Multiple Integrals (Inversion of Order / Change of Coordinates)

6.1 Double Integrals

Just as the definite integral of a positive function of one variable represents the area of the region
between the graph of the function and the 𝑥-axis, the double integral of a positive function of two variables
represents the volume of the region between the surface defined by the function (on the three-dimensional
Cartesian plane where 𝑧 = 𝑓(𝑥, 𝑦) and the plane which contains its domain. If there are more variables, a
multiple integral will yield hypervolumes of multidimensional functions.

Example 1: Double integral over rectangular region.

Compute the integral
𝑑

∬ 𝑥𝑦 2 𝑑𝐴
𝐷

where D is the rectangle defined by 0 ≤ 𝑥 ≤ 2 and 0 ≤ 𝑦 ≤ 1.

x
1 2

Solution:
We will compute the double integral as the iterated integral,

𝑑
1 2
∬ 𝑥𝑦 2 𝑑𝐴 = ∫ (∫ 𝑥𝑦 2 𝑑𝑥) 𝑑𝑦
0 0
𝐷

1
𝑥2 2 𝑥 = 2
=∫ ( 𝑦 | ) 𝑑𝑦
0 2
𝑥=0
1
22 2 02 2
=∫ ( 𝑦 − 𝑦 ) 𝑑𝑦
0 2 2

1
= ∫ 2𝑦 2 𝑑𝑦
0

Note that in the first line above, we wrote the limits as 𝑥 = 2 and 𝑥 = 0 so it is unambiguous that 𝑥
is the variable we just integrated.
To finish, we need to compute the integral with respect to 𝑦, which is simple. Since 𝑥 is gone, it's
just a regular one-variable integral. We calculate that our double integral is

𝑑
1
∬ 𝑥𝑦 2 𝑑𝐴 = ∫ 2𝑦 2 𝑑𝑦
0
𝐷
 2𝑦 3 1
= |
3
0

2(1)3 2(0)3
= −
3 3

2
=
3

To double check our answer, we can compute the integral in the other direction, integrating first
with respect to 𝑦 and then with respect to 𝑥. The only trick is to remember that when integrating with
respect to 𝑦, we must think of 𝑥 as a constant. Since for any constant c, the integral of 𝑐𝑦 2 is 𝑐𝑦 3 /3, we
calculate

𝑑
2 1
∬ 𝑥𝑦 2 𝑑𝐴 = ∫ (∫ 𝑥𝑦 2 𝑑𝑦 ) 𝑑𝑥
0 0
𝐷

𝑑 𝑦=1
2
𝑦3
∬ 𝑥𝑦 2 𝑑𝐴 = ∫ (𝑥 | ) 𝑑𝑥
0 3
𝐷 𝑦=0

2
𝑥
=∫ 𝑑𝑥
0 3

𝑥2 2
= |
6
0

(2)2
= −0
6

2
=
3

As it must, this iterated integral gives the same answer.

Example 2: Double integral over triangular region.

Compute the integral
𝑑

∬ 𝑥𝑦 2 𝑑𝐴
𝐷

where D is the above triangle given below.

1
y = x/2

(x, x/2)

x
(x, 0) 1 2

Solution:
 We integrate over the region defined by 0 ≤ 𝑥 ≤ 2 and 0 ≤ 𝑦 ≤ 𝑥/2. The fact that the range of 𝑦
depends on 𝑥 means this region is not a rectangle.
For the triangle defined by 0 ≤ 𝑥 ≤ 2 and 0 ≤ 𝑦 ≤ 𝑥/2, the limits of y depend on 𝑥. For a given
value of 𝑥, 𝑦 ranges from 0 to 𝑥/2, as illustrated above by the vertical dashed line from (𝑥,0) to (𝑥, 𝑥/2).
In a double integral, the outer limits must be constant, but the inner limits can depend on the
outer variable. This means, we must put 𝑦 as the inner integration variables, as was done in the second
way of computing Example 1. The only difference from Example 1 is that the upper limit of 𝑦 is 𝑥/2. The
double integral is

𝑑
2 𝑥/2
∬ 𝑥𝑦 2 𝑑𝐴 = ∫ (∫ 𝑥𝑦 2 𝑑𝑦) 𝑑𝑥
0 0
𝐷

2 𝑦 = 𝑥/2
𝑦3
= ∫ (𝑥 | ) 𝑑𝑥
0 3
𝑦=0
2
𝑥4
=∫ 𝑑𝑥
0 24

2
𝑥4
=∫ 𝑑𝑥
0 24

𝑥5 2
= |
120
0

25
= −0
120

4
=
15

6.1.1 Inversion of Order

Given a double integral
𝑑

∬ 𝑓(𝑥, 𝑦)𝑑𝐴
𝐷

of a function 𝑓(𝑥, 𝑦) over a region D. We can integrate with respect to 𝑥 first, or we can integrate with
respect to 𝑦 first. If we integrate with respect to 𝑥 first, we will obtain an integral that looks something like

∬ 𝑓(𝑥, 𝑦)𝑑𝐴 = ∫ (∫ 𝑓( 𝑥, 𝑦) 𝑑𝑥) 𝑑𝑦

𝐷

and if you integrate with respect to 𝑦 first, you will obtain an integral that looks something like

∬ 𝑓(𝑥, 𝑦)𝑑𝐴 = ∫ (∫ 𝑓( 𝑥, 𝑦) 𝑑𝑦) 𝑑𝑥.

𝐷

We often say that the first integral is in 𝑑𝑥𝑑𝑦 order and the second integral is in 𝑑𝑦𝑑𝑥 order.
One difficult part of computing double integrals is determining the limits of integration, i.e.,
determining what to put in place of the boxes □ in the above integrals. In some situations, we know the
limits of integration the 𝑑𝑥𝑑𝑦 order and need to determine the limits of integration for the equivalent
integral in 𝑑𝑦𝑑𝑥 order (or vice versa). The process of switching between 𝑑𝑥𝑑𝑦 order and 𝑑𝑦𝑑𝑥 order in
double integrals is called changing the order of integration (or reversing the order of integration).
Changing the order of integration is slightly tricky because it’s hard to write down a specific
algorithm for the procedure. The easiest way to accomplish the task is through drawing a picture of the
region D. From the picture, you can determine the corners and edges of the region D, which is what you
need to write down the limits of integration.

Example 1: Change order of integration region with exponential, y first

Change the order of integration in the following integral

1 𝑒𝑦
∫ ∫ 𝑓(𝑥, 𝑦)𝑑𝑥 𝑑𝑦.
0 1

Solution:
Since the focus of this example is the limits of integration, we won't specify the function 𝑓(𝑥, 𝑦).
The procedure doesn't depend on the identity of 𝑓.
In the original integral, the integration order is 𝑑𝑥𝑑𝑦 . This integration order corresponds to
integrating first with respect to 𝑥 (i.e., summing along rows in the picture below), and afterwards
integrating with respect to y (i.e., summing up the values for each row). Our task is to change the
integration to be 𝑑𝑦𝑑𝑥, which means integrating first with respect to 𝑦.
We begin by transforming the limits of integration into the domain D. The limits of the outer 𝑑𝑦
integral mean that 0 ≤ 𝑦 ≤ 1, and the limits on the inner 𝑑𝑥 integral mean that for each value of 𝑦 the
range of 𝑥 is 1 ≤ 𝑥 ≤ 𝑒 𝑦 . The region D is shown in the following figure.

The maximum range of y over the region is from 0 to 1, as indicated by the gray bar to the left of
the figure. The horizontal hashing within the figure indicates the range of 𝑥 for each value of 𝑦, beginning
at the left edge 𝑥 = 1 (blue line) and ending at the right curve edge 𝑥 = 𝑒 𝑦 (red curve).

We have also labeled all the corners of the region. The upper-right corner is the intersection of the line
𝑦 = 1 with the curve 𝑥 = 𝑒 𝑦 . Therefore, the value of 𝑥 at this corner must be 𝑒 = 𝑒 1 = 𝑒, and the point is
(e,1).

To change order of integration, we need to write an integral with order 𝑑𝑦𝑑𝑥. This means that 𝑥 is the
variable of the outer integral. Its limits must be constant and correspond to the total range of 𝑥 over the
region D. The total range of 𝑥 is 1 ≤ 𝑥 ≤ 𝑒, as indicated by the gray bar below the region in the following
figure.
Since 𝑦 will be the variable for the inner integration, we need to integrate with respect to 𝑦 first. The vertical
hashing indicates how, for each value of 𝑥, we will integrate from the lower boundary (red curve) to the
upper boundary (purple line). These two boundaries determine the range of 𝑦. Since we can rewrite the
equation 𝑥 = 𝑒 𝑦 for the red curve as 𝑦 = log 𝑥 , the range of 𝑦 is log 𝑥 ≤ 𝑦 ≤ 1 . (The function log 𝑥
indicates the natural logarithm, which sometimes we write as ln 𝑥.)

In summary, the region D can be described not only by

0≤𝑦≤1
1 ≤ 𝑥 ≤ 𝑒𝑦

as it was for the original 𝑑𝑥𝑑𝑦 integral, but also by

1≤𝑥≤𝑒
log 𝑥 ≤ 𝑦 ≤ 1,

which is the description we need for the new 𝑑𝑦𝑑𝑥 integration order. This latter pair of inequalites
determine the bounds for integral.

1 𝑒𝑦
We conclude that the integral ∫0 ∫1 𝑓(𝑥. 𝑦)𝑑𝑦𝑑𝑥 with integration order reversed is

𝑒 1
∫ ∫ 𝑓(𝑥, 𝑦)𝑑𝑦 𝑑𝑥.
1 log 𝑥

Example 2: Change order of integration triangular region

Change the order of integration in the following integral

1 1
2
∫ ∫ 𝑒 𝑦 𝑑𝑦 𝑑𝑥.
0 𝑥

Solution:
Sometimes you need to change the order of integration to get a tractable integral. For example, if you
tried to evaluate the integral of Example 2 directly, you would run into trouble. There is no antiderivative
2
of 𝑒 𝑦 , so you get stuck trying to compute the integral with respect to 𝑦. But, if we change the order of
integration, then we can integrate with respect to 𝑥 first, which is doable. And, it turns out that the integral
with respect to y also becomes possible after we finish integrating with respect to 𝑥.
According to the limits of integration of the given integral, the region of integration is

0≤𝑥≤1
𝑥≤𝑦≤1

Since we can also describe the region by

0≤𝑦≤1
 0 ≤ 𝑥 ≤ 𝑦,

the integral with the order changed is

1 1 1 𝑦
2 2
∫ ∫ 𝑒 𝑦 𝑑𝑦𝑑𝑥 = ∫ ∫ 𝑒 𝑦 𝑑𝑥𝑑𝑦
0 𝑥 0 0

With this new 𝑑𝑥𝑑𝑦 order, we integrate first with respect to 𝑥

1 𝑦 1 𝑥=𝑦 1
2 2 2
∫ ∫ 𝑒𝑦 𝑑𝑥𝑑𝑦 = ∫ 𝑥𝑒 𝑦 | 𝑑𝑦 = ∫ 𝑦𝑒 𝑦 𝑑𝑦.
0 0 0 𝑥=0 0

Since the integration with respect to 𝑥 gave us an extra factor of 𝑦, we can compute the integral
with respect to 𝑦 by using a 𝑢-substitution, 𝑢 = 𝑦 2, so 𝑑𝑢 = 2𝑦𝑑𝑦. With this substitution, 𝑢 ranges from 0
to 1, and we calculate the integral as

1 𝑦 1
2 2
∫ ∫ 𝑒 𝑦 𝑑𝑥𝑑𝑦 = ∫ 𝑦𝑒 𝑦 𝑑𝑦
0 0 0

1
1 𝑢
=∫ 𝑒 𝑑𝑢
0 2

1 𝑢 1
= 𝑒 |
2
0
1
= (𝑒 − 1)
2

6.1.2 Change of Coordinates

We had to compute the double integral

∬ 𝑔(𝑥, 𝑦)𝑑𝐴
𝐷

where 𝑔(𝑥, 𝑦) = 𝑥 2 + 𝑦 2 and D is the disk of radius 6 centered at the origin.

In terms of the standard rectangular (or Cartesian) coordinates 𝑥 and 𝑦, the disk is given by

−6 ≤ 𝑥 ≤ 6

−√36 − 𝑥 2 ≤ 𝑦 ≤ √36 − 𝑥 2

We could start to calculate the integral in terms of 𝑥 and 𝑦 as

 1
6 √36−𝑥 2
∬ 𝑔(𝑥, 𝑦)𝑑𝐴 = ∫ ∫ (𝑥 2 + 𝑦 2 )𝑑𝑦𝑑𝑥
−6 −√36−𝑥 2
𝐷

which is a mess.
It turns out that this integral would be a lot easier if we could change variables to polar coordinates.
In polar coordinates, the disk is the region we'll call D∗ defined by 0 ≤ 𝑟 ≤ 6 and 0 ≤ 𝜃 ≤ 2𝜋. Hence the
region of integration is simpler to describe using polar coordinates.
Moreover, the integrand 𝑥 2 + 𝑦 2 is simple in polar coordinates because 𝑥 2 + 𝑦 2 = 𝑟 2 . Using polar
coordinates, our lives will be a lot easier because it seems that all we need to do is integrate 𝑟 2 over the
region D∗ defined by 0 ≤ 𝑟 ≤ 6 and 0 ≤ 𝜃 ≤ 2𝜋.
Unfortunately, it's not quite that easy. We need to account for one more consequence of changing
variables, which is how changing variables changes area. You may recall that 𝑑𝐴 stands for the area of a
little bit of the region D. In rectangular coordinates, we replaced 𝑑𝐴 by 𝑑𝑥𝑑𝑦 (or 𝑑𝑦𝑑𝑥). We need to
determine what 𝑑𝐴 becomes when we change variables. As you will see, in polar coordinates, 𝑑𝐴 does
not becomes 𝑑𝑟𝑑𝜃.
The relationship between rectangular (𝑥, 𝑦) and polar (𝑟, 𝜃) coordinates is given by 𝑥 = 𝑟 cos 𝜃, 𝑦 =
𝑟 sin 𝜃. To see how area gets changed, let's write the change of variables as the function

(𝑥, 𝑦) = 𝑇(𝑟, 𝜃) = (𝑟 cos 𝜃, 𝑟 sin 𝜃).

The function 𝑇(𝑟, 𝜃) gives rectangular coordinates in terms of polar coordinates. The
transformation T gives the perspective of polar coordinates as a mapping from the polar plane to the
Cartesian plane.
If we use (𝑥, 𝑦) = 𝑇(𝑟, 𝜃) to change variables, we can instead integrate the function 𝑔(𝑇(𝑟, 𝜃)) =
2
𝑟 over the region D∗.

1 1

∬ 𝑔(𝑥, 𝑦)𝑑𝐴 = ∬(𝑥 2 + 𝑦 2 )𝑑𝐴

𝐷 𝐷

2𝜋 6
= ∫ ∫ 𝑟 2 𝑟 𝑑𝑟 𝑑𝜃
0 0

2𝜋 6
= ∫ ∫ 𝑟 3 𝑑𝑟 𝑑𝜃
0 0

2𝜋
𝑟4 6
= ∫ ( | ) 𝑑𝜃
0 4
0
2𝜋
= ∫ 324 𝑑𝜃
0

2𝜋
= 324 𝜃 |
0

= 648𝜋

6.2 Triple Integrals

Let 𝑓(𝑥, 𝑦, 𝑧) be the density of a three-dimensional solid 𝑊 at the point (𝑥, 𝑦, 𝑧 inside the solid. We
want to define the triple integral of 𝑓 over 𝑊 to be the total mass of 𝑊.
 If the solid 𝑊 is a cube defined by𝑎 ≤ 𝑥 ≤ 𝑏, 𝑐 ≤ 𝑦 ≤ 𝑑, and𝑝 ≤ 𝑧 ≤ 𝑞, then we can easily write
the triple integral as an iterated integral. We could first integrate 𝑥 from a to b, then integrate 𝑦 from c to
d, and finally integrate 𝑧 from p to q,

𝑠
𝑝 𝑑 𝑏
∭ 𝑓𝑑𝑉 = ∫ (∫ (∫ 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑧) 𝑑𝑦) 𝑑𝑥
𝑞 𝑐 𝑎
𝑊
Just as with double integrals, other orders of integration are possible. We could, for example, first
integrate with respect to 𝑧, then integrate with respect to 𝑥, and lastly integrate with respect to 𝑦,

𝑠
𝑑 𝑏 𝑞
∭ 𝑓𝑑𝑉 = ∫ (∫ (∫ 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑧) 𝑑𝑦) 𝑑𝑥
𝑐 𝑎 𝑝
𝑊

The integration always proceeds from the inside to the outside. The integration order coincides
with the order that the differentials (e.g.𝑑𝑥) appear, so we can specify the integration order by a list of the
differentials. The previous integral is in the order𝑑𝑧 𝑑𝑥 𝑑𝑦. With this understanding, the parentheses
become optional, and we usually omit them, writing the previous integral as

𝑠
𝑑 𝑏 𝑝
∭ 𝑓𝑑𝑉 = ∫ ∫ ∫ 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑧 𝑑𝑦 𝑑𝑥
𝑐 𝑎 𝑞
𝑊

The iterated integral is simple when the solid W is a rectangular solid (like a cube, but where all edges
aren't necessarily the same length). For more complicated shapes, finding the limits of integration can be
tricky.

As a first step, just remember these rules,

1. The outer limits have to be constant. They cannot depend on any of the variables.
2. The middle limits can depend on the variable from the outer integral only. They cannot depend on
the variable from the inner integral.
3. The inner limits can depend on the variable from the outer integral and the variable from the middle
integral.

Example:
Evaluate

3 4 0
∫ ∫ ∫ (4𝑥 2 𝑦 − 𝑧 3 )𝑑𝑧 𝑑𝑦 𝑑𝑥 .
2 1 1

Solution:
All we need to do is integrate following the given order and recall that just like with double integrals
we start with the “inside” integral and work our way out.

3 4 0 3 4
𝑧4 0
∫ ∫ ∫ (4𝑥 2 𝑦 − 𝑧 3 )𝑑𝑧 𝑑𝑥 𝑑𝑦 = ∫ ∫ (4𝑥 2 𝑦𝑧 − ) | 𝑑𝑦 𝑑𝑥
2 −1 1 2 −1 4
1
3 4
1
= ∫ ∫ ( − 4𝑥 2 𝑦) 𝑑𝑦 𝑑𝑥
2 1 4

3
𝑦 4
= ∫ ( − 2𝑥 2 𝑦 2 ) | 𝑑𝑥
2 4 −1
 3
5
= ∫ ( − 30𝑥 2 ) 𝑑𝑥
2 4

5 3
= ( 𝑥 − 10𝑥 3 ) |
4
2
755
=−
4