a region in the plane, it is not
hard to see how to generalize to triple integrals and integrating a function over a region in 3-space: we just
increase the number of variables by 1, and integrate functions f (x, y, z) over regions in 3-space instead of
functions f (x, y) over regions in the plane.

◦ Denition: A region in 3-dimensional space is a closed, bounded subset of 3-space.

◦ For clarity, we will use D to denote solid regions in 3-space, and reserve R for regions in the plane.

• The motivating problem for integration in three variables is somewhat less clear: for single integrals we wanted
to nd the area under a curve y = f (x), and for double integrals we wanted to nd the volume under a surface
z = f (x, y).

7
 ◦ For triple integrals it is somewhat harder to envision what happens when we move up by 1 dimension: we
are then nding the 4-dimensional volume under a 3-dimensional hypersurface (whatever that means!).

◦ One way to interpret what a triple integral represents is to think of a function f (x, y, z) as being the
density of a solid object D at a given point (x, y, z): then the triple integral of f (x, y, z) on the region D
represents the total mass of the solid.

3.2.1 Triple Integrals via Riemann Sums

• The denition of a triple integral via Riemann sums is essentially same as with double integrals: we approx-
imate the region D by many small rectangular pieces, sum the function over all of the pieces, and then take
the limit as the size of the pieces gets smaller.

◦ Denition: For a region D a (tagged) partition of D into n pieces is a list of disjoint rectangular boxes
inside D, k th rectangle contains the point (xk , yk , zk ), has length ∆xk , width ∆yk , height
where the
∆zk , and volume ∆Vk = ∆zk · ∆yk · ∆xk . The norm of the partition P is the largest number among the
dimensions of all of the boxes in P .

◦ Denition: For f (x, y, z) a continuous function and P a partition of the region D, we dene the
n
X
Riemann sum of f (x, y, z) on D corresponding to P to be RSP (f ) = f (xk , yk , zk ) ∆Vk .
k=1
˝
◦ Denition: For f (x, y, z) a continuous function, we dene f (x, y, z) dV ,  the (triple) integral of f on the region D,
D
to be the value of L such that, for every  > 0, there exists a δ>0 (depending on ) such that for every
partition P with norm(P ) < δ , we have |RSP (f ) − L| < .
◦ Remark: It can be proven (with signicant eort) that, if f (x, y, z) is continuous, then a value of L
satisfying the hypotheses actually does exist.

• Triple integrals share the same formal properties as double integrals. Specically, for C an arbitrary constant
and f (x, y, z) g(x, y, z) continuous functions, the following properties hold:
and
˝
◦ Integral of constant: C dV = C · Volume(D).
D ˝ ˝
◦ Constant multiple of a function: C f (x, y, z) dV = C · D f (x, y, z) dV .
˝ D ˝ ˝
◦ Addition of functions: f (x, y, z) dV + D g(x, y, z) dV = D [f (x, y, z) + g(x, y, z)] dV .
˝
D ˝ ˝
◦ Subtraction of functions: f (x, y, z) dV − D g(x, y, z) dV = D [f (x, y, z) − g(x, y, z)] dV .
D ˝
◦ Nonnegativity: if f (x, y, z) ≥ 0, then f (x, y, z) dV ≥ 0.
D ˝ ˝ ˝
◦ Union: If D1 and D2 don't overlap and have union D ,
D1
f (x, y, z) dV + D2 f (x, y, z) dV = D f (x, y, z) dV .

3.2.2 Fubini's Theorem and Methods of Evaluation

• Like with double integrals, we generally want to write triple integrals as iterated integrals, and avoid using
Riemann sums. Once we have the integrals, the procedure for evaluating them is exactly the same:
ˆ 1 ˆ y ˆ y
• Example: Evaluate the integral xyz dz dx dy .
0 0 x

◦ We just work one step at a time:

ˆ 1 ˆ y ˆ y ˆ 1 ˆ y 
1 2 y

xyz dz dx dy = xyz z=x
dx dy
0 0 x 0 0 2
ˆ 1ˆ y 
1 3 1 3
= xy − x y dx dy
0 0 2 2
ˆ 1 
1 2 3 1 4 y
= x y − x y x=0 dy
0 4 8
ˆ 1
1 5 1
= y dy = .
0 8 48

8
• Much of the diculty in setting up triple integrals is converting the description of the region into explicit
bounds of integration. To do this, we choose an order of integration, and then slice up the region of integration
accordingly; however, things are complicated by the fact that we now have 3 variables instead of 2.

• We might also worry that the value of a triple integral might depend on the order of integration, but as we
would hope, there is a version of Fubini's Theorem that guarantees the order of integration does not matter
as long as the function is continuous:

• Theorem (Fubini): If f (x, y, z) is continuous on D = {(x, y, z) : a ≤ x ≤ b, g1 (x) ≤ y ≤ g2 (x), h1 (x, y) ≤ z ≤ h2 (x, y)},
˚ ˆ b ˆ g2 (x) ˆ h2 (x,y)
then f (x, y, z) dV = f (x, y, z) dz dy dx, and all other orders of integration will also
D a g1 (x) h1 (x,y)
yield the same value.

• Here is the general process for setting up and evaluating triple integrals:

◦ Step 1: If necessary, determine the region of integration, and sketch it.

◦ Step 2: Decide on an order of integration and determine the limits of integration. Note that the region
may need to be split into several pieces, depending on the order chosen.

∗ The simplest method is to project the solid region into the plane spanned by the outer and middle
variables, obtaining a region in that plane: then set up the outer and middle limits in the same way
as for a double integral on that planar region.

∗ Specically: with the integration orders dz dy dx or dz dx dy we project into the xy -plane, with the
orders dy dz dx or dy dx dz we project into the xz -plane, and with dx dz dy or dx dy dz we project
into the yz -plane.

∗ Then to nd the bounds on the inner limit, we imagine moving parallel to the direction of the inner
variable until we enter the region (the surface we pass through giving the lower bound of integration)
and continuing until we leave the region (the surface we pass through giving the upper bound of
integration).

∗ Note that the bounds for the outer variable must be constants, the bounds for the middle variable
can only depend on the outer variable, and the bounds for the inner variable can depend on both of
the others.

◦ Step 3: Evaluate each iterated integral as a single-variable integral in the appropriate variable.

∗ Remember that all variables except the current variable of integration are to be treated as constants.

˝
• Example: Find
D
x dV where D is the tetrahedron bounded by the planes x = 0, y = 0, z = 0, and
x + 2y + 3z = 6.

◦ First, we sketch the region. It is a tetrahedron (triangular pyramid) whose vertices are (0, 0, 0), (6, 0, 0),
(0, 3, 0), and (0, 0, 2):

◦ There are six possible orders of integration, and we could use any of the six. We will set up the integral
in the order dz dy dx, which requires us to project this solid into the xy -plane.

9
 ◦ Motivated by the picture, we can see that the projection of the solid yields the interior of a triangle in
the xy -plane (namely, the base of the pyramid):

◦ We then cut up plane region with vertical slices (corresponding to the order dy dx). Since the diagonal
line has equation x + 2y = 6, which is the same as y = (6 − x)/2, we see that the region is described by
0≤x≤6 and 0 ≤ y ≤ (6 − x)/2.
◦ For the bounds on z, x and y , as we
we need to use the 3-dimensional picture of the solid. For xed
move in the direction of increasingz , we enter the solid through the xy -plane z = 0 and exit the solid
through the tilted plane z = (6 − x − 2y)/3, so the bounds on z are 0 ≤ z ≤ (6 − x − 2y)/3.
´ 6 ´ (6−x)/2 ´ (6−x−2y)/3
◦ Hence the desired integral is 0 0 0
x dz dy dx. Now we can evaluate it:
ˆ 6 ˆ (6−x)/2 ˆ (6−x−2y)/3 ˆ 6 ˆ (6−x)/2 (6−x−2y)/3
x dz dy dx = [xz] z=0 dy dx
0 0 0 0 0
ˆ 6 ˆ  (6−x)/2 
1 2
= 2x − x2 − xy dy dx
0 0 3 3
ˆ 6 
1 1 (6−x)/2
2xy − x2 y − xy 2 y=0

= dx
0 3 3
ˆ 6 
1 2 1 2
= x(6 − x) − x (6 − x) − x(6 − x) dx
0 6 12
ˆ 6    
2 1 3 3 2 1 3 1 4 6
= 3x − x + x dx = x − x + x x=0 = 9 .
0 12 2 3 48
˝
• Example: Find
D
x2 dV , where D is the triangular wedge bounded by the planes x = 0, x = 1, z = x ,
z = 2x, y = 0, and y = 1.

◦ Upon sketching the region, we see that it has a prism shape:

◦ Aided by the picture, it is easy to describe the projection of the region into the xz -plane as 0≤x≤1
and x ≤ z ≤ 2x. Then, if we move in the direction of increasing y, we enter the region through y=0
and leave through y = 1.

10
 ´ 1 ´ 2x ´ 1
◦ So the bounds are 0 ≤ x ≤ 1, x ≤ z ≤ 2x, and 0≤y≤1 and thus the integral is
0 x 0
x2 dy dz dx:
ˆ 1 ˆ 2x ˆ 1 ˆ 1 ˆ 2x ˆ 1 ˆ 2x
 2  1
x2 dy dz dx = x y y=0 dz dx = x2 dz dx
0 x 0 0 x 0 x
ˆ 1 ˆ 1
 2  2x 1
= x z z=x dx = x3 dx = .
0 0 4
˝
• Example: Set up, but do not evaluate, an iterated triple integral for
D
x2 dV , where D is the region below
2 2 2 2
the surface z = 12 − x − y and above the surface z = 2x + y .

◦ Each of the surfaces is a parabolic bowl (the rst opening downward and the second opening upward);
here is a plot of each surface along with the region D:

◦ Since both surfaces are of the form z = f (x, y), it will be easiest to project into the xy -plane. We can
see (per the pictures) that the widest part of the solid occurs when the two surfaces intersect. We see
thatz = 12 − x2 − y 2 intersects z = 2x2 + y 2 when 12 − x2 − y 2 = 2x2 + y 2 , which is equivalent to
3x + 2y 2 = 12.
2

◦ Thus, we see that the projection of the solid into the xy -plane is the interior of the ellipse 3x2 + 2y 2 = 12,
pictured below:

◦ With vertical slices (corresponding to the integration order

√ √ dy dx), we see that the corresponding range
is −2 ≤ x ≤ 2, − 12 − 3x2 /2 ≤ y ≤ 12 − 3x2 /2.
◦ Furthermore, since the upper surface is z = 12 − x2 − y 2 and the lower surface is z = 2x2 + y 2 , the range
2 2 2 2
for z is 2x + y ≤ z ≤ 12 − x − y .
ˆ 2 ˆ √12−3x2 /2 ˆ 2x2 +y 2 √
◦ Thus, the full integral is
√ x2 dz dy dx . (The value is 10π 3, but the actual
−2 − 12−3x2 /2 12−x2 −y 2
computation required to obtain this value involves a lot of algebra!)

11
 • As a nal remark, we will note that there exists a computational algorithm known as the cylindrical algebraic decomposition
for converting a description of a region in space bounded by polynomial inequalities (such as x2 ≤ y + z or
2 2 2
x + y + z < 4) to a union of regions described as a ≤ x ≤ b, c(x) ≤ y ≤ d(x), e(x, y) ≤ z ≤ f (x, y).

◦ This algorithm, implemented in some computer algebra systems, provide a method for setting up inte-
gration problems over regions of these forms.

◦ For example, applying this algorithm to the region dened by the inequalities x2 + y 2 < 1, x2 + z 2 < 1,
1
y 2 + z 2 < 1, and 0 < x < y < z shows that it consists of a single piece dened by 0 < x < √ ,
2
1 p
x<y< √ , and y < z < 1 − y2 .
2

3.3 Alternative Coordinate Systems and Changes of Variable

• We might wonder if there is a multivariable equivalent of the one-variable integration technique of substitution.
If we think about substitution as being a change of variables to a new system of coordinates, then the answer
is yes: we can rewrite multiple integrals in dierent coordinate systems.

• In order to simplify the computation of integrals on regions that have circular or spherical symmetries, which
arise very often in physics due to the spherical symmetry of gravitational and electrical elds, there are several
alternative coordinate systems we frequently employ beyond our usual rectangular xy and xyz coordinate
systems: polar coordinates (r, θ) in 2-dimensional space, and cylindrical (r, θ, z) coordinates and spherical
coordinates (ρ, θ, ϕ) in 3-dimensional space.

3.3.1 General Changes of Variable

• Here are the general theorems on changing coordinates in a double or triple integral (we will not prove these):

• Theorem (General Substitution, 2 variables): If

¨ ¨ f (x, y) is continuous
 on R, and x = x(s, t) and y = y(s, t) are
 ∂(x, y) 
functions of s and t, then f (x, y) dy dx = g(s, t)   dt ds, where R0 is the region R expressed in
R R 0 ∂(s, t) 
 
∂(x, y)  ∂x/∂s ∂x/∂t 
st-coordinates, g(s, t) = f (x(s, t), y(s, t)), and = J(x, y) =    is the Jacobian of the
∂(s, t) ∂y/∂s ∂y/∂t 
coordinate transformation.

• Theorem (General Substitution, 3 variables): If

¨ f (x, y, z) is continuous¨on D, and x = x(s, t, u)  , y = y(s, t, u),
 ∂(x, y, z) 
and z = z(s, t, u) are functions of s,t,u, then f (x, y, z) dz dy dx = g(s, t, u)   du dt ds, where
D D0 ∂(s, t, u) 
∂(x, y, z)
D0 is the region D expressed in stu-coordinates, g(s, t, u) = f (x(s, t, u), y(s, t, u), z(s, t, u)), and =
  ∂(s, t, u)
 ∂x/∂s ∂x/∂t ∂x/∂u 
 
J(x, y, z) =  ∂y/∂s ∂y/∂t ∂y/∂u  is the Jacobian of the coordinate transformation.
 ∂z/∂s ∂z/∂t ∂z/∂u 

◦ Warning: For this theorem to apply, the change of coordinates needs to be injective (one-to-one) on the
original region of integration: if the image of the old region in the new set of coordinates ranges over
some parts of the new region of integration more than once, the formula will be incorrect .
1

◦ The way in which the region of integration and the function transform under a change of coordinates are
very natural. The only signicant content of the substitution theorems are the statements about how
the dierential of area dA = dy dx and the dierential of volume dV = dz dy dx transform.

◦ The new dierential is related to the old one by the Jacobian, which measures how the size of a small
box in xy (or xyz )-coordinates changes when we switch to the new st (or stu)-coordinates. Giving a
precise proof requires the geometric denition of the determinant of a matrix, which in fact measures
precisely how a (linear) transformation changes areas/volumes.

1 For example, the change of coordinates s = sin(x), t = sin(y) fails this criterion if applied to the entire xy -plane, because as x and y
range over, say, the square 0 ≤ x ≤ 4π , 0 ≤ y ≤ 4π , every point in the image square −1 ≤ s ≤ 1, −1 ≤ t ≤ 1 will be covered four times.

12
 ˜
• Example: Evaluate the integral
R
(y + x) dA where R is the region bounded by the lines y = x, y = 2 + x,
y = −x, and y = 3 − x.

◦ The region is a parallelogram:

◦ In principle, we could divide this region into three pieces (using horizontal or vertical slices) and then
compute the integral separately on each one. This would be rather lengthy since we would need to
compute all of the points of intersection, then nd the necessary bounds to describe each region, and so
on.

◦ We will do it a dierent way: notice that the bounding lines can be written as y − x = 0, y − x = 2,
y + x = 0, and y + x = 3.
◦ This suggests making the change of variables s = y + x, t = y − x, since in st-coordinates the equations
become the much simpler t = 0, t = 2, s = 0, and s = 3. (This change of variables is one-to-one since it
is linear.)

◦ The new region is then 0 ≤ t ≤ 2, 0 ≤ s ≤ 3, and the new function is x + y = s.

 
s−t s+t  1/2 −1/2  1
◦ For the Jacobian, we have x = and y =
 1/2 1/2  = 2 .
, so J(s, t) =  
2 2
˜ ´2´3 1 ´2 9 9
◦ The desired integral is then
R
(y + x) dA = 0 0
s· ds dt = 0 dt = .
2 4 2

3.3.2 Double Integrals in Polar Coordinates

• Polar coordinates may already be familiar, as they are often discussed in precalculus and single-variable
calculus.

• Denition: The polar coordinates (r, θ) of a point (x, y) satisfy x = r cos(θ), y = r sin(θ), for r ≥ 0 and
0 ≤ θ ≤ 2π .
p
◦ We have r= x2 + y 2 and θ = tan−1 (y/x), possibly plus π depending on the signs of x and y.
◦ The parameter r is a distance giving the radius from the origin, and the parameter θ measures the
angle with respect to the positive x-axis.
◦ Polar coordinates are useful in simplifying regions that are circular: the circle x2 + y 2 = a2 in rectangular
coordinates (over which it is cumbersome to set up double integrals) becomes the much simpler equation
r=a in polar coordinates.
p
◦ Polar coordinates are also useful in simplifying functions which involve x2 + y 2 or (especially) x2 + y 2 .
◦ Lines through the origin also have reasonably simple descriptions: the line y = mx becomes the pair of
rays θ = tan−1 (m) and θ = tan−1 (m) + π when written in polar coordinates. (The two rays point in
opposite directions.)
   
 ∂(x, y)   cos θ −r sin θ 
◦ For polar coordinates, we have J =   =   = r.
∂(r, θ)   sin θ r cos θ 
◦ Thus the dierential in polar coordinates is dA = r dr dθ. We typically set up polar integrals with this
integration order, because most of the nice curves in polar coordinates have equation r = f (θ) for some
function f.

• Example: Integrate the function f (x, y) = x + 2y on the region R lying above the lines y=x and y = −x
and inside the circle x2 + y 2 = 4.

13
 ◦ Upon sketching the region, we see that it is a quarter-disc:

◦ In principle we could do this integral in rectangular coordinates, but it would be messy: using either
horizontal or vertical slices would require splitting the region into two pieces, since the boundary curves
change partway through the region in each case. So instead we will use polar coordinates.

◦ To describe the region, remember that in polar coordinates the line y = mx becomes the pair of rays
θ = tan−1 (m) and θ = tan−1 (m) + π . (The two rays point in opposite directions.) The circle x2 + y 2 = 4
also becomes r = 2.

◦ Thus, the line y = x gives the right boundary θ = π/4 and the line y = −x gives the left boundary
θ = 3π/4. (The other two rays do not bound the region we are interested in.)

◦ So we see that the range for θ is π/4 ≤ θ ≤ 3π/4, and the range for r is 0 ≤ r ≤ 2.
◦ Also, the function is f = x + 2y = r cos θ + 2r sin θ, and the dierential is r dr dθ.
ˆ 3π/4 ˆ 2
◦ Thus, in polar coordinates the integral becomes (r cos θ + 2r sin θ) · r dr dθ, so we get
π/4 0
ˆ 3π/4 ˆ 2 ˆ 3π/4
1 2
(r cos θ + 2r sin θ) · r dr dθ = (cos θ + 2 sin θ) · r3 r=0 dθ
π/4 0 π/4 3
ˆ 2π
8
= (cos θ + 2 sin θ) dθ
0 3
√
8 3π/4 16 2
= (− sin θ + 2 cos θ) θ=π/4 =
 .
3 3
ˆ 1 ˆ √
1−x2 √ 2 2
• Example: Evaluate the integral
√ e x +y dy dx.
−1 − 1−x2

◦ As written the integral is completely intractable, so we will try switching to polar coordinates.
√ √
◦ The region of integration is dened by the inequalities −1 ≤ x ≤ 1, − 1 − x2 ≤ y ≤ 1 − x2 , which we
can recognize as the interior of the unit circle x2 + y 2 ≤ 1 .
◦ 0 ≤ r ≤ 1 and 0 ≤ θ ≤ 2π (since there are
In polar coordinates, the region is no restrictions on θ).
√ 2 2
x +y
◦ Also, the function is f (r, θ) = e = er , and the dierential is r dr dθ.
ˆ 2π ˆ 1
◦ Thus, in polar coordinates the integral becomes er · r dr dθ.
0 0
◦ Now we can evaluate it (using integration by parts to compute the inner integral in r):
ˆ 2π ˆ 1 ˆ 2π ˆ 2π
1
er · r dr dθ = [r er − er ] r=0 dθ = 1 dθ = 2π .
0 0 0 0

• As an application of integration in polar coordinates, we can describe how to evaluate the famous Gaussian
´∞ 2 2
integral I =−∞
e−x dx, which is quite dicult to compute because the function e−x does not have an
elementary antiderivative.

2
◦ This integral is fundamental in statistics, since the function p(x) = e−x arises (after a suitable change of
variables) as the probability density function of the famous Gaussian normal distribution, which describes
the distributions of quantities arising as the sum of independent small variations, such as human heights,
errors in measurements, exam grades, and many other physical phenomena.

14
 ´∞ 2 ´∞ 2
◦ To compute this integral, we note that if I = −∞
e−x dx then I = −∞
e−y dy , and so I2 =
h´ i h´ i ´∞ ´∞
∞ 2 ∞ 2 2
+y 2 )
−∞
e−x dx −∞
e−y dy = −∞ −∞
e−(x dy dx.
◦ 0 ≤ θ ≤ 2π
Now convert to polar coordinates: the region is the entire plane, with integration bounds
2 2 2
and 0 ≤ r < ∞, e−(x +y ) = e−r .
and the function becomes
´ 2π ´ ∞ 2
◦ Thus, in polar coordinates we see I 2 = 0 0 e−r r dr dθ, which we can now evaluate using a substi-
´ 2π 1 −r2 ∞ ´ 2π
2 2
tution u = r to see I =
√ 0
[ 2 e ]|r=0 dθ = 0 12 dθ = π . Therefore, since I > 0, we deduce that
I = π.

3.3.3 Triple Integrals in Cylindrical Coordinates

• Cylindrical coordinates are a simple three-dimensional version of polar coordinates: we merely include the
z -coordinate along with the polar coordinates r and θ.
• Denition: The cylindrical coordinates (r, θ, z) of a point (x, y, z) satisfy x = r cos(θ), y = r sin(θ), z = z for
r≥0 0 ≤ θ ≤ 2π .
and
p
◦ We have r = x2 + y 2 and θ = tan−1 (y/x) (possibly plus π depending on the signs of x and y ), and
obviously z = z .

◦ The interpretations of r and θ are essentially the same as in polar coordinates: r measures the distance
of a point to the z -axis and θ measures the angle (in a horizontal plane) from the positive x-direction.
◦ Cylindrical coordinates are useful in simplifying regions that have a circular symmetry. In particular,
the cylinder x2 + y 2 = a 2 in 3-dimensional rectangular coordinates (over which it is cumbersome to set
up double integrals) becomes the simpler equation r = a in cylindrical.
 
 ∂(x, y, z)   cos θ −r sin θ 0
   

◦ For cylindrical coordinates, we have J =   =  sin θ r cos θ 0  = r.
∂(r, θ, z)   
0 0 1 

◦ Thus the dierential in cylindrical coordinates is dV = r dz dr dθ . We typically set up cylindrical

integrals with this integration order (since typically the z -bounds are the most complicated), although
other orders are possible.
p
• Example: Integrate the function f (x, y, z) = x2 + y 2 on the solid enclosed by the cylinder x2 + y 2 = 4 ,
2 2
bounded above by the paraboloid z = 5 − x − y , and bounded below by z = −1.

◦ The surfaces involved suggest using cylindrical coordinates, since they both have reasonably simple
descriptions in that coordinate system, as does the function f itself. Here is a plot of the region:

◦ In cylindrical coordinates, the cylinder has equation r = 2, the upper paraboloid has equation z = 5 − r2 ,
and the lower plane is z = −1.
∗ 0 ≤ r ≤ 2, the points below the paraboloid satisfy z ≤ 16 − r2 ,
The points inside the cylinder satisfy
and the points above the plane satisfy z ≥ −2.

∗ Since there are no restrictions on θ, putting all of this together indicates that the integration bounds
2
are 0 ≤ θ ≤ 2π , 0 ≤ r ≤ 2, −1 ≤ z ≤ 5 − r .
p
◦ Since x2 + y 2 = r, the function is simply f (r, θ, z) = r, and the cylindrical dierential is r dz dr dθ.

15
 ◦ The integral is therefore equal to

ˆ 2π ˆ 2 ˆ 5−r 2 ˆ 2π ˆ 2 ˆ 2π ˆ 2
 2  5−r2  2
6r − r4 dr dθ =

r · r dz dr dθ = r z z=−1 dr dθ =
0 0 −1 0 0 0 0
ˆ 2π   ˆ 2π
1 2 32
= 2r3 − r5 r=0 dθ = (16 − ) dθ
0 5 0 5
ˆ 2π
48 96π
= dθ = .
0 5 5

ˆ 2 ˆ √
4−x2 ˆ 2 p
• Example: Evaluate the integral √ z x2 + y 2 dz dy dx.
0 0 x2 +y 2

p
◦ 2 2
This is an iterated integral of the function
√ pf (x, y, z) = z x + y over the solid region D dened by the

inequalities 0 ≤ x ≤ 2, 0 ≤ y ≤ 4 − x2 , 2 2
x + y ≤ z ≤ 2.
√
◦ Notice that the projection of D into the xy -planepis the region 0 ≤ x ≤ 2, 0 ≤ y ≤ 4 − x2 , which is
a quarter-disc. This, along with the presence of x2 + y 2 in the z -limit and in the function, strongly
suggest converting to cylindrical coordinates.

◦ In cylindrical coordinates, we can see that the xy -region becomes 0 ≤ r ≤ 2, 0 ≤ θ ≤ π/2. Also, the
range for z becomes r ≤ z ≤ 2.
p
◦ Since x2 + y 2 = r , the function is simply f = zr, and the cylindrical dierential is r dz dr dθ.
◦ The integral is therefore equal to

ˆ π/2 ˆ 2 ˆ 2 ˆ π/2 ˆ  2  ˆ π/2 ˆ 2  

1 2 2 2 2 1 4
zr · r dz dr dθ = z r z=r
dr dθ = 2r − r dr dθ =
0 0 r 0 0 2 0 0 2
ˆ π/2   ˆ π/2
2 3 1 5 2 16 32
= r − r r=0
dθ = ( − ) dθ
0 3 10 0 3 10
ˆ π/2
32 16π
= dθ = .
0 15 15

1
• Example: Integrate the function f (x, y, z) = p on the region underneath z = 9 − x2 − y 2 and above
x2 + y 2
the xy -plane.

◦ We set up this problem in cylindrical coordinates: the paraboloid has equation z = 9 − r2 , so the portion
with z ≥ 0 corresponds to 0 ≤ r ≤ 3. There are no restrictions on θ , so we have 0 ≤ θ ≤ 2π , and also
0 ≤ z ≤ 9 − r2 .
p
◦ Since x2 + y 2 = r, the function is simply f = 1/r, and the cylindrical dierential is r dz dr dθ.
ˆ 2π ˆ 3 ˆ 9−r2
◦ The desired integral is thus 1 dz dr dθ. Now we can evaluate it:
0 0 0

ˆ 2π ˆ 3 ˆ 9−r 2 ˆ 2π ˆ 3 ˆ 2π
1 dz dr dθ = (9 − r2 ) dr dθ = 18 dθ = 36π .
0 0 0 0 0 0

3.3.4 Triple Integrals in Spherical Coordinates

• Spherical coordinates are another 3-dimensional version of polar coordinates:

• Denition: The spherical coordinates (ρ, θ, ϕ) of a point (x, y, z) satisfy x = ρ sin ϕ cos θ, y = ρ sin ϕ sin θ, z =
ρ cos ϕ for ρ ≥ 0, 0 ≤ θ ≤ 2π ,
0 ≤ ϕ ≤ π. and
p √
◦ With r and θ as dened for cylindrical coordinates, we also have ρ = x2 + y 2 + z 2 = r 2 + z 2 ,
r = ρ sin ϕ, and ϕ = tan−1 (r/z).

16
 ◦ The parameter ρ measures the distance from the origin. The angle θ has the same interpretation as
in cylindrical coordinates, while the angle ϕ measures the angle with the positive z -axis (often called
declination from the z -axis): thus, for example, the graph of ϕ = π/2 consists of the points in the
xy -plane, since the vectors joining these points with the origin make a right angle with the x-axis.
◦ Important Notational Remark: Most applied scientic elds that make use of spherical coordinates (e.g.,
chemistry and physics) have their own standard notations for spherical coordinates which dier from the
standard used in most of mathematics.

∗ Specically, the angles ϕ and θ are generally reversed from the above, and r and ρ are often reversed
as well. In addition, the angle ϕ is sometimes measured as an inclination from the xy -plane rather
than a declination from the z -axis, thus giving it a range of [− π2 , π2 ] rather than [0, π].

∗ In mathematics, we set up the angles this way in order to maintain consistency with polar coordinates,
so that θ means the same thing in both contexts.

◦ Spherical coordinates are most useful when integrating over regions with spherical symmetries: the sphere
x2 + y 2 + z 2 = a2 , over which it is typically very dicult to set up triple integrals, becomes the much
simpler ρ = a in spherical coordinates.
p
◦ The cone az = x2 + y 2 with vertex at the origin also has a simple expression in spherical coordinates,
−1
namely as ϕ = tan (a).
 
 ∂(x, y, z)   cos θ sin ϕ ρ cos θ cos ϕ −ρ sin θ sin ϕ 
   
◦ For spherical coordinates, J =    = sin θ sin ϕ ρ sin θ cos ϕ ρ cos θ sin ϕ  = ρ2 sin ϕ.

∂(ρ, ϕ, θ)  
cos ϕ −ρ sin ϕ 0 

◦ Thus the dierential in spherical coordinates is dV = ρ2 sin ϕ dρ dϕ dθ . We typically set up spherical

integrals with this integration order, because typically the ρ bounds are the most complicated, while the
θ bounds are the simplest.

• Example: Integrate the function f (x, y, z) = z 2 over the region 1 ≤ x2 + y 2 + z 2 ≤ 4.

◦ Here is a plot of the region of integration, which is bounded by two spheres:

◦ Given that the region is bounded by the two spheres x2 + y 2 + z 2 = 1 and x2 + y 2 + z 2 = 4, we will
switch to spherical. The rst sphere is ρ = 1 and the second is ρ = 2, and there are no restrictions on
the angles ϕ and θ . Thus, the region of integration is 1 ≤ ρ ≤ 2, 0 ≤ ϕ ≤ π , and 0 ≤ θ ≤ 2π .

◦ In spherical coordinates, the function is f (ρ, θ, ϕ) = z 2 = ρ2 cos2 (ϕ) and the dierential is ρ2 sin(ϕ) dρ dϕ dθ.
ˆ 2π ˆ π ˆ 2
◦ The integral in spherical coordinates is therefore ρ4 cos2 (ϕ) sin(ϕ) dρ dϕ dθ. Now we can
0 0 1
evaluate it:
ˆ 2π ˆ π ˆ 2 ˆ 2π ˆ π
ρ5
 
4 2 2
2
ρ cos (ϕ) sin(ϕ) dρ dϕ dθ = cos (ϕ) sin(ϕ) ρ=1 dϕ dθ
0 0 1 0 0 5
ˆ 2π ˆ π
31
= cos2 (ϕ) sin(ϕ) dϕ dθ
0 0 5
ˆ 2π  
31 1 π
= · − cos3 (ϕ) ϕ=0 dθ
0 5 3
ˆ 2π
31 2 124
= · dθ = π.
0 5 3 15

17
 ˆ 1 ˆ √
1−x2 ˆ √2−x2 −y2 p
• Example: Evaluate the integral √ x2 + y 2 + z 2 dz dy dx.
0 0 x2 +y 2

p
◦ f (x, y, z)
This is an iterated integral of the function = z x2 + y 2 over
p the solid region D dened by the
√ √ p
inequalities 0 ≤ x ≤ 1, − 1 − x ≤ y ≤ 1 − x , x + y ≤ z ≤ 2 − x2 − y 2 .
2 2 2 2

◦ Notice that thisp

region D is the region bounded by the p planes x = 0 and y = 0 that lies below the
hemisphere z=2 − x2 − y 2 and above the cone z = x2 + y 2 . All of pthese surfaces have very simple
descriptions in spherical coordinates, as does the function f (x, y, z) = x2 + y 2 + z 2 .
√
◦ Specically, the hemisphere becomes ρ = 2 and the cone becomes ϕ = π/4. Since we want the region
above the cone, we need 0 ≤ ϕ ≤ π/4, and since we want the region inside the hemisphere we want
√
0 ≤ ρ ≤ 2. Also, since we want the region where x ≥ 0 and y ≥ 0, we want 0 ≤ θ ≤ π/2.
p
◦ Since x2 + y 2 + z 2 = ρ, the function is simply f = ρ, and the dierential is ρ2 sin(ϕ) dρ dϕ dθ.
ˆ π/2 ˆ π/4 ˆ √2
◦ The integral in spherical coordinates is therefore ρ · ρ2 sin(ϕ) dρ dϕ dθ. Now we can
0 0 0
evaluate it:

ˆ π/2 ˆ π/4 ˆ √
2 ˆ π/2 ˆ π/4  √
ρ4

2
 2
ρ · ρ sin(ϕ) dρ dϕ dθ = sin(ϕ) ρ=0 dϕ dθ
0 0 0 0 0 4
ˆ π/2 ˆ π/4
= sin(ϕ) dϕ dθ
0 0
ˆ π/2 π/4
= [− cos(ϕ)] ϕ=0 dθ
0
ˆ π/2
√ ! √ !
2 2 π
= 1− dθ = 1− .
0 2 2 2

z
• Example: Integrate the function f (x, y, z) = p over the lower half of the sphere x2 + y 2 + z 2 = 9.
x2 + y 2

◦ We set up this problem in spherical coordinates: the sphere is ρ = 3, and the lower half corresponds to
π/2 ≤ ϕ ≤ π . Since there are no restrictions on θ, the region of integration is 0 ≤ ρ ≤ 3, π/2 ≤ ϕ ≤ π ,
and 0 ≤ θ ≤ 2π .

ρ cos(ϕ) cos(ϕ)
◦ In spherical coordinates, the function is f (ρ, θ, ϕ) = q = and the dierential is
ρ2 sin2 (ϕ) sin(ϕ)
ρ2 sin(ϕ) dρ dϕ dθ.
ˆ 2π ˆ π ˆ 3
◦ The desired integral is thus ρ2 cos(ϕ) dρ dϕ dθ. Now we can evaluate it:
0 π/2 0

ˆ 2π ˆ π ˆ 3 ˆ 2π ˆ π ˆ 2π
ρ2 cos(ϕ) dρ dϕ dθ = 9 cos(ϕ) dϕ dθ = 9dθ = 18π .
0 π/2 0 0 π/2 0

3.3.5 Additional Examples of Coordinate Changes

• To change a multiple integral into a dierent coordinate system (polar, cylindrical, spherical, or something
more general), follow these steps:

1. Sketch the region of integration.

2. Decide on the new coordinate system to use, and express the region in the new coordinate system.

◦ Usually the choice will be decided by shape of the region of integration, or sometimes the function
being integrated.

18
 ◦ For example, an integral over x2 + y 2 + z 2 ≤ 1 is a good choice for spherical coordinates, since
the region takes the much simpler form 0 ≤ ρ ≤ 1, while an integral over the region x2 + y 2 ≤ 4,
−2 ≤ z ≤ 2 is probably better in cylindrical coordinates.

◦ In general, the best coordinate system is one that turns the region into something simple to describe,
and that simplies the function.

◦ Cylinders x2 +y 2 = a2 and other surfaces of the form z = f (x, y) are easiest to describe in cylindrical
coordinates,
p while spheres x2 + y 2 + z 2 = a2 are easiest in spherical coordinates. Cones of the form
z=α x2 + y 2 have easy descriptions in both systems: as z = αr and as ϕ = tan−1 (α) respectively.
3. Express the function in the new coordinates.

◦ For polar and cylindrical, substitute x = r cos(θ), y = r sin(θ). (And z = z , in cylindrical.)

◦ For spherical, substitute x = ρ cos(θ) sin(ϕ), y = ρ sin(θ) sin(ϕ), z = ρ cos(ϕ).

4. Change the dierential into the new coordinates. For a general transformation, this requires multiplying
∂[old]
by the Jacobian, a determinant whose terms are .
∂[new]
◦ For polar the new dierential is r dr dθ, for cylindrical it is r dr dθ dz , and for spherical, it is
ρ2 sin(ϕ) dρ dϕ dθ.
5. Evaluate the new integral.
p
• Example: Integrate f (x, y, z) = x2 + y 2 over the region inside x2 + y 2 = 1 , below the plane z = x, and
above the plane z = −2 − y .

◦ Here is a plot of the region:

◦ The surface x2 + y 2 = 1 is a cylinder, and the other two bounding curves are functions of z . The function
2 2
also involves x + y ; all of these things indicate that we should switch to cylindrical coordinates.

◦ The inside of the cylinder is given by 0≤r≤1 in cylindrical coordinates.

◦ As we move in the direction of increasing z , we enter the region through the plane z = −2 − y =
−2 − r sin(θ) and exit via the plane z = x = r cos(θ), so the z -bounds are −2 − r sin(θ) ≤ z ≤ r cos(θ).
◦ There are no restrictions on θ so we have the simple bounds 0 ≤ θ ≤ 2π .
p
◦ In cylindrical coordinates, the function is f (x, y) = x2 y2
+ = r, and the dierential is r dz dr dθ.
´ 2π ´ 1 ´ r cos(θ)
◦ Thus, in cylindrical coordinates the integral is
0 0 −2−r sin(θ)
r · r dz dr dθ. Now we can evaluate it:

ˆ 2π ˆ 1 ˆ r cos(θ) ˆ 2π ˆ 1
r · r dz dr dθ = r2 [r cos(θ) − (−2 − r sin(θ))] dr dθ
0 0 −2−r sin(θ) 0 0
ˆ 2π ˆ 1  3
r cos(θ) + 2r2 + r3 sin(θ) dr dθ

=
0 0
ˆ 2π  
1 2 1 4π
= cos(θ) + + sin(θ) dθ = .
0 4 3 4 3

19
 ˆ 1 ˆ 0
2
• Example: Evaluate the integral √ dx dy .
0 − 1−y 2 (1 + x2 + y 2 )2
p
◦ The region is dened by 0≤y≤1 and − 1 − y 2 ≤ x ≤ 0. Upon sketching, we see that it is a quarter-
circle of radius 1:

◦ In theory, it is possible to evaluate the integral as written, but it would be rather unpleasant, so we look
for an alternative.
p
◦ The − 1 − y2 x2 + y 2 terms in the function strongly suggest
term in the limits of integration and the
π
using polar coordinates, as the limits in polar are very simple: ≤ θ ≤ π and 0 ≤ r ≤ 1.
2
◦ To convert the function into polar coordinates, we substitute x = r cos(θ) and y = r sin(θ): then
2 2
= .
(1 + x2 + y 2 )2 (1 + r2 )2
ˆ π ˆ 1
2
◦ Since the polar dierential is r dr dθ, the new integral is r dr dθ.
π/2 0 (1 + r2 )2
◦ Now it is much easier to evaluate: we can substitute u = 1 + r2 with du = 2r dr to obtain
ˆ π ˆ 1 ˆ π ˆ 2 ˆ π ˆ π
2 1  −1  2 1 π
r dr dθ = du dθ = u  dθ = dθ = .
π/2 0 (1 + r2 )2 π/2 1 u2 π/2
u=1
π/2 2 4
p
• Example: Integrate the function f (x, y, z) = 1 on the region inside x2 + y 2 + z 2 = 9, below z= x2 + y 2 ,
and above the xy -plane.

◦ The rst surface x2 + y 2 + z 2 = 9 is a sphere of radius 3, and the second surface is a cone. Spheres and
cones have simple descriptions in spherical coordinates, so we will use spherical coordinates.
p
◦ In spherical coordinates, the equation x2 + y 2 + z 2 = 9 becomes ρ = 3, the equation z = x2 + y 2
π
becomes ρ cos ϕ = ρ sin ϕ which is easily seen to be the same as ϕ = , and the xy -plane has equation
4
π
ϕ= .
2
π π
◦ The bounds of integration in spherical coordinates are therefore 0 ≤ θ ≤ 2π , ≤ ϕ ≤ , 0 ≤ ρ ≤ 3.
4 2
◦ Furthermore, the function is simply 1, and the dierential is ρ2 sin ϕ dρ dϕ dθ.
ˆ 2π ˆ π/2 ˆ 3 ˆ 2π ˆ π/2 ˆ 2π
√
√
2
◦ The integral is therefore ρ2 sin φ dρ dϕ dθ = 9 sin φ dϕ dθ = 9 dθ = 9π 2 .
0 π/4 0 0 π/4 0 2
˜
• Example: Evaluate
R
f (x, y) = x2 + y 2 ,
f (x, y) dA, where and R is the region in the rst quadrant dened
y
by the inequalities 1 ≤ xy ≤ 9, 1 ≤ ≤ 4.
x
◦ The region is as follows:

20
 ◦ We could in principle divide the region into three pieces (horizontally or vertically), but this would be
somewhat cumbersome.
y
◦ Instead, based on the bounds of integration, we will make a change of variables to set s = xy , t = .
x
√
r
s
Solving for x and y in terms of s and t yields x= and y= st.
t
s
◦ The new bounds of integration are 1≤s≤9 and 1 ≤ t ≤ 4, and the new function is x2 + y 2 = + st.
 t 
  1 −1/2 −1/2 1
− s1/2 t−3/2

 ∂x/∂s ∂x/∂t   s t
 
◦ For the new dierential, we compute the Jacobian: J =  = 2 2 =

∂y/∂s ∂y/∂t   1 s−1/2 t1/2 1 1/2 −1/2
 
s t 
2 2
 
     
1 −1/2 −1/2 1 1/2 −1/2 1 1 1 1
s t s t − s−1/2 t1/2 − s1/2 t−3/2 = . Hence the new dierential is dt ds.
2 2 2 2 2t 2t
ˆ 9ˆ 4  1
s
◦ The integral in st-coordinates is thus + st · dt ds. Now we evaluate:
0 1 t 2t
ˆ 9ˆ 4  1 ˆ 9ˆ 4 ˆ 9
s s  −2  s  −1  4
+ st · dt ds = t + 1 dt ds = −t + t t=1 ds
1 1 t 2t 1 1 2 1 2
ˆ 9   ˆ 9
s 3 15
= + 3 ds = s ds = 75 .
1 2 4 1 8

ˆ ˆ √ 2ˆ √ 2 2 2 4−x 4−x −y
• Example: Evaluate the integral (x2 + y 2 + z 2 )−3/4 dz dy dx.
0 0 0
√ p
◦ The region is dened by the inequalities 0 ≤ x ≤ 2, 0 ≤ y ≤ 4 − x2 , and 0 ≤ z ≤ 4 − x2 − y 2 , which
2 2 2
collectively describe the region inside the sphere x + y + z = 4 inside the rst octant (i.e., with x, y, z
nonnegative).

◦ Furthermore, the integrand has a very simple description in spherical coordinates, since it is simply
(ρ2 )−3/4 = ρ−3/2 .
◦ Therefore, we can evaluate this integral by converting to spherical coordinates. The region is dened by
0 ≤ θ ≤ π/2, 0 ≤ ϕ ≤ π/2, and 0 ≤ ρ ≤ 2, while the function is ρ−5 .
◦ Thus, in spherical coordinates, the integral is
ˆ π/2 ˆ π/2 ˆ 2 ˆ π/2 ˆ π/2 ˆ 2
ρ−3/2 · ρ2 sin ϕ dρ dϕ dθ = ρ1/2 sin ϕ dρ dϕ dθ
0 0 0 0 0 0
ˆ π/2 ˆ π/2 √
ˆ π/2 √ √
4 2 4 2
= 3 sin ϕ dϕ dθ = 3 = 2π 2/3 .
0 0 0

ˆ 1 ˆ 0 ˆ √x2 +y2
z
• Example: Evaluate the integral
√
pdz dy dx.
−1 − 1−x2 0 + y2 x2
√ p
◦ The region is dened by the inequalities −1 ≤ x ≤ 1, − 1 − x2 ≤ y ≤ 0, and0≤z≤ x2 + y 2 . The x
and y inequalities describe the lower half of the interior of the unit circle, while the z inequalities have
simple descriptions in cylindrical coordinates, as does the integrand.

◦ Therefore, we can evaluate this integral by converting to cylindrical coordinates. The region is dened
z
π ≤ θ ≤ 2π , 0 ≤ r ≤ 1, and 0 ≤ z ≤ r, while the function is .
by
r
´ 2π ´ 1 ´ r z ´ 2π ´ 1 1 2 ´ 2π
◦ Thus, the integral is π 0 0 r dz dr dθ = π 0 2 r dr dθ = π 16 dθ = π/6 .
r

3.4 Applications of Multiple Integration

• There are a number of applications of multiple integration to computing quantities like areas and volumes,
the average value of a function on a region, and the mass of a lamina (a solid plate with variable density) in
the plane or of a solid having variable density in space.

21
3.4.1 Areas, Volumes, and Average Values

• One straightforward but still very useful application of multiple integration is to computing areas of regions
in the plane, and volumes of regions in space.

˜ ˝
◦ The central ideas are that Area(R) = R
1 dA and Volume(D) = D
1 dV .
◦ Thus, if we can describe a region in a form that lends itself to integration, we can calculate the region's
area (or volume).

◦ A closely related problem is to calculate the average value of a function on a region. To calculate this,
we simply integrate the function over the region, and then divide by the region's area or volume.

• Example: Find the area of the region bounded by the curves x = y2 − 1 and y = 1 − x. Then nd the average
value of y on this region.

◦ A sketch of the region indicates its shape is as below:

◦ We can set up the area integral with either integration order, but if we use vertical slices we would need
to divide the region into two pieces, since the top curve changes from the upper half of the parabola to
the line in the middle of the region of integration.

◦ Therefore, we use horizontal slices with the integration order dx dy : we see that the range for y is
−2 ≤ y ≤ 1, and then the corresponding range for x is y2 − 1 ≤ x ≤ 1 − y.
´ 1 ´ 1−y
◦ The area is then given by
−2 y 2 −1
1 dx dy , which we can evaluate:

ˆ 1 ˆ 1−y ˆ 1 1−y ˆ 1
(1 − y) − (y 2 − 1) dy
 
1 dx dy = [x]  dy =

−2 y 2 −1 −2 x=y 2 −1 −2
ˆ 1  
 2
 1 2 1 3 1 9
= 2−y−y dy = 2y − y − y  = .
−2 2 3 y=−2 2

1 ´ 1 ´ 1−y
◦ Now to calculate the average value of y, we evaluate · −2 y 2 −1
y dx dy :
Area(R)

ˆ 1 ˆ 1−y ˆ 1 1−y ˆ 1
1 2 2
y(1 − y) − y(y 2 − 1) dy
 
· y dx dy = [xy]  2 dy =

Area(R) −2 y 2 −1 9 −2 x=y −1 9 −2
ˆ 1  
2 2 2 1 3 1 4 1 1
2y − y 2 − y 3 dy =
 
= y − y − y  = − .
9 −2 9 3 4 y=−2 2

• Example: Find
pthe volume of the ice cream cone solid that lies inside the sphere x2 + y 2 + z 2 = 1 and above
the cone z= x2 + y 2 . Then nd the average value of z on this region.

◦ Here is the plot of the solid:

◦ Since we are dealing with a sphere, we set up the problem in spherical coordinates.

22
 ◦ The sphere has equation ρ = 1, and the interior is 0 ≤ ρ ≤ 1.
p p
◦ For the cone, we know x2 + y 2 = ρ sin ϕ and z = ρ cos ϕ, so the cone z = x2 + y 2 in spherical is the
π
same as sin ϕ = cos ϕ, or ϕ = . Since ϕ measures declination from the z -axis, the region above the
4
π
cone is therefore 0 ≤ ϕ ≤ .
4
´ 2π ´ π/4 ´ 1
◦ Hence the volume of this region is given by 0 0 0
1 · ρ2 sin ϕ dρ dϕ dθ. We evaluate:

ˆ 2π ˆ π/4 ˆ 1 ˆ 2π ˆ π/4 ˆ 2π
√ ! √
1 1 2 π(2 − 2)
1 · ρ2 sin ϕ dρ dϕ dθ = sin ϕ dϕ dθ = 1− dθ = .
0 0 0 0 0 3 0 3 2 3

1 ´ 2π ´ π/4 ´ 1
◦ Now to calculate the average value of z = ρ cos φ, we evaluate · 0 0 0
ρ cos ϕ·ρ2 sin ϕ dρ dϕ dθ:
Volume(D)

ˆ 2π ˆ π/4 ˆ 1 ˆ 2π ˆ π/4 ˆ 1
1 3 1 3
· ρ cos ϕ · ρ2 sin ϕ dρ dϕ dθ = √ ρ sin(2ϕ) dρ dϕ dθ
Volume(D) 0 0 0 π(2 − 2) 0 0 0 2
ˆ 2π ˆ π/4
3 1
= √ sin(2ϕ) dϕ dθ
π(2 − 2) 0 0 8
ˆ 2π
3 1 3
= √ dθ = √ .
π(2 − 2) 0 16 8(2 − 2)

3.4.2 Masses, Centroids, and Moments of a Plate and Solid

• The center of mass of a physical object is its balancing point, where, if the object is supported only at that
point, gravity will not cause it to tip over. The center of mass is also called the centroid of an object.

• If we are given the shape and density of an object in 2 or 3 dimensions, then we can nd its mass and the
location of its center of mass. We can also calculate its moments about an axis or a point, which show
up in physics calculations involving angular momentum. Each of these quantities is the integral of a simple
function, weighted by the density of the object.

• Center of Mass and Moment Formulas (2 Dimensions): We are given a 2-dimensional plate of variable density
δ(x, y) on a region R.
˜
◦ The total mass M is given by M= δ(x, y)dA.
˜
R
◦ The x-moment My is given by My = x δ(x, y)dA.
˜R
◦ The y -moment Mx is given by Mx = R
y δ(x, y)dA.
∗ The notation for these moments is rather confusing in this instance: the phrase  x-moment means
the moment of the object about the x-axis, which is given by integrating the distance to the x-axis
(namely, y) rather than the x-coordinate.
 
My M x
◦ The center of mass (x̄, ȳ) has coordinates , .
M M
∗ Equivalently, each coordinate of the centroid is given by the average value of that coordinate over
the plate  hence the notation x̄, which means the average value of x.

• Center of Mass and Moment Formulas (3 Dimensions): We are given a 3-dimensional plate of variable density
δ(x, y, z) on a region D.
˝
◦ The total mass M is given by M= δ(x, y, z)dV .
D
˝
◦ The moment Myz about the yz -plane is given by Myz = x δ(x, y, z)dV .
˝D
◦ The moment Mxz about the xz -plane is given by Mxz = y δ(x, y, z)dV .
˝D
◦ The moment Mxy about the xy -plane is given by Mxy = D
z δ(x, y, z)dV .

23
  
Myz Mxz Mxy
◦ The center of mass (x̄, ȳ, z̄) has coordinates , , .
M M M

• Example: Find the mass and center of mass of a plate on the triangle 0 ≤ x ≤ 1, 0 ≤ y ≤ x whose density is
δ(x, y) = y 2 .

˜
´1 1 3 1 ´1´x
◦ The total mass M is given by M= 0 3 R
δ(x, y)dA =
x dx = .
0 0
y 2 dy dx =
12
1 ˜ ´1´x ´1 1 12
◦ The x-coordinate of the centroid is x δ(x, y)dA = 12 0 0 xy 2 dy dx = 12 0 x4 dx = .
M R 3 15
1 ˜ ´1´x ´1 1 12
◦ The y -coordinate of the centroid is R
y δ(x, y)dA = 12 0 0 y 3 dy dx = 12 0 x4 dx = .
M 4 20
 
4 3
◦ Therefore, the center of mass has coordinates (x̄, ȳ) = , .
5 5

• Example: Find the mass and center of mass of a solid in the shape of the cube bounded 0 ≤ x ≤ 1, 0 ≤ y ≤ 1,
0 ≤ z ≤ 1, whose density is δ(x, y) = 1 + xyz .

◦ The total mass M is given by

˚ ˆ 1 ˆ 1 ˆ 1 ˆ 1 ˆ 1 ˆ 1
1 1 9
M= δ(x, y, z)dV = (1 + xyz) dz dy dx = (1 + xy) dy dx = (1 + x) dx = .
D 0 0 0 0 0 2 0 4 8

◦ The x-coordinate of the centroid is given by

˚ ˆ 1 ˆ 1 ˆ 1 ˆ 1 ˆ 1 ˆ 1
1 8 2 8 1 8 1 21
x δ(x, y, z)dV = (x+x yz) dz dy dx = (x+ x2 y) dy dx = (x+ x2 ) dx = .
M D 9 0 0 0 9 0 0 2 9 0 4 32

◦ The y -coordinate of the centroid is given by

˚ ˆ 1 ˆ 1 ˆ 1 ˆ 1 ˆ 1 ˆ 1
1 8 2 8 1 8 1 1 21
y δ(x, y, z)dV = (y+xy z) dz dy dx = (y+ xy 2 ) dy dx = ( + x) dx = .
M D 9 0 0 0 9 0 0 2 9 0 2 6 32

◦ The x-coordinate of the centroid is given by

˚ ˆ 1 ˆ 1 ˆ 1 ˆ 1 ˆ 1 ˆ 1
1 8 8 1 1 8 1 1 21
x δ(x, y, z)dV = (z+xyz 2 ) dz dy dx = ( + xy) dy dx = ( + x) dx = .
M D 9 0 0 0 9 0 0 2 3 9 0 2 6 32

 
21 21 21
◦ Therefore, the center of mass has coordinates (x̄, ȳ, z̄) = , , .
32 32 32

Well, you're at the end of my handout. Hope it was helpful.

Copyright notice: This material is copyright Evan Dummit, 2012-2021. You may not reproduce or distribute this
material without my express permission.

24