Definition of double integration

In this note we will work abstractly, defining double integration as a sum, technically a limit
of Riemann sums. It is best to learn this first before getting into the details of computing
the value of a double integral –we will learn how to do that next.
Definition of double integrals
Suppose we have a region in the plane R and a function f (x, y), Then the double integral
� �
f (x, y) dA
R

is defined as follows.
Divide the region R into small pieces, numbered from 1 to n. Let ΔAi be the area of the ith
piece and also pick a point (xi , yi ) in that piece. The figure shows a region R divided into
small pieces and shows the ith piece with its area, and choice of a point in the little region.
y

• (xi , yi )
area = ΔAi

Now form the sum

n
�
f (xi , yi ) ΔAi ,
i=1
and then, finally
� � n
�
f (x, y) dA = lim f (xi , yi ) ΔAi .
R ΔA→0
i=1
Here, the limit is taken by letting the number of pieces go to infinity and the area of each
piece go to 0. There are technical requirements that the limit exist and be independent of
the specific limiting process. In 18.02 these requirements are always met. (Later you might
study fractals and other strange objects which don’t satisfy them.)

Interpretations of the double integral

As you saw in single variable calculus, these sums can be used to compute areas, volumes,
mass, work, moment of inertia and many other quantities. Again, before focusung on some
computational issues we will show you how easy it is to setup a double integral to compute
certain quantities.

Example 1: Set up a double integral to compute the area of a region R in the plane.
Answer: Use the figure above for visualization. The area of R is just the sum of the areas
of the pieces. That is, � �
area = dA.
R

Example 2: Set up a double integral to compute the volume of the solid below the graph
of z = f (x, y) = 2 − .5(x + y) and above the unit square in the xy-plane.
Answer: The figure below shows the graph of f (x, y) above the unit square in the plane.
The unit square is labeled R. We also show a little piece of the R and the solid region above
that piece. We are imagining we’ve divided R into n small pieces and this is the ith one. It
contains the point (xi , yi ) and has area ΔAi .
z

1
y
• (xi , yi ), area = ΔAi

1 R

The small solid region is almost a box and so its volume, ΔVi , is roughly its base times its
height, i.e.,
ΔVi ≈ ΔAi × f (xi , yi ).
The total volume is the sum of the volumes of all the small pieces, i.e.,
n
� n
�
volume = ΔVi ≈ ΔAi × f (xi , yi ).
i=1 i=1

In the limit this becomes an exact integral for volume

� �
volume = f (x, y) dA.
R
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010 ��

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 Limits in Iterated Integrals

For most students, the trickiest part of evaluating multiple integrals by iteration is to
put in the limits of integration. Fortunately, a fairly uniform procedure is available which
works in any coordinate system. You must always begin by sketching the region; in what
follows we’ll assume you’ve done this.

x 2+y 2= 1
1. Double integrals in rectangular coordinates.
Let’s illustrate this procedure on the first case that’s usually
taken up: double integrals in rectangular coordinates. Suppose we x+y= 1
want to evaluate over the region R pictured the integral
��
f (x, y) dy dx , R = region between x2 + y 2 = 1 and x+y =1 ;
R

we are integrating first with respect to y. Then to put in the limits,

1. Hold x fixed, and let y increase (since we are integrating with respect to y).
As the point (x, y) moves, it traces out a vertical line.
2. Integrate from the y-value where this vertical line enters the region R, to
the y-value where it leaves R.
3. Then let x increase, integrating from the lowest x-value for which the vertical
line intersects R, to the highest such x-value.
Carrying out this program for√the region R pictured, the vertical line enters R where
y = 1 − x, and leaves where y = 1 − x2 .
The vertical lines which intersect R are those between x = 0 and y=1-x 2
x = 1. Thus we get for the limits:
y=1-x
�� � 1 � √
1−x2
f (x, y) dy dx = f (x, y) dy dx.
R 0 1−x

��
To calculate the double integral, integrating in the reverse order R
f (x, y) dx dy,
1. Hold y fixed, let x increase (since we are integrating first with respect to x).
This traces out a horizontal line.
2. Integrate from the x-value where the horizontal line enters R to the x-value
where it leaves.
3. Choose the y-limits to include all of the horizontal lines which intersect R.
Following this prescription with our integral we get: x=1-y 2
x=1-y
�� � 1 � √1−y2
f (x, y) dx dy = f (x, y) dx dy.
R 0 1−y

1
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 Limits for double integrals
��
1. Evaluate R x dA, where R is the finite region bounded by the axes and 2y + x = 2.
Answer:
First we sketch the region.
y

1
R
x
2

Next, we find limits of integration. By using vertical stripes we get limits

Inner: y goes from 0 to 1 − x/2; outer: x goes from 0 to 2.
Thus the integral is
� 2 � 1−x/2
x dy dx
0 0
Finally, we compute the inner, then the outer integrals.
1−x/2 x2
Inner: xy|0 =x− .
2
2 3
�2
x x � 2
Outer: − �� = .
2 6 0 3
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 Changing the order of integration

1. Evaluate � π/2 � π/2

sin y
I= dy dx
0 x y
by changing the order of integration.
Answer:
The given limits are (inner) y from x to π/2; (outer) x from 0 to π/2.
We use these to sketch the region of integration.
y
π/2
y=x

x
π/2

The given limits have inner variable y. To reverse the order of integration we use horizontal
stripes. The limits in this order are
(inner) x from 0 to y; (outer) y from 0 to π/2.
So the integral becomes
� π/2 � y
sin y
I= dx dy
0 0 y
We compute the inner, then the outer integrals.
sin y ��y
�
π/2
Inner: x� = sin y. Outer: − cos y|0 = 1.
y 0
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 Integration in polar coordinates

Polar Coordinates
Polar coordinates are a different way of describing points in the plane. The polar coordinates
(r, θ) are related to the usual rectangular coordinates (x, y) by by

x = r cos θ, y = r sin θ

The figure below shows the standard polar triangle relating x, y, r and θ.
y�

r y
θ �x
x

Because cos and sin are periodic, different (r, θ) can represent the same point in the plane.
The table below shows this for a few points.
(x, y) (1, 0) (0, 1) (2, 0) (1,
√ 1) (−1,
√ 1) √ −1)
(−1, (0, 0)
(r, θ) (1, 0) (1, π/2) (2, 0) (√2, π/4) ( 2, 3π/4) ( √ 2, 5π/4) (0, π/2)
(r, θ) (1, 2π) ( 2, 9π/4) (− 2, π/4) (0, −7.2)
(r, θ) (1, 4π)
In fact, you can add any multiple of 2π to θ and the polar coordinates will still represent
the same point.
Because θ is not uniquely specified it’s a little trickier going from rectangular to polar
coordinates. The equations are easily deduced from the standard polar triangle.
�
r = x2 + y 2 , ”θ = tan−1 (y/x)”.

We use quotes around tan−1 to indicate it is not a single valued function.

The area element in polar coordinates

In polar coordinates the area element is given by

dA = r dr dθ.

The geometric justification for this is shown in by the following figure.

y�

ΔA
r rΔθ
• Δr
Δθ r
�x

The small curvy rectangle has sides Δr and rΔθ, thus its area satisfies ΔA ≈ (Δr)(r Δθ).
As usual, in the limit this becomes dA = r dr dθ.
Double integrals in polar coordinates
The area element is one piece of a double integral, the other piece is the limits of integration
which describe the region being integrated over.
Finding procedure for finding the limits in �polar
� coordinates is the same as for rectangular
coordinates. Suppose we want to evaluate dr dθ over the region R shown.
R
y� y�

R
�
x+y =1 R r = 1/(cos θ + sin θ) •r = 1
�x �x
(The integrand, including the r that usually goes with r dr dθ, is irrelevant here, and there­
fore omitted.)
As usual, we integrate first with respect to r. Therefore, we
1. Hold θ fixed, and let r increase (since we are integrating with respect to r). As the point
moves, it traces out a ray going out from the origin.
2. Integrate from the r-value where the ray enters R to the r-value where it leaves. This
gives the limits on r.
3. Integrate from the lowest value of θ for which the corresponding ray intersects R to the
highest value of θ.
To follow this procedure, we need the equation of the line in polar coordinates. We have
1
x+y =1 → r cos θ + r sin θ = 1, or r= .
cos θ + sin θ
This is the r value where the ray enters the region; it leaves where r = 1. The rays which
intersect R lie between θ = 0 and θ = π/2. Thus the double iterated integral in polar
coordinates has the limits � � π/2 1
dr dθ.
0 1/(cos θ+sin θ)

Example: Find the mass of the region R shown if it has density δ(x, y) = xy (in units of
mass/unit area)
In polar coordinates: δ = r2 cos θ sin θ.
Limits of integration: (radial lines sweep out R):
y�
inner (fix θ): 0 < r < 2, outer: 0 < θ < π/3.
� � � π/3 � 2
⇒ Mass M = δ(x, y) dA = r2 cos θ sin θ r dθ dr 2
R θ=0 r=0
� 2 2
r4
�
r3 cos θ sin θ dr = π/3
�
Inner: cos θ sin θ�� = 4 cos θ sin θ R �x
0 4 0
� π/3
�π/3 3
Outer: M = 4 cos θ sin θ dθ = 2 sin2 θ�0 = .
0 2
 �2�x 1
Example: Let I = 1 0 (x2 +y 2 )3/2 dy dx. Compute I using polar coordinates.
Answer: Here are the steps we take.
Draw the region.
Find limits in polar coordinates: y�
Inner (fix θ): sec θ < r < 2 sec θ, outer: 0 < θ < π/4.
� π/4 � 2 sec θ y=x
1 �
⇒ I= r dr dθ. y = 2 or r = 2 sec θ
3
θ=0 r=sec θ r r
Compute the integral: π/4
� 2 sec θ � �x
1 1 �� 1 1 2
Inner: 2
dr = − � 2 sec θ = cos θ.
sec θ r r sec θ 2
� π/4 �π/4 √
1 1 � 2
Outer: I = cos θ dθ = sin θ�� = .
0 2 2 0 4

Example: Find the volume of the region above the xy-plane and below the graph of
z = 1 − x2 − y 2 .
You should draw a picture of this.
In polar coordinates we have z = 1 − r2 and we want the volume under the graph and
above the inside �of the
� unit disk.
2π 1
⇒ volume V = (1 − r2 ) rdr dθ.
0 0
� 1
1 1 1
Inner integral: (1 − r2 ) rdr = − = .
0 2 4 4
� 2π
1 π
Outer integral: V = dθ = .
0 4 2

Gallery of polar graphs (r = f (θ))

A point P is on the graph if any representation of P satisfies the equation.

Examples:
y�
y� y�
y� � 2 �
� 2
�x �x �x
π/3 � x 2

Circle centered on 0: Vertical line x = 2 ⇔ Horizontal line y = 2 ⇔

Ray: θ = π/3 r=2 r = 2 sec θ. r = 2/ sin θ.
Example: Show the graph of r = 2a cos θ is a circle of radius a centered at (a, 0).
Some simple algebra gives r2 = 2ar cos θ = 2ax ⇒ x2 + y 2 = 2ax ⇒ (x − a)2 + y 2 = a2 .
This is a circle or radius a centered at (a, 0).
Note: we can determine from the graph that the range of theta is −π/2 ≤ θ ≤ π/2.
y� y
�

• �x •a
a

�x

r = 2a cos θ r = 2a sin θ
−π/2 ≤ θ ≤ π/2. 0 ≤ θ ≤ π.

Warning: We can use negative values of r for plotting. You should never use it in
integration. In integration it is better to make use of symmetry and only integrate over
regions where r is positive.
Here are a few more curves.
y�
y�

�x �x

Cardiod: r = a(1 + cos θ) Limaçon: r = a(1 + b cos θ) (b > 1)

y�
y�

�x
�x

Lemniscate: r2 = 2a2 cos 2θ Four leaved rose: r = a sin 2θ

MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010 ��

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 Mass and average value

Center of Mass
Example 1: For two equal masses, the center of mass is at the midpoint between them.
m1 = 1 m2 = 1
• •
xcm • xcm = x1 +x2 .
2

x1 x2
Example 2: For unequal masses the center of mass is a weighted average of their positions.
m1 = 2 m2 = 1
• x• • xcm = 2x13+x2 .
cm
x1 x2
�
In general, xcm = weighted average of position = �mi xi .
mi
For a continuous density, δ(x), on the segment [a, b] (units of density are mass/unit length)
the sums become integrals. We will skip running through the logic of this since we are
about to show it for two dimensions.
� b
1
� b δ(x)
M= δ(x) dx, xcm = xδ(x) dx. a
a M a b

In 2 dimensions we label the center of mass as (xcm , ycm ) and we have the following formulas
� � � � � �
1 1
M= δ(x, y) dA, xcm = xδ(x, y) dA, ycm = yδ(x, y) dA.
R M R M R

These formulas are easy to justify using our usual method for building integrals.
In this case, we divide our region into little pieces and we sum up the contributions of each
piece using an integral. To keep the figure below uncluttered we only show one piece and
we don’t bother to label it as the ith . In the end we will go directly to the integral, by
thinking if it as a sum.
y�

(x, y), area = dA = dx dy

• dy
dx

�x
The little piece shown has mass δ(x, y) dA and the total mass is just the sum the pieces.
That is, it’s just the integral � �
M= δ dA.
R
Likewise the x and y coordinates of the center of mass are just the weighted average of the
x and y coordinates of each of the pieces. So, we get the formulas given above.
Example: Suppose the unit square has density δ = xy; Find its mass and center of mass.
Answer:
� � � 1� 1
1
M= δ dA = xy dx dy = . �
R 0 0 4
1
� �
1
� 1� 1 1
xcm = xδ dA = x2 y dy dx.
M M 0 0
2
This is easily computed as xcm = .
3 �
By symmetry of the region and the density δ, we also have O 1
2
ycm = .
3
Average Value
We can think of center of mass as the average position of the mass. That is, it’s the average
of position with respect to mass. We can also take averages of functions with respect to
other things. For instance, the average value of f (x, y) with respect to area on a region R is
� �
1
f (x, y) dA.
area R R

In general, if we simply say the average value of a function, it means average value with
respect to area (or later, when we do triple integrals it can mean with respect to volume).

Example: What’s the average distance of a point in a unit square from the center?
Answer: We center the square on the origin. Notice each side has length 2a and the area
of the square is 4a2 . So,
� �
1 �
average distance = 2 x2 + y 2 dx dy.
4a square

By symmetry the integral is 8 times the integral of the triangular region R shown. We
actually compute the integral in polar coordinates.

y�
� �
8
average = r r dr dθ
4a2 R
� π/4 � a sec θ a
2
= r2 dr dθ �
a2 0 0 r = a sec θ
R
2
� π/4 3 3
a sec θ �x
= dθ a
2
a 0 3
a √ √
= ( 2 + ln( 2 + 1)).
3

This last integral was computed using integration by parts as

� �
3 sec θ dθ + sec θ tan θ ln(sec θ + tan θ) + sec θ tan θ
sec θ dθ = = .
2 2
Note: the center of mass is the average value of x and y with respect to mass.
The geometric center has coordinates given by the average value of x and y with respect to
area, i.e., the center of mass when δ = 1.
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010 ��

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 Moment of Inertia

Moment of inertia
We will leave it to your physics class to really explain what moment of inertia means. Very
briefly it measures an object’s resistance (inertia) to a change in its rotational motion. It is
analogous to the way mass measure the resistance to changes in the object’s linear motion.
Because it has to do with rotational motion the moment of inertia is always measured about
a reference line, which is thought of as the axis of rotation.
For a point mass, m, the moment of inertia about the line is

I = m d2 ,

where d is the distance from the mass to the line. (The letter I is a standard notation for
moment of inertia.)
If we have a distributed mass we compute the moment of inertia by summing the contribu­
tions of each of its parts. If the mass has a continuous distribution, this sum is, of course,
an integral.

Example 1: Suppose the unit square, R, has density δ = xy.

Find its moment of inertia about the y-axis.
yy (x, y)
1
x •
dy
dx
-x
O 1
Answer: The distance from the small piece of the square (shown in the figure) to the y-axis
is x. If the piece has mass dm then its moment of inertia is

dI = x2 dm = x2 δ(x, y) dA = x3 y dx dy.

We took a shortcut here: we went straight to the notation for infinitesimal pieces, dI, dA
and used equalities. rather than using more formal notation ΔI, ΔA, using approximations
and then taking limits.
In the equation above, we used the notation dI to indicate it is just a small bit of moment
of inertia. We also used that the mass of a piece is density times area. Now it’s a simple
matter to sum up all the bits of moment of inertia using an integral
� � � 1� 1
1
I= dI = x3 y dy dx = .
R 0 0 8

(Note: this integral is so easy to compute that we don’t give the details.)

1
Example 2: For the same square as in example 1, find the polar moment of inertia.
yOy
1
• (x, y)
r
/- x
O 1
Answer: The polar moment of inertia of a planar region is the moment of inertia about
the origin (the axis of rotation is the z-axis). Finding this is exactly the same as in example
1, except the distance to the axis is now the polar distance r. We get,

dI = r2 dm = (x2 + y 2 )δ(x, y) dA = (x3 y + xy 3 ) dx dy.

Summing, using an integral gives

� Z
Z � �
Z 1�
Z 1
1
I= dI = x3 y + xy 3 dy dx = .
R 0 0 4

2
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 Changing Variables in Multiple Integrals

1. Changing variables.
Double integrals in x, y coordinates which are taken over circular regions, or have inte­
grands involving the combination x2 + y 2 , are often better done in polar coordinates:
�� ��
(1) f (x, y) dA = g(r, θ) r drdθ .
R R

This involves introducing the new variables r and θ, together with the equations relating
them to x, y in both the forward and backward directions:
�
(2) r= x2 + y 2 , θ = tan−1 (y/x); x = r cos θ, y = r sin θ .

Changing the integral to polar coordinates then requires three steps:

A. Changing the integrand f (x, y) to g(r, θ), by using (2);
B. Supplying the area element in the r, θ system: dA = r dr dθ ;
C. Using the region R to determine the limits of integration in the r, θ system.

In the same way, double integrals involving other types of regions or integrands can
sometimes be simplified by changing the coordinate system from x, y to one better adapted
to the region or integrand. Let’s call the new coordinates u and v; then there will be
equations introducing the new coordinates, going in both directions:

(3) u = u(x, y), v = v(x, y); x = x(u, v), y = y(u, v)

(often one will only get or use the equations in one of these directions). To change the
integral to u, v-coordinates, we then have to carry out the three steps A, B, C above. A
first step is to picture the new coordinate system; for this we use the same idea as for polar
coordinates, namely, we consider the grid formed by the level curves of the new coordinate
functions:
u=u0
(4) u(x, y) = u0 , v(x, y) = v0 . u=u1
u=u 2

Once we have this, algebraic and geometric intuition will usually handle
steps A and C, but for B we will need a formula: it uses a determinant v=v0
called the Jacobian, whose notation and definition are v=v1
v=v2
� �
∂(x, y) �x xv ��
(5) = �� u .
∂(u, v) yu yv �

Using it, the formula for the area element in the u, v-system is
� �
� ∂(x, y) �
(6) dA = �
� � du dv ,
∂(u, v) �

1
2 CHANGING VARIABLES IN MULTIPLE INTEGRALS

so the change of variable formula is

� �
� ∂(x, y) �
�� ��
(7) f (x, y) dx dy = g(u, v) � ∂(u, v) � du dv ,
� �
R R

where g(u, v) is obtained from f (x, y) by substitution, using the equations (3).
We will derive the formula (5) for the new area element in the next section; for now let’s
check that it works for polar coordinates.

Example 1. Verify (1) using the general formulas (5) and (6).
Solution. Using (2), we calculate:
� � � �
∂(x, y) � x xθ �
� = � cos θ −r sin θ ��
= �� r
�
= r(cos2 θ + sin2 θ) = r ,
∂(r, θ) yr yθ � � sin θ r cos θ �

so that dA = r dr dθ, according to (5) and (6); note that we can omit the absolute value,
since by convention, in integration problems we always assume r ≥ 0, as is implied already
by the equations (2).
We now work an example illustrating why the general formula is needed and how it is
used; it illustrates step C also — putting in the new limits of integration. 1

�� � �2
x−y -1
Example 2. Evaluate dx dy over the region R pictured. 1
R x+y+2
Solution. This would be a painful integral to work out in rectangular coordinates. -1
But the region is bounded by the lines

(8) x + y = ±1, x − y = ±1

and the integrand also contains the combinations x − y and x + y. These powerfully suggest
that the integral will be simplified by the change of variable (we give it also in the inverse
direction, by solving the first pair of equations for x and y):
u+v u−v
(9) u = x + y, v = x − y; x= , y= .
2 2
We will also need the new area element; using (5) and (9) above. we get
� �
∂(x, y) � 1/2 1/2 �� 1
(10) = �� = − ;
∂(u, v) 1/2 −1/2 � 2
note that it is the second pair of equations in (9) that were used, not the ones introducing
u and v. Thus the new area element is (this time we do need the absolute value sign in (6))
1
(11) dA = du dv .
2

We now combine steps A and B to get the new double integral; substituting into the
integrand by using the first pair of equations in (9), we get
�� � �2 �� � �2
x−y v 1
(12) dx dy = du dv .
R x+y+2 R u+2 2
 CHANGING VARIABLES IN MULTIPLE INTEGRALS 3

In uv-coordinates, the boundaries (8) of the region are simply u = ±1, v = ±1, so the
integral (12) becomes
�� � �2 1 1 � �2
v 1 v 1
� �
du dv = du dv
R u+2 2 −1 −1 u+2 2

We have
�u=1 �1
v2 v2 v3 2
inner integral = − = ; outer integral = = .
2(u + 2) u=−1 3 9 −1 9
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 Changing Variables in Multiple Integrals
y
rd θ
2. The area element. dA
dr
In polar coordinates, we found the formula dA = r dr dθ for the area element by dθ
drawing the grid curves r = r0 and θ = θ0 for the r, θ-system, and determining (see x
the picture) the infinitesimal area of one of the little elements of the grid.
For general u, v-coordinates, we do the same thing. The grid curves (4) divide up the
plane into small regions ΔA bounded by these contour curves. If the contour curves are close
together, they will be approximately parallel, so that the grid element will be approximately
a small parallelogram, and y
v=v 0 + Δv
(13) ΔA ≈ area of parallelogram PQRS = |P Q × P R| S
v=v 0
In the uv-system, the points P, Q, R have the coordinates R ΔA
Q u=u 0 + Δ u
P : (u0 , v0 ), Q : (u0 + Δu, v0 ), R : (u0 , v0 + Δv) ; P u=u 0
to use the cross-product however in (13), we need PQ and PR in i j - coordinates. x
Consider PQ first; we have
(14) P Q = Δx i + Δy j ,
where Δx and Δy are the changes in x and y as you hold v = v0 and change u0 to u0 + Δu.
According to the definition of partial derivative,
� � � �
∂x ∂y
Δx ≈ Δu, Δy ≈ Δu;
∂u 0 ∂u 0
so that by (14),
� � � �
∂x ∂y
(15) PQ ≈ Δu i + Δu j .
∂u 0 ∂u 0
In the same way, since in moving from P to R we hold u fixed and increase v0 by Δv,
� � � �
∂x ∂y
(16) PR ≈ Δv i + Δv j .
∂v 0 ∂v 0
We now use (13); since the vectors are in the xy-plane, P Q × P R has only a k -component,
and we calculate from (15) and (16) that
� �
� xu Δu yu Δu �
k -component of P Q × P R ≈ � � �
xv Δv yv Δv �0
� �
� x u xv �
(17) = �� � ΔuΔv ,
y u yv � 0
where we have first taken the transpose of the determinant (which doesn’t change its value),
and then factored the Δu and Δv out of the two columns. Finally, taking the absolute
value, we get from (13) and (17), and the definition (5) of Jacobian,
� �
� ∂(x, y) �
ΔA ≈ � � � ΔuΔv ;
∂(u, v) � 0
passing to the limit as Δu, Δv → 0 and dropping the subscript 0 (so that P becomes any
point in the plane), we get the desired formula for the area element,
� �
� ∂(x, y) �
dA = � � � du dv .
∂(u, v) �
1
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 Changing Variables in Multiple Integrals

3. Examples and comments; putting in limits.

If we write the change of variable formula as

∂(x, y)
ZZ ZZ
(18) f (x, y) dx dy = g(u, v) du dv ,
R R ∂(u, v)
where
∂(x, y) xu xv
(19) = , g(u, v) = f (x(u, v), y(u, v)),
∂(u, v) yu yv

it looks as if the essential equations we need are the inverse equations:

(20) x = x(u, v), y = y(u, v)

rather than the direct equations we are usually given:

(21) u = u(x, y), v = v(x, y) .

If it is awkward to get (20) by solving (21) simultaneously for x and y in terms of u and v,
sometimes one can avoid having to do this by using the following relation (whose proof is
an application of the chain rule, and left for the Exercises):

∂(x, y) ∂(u, v)
(22) = 1
∂(u, v) ∂(x, y)

The right-hand Jacobian is easy to calculate if you know u(x, y) and v(x, y); then the left­
hand one — the one needed in (19) — will be its reciprocal. Unfortunately, it will be in
terms of x and y instead of u and v, so (20) still ought to be needed, but sometimes one
gets lucky. The next example illustrates. 2 2 1 y=x/ 2
x -y =

y 2
ZZ
R x2-y = 4
Example 3. Evaluate dx dy, where R is the region pictured, having
R x
as boundaries the curves x2 − y 2 = 1, x2 − y 2 = 4, y = 0, y = x/2 .
Solution. Since the boundaries of the region are contour curves of x2 − y 2 and y/x ,
and the integrand is y/x, this suggests making the change of variable

y
(23) u = x2 − y 2 , v = .
x

We will try to get through without solving these backwards for x, y in terms of u, v. Since
changing the integrand to the u, v variables will give no trouble, the question is whether we
can get the Jacobian in terms of u and v easily. It all works out, using (22):

∂(u, v) 2x −2y ∂(x, y) 1

= = 2 − 2y 2 /x2 = 2 − 2v 2 ; so = ,
∂(x, y) −y/x2 1/x ∂(u, v) 2(1 − v 2 )
1
2 CHANGING VARIABLES IN MULTIPLE INTEGRALS

according to (22). We use now (18), put in the limits, and evaluate; note that the answer is
positive, as it should be, since the integrand is positive.

y v
ZZ ZZ
dx dy = 2
du dv
R x R 2(1 − v )
Z 1/2 Z 4
v
= du dv
0 1 2(1 − v2 )
1/2
3 2 3 3
= − ln(1 − v ) = − ln .
4 0 4 4

Putting in the limits

In the examples worked out so far, we had no trouble finding the limits of integration,
since the region R was bounded by contour curves of u and v, which meant that the limits
were constants.
If the region is not bounded by contour curves, maybe you should use a different change of
variables, but if this isn’t possible, you’ll have to figure out the uv-equations of the boundary
curves. The two examples below illustrate.
Z 1 Z x
Example 4. Let u = x + y, v = x − y; change dy dx to an iterated integral
0 0
du dv.
∂(x, y)
Solution. Using (19) and (22), we calculate = −1/2, so the Jacobian factor
∂(u.v)
in the area element will be 1/2.
1
To put in the new limits, we sketch the region of integration, as shown at the
right. The diagonal boundary is the contour curve v = 0; the horizontal and vertical
boundaries are not contour curves — what are their uv-equations? There are two
ways to answer this; the first is more widely applicable, but requires a separate
1
calculation for each boundary curve.

Method 1 Eliminate x and y from the three simultaneous equations u = u(x, y), v = v(x, y),
and the xy-equation of the boundary curve. For the x-axis and x = 1, this gives
 
 u=x+y
  u=x+y
 
u=1+y
v =x−y ⇒ u = v; v =x−y ⇒ ⇒ u + v = 2.
  v =1−y
y=0 x=1
 

Method 2 Solve for x and y in terms of u, v; then substitute x = x(u, v), y = y(u, v) into
the xy-equation of the curve.
Using this method, we get x = 12 (u+v), y = 12 (u−v); substituting into the xy-equations:

1 1
y=0 ⇒ (u − v) = 0 ⇒ u = v; x=1 ⇒ (u + v) = 1 ⇒ u + v = 2.
2 2
RR
To supply the limits for the integration order du dv, we
 CHANGING VARIABLES IN MULTIPLE INTEGRALS 3

1. first hold v fixed, let u increase; this gives us the dashed lines shown; v= 0
2. integrate with respect to u from the u-value where a dashed line enters 1
R (namely, u = v), to the u-value where it leaves (namely, u = 2 − v). v=v 0
3. integrate with respect to v from the lowest v-values for which the u=2-v
dashed lines intersect the region R (namely, v = 0), to the highest such v-
value (namely, v = 1). v= 1
Z 1 Z 2−v u=v
1 1
Therefore the integral is du dv .
0 v 2
(As a check, evaluate it, and confirm that its value is the area of R. Then try setting up
the iterated integral in the order dv du; you’ll have to break it into two parts.)

2 2
Example 5. Using the changeZ of
Z coordinates u = x − y , v = y/x of Example
dxdy
3, supply limits and integrand for 2
, where R is the infinite region in the first
R x
quadrant under y = 1/x and to the right of x − y 2 = 1.
2

Solution. We have to change the integrand, supply the Jacobian factor, and put in the
right limits.
To change the integrand, we want to express x2 in terms of u and v; this suggests
eliminating y from the u, v equations; we get
u
u = x2 − y 2 , y = vx ⇒ u = x2 − v 2 x 2 ⇒ x2 = .
1 − v2
1
From Example 3, we know that the Jacobian factor is ; since in the region R we
2(1 − v 2 )
have by inspection 0 ≤ v < 1, the Jacobian factor is always positive and we don’t need the
absolute value sign. So by (18) our integral becomes

dx dxy 1 − v2 du dv
ZZ ZZ ZZ
2
= 2
du dv =
R x R 2u(1 − v ) R 2u

Finally, we have to put in the limits. The x-axis and the left-hand boundary curve
x2 − y 2 = 1 are respectively the contour curves v = 0 and u = 1; our problem is the upper
boundary curve xy = 1. To change this to u − v coordinates, we follow Method 1:
 2 2
 u=x −y
 
u = x2 − 1/x2 1
y = vx ⇒ 2
⇒ u= −v .
 v = 1/x v 2 1
x2-y =
xy = 1

u=1
The form of this upper limit suggests that we should integrate first with P
v=a u=1/v-v
respect to u. Therefore we hold v fixed, and let u increase; this gives the xy= 1
dashed ray shown in the picture; we integrate from where it enters R at
1 O
u = 1 to where it leaves, at u = − v.
v
The rays we use are those intersecting R: they start from the lowest ray, corresponding
to v = 0, and go to the ray v = a, where a is the slope of OP. Thus our integral is
a 1/v−v
du dv
Z Z
.
0 1 2u
4 CHANGING VARIABLES IN MULTIPLE INTEGRALS

To complete the work, we should determine a explicitly. This can be done by solving
xy = 1 and x2 − y 2 = 1 simultaneously to find the coordinates of P . A more elegant
approach is to add y = ax (representing the line OP) to the list of equations, and solve all
three simultaneously for the slope a. We substitute y = ax into the other two equations,
and get  2 √
ax = 1 2 −1 + 5
⇒ a = 1 − a ⇒ a = ,
x2 (1 − a2 ) = 1 2
by the quadratic formula.
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 D

Limits in Iterated Integrals

3. Triple integrals in rectangular and cylindrical coordinates.

���
You do these the same way, basically. To supply limits for D
dz dy dx over R
the region D, we integrate first with respect to z. Therefore we
1. Hold x and y fixed, and let z increase. This gives us a vertical line.
2. Integrate from the z-value where the vertical line enters the region D to the
z-value where it leaves D.
3. Supply the remaining limits (in either xy-coordinates or polar coordinates)
so that you include all vertical lines which intersect D. This means that you will
be integrating the remaining double integral over the region R in the xy-plane
which D projects onto. z=4-x2-y 2

For example, if D is the region lying between the two paraboloids

z = x2 + y 2 z = 4 − x2 − y 2 ,
z=x2 +y2
we get by following steps 1 and 2,
��� �� � 4−x2 −y 2
dz dy dx = dz dA
D R x2 +y 2

where R is the projection of D onto the xy-plane. To finish the job, we have to determine
what this projection is. From the picture, what we should determine is the xy-curve over
which the two surfaces intersect. We find this curve by eliminating z from the two equations,
getting

x2 + y 2 = 4 − x2 − y 2 , which implies
2 2
x +y = 2.

Thus the
√ xy-curve bounding R is the circle in the xy-plane with center at the origin and
radius 2 .
This makes it natural to finish the integral in polar coordinates. We get
��� � 2π � √
2 � 4−x2 −y 2
dz dy dx = dz r dr dθ ;
D 0 0 x2 +y 2

the limits on z will be replaced by r2 and 4 − r2 when the integration is carried out.

1
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 Limits in Spherical Coordinates
Definition of spherical coordinates
ρ = distance to origin, ρ ≥ 0
φ = angle to z-axis, 0 ≤ φ ≤ π
θ = usual θ = angle of projection to xy-plane with x-axis, 0 ≤ θ ≤ 2π
Easy trigonometry gives:

z = ρ cos φ
x = ρ sin φ cos θ
y = ρ sin φ sin θ.

The equations for x and y are most easily deduced by noticing that for r from polar coor-
dinates we have
r = ρ sin φ.
This implies
x = r cos θ = ρ sin φ cos θ, and y = r sin θ = ρ sin φ sin θ.

Going the other way:

p
ρ = z 2 + y 2 + z 2 φ = cos−1 (z/ρ) θ = tan−1 (y/x).

Example: (x, y, z) = (1, 0, 0) ⇒ ρ = 1, φ = π/2, θ = 0

(x, y, z) = (0, 1, 0) ⇒ ρ = 1, φ = π/2, θ = π/2
(x, y, z) = (0, 0, 1) ⇒ ρ = 1, φ = 0, θ –can be anything

The volume element in spherical coordinates

dV = ρ2 sin φ dρ dφ dθ.

The figure at right shows how we get this. The volume of the curved
box is

∆V ≈ ∆ρ · ρ∆φ · ρ sin φ∆θ = ρ2 sin φ ∆ρ ∆φ ∆θ.

Finding limits in spherical coordinates

We use the same procedure asRforR Rrectangular and cylindrical coordinates. To calculate the
limits for an iterated integral D dρ dφ dθ over a region D in 3-space, we are integrating
first with respect to ρ. Therefore we q

1. Hold φ and θ fixed, and let ρ increase. This gives us a ray going out from the origin.
2. Integrate from the ρ-value where the ray enters D to the ρ-value where the ray leaves e
D. This gives the limits on ρ.

1
3. Hold θ fixed and let φ increase. This gives a family of rays, that form a sort
of fan. Integrate over those φ-values for which the rays intersect the region D.
e
4. Finally, supply limits on θ so as to include all of the fans which intersect the
region D.
z P
For example, suppose we start with the circle in the yz-plane of radius 1 and q l
center at (1, 0), rotate it about the z-axis, and take D to be that part of the q y
resulting solid lying in the first octant. O 1 A

First of all, we have to determine the equation of the surface formed by the
rotated circle. In the yz-plane, the two coordinates ρ and φ are indicated. To
see the relation between them when P is on the circle, we see that also angle
OAP = φ, since both the angle φ and OAP are complements of the same angle,
AOP . From the right triangle, this shows the relation is ρ = 2 sin φ.
As the circle is rotated around the z-axis, the relationship stays the same, so ρ = 2 sin φ is
the equation of the whole surface.
To determine the limits of integration, when φ and θ are fixed, the corresponding ray enters
the region where ρ = 0 and leaves where ρ = 2 sin φ.
As φ increases, with θ fixed, it is the rays between φ = 0 and φ = π/2 that intersect D,
since we are only considering the portion of the surface lying in the first octant (and thus
above the xy-plane).
Again, since we only want the part in the first octant, we only use θ values from 0 to π/2.
So the iterated integral is
Z π/2 Z π/2 Z 2 sin φ
dρ dφ dθ.
0 0 0

2
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 Changing Variables in Multiple Integrals

4. Changing coordinates in triple integrals

Here the coordinate change will involve three functions

u = u(x, y, z), v = v(x, y, z) w = w(x, y, z)

but the general principles remain the same. The new coordinates u, v, and w give a three-
dimensional grid, made up of the three families of contour surfaces of u, v, and w. Limits
are put in by the kind of reasoning we used for double integrals. What we still need is the
formula for the new volume element dV .
To get the volume of the little six-sided region ΔV of space bounded by three pairs of
these contour surfaces, we note that nearby contour surfaces are approximately parallel,
so that ΔV is approximately a parallelepiped, whose volume is (up to sign) the 3 × 3
determinant whose rows are the vectors forming the three edges of ΔV meeting at a corner.
These vectors are calculated as in section 2; after passing to the limit we get
� �
� ∂(x, y, z) �
(24) dV = � � � du dv dw ,
∂(u, v, w) �

where the key factor is the Jacobian

� �
� xu xv xw ��
∂(x, y, z) �
(25) = �� yu yv yw �� .
∂(u, v, w) � zu zv zw �

As an example, you can verify that this gives the correct volume element for the change
from rectangular to spherical coordinates:

(26) x = ρ sin φ cos θ, y = ρ sin φ sin θ, z = ρ cos φ;

while this is a good exercise, it will make you realize why most people prefer to derive the
volume element in spherical coordinates by geometric reasoning.

1
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 Integrals in Spherical Coordinates

1. Find the volume of a sphere of radius a.

Answer: From the problems on limits in spherical coordinates (Session 76), we have
limits: inner ρ: 0 to a –radial segments
middle φ: 0 to π –fan of rays.
outer θ: 0 to 2π –volume.
2π π a
V = dV = ρ2 sin φ dρ dφ dθ
D 0 0 0
a
ρ3 a3
Inner: sin φ = sin φ
3 0 3
π
a3 2
Middle: − cos φ = a3
3 0 3
4 3
Outer: 3 πa –as it should be.
fan of rays

2. Find the centroid of the region bounded by the sphere ρ = a and the
cone φ = α.
Answer: In Session 76 we computed the limits:
ρ: 0 to a, φ: 0 to α, θ: 0 to 2π.
By symmetry, xcm = ycm = 0.
1 1 2π π a
zcm = z dV = ρ cos φ · ρ2 sin φ dρ dφ dθ
V D V 0 0 0
1 2π π a 3
= ρ cos φ sin φ dρ dφ dθ.
V 0 0 0
1 πa4 sin2 α
Inner, middle and outer integrals are easy to compute: zcm = · .
V 4
2π π a
2 3
V = dV = ρ2 sin φ dρ dφ dθ =
πa (1 − cos α).
D 0 0 0 3
a4 sin2 α π 3 3a sin2 α 3
⇒ zcm = · 3
= · = a(1 + cos α).
4 2πa (1 − cos α) 8 1 − cos α 8
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 Gravitational Attraction

We use triple integration to calculate the gravitational attraction that a solid body V of
mass M exerts on a unit point mass placed at the origin.
If the solid V is also a point mass, then according to Newton’s law of gravitation, the
force it exerts is given by

GM V
(1) F = r,
|R |2
R
where R is the position vector from the origin 0 to the point V , and the
unit vector r = R/|R| is its direction. O

If however the solid body V is not a point mass, we have to use integration. We concen­
trate on finding just the k component of the gravitational attraction — all our examples
will have the solid body V placed symmetrically so that its pull is all in the k direction
anyway.
To calculate this force, we divide up the solid V into small pieces having volume ΔV and
mass Δm. If the density function is δ(x, y, z), we have for the piece containing the point
(x, y, z)

(2) Δm ≈ δ(x, y, z)ΔV,

Thinking of this small piece as being essentially a point mass at (x, y, z), the force ΔF it
exerts on the unit mass at the origin is given by (1), and its k component ΔFz is therefore

Δm z
ΔFz = G r· k ,
|R|2 Δm
φ
ρ
which in spherical coordinates becomes, using (2), and the picture,

cos φ δ ΔV
ΔFz = G 2
δ ΔV = G 2 cos φ .
ρ ρ

If we sum all the contributions to the force from each of the mass elements Δm and pass to
the limit, we get for the k -component of the gravitational force

cos φ
���
(3) Fz = G δ dV .
V ρ2

If the integral is in spherical coordinates, then dV = ρ2 sin φ dρ dφ dθ, and the integral
becomes
���
(4) Fz = G δ cos φ sin φ dρ dφ dθ .
V

1
2 GRAVITATIONAL ATTRACTION

Example 1. Find the gravitational attraction of the upper

� half of a solid sphere of
radius a centered at the origin, if its density is given by δ = x2 + y 2 .

Solution. Since the solid and its density are symmetric about the z-axis, the force
will be in the k -direction, and we can use (3) or (4). Since
�
x2 + y 2 = r = ρ sin φ ,

the integral is
� 2π � π/2 � a
Fz = G ρ sin2 φ cos φ dρ dφ dθ
0 0 0

which evaluates easily to πGa2 /3 . V

a
a 2
Example 2. Let V√ be the solid spherical cap obtained by slicing a
solid sphere of radius a 2 by a plane at a distance a from the center of the
sphere. Find the gravitational attraction of V on a unit point mass at the
center of the sphere. (Take the density to be 1.)

Solution. To take advantage of the symmetry, place the origin at the center of the
sphere, and align the axis of the cap along the z-axis (so the flat side of the cap is parallel
to the xy-plane).
We use spherical coordinates; the main problem is determining the limits of integration.
If we fix φ and θ and let ρ vary, we get a ray which enters V at its flat side

z = a,
or ρ cos φ = a,
√
and leaves V on its spherical side, ρ = a 2. The rays which intersect V in this way are
those for which 0 ≤ φ ≤ π/4, as one sees from the picture. Thus by (4),
√
� 2π � π/4 � a 2
Fz = G sin φ cos φ dρ dφ dθ,
0 0 a/ cos φ

which after integrating with respect to ρ (and θ) becomes

π/4 �√ 1 �
�
= 2πG a 2− sin φ cos φ dφ
0 cos φ
� 3√2 �
= 2πGa −1 .
4

Remark. Newton proved that a solid sphere of uniform density and mass M exerts the
same force on an external point mass as would a point mass M placed at the center of the
sphere. (See Problem 6a in problem section 5C).
This does not however generalize to other uniform solids of mass M — it is not true that
the gravitational force they exert is the same as that of a point mass M at their center
of mass. For if this were so, a unit test mass placed on the axis between two equal point
masses M and M ′ ought to be pulled toward the midposition, whereas actually it will be
pulled toward the closer of the two masses.
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 V1. Plane Vector Fields

1. Vector fields in the plane; gradient fields.

We consider a function of the type

(1) F(x, y) = M (x, y) i + N (x, y) j .

where M and N are both functions of two variables. To each pair of values (x0 , y0 ) for
which both M and N are defined, such a function assigns a vector F(x0 , y0 ) in the plane.
F is therefore called a vector function of two variables. The set of points (x, y) for which
F is defined is called the domain of F.
To visualize the function F(x, y), at each point (x0 , y0 ) in the domain we
place the corresponding vector F(x0 , y0 ) so that its tail is at (x0 , y0 ). Thus
each point of the domain is the tail end of a vector, and what we get is
called a vector field. This vector field gives a picture of the vector function
F(x, y).
Conversely, given a vector field in a region of the xy-plane, it determines a vector function
of the type (1), by expressing each vector of the field in terms of its i and j components.
Thus there is no real distinction between “vector function” and “vector field”. Mindful of
the applications to physics, in these notes we will mostly use “vector field”. We will use
the same symbol F to denote both the field and the function, saying “the vector field F”,
rather than “the vector field corresponding to the vector function F”.
We say the vector field F is continuous in some region of the plane if both M (x, y) and
N (x, y) are continuous functions in that region. The intuitive picture of a continuous vector
field is that the vectors associated to points sufficiently near (x0 , y0 ) should have direction
and magnitude very close to that of F(x0 , y0 ) — in other words, as you move around the
field, the vectors should change direction and magnitude smoothly, without sudden jumps
in size or direction.
In the same way, we say F is differentiable in some region if M and N are differentiable,
that is, if all the partial derivatives

∂M ∂M ∂N ∂N
, , ,
∂x ∂y ∂x ∂y

exist in the region. We say F is continuously differentiable in the region if all these partial
derivatives are themselves continuous there. In general, all the commonly used vector fields
are continuously differentiable, except perhaps at isolated points, or along certain curves.
But as you will see, these points or curves affect the properties of the field in very important
ways.

1
2 V1. PLANE VECTOR FIELDS

Where do vector fields arise in science and engineering?

One important way is as gradient vector fields. If

(2) w = f (x, y)

is a differentiable function of two variables, then its gradient

∂w ∂w
(3) ∇w = i+ j
∂x ∂y

is a vector field, since both partial derivatives are functions of x and y. We recall the
geometric interpretation of the gradient:
�
dw ��
(4) dir ∇w = the direction u in which is greatest;
ds �u
�
dw ��
|∇w| = this greatest value of ,
ds �u
�
dw ��
where = ∇w · u is the directional derivative of w in the direction u.
ds �u
Another important fact about the gradient is that if one draws the contour curves of
f (x, y), which by definition are the curves

f (x, y) = c, c constant,

then at every point (x0 , y0 ), the gradient vector ∇w at this point is per­
pendicular to the contour line passing through this point, i.e.,

(5) the gradient field of f is perpendicular to the contour curves of f .

�
Example 1. Let w = x2 + y 2 = r. Using the definition (3) of gradient, we find
x y xi + yj
∇w = i+ j = .
r r r
The domain of ∇w is the xy-plane with (0, 0) deleted, and it is con­
tinuously differentiable in this region. Since |x i + y j| = r, we see that
|∇w| = 1 . Thus all the vectors of the vector field ∇w are unit vectors,
and they point radially outward from the origin. This makes sense by (4),
since the definition of w shows that dw/ds should be greatest in the radially
outward direction, and have the value 1 in that direction.
Finally, the contour curves for w are circles centered at (0, 0), which are perpendicular
to the vectors ∇w everywhere, as (5) predicts.

2. Force and velocity fields.

Continuing our search for ways in which vector fields arise, here are two physical situations
which are described mathematically by vector fields. We shall refer to them often in the
sequel, using our physical intuition to suggest the sort of mathematical properties that
vector fields ought to have.
 V1. PLANE VECTOR FIELDS 3

Force fields.
From physics, we have the two-dimensional electrostatic force fields arising from a distri­
bution of static (i.e., not moving) charges in the plane. At each point (x0 , y0 ) of the plane,
we put a vector representing the force which would act on a unit positive charge placed at
that point.
In the same way, we get vector fields arising from a distribution of masses in the xy-plane,
representing the gravitational force acting at each point on a unit mass. There are also the
electromagnetic fields arising from moving electric charges and/or a distribution of magnets,
representing the magnetic force at each point.
Any of these we shall simply refer to as a force field.
Example 2. Express in i − j form the electrostatic force field F in the xy-plane arising
from a unit positive charge placed at the origin, given that the force vector at (x, y) is
directed radially away from the origin and that it has magnitude c/r2 , c constant.
Solution. Since the vector x i + y j with tail at (x, y) is directed radially outward and
has magnitude r, it has the right direction, and we need only change its magnitude to c/r2 .
We do this by multiplying it by c/r3 , which gives

cx cy xi + yj
F = i+ 3 j = c 2 .
r3 r (x + y 2 )3/2

Flow fields and velocity fields

A second way vector fields arise is as the steady-state flow fields and velocity fields.
Imagine a fluid flowing in a horizontal shallow tank of uniform depth, and assume that
the flow pattern at any point is purely horizontal and not changing with time. We will call
this a two-dimensional steady-state flow or for short, simply a flow. The fluid can either be
compressible (like a gas), or incompressible (like water). We also allow for the possibility
that at various points, fluid is beiong added to or subtracted from the flow; for instance,
someone could be standing over the tank pouring in water at a certain point, or over a
certain area. We also allow the density to vary from point to point, as it would for an
unevenly heated gas.
With such a flow we can associate two vector fields.
There is the velocity field v(x, y) where the vector v(x, y) at the point (x, y) represents
the velocity vector of the flow at that point — that is, its direction gives the direction of
flow, and its magnitude gives the speed of the flow.
Then there is the flow field, defined by

(6) F = δ(x, y) v(x, y)

where δ(x, y) gives the density of the fluid at the point (x, y), in terms of mass per unit
area. Assuming it is not 0 at a point (x, y), we can interpret F(x, y) as follows:

(7) dir F = direction of fluid flow at (x, y);

Δl
rate (per unit length per second) of mass transport ΔA
�
|F| = F
across a line perpendicular to the flow direction at (x, y).

Namely, we see that first by (6) and then by the picture, v Δt

4 V1. PLANE VECTOR FIELDS

|F| Δl Δt = δ |v| Δt Δl = mass in ΔA,

from which (7) follows by dividing by Δl Δt and letting Δl and Δt → 0.
If the density is a constant δ0 , as it would be for an incompressible fluid at a uniform
temperature, then the flow field and velocity field are essentially the same, by (6) — the
vectors of one are just a constant scalar multiple of the vectors of the other.
xi + yj
Example 3. Describe and interpret F = as a flow field and a force field.
x2 + y 2
Solution. As in Example 2, the field F is defined everywhere except (0, 0) and its
direction is radially outward; now, however, its magnitude is r/r2 , i.e., |F| = 1/r.
F is the flow field for a source of magnitude 2π at the origin. To see this, look at a circle
of radius a centered at the origin. At each point P on this circle, the flow is radially outward
and by (7),
1
mass transport rate at P = , so that
a
1
mass transport rate across circle = · 2πa = 2π .
a
This shows that in one second, 2π mass flows out through every circle centered at the origin.
This is the flow field for a source of magnitude 2π at the origin — for example, one could
imagine a narrow pipe placed over the tank, introducting 2π mass units per second at the
point (0, 0).
We know that |F| = δ |v| = 1/r. Two important cases are:
• if the fluid is incompressible, like water, then its density is constant, so the flow speed
must decrease like 1/r — the flow outward gets slower the further you are from the origin;
• if it is compressible like a gas, and its flow speed stays constant, then the density must
decrease like 1/r.
We now interpret the same field as a force field.
Suppose we think of the z-axis in space as a long straight wire, bearing a uniform positive
electrostatic charge. This gives us a vector field in space, representing the electrostatic force
field.
Since one part of the wire looks just like any other part, radial symmetry shows first that
the vectors in the force field have 0 as their k -component, i.e., they are pointed radially
outward from the wire, and second that their magnitude depends only on their distance r
from the wire. It can be shown in fact that the resulting force field is F, up to a constant
factor.
Such a field is called “two-dimensional”, even though it is a vector field in space, because
z and k don’t enter into its description — once you know how it looks in the xy-plane, you
know how it looks all through space.
The important thing to notice is that the magnitude of the force field in the xy-plane
decreases like 1/r, not like 1/r2 , as it would if the charge were all at a point.
In the same way, the gravitational field of a uniform mass distribution along the z-axis
would be −F, up to a constant factor, and would be called a “two-dimensional gravitational
 V1. PLANE VECTOR FIELDS 5

field”. Naturally, we don’t have infinite long straight wires, but if you have a long straight
wire, and stay away from its ends, or have only a short straight wire, but stay close to it,
the force field will look like F near the wire.
Example 4. Find the velocity field of a fluid with density 1 in a shallow tank, rotating
with constant angular velocity ω counterclockwise around the origin.
Solution. First we find the field direction at each point (x, y). x y
-y x
We know the vector x i + y j is directed radially outward. Therefore a
vector perpendicular to it in the counterclockwise direction (see picture)
will be −y i + x j (since its scalar product with x i + y j is 0 and the signs
are correct).
The preceding vector has magnitude r. If the angular velocity is ω, then the linear
velocity is given by

|v| = ωr,

so to get the velocity field, we should multiply the above field by ω :

v = −ωy i + ωx j .
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 Work and line integrals

Line integrals: (also called path integrals)

Ingredients:
Field F = M i + N j = �M, N �
Curve C: r(t) = x(t)i + y(t)j = �x, y� ⇒ dr = �dx, dy�.
Line integral: � � �
F · dr = �M, N � · �dx, dy� = M dx + N dy.
C C C
We need to discuss:
a) What this notation means and how line integrals arise.
b) How to compute them.
c) Their properties and notation.

a) How line integrals arise.

The figure on the left shows a force F being applied over a displacement Δr. Work is force
times distance, but only the component of the force in the direction of the displacement
does any work. So, work = |F| cos θ |Δr| = F · Δr.
dr
F
F
F
dr
θ F
Δr dr
For a variable force applied over a curve the total work is found by ’summing’ the infinites­
imal pieces. We call this a line integral and denote it
�
F · dr.
C

It is illustrated in the above figure on the right.

y
b) Computing line integrals.
We show this by steps by example. 1
�
Example 1: Evaluate I = x2 y dx + (x − 2y) dy
C C
over the part of the parabola y = x2 from (0, 0) to (1, 1).
First, parametrize the curve: x
1
x = t, y = t2 , 0 ≤ t ≤ 1.

Note, we specified the range of t to get exactly the part of the curve we wanted.
Next, compute the differentials of x and y:

dx = dt, dy = 2t dt.
Finally substitute everything in the integral and compute the standard single variable in­
tegral: � 1 � 1
2 2 2 2
I= t (t ) dt + (t − 2t )2t dt = t4 + 2t2 − 4t3 dt = − .
0 0 15
y
Example 2: (Line integrals depend on the path.) 1
Same integral as previous example except C is the
straight line from (0, 0) to (1, 1). C
Parametrize curve: x = t, y = t, 0 ≤ t ≤ 1. ⇒ dx = dt, dy = dt
� 1 � 1
2 1 x
⇒ I= t · t dt + (t − 2t) dt = t3 − t dt = − .
0 0 4 1

Note: this is a different value from example 1 and illustrates the very important fact that,
in general, the line integral depends on the path. Later we will learn how to spot the cases
when the line integral will be independent of path.

Example 3: (Line integrals are independent of the parametrization.)

Here we do the same integral as in example 1 except use a different parametrization of C.
Parametrize C: x = sin t, y = sin2 t, 0 ≤ t ≤ π/2 ⇒ dx = cos t dt, dy = 2 sin t cos t dt.
� π/2
⇒ I = sin4 t cos t dt + (sin t − 2 sin2 t)2 sin t cos t dt
0 �π/2
� π/2
4 2 3 sin5 t 2 3 4 �
� 2
= (sin t + 2 sin t − 4 sin t) cos t dt = + sin t − sin t� =− .
0 5 3 0 15
This is same value as example 1 and illustrates the very important point that the line
integral is independent of how the curve is parametrized.

c) Properties and notation of line integrals

1. They are independent of parametrization.
2. If you reverse direction on curve then the line integral changes sign. That is,
� �
F · dr = − F · dr.
−C C

(−C means the same curve traversed in the opposite direction.)

� �
3. If C is closed we use the notation F · dr = M dx + N dy.
C C
 �
Example 4: Evaluate I = y dx + (x + 2y) dy where C is the curve shown.
C
Answer: The curve has two pieces so the integral will also y
C1
1 (1, 1)
� �
I= y dx + (x + 2y) dy + y dx + (x + 2y) dy.
C1 C2
C2
Here we see that we don’t always need to introduce a new variable t.
C1 : y = 1, use x as parameter. 0 ≤ x ≤ 1 ⇒ dx = dx, dy = 0.
� � 1 x
Substituting in the integral ⇒ y dx + (x + 2y) dy = dx = 1. 1
C1 0
C2 : x = 1, use y as parameter. y goes from 1 to 0 and dx = 0
� � 0 � 1
⇒ y dx + (x + 2y) dy = (1 + 2y) dy = − 1 + 2y dy = −2.
C2 1 0
We get I = 1 − 2 = −1.
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010 ��

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 Geometric Approach to Line Integrals

Line integrals are intrinsically geometric, so we should sometimes be able to use geomet­
ric reasoning to avoid the tedious calculations used in computing certain line integrals.
The geometry can also give us some insight into the situation that calculation sometimes
obscures. y
We start with a line integral that we compute directly, 1
Example 1: Evaluate I = −y dx + x dy where C is the
C
unit circle traversed in a counterclockwise (CCW) direction.
Parametrization: x = cos t, y = sin t, 0 ≤ t ≤ 2π. x
1
⇒ dx = − sin t dt, dy = cos t dt.
2π 2π
⇒ I= − sin t(− sin t) dt + cos t(cos t) dt = dt = 2π.
0 0

dr ds
The intrinsic formula Recall that we know = T , where T = unit tangent and s
dt dt
= arclength. Removing the dt gives dr = T ds. We can use this in our formula for line
integrals and get a form that we call the intrinsic formula

F · dr = F · T ds.
C C

Example 2: Redo example 1 using the intrinsic formula: F · dr = F · T ds.

C C
T = unit tangent ⇒ T = −y i + x j (on the unit circle x2 + y2 = 1).
⇒ F · T = y 2 + x2 = 1 (on the unit circle) ⇒ I = ds = arclength of circle = 2π.
C
Lesson: it can pay to think geometrically.
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 Fundamental Theorem for Line Integrals
Gradient fields and potential functions
Earlier we learned about the gradient of a scalar valued function
vf (x, y) = Ufx , fy ).
For example, vx3 y 4 = U3x2 y 4 , 4x3 y 3 ).
Now that we know about vector fields, we recognize this as a special case. We will call it a
gradient field. The function f will be called a potential function for the field.
For gradient fields we get the following theorem, which you should recognize as being similar
to the fundamental theorem of calculus. y
Theorem (Fundamental Theorem for line integrals)
If F = vf is a gradient field and C is any curve with endpoints P1 = (x1 , y1 )
P0 = (x0 , y0 ) and P1 = (x1 , y1 ) then C

F · dr = f (x, y)|PP10 = f (x1 , y1 ) − f (x0 , y0 ).

C P0 = (x0 , y0 )
x
That is, for gradient fields the line integral is independent of the
path taken, i.e., it depends only on the endpoints of C.
Example 1: Let f (x, y) = xy 3 + x2 ⇒ F = vf = Uy 3 + 2x, 3xy 2 )
Let C be the curve shown and compute I = F · dr.
C y
Do this both directly (as in the previous topic) and using the above formula.
(1, 2)
Method 1: parametrize C: x = x, y = 2x, 0 ≤ x ≤ 1.
1
⇒ I= (y 3 + 2x) dx + 3xy 2 dy = (8x3 + 2x) dx + 12x3 2 dx C
C 0
1
= 32x3 + 2x dx = 9.
0

Method 2: vf · dr = f (1, 2) − f (0, 0) = 9. x

C (0, 0)
Proof of the fundamental theorem
t1
dx dy
vf · dr = fx dx + fy dy = fx (x(t), y(t)) + fy (x(t), y(t)) dt
C C t0 dt dt
t1
d
= f (x(t), y(t)) dt = f (x(t), y(t))|tt10 = f (P1 ) − f (P0 )
t0 dt
The third equality above follows from the chain rule.

Significance of the fundamental theorem

For gradient fields F the work integral F · dr depends only on the endpoints of the path.
C
We call such a line integral path independent.
The special case of this for closed curves C gives:

F = vf ⇒ F · dr = 0 (proof below).
C

1
Following physics, where a conservative force does no work around a closed loop, we say
F = vf is a conservative field.

Example 1: If F is the electric field of an electric charge it is conservative.

Example 2: The gravitational field of a mass is conservative.

Differentials: Here we can use differentials to rephrase what we’ve done before. First
recall:
a) vf = fx i + fy j ⇒ vf · dr = fx dx + fy dy.
Z
b) vf · dr = f (P1 ) − f (p0 ).
C
Using differentials we have df = fx dx + fy dy. (This is the same as vf · dr.) We say
M dx + N dy is an exact differential if M dx + N dy = df for some function f .
Z Z
As in (b) above we have M dx + N dy = df = f (P1 ) − f (P0 ).
C C

Proof that path independence is equivalent to conservative

We show that Z
F · dr is path independent for any curve C
C
is equivalent to
I
F · dr = 0 for any closed path.
C
This is not hard, it is really an exercise to demonstrate the logical structure of a proof
showing equivalence. We have to show:
i) Path independence ⇒ the line integral around any closed path is 0.
ii) The line integral around all closed paths is 0 ⇒ path independence.

i) Assume path independence and consider the closed path C shown

I in figure (i) below. Since
the starting point P0 is the same as the endpoint P1 we get F · dr = f (P1 ) − f (P0 ) = 0
C
(this proves (i)).
I
ii) Assume F · dr = 0 for any closed curve. If C1 and C2 are both paths between P0
C
and P1 (see fig. 2) then C1 − C2 is a closed path. So by hypothesis
I Z Z Z Z
F · dr = F · dr − F · dr = 0 ⇒ F · dr = F · dr.
C1 −C2 C1 C2 C1 C2

That is the line integral is path independent, which proves (ii).

y y
P0 = P1
C1 P1
C
C2

P0
x x

Figure (i) Figure (ii)

2
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 V2.1 Gradient Fields and Exact Differentials

1. Criterion for gradient fields.

Let F = M (x, y) i + N (x, y) j be a two-dimensional vector field, where M and N are
continuous functions. There are three equivalent ways of saying that F is conservative, i.e.,
a gradient field:
� Q �
(1) F = ∇f ⇔ F · dr is path-independent ⇔ F · dr = 0 for any closed C
P C

Unfortunately, these equivalent formulations don’t give us any effective way of deciding
if a given field F is a conservative field or not. However, if we assume that F is not just
continuous but is even continuously differentiable (meaning: Mx , My , Nx , Ny all exist and
are continuous), then there is a simple and elegant criterion for deciding whether or not F
is a gradient field in some region.
Criterion. Let F = M i + N j be continuously differentiable in a region D. Then, in
D,

(2) F = ∇f for some f (x, y) ⇒ M y = Nx .

Proof. Since F = ∇f , this means

M = fx and N = fy . Therefore,
My = fxy and Nx = fyx .

But since these two mixed partial derivatives are continuous (since My and Nx are, by
hypothesis), a standard 18.02 theorem says they are equal. Thus My = Nx . �
This theorem may be expressed in a slightly different form, if we define the scalar function
called the two-dimensional curl of F by

(3) curl F = Nx − My .

Then (2) becomes

(2′ ) F = ∇f ⇒ curl F = 0 .

This criterion allows us to test F to see if it is a gradient field. Naturally, we would also
like to know that the converse is true: if curl F = 0, then F is a gradient field. As we shall
see, however, this requires some additional hypotheses on the region D. For now, we will
assume D is the whole plane. Then we have
Converse to Criterion. Let F = M i + N j be continuously differentiable for all x, y.

(4) My = Nx for all x, y ⇒ F = ∇f for some differentiable f and all x, y.

The proof of (4) will be postponed until we have more technique. For now we will illustrate
the use of the criterion and its converse.
1
2 V2.1 GRADIENT FIELDS AND EXACT DIFFERENTIALS

Example 1. For which value(s), if any of the constants a, b will axy i + (x2 + by) j be
a gradient field?
Solution. The partial derivatives are continuous for all x, y and My = ax, Nx = 2x.
Thus by (2) and (4), F = ∇f ⇔ a = 2; b is arbitrary.

xi + yj −y i + x j
Example 2. Are the fields F = , G = conservative?
x2 + y 2 x2 + y 2
Solution. We have (the second line follows from the first by interchanging x and y):

−x2 + y 2
� � � �
∂ x ∂ y 2xy
= ; = ;
∂x x2 + y 2 (x2 + y 2 )2 ∂x x2 + y 2 (x2 + y 2 )2
x2 − y 2
� � � �
∂ y ∂ x 2yx
= ; = ;
∂y x2 + y 2 (x2 + y 2 )2 ∂y x2 + y 2 (x2 + y 2 )2

from this, we see immediately that

� � � � � � � �
∂ x ∂ y ∂ y ∂ x
= ; − 2 = ;
∂y x + y2
2 ∂x x2 + y 2 ∂y x + y2 ∂x x2 + y 2

the two equations in the last line show respectively that F and G satisfy the criterion (2).
However, neither field is defined at (0, 0), so that the converse (4) is not applicable. So the
question cannot be decided just on the basis of (2) and (4). In fact, it turns out that F is a
gradient field, since one can check that
�
F = ∇ ln x2 + y 2 = ∇ ln r .

On the other hand, G is not conservative, since if C is the unit circle x = cos t, y = sin t,
we have
2π
−y dx + x dy sin2 t dt + cos2 t dt
� � �
G · dr = = = 2π =
6 0.
C C x2 + y 2 0 sin2 t + cos2 t

We will return to this example later on.

MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 V2.2-3 Gradient Fields and Exact Differentials

2. Finding the potential function.

In example 2 in the previous reading we saw that

xi + yj p
F= 2 2
= ∇ ln x2 + y 2 = ∇ ln r.
x +y

This raises the question of how we found the function 12 ln(x2 + y 2 ). More generally, if we
know that F = ∇f — for example if curl F = 0 in the whole xy-plane — how do we find
the function f (x, y)? There are two methods; some students prefer one, some the other.
Method 1. Suppose F = ∇f . By the Fundamental Theorem for Line Integrals,
Z (x,y)
(5) F · dr = f (x, y) − f (x0 , y0 ) .
(x0 ,y0 )

Read from left to right, (5) gives us an easy way of finding the line integral in terms of
f (x, y). But read right to left, it gives us a way of finding f (x, y) by using the line integral:
Z (x,y)
(5′ ) f (x, y) = F · dr + c .
(x0 ,y0 )

(Here c is an arbitrary constant of integration; as (5′ ) shows, c = f (x0 , y0 ).)

Remark. It is common to refer to f (x, y) as the (mathematical) potential function.
The potential function used in physics is −f (x, y). The negative sign is used by physicists
so that the potential difference will represent work done against the field F, rather than
work done by the field, as the convention is in mathematics.
Example 3. Let F = (x + y 2 ) i + (2xy + 3y 2 ) j . Verify that F satisfies the Criterion
(2), and use method 1 above to find the potential function f (x, y).
∂(y 2 ) ∂(2xy)
Solution. We verify (2) immediately: = 2y = .
∂y ∂x
We use (5′ ). The point (x0 , y0 ) can be any convenient starting point; (0, 0) is the usual
choice, if the integrand is defined there. (We will subscript the variables to avoid confusion
with the variables of integration, but you don’t have to.) By (5′ ),
Z (x1 ,y1 )
(6) f (x1 , y1 ) = (x + y 2 ) dx + (2xy + 3y 2 ) dy .
(0,0) (x 1, y 1 )

Since the integral is path-independent, we can choose any path we like. C2

The usual choice is the one on the right, as it simplifies the computations. x1
C
(Most of what follows you can do mentally, with a little practice.) 1

1 2
Z x1
On C1 , we have y = 0, dy = 0, so the integral on C1 becomes x dx = x .
0 2 1
Z y1
On C2 , we have x = x1 , dx = 0, so the integral is (2x1 y + 3y 2 ) dy = x1 y12 + y13 .
0
1
2 V2.2-3 GRADIENT FIELDS AND EXACT DIFFERENTIALS

Adding the integrals on C1 and C2 to get the integral along the entire path, and dropping
the subscripts, we get by (6) and (5′ )

1 2
f (x, y) = x + xy 2 + y 3 + c .
2

(The constant of integration is added by (5′ ), since the choice of starting point was arbitrary.
You should always confirm the answer by checking that ∇f = F.) 

Method 2. Once again suppose F = ∇f , that is M i + N j = fx i + fy j . It follows

that

(7) fx = M and fy = N .

These are two equations involving partial derivatives, which we can solve simultaneously by
integration. We illustrate using the previous example: F = (x + y 2 , 2xy + 3y 2 ).
Solution by Method 2. Using the first equation in (7),

∂f
= x + y2 . Hold y fixed, integrate with respect to x:
∂x
1 2
(8) f = x + y 2 x + g(y). where g(y) is an arbitrary function of y.
2

∂f
To find g(y), we calculate two ways:
∂y

∂f
= 2yx + g ′ (y) by (8), while
∂y
∂f
= 2xy + 3y 2 from (7), second equation.
∂y

Comparing these two expressions, we see that g ′ (y) = 3y 2 , so g(y) = y 3 + c. Putting it all
together, using (8), we get f (x, y) = 12 x2 + y 2 x + y 3 + c, as before. 
In the first method, the answer is written down immediately as a line integral; the rest of
the work is in evaluating the integral, which goes quickly, since on a horizontal or vertical
path either dx = 0 or dy = 0.
In the second method, the answer is obtained by an algorithm involving several steps
which should be carried out in the right order.
The first method has the advantage of reminding you each time how f (x, y) is defined and
what it means, facts of theoretical and practical importance. The second method has the
advantage of requiring no knowledge of line integrals, which makes it popular with students;
on the other hand, when done in three dimensions, the bookkeeping gets more complicated,
whereas in the first method it does not; overall, the first method is faster, provided you are
confident enough to do some of the work mentally.
 V2.2-3 GRADIENT FIELDS AND EXACT DIFFERENTIALS 3

3. Exact differentials.
The formal expressions M (x, y) dx + N (x, y) dy which have appeared as the integrands
in our line integrals are called differentials. In some applications, most notably thermody-
namics,
R one usually works directly with the differential M dx + N dy and its line integral
M dx + N dy, without considering or using the associated vector field F = M i + N j .
Therefore it is important to have the results about gradient fields in this section translated
into the language of differentials. We do this now.
If f (x, y) is a differentiable function, its total differential df (or simply differential) is by
definition the expression

(9) df = fx dx + fy dy .

For example, if f (x, y) = x2 y 3 , then d(x2 y 3 ) = 2xy 3 dx + 3x2 y 2 dy .

We call the differential M dx + N dy exact, in a region D where M and N are defined,
if it is the total differential of some function f (x, y) in this region, i.e., if in D,

(10) M = fx and N = fy , for some f (x, y).

From this we see that the relation between differentials and vector fields is

M dx + N dy is exact ⇔ Mi +Nj is a gradient field

M dx + N dy = df ⇔ M i + N j = ∇f .

In this language, the criterion (we use the same equation number as in the section where
they were first presented)

(2) F = ∇f for some f (x, y) ⇒ M y = Nx .

and its partial converse:

Let F = M i + N j be continuously differentiable for all x, y.

(4) My = Nx for all x, y ⇒ F = ∇f for some differentiable f and all x, y.

become the

Exactness Criterion. Assume M and N are continuously differentiable in a region D

of the plane. Then in this region,

(11) M dx + N dy exact ⇒ M y = Nx ;
(12) if D is the whole xy-plane, My = Nx ⇒ M dx + N dy exact.

If the exactness criterion shows that M dx + N dy is exact, then the function f (x, y) may
be found by either of the two methods previously described.
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 18.02 Problem Set 7
At MIT problem sets are referred to as ’psets’. You will see this term used occasionally
within the problems sets.
The 18.02 psets are split into two parts ’part I’ and ’part II’. The part I are all taken
from the supplementary problems. You will find a link to the supplementary problems
and solutions on this website. The intention is that these help the student develop
some fluency with concepts and techniques. Students have access to the solutions
while they do the problems, so they can check their work or get a little help as they
do the problems. After you finish the problems go back and redo the ones for which
you needed help from the solutions.
The part II problems are more involved. At MIT the students do not have access
to the solutions while they work on the problems. They are encouraged to work
together, but they have to write their solutions independently.

Part I (15 points)

At MIT the underlined problems must be done and turned in for grading.
The ‘Others’ are some suggested choices for more practice.
A listing like ’§1B : 2, 5b, 10’ means do the indicated problems from supplementary
problems section 1B.
1 Double Integrals.
§3A: 2c, 3b, 4c, 5a; Others: 1, 2ab, 3a, 4b, 5bc, 6, 7

2 Double Integrals in polar coordinates. Applications.

§3B: 1a, 2d, 3ac; Others: 1cd, 2b, 3b
§3C: 1, 2a, 4; Others: 2b, 3

3 Change of variables in double integrals.

§3D: 1, 2, 4; Others: 3
 Part II (25 points)

Problem 1 (3: 1,2)

a) Sketch the solid in the first octant bounded√by the xy, yz and xz coordinate planes,
the plane x + y = 4 and the surface z = 4 − x.
b) Find the volume of the solid of part (a).

Problem 2 (5: 3,2)

The equation of a surface of revolution obtained by spinning the (1-D) graph of the
function z = f (y) in the y-z plane around the z-axis is given in polar coordinates by
z = f (r, θ) = f (r) (that is, there is no dependence on θ). Assume that h := f (0) > 0.
a) Show that the formula which you get by using the double integral in polar coor­
dinates for the volume under the graph of this surface of revolution and over the x-y
plane is the same as that given by the “shell method” in single-variable calculus.
b) Illustrate with a sketch.

Problem 3 (4: 2,1,1)

a) Find the formula for the centroid (center of mass assuming density = 1) of a
uniform plane region in the shape of circular sector of radius a and central angle θ
(in terms of a and θ).
b) Write down the formula for the centroid of a uniform plane region in the shape of
a isoceles triangle with height a and angle θ between the two equal sides. (You may
use the formula from elementary geometry or compute it out using the integral.)
c) If you align the two regions above using the same values for a and θ and with
the central angle of the sector coinciding with the apex angle of the triangle, which
centroid will be closer to the center (= the apex)? What do the math and the physics
predict and why, and are they consistent?

Background for problems 4 and 5 :

In fluid mechanics the fluid flow map ϕ is defined as follows: if (x, y, z) is the position
of a point mass in the flow at time t = 0, then (X, Y, Z) = ϕ(x, y, z, t) is the
downstream position of that same point mass after an elapsed time t.
The standard assumptions on ϕ are that it is smooth and one-to-one.
We will call a flow volume incompressible if for any bounded space region R in the
flow, the volume of ϕ(R, t) is the same as the volume of R for all t. In other words,
if Rt = ϕ(R, t) = {ϕ(x, y, z, t) | (x, y, z) in R} is the the region formed by the points
from R which have been carried downstream by the flow, then Rt can have a different
shape but must have the same volume at all times, if the flow is ‘v-i’.
In problems 4 & 5, we’ll take the simpler case of a 2D flow (which could be e.g. a 2D
section of a flow in 3D). Let (X(x, y, t), Y (x, y, t)) = ϕ(x, y, t). Note that then by
definition, the velocity vectors v(x, y, t) of the flow are given by v(x, y, t) = � ∂X , ∂Y �.
∂t ∂t
A v-i flow in this case is one that preserves area, since area is the 2D version of
volume.
 ∂(X, Y )
For a fixed value of t, let J(x, y, t) = be the Jacobian of the transformation
∂(x, y)
(x, y) �→ (X(x, y, t), Y (x, y, t)). The general change-of-variables formula says that if
a region R goes to a region R� by a transformation (x, y) �→ (X, Y ) with Jacobian
∂(X,Y ) � �
��
∂(x,y)
, then the areas of R and R are related by A (R ) = R
| J(x, y) | dA. Here
��
this gives that A (Rt ) = R | J(x, y, t) | dA, and therefore that a 2D flow is v-i if
and only if | J(x, y, t) | = 1 for all (x, y, t).
In problems 5 & 6 we will look at three examples of 2D flows, v-i and non-vi, in order
to illustrate this idea.
Example A: ϕ(x, y, t) = ((1 + t)x, (1 + t)y);
R = the triangle with vertices at (0, 0), (1, 1) and (1, −1).
Example B: ϕ(x, y, t) = (x cos t − y sin t, x sin t + y cos t);
R = the triangle with vertices at (0, 0), (2, 0) and (2, 1).
1
� �
Example C: ϕ(x, y, t) = (1 + t)x, ( 1+t )y ;
R = the rectangle with vertices at (1, 1), (1, 4), (2, 1) and (2, 4).

Problem 4 (4: 2,2)

In each of the cases A, B and C:
i) compute the Jacobian J(x, y, t)
ii) compute the area A (Rt )

Problem 5 (9: 3,3,3)

In each of the cases A, B and C:
i) Sketch the pattern of the flow paths over time, including some starting from points
in R.
ii) Compute the velocity vectors of the flow and sketch in a few on the flow lines.
iii) Sketch the regions R and Rt and check this against the areas calculated in problem
5 to see if it looks correct in each case. Use the following values for t:
A: t = 2, B: t = π2 , and C: t = 3.
Suggestion for sketching Rt : see where the corners of R end up on Rt .
For the flow lines in case C, also note that X(x, y, t)Y (x, y, t) = xy for all values of t.
iv) Identify which flows are v-i, using the computed results (and the sketches).
v) In the cases A and B, describe the what the flow is doing.

Note: after we learn Green’s Theorem in normal form in 2D, or equivalently the
Divergence Theorem in 3D, we’ll be able to prove that incompressibility and (our
ad-hoc terminology) ‘volume-incompressibility’ are in fact equivalent conditions.
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 18.02 Problem Set 7, Part II Solutions

1.(a)

(b)
4 4−x √
� �
V = 4 − xdy dx
�0 4 0
� √ �y=4−x
= y 4 − x y=0 dx
0
� 4
�4
3/2 2 � 2 64
= (4 − x) dx = − (4 − x) � = 45/2 = .
5/2 �
0 5 0 5 5

2. (a) For simplicity let us assume we are integrating the volume of revo­
lution out to some radius a. We also assume that f (r) ≥ 0 for 0 ≤ r ≤ a.
Then if R is the disc x2 + y 2 ≤ a, we want
� �
V = f dA.
R

In polar coordinates this is

� 2π � a
V = f (r)rdrdθ.
0 0

We may write the integral in the other order as well, because the limits to
each integral are constants.
� a �� 2π �
V = f (r)dθ rdr.
0 0

1
Evaluating the inner integral gives
� a
2πrf (r)dr
0

which is the shell method formula.

(b)

3. (a) For a circular sector Sθ with center angle 2θ and radius a,

1
A(Sθ ) = a2 (2θ) = a2 θ
2
and its centroid is at (x̄S (θ), 0) where
� θ � a
1
x̄S (θ) = x̄(Sθ ) = 2 (r cos ϕ) r dr dϕ.
aθ −θ 0

This comes out to � �

2 sin θ
x̄S (θ) = a.
3θ
2 sin θ
So we observe a factor fs (θ) = 3θ
governing at what multiple of the radius
the centroid must occur.

2
(b) The result from elementary geometry is that the centroid of a triangular
region with uniform density is located at the intersection of the three side-
bisectors or ‘medians’, and that this point divides the medians in a ratio of 2
to 1, with the shorter segment nearest the bisection point. Thus we get that
for the triangle given and positioned in the same way as the circular sector
on the x-axis � �
2
x̄Δ = a.
3

So fΔ = the factor which multiplies a is equal to 23 , independent of θ.

(c) The circular sector region is a subset of the triangular region, with the
excess part of the triangle farther away from the origin. Thus we should have
x̄S (θ) < x̄Δ . But in fact the math agrees, since sinθ θ < 1, and so the fs (θ),
the factor of a for the sector, which we found in part(a) to be fs (θ) = 2 sin
3θ
θ
2
thus satisfies the inequality fs (θ) < 3 = fΔ .

4. Case A: (X(x, y, t), Y (x, y, t)) = ((1 + t)x, (1 + t)y).

� �
∂(X, Y ) 1+t 0
J(x, y, t) = =
∂(x, y) 0 1+t
and so (a) | J(x, y, t) | = (1 + t)2 and (b) A (Rt ) = (1 + t)2 A (R)
Case B: (X(x, y, t), Y (x, y, t)) = (x cos t − y sin t, x sin t + y cos t),
� �
cos t − sin t
J(x, y, t) =
sin t cos t
and so (a) | J(x, y, t) | = 1 and (b) A (Rt ) = A (R) for all t.
1
� �
Case C: (X(x, y, t), Y (x, y, t)) = (1 + t)x, ( 1+t )y ,
� �
1+t 0
J(x, y, t) = 1
0 1+t
and so (a) | J(x, y, t) | = 1 and (b) A (Rt ) = A (R) for all t.

5. Case A: (X(x, y, t), Y (x, y, t)) = ((1 + t)x, (1 + t)y).

v(x, y, t) = � ∂X , ∂Y � = �x, y�. The flow lines are straight lines fanning out
∂t ∂t
from the origin. The velocity vectors depend only on the position, and their
magnitude increases with the distance from the origin; thus the flow gets
faster as it moves away from O.
R2 , the points downstream at t = 2 from the triangular region R, form a
triangular region with vertices at (0, 0), (3, 3) and (3, −3). Thus A (R2 ) =
1
2
· 3 · 6 = 9 = (1 + 2)2 A (R) as predicted in problem 5, since A (R) = 12 · 1 · 2

3
= 1. The flow is not v-i.
Case B: (X(x, y, t), Y (x, y, t)) = (x cos t − y sin t, x sin t + y cos t).
v(x, y, t) = �−x sin t − y cos t, x cos t − y sin t�. The flow lines are circular
paths centered at the origin. The velocity � vectors depend on position and
time; however the speed | v(x, y, t) | = x2 + y 2 does not depend explicitly
on time; its magnitude increases with the distance from the origin, but the
angular velocity ω = |vr | = 1 is constant. So the flow is a ‘pure rotating’
circular flow moving counter-clockwise around the origin at 1 rad./unit time.
R π2 , the points downstream at t = π2 from the triangular region R, form a
triangular region with vertices at (0, 0), (0, 2) and (−1, � 2),� i.e. the triangular
region R rotated by π2 counter-clockwise. Thus A R π2 = A (R) = 1, as
predicted in problem 4, since in general the flow is v-i.
1
� �
Case C: (X(x, y, t), Y (x, y, t)) = (1 + t)x, ( 1+t )y .
−y
v(x, y, t) = �x, (1+t)2 �. The flow lines are the hyberbolas XY = xy = con­

stant, with the x and y-axes as asymptotes. The velocity vectors depend on
position and time. (The j-component of the velocity goes to zero as t > 0
increases, which is consistent with the fact that the x-axis is a horizontal
asymptote.) The flow comes ‘screaming in’ at high speed from (0, ∞) for
t > −1, and then slows down as t increases.
R3 , the points downstream at t = 3 from the rectangular region R, form
a rectangular region with vertices at (4, 14 ), (4, 1), (8, 14 ), and (8, 1). Thus
A (R3 ) = A (R) = 3, as predicted in problem 5, since in general the flow is
v-i. However, it is not as easy to see why this is the case as it was in case
B, where the flow just rotates a region into a congruent region.

4
5
6
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 18.02 Problem Set 10
At MIT problem sets are referred to as ’psets’. You will see this term used occasionally
within the problems sets.
The 18.02 psets are split into two parts ’part I’ and ’part II’. The part I are all taken
from the supplementary problems. You will find a link to the supplementary problems
and solutions on this website. The intention is that these help the student develop
some fluency with concepts and techniques. Students have access to the solutions
while they do the problems, so they can check their work or get a little help as they
do the problems. After you finish the problems go back and redo the ones for which
you needed help from the solutions.
The part II problems are more involved. At MIT the students do not have access
to the solutions while they work on the problems. They are encouraged to work
together, but they have to write their solutions independently.

Part I (11 points)

At MIT the underlined problems must be done and turned in for grading.
The ‘Others’ are some suggested choices for more practice.
A listing like ’§1B : 2, 5b, 10’ means do the indicated problems from supplementary
problems section 1B.

1 Triple Integrals in rectangular and cylindrical coordinates.

§5A: 2d, 3, 4, 5; Others: 1, 2ac, 3, 6, 7

2 Triple integrals in spherical coordinates. Gravitational attraction.

§5B: 1bc, 2, 3, 4b; Others: 1a, 4ac
§5C: 2; Others: 3, 4

1
 Part II (19 points)

Problem 1 (6: 3,3)

2 2
Let G be the solid region in 3-space which
hlies inside the surface x + (y − 1) = 1,
above z = 0, and below the surface z = x2 + y 2 .
a) Find the volume of G.
b) Find the z-coordinate of the centroid of G.

Problem 2 (8: 2,2,2,2)

√
2 2x 4−x2
f (x, y, z) dV = f (x, y, z) dz dy dx .
G 0 0 0
a) Sketch the 3-D solid region G.
Rewrite f (x, y, z) dV as a triple integral with the correct limits of integration
G
in the order:
b) f (x, y, z) dy dx dz c) f (x, y, z) dx dz dy

d) Take f = 1 (constant) and verify that get you the same answer (which is the
volume of G) from (i) the iterated integrals in the given order, and (ii) the iterated
integrals you found in part(b).
(Take a look at the integration you’d need to do for the order of part(c) too, just to
convince yourself that you don’t want to mess with it – unless of course you do.)

Problem 3 (5)
In this problem we use triple integrals to compute the amount of work needed to fill a
tank. (This is the calculation one uses for example to compute the energy which can
be stored and then used to generate electricity as needed; for more on this application,
see ... .)
As an example, suppose that a storage device uses electricity to pump water into a
storage vessel which is shaped like a parabolic trough, with cross-section profile given
1
by the curve z = 100 x4 with z = 0 to 81 meters, and with length is L = 200 meters,
as shown:

2
Compute the total amount of energy that can be stored in this storage vessel, using
1 kg/m3 as the mass density of water. Express answer in Joules.
(Hint: The potential energy of a mass m increases by mgh when it is lifted a height
h, where g = 9.8 m/s2 . Assume that all of the water was initially at height z = 0
before being pumped into the vessel).

3
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 18.02 Problem Set 10, Part II Solutions

1. Base: R: x2 + (y − 1)2 ≤ 1. Top: z = f (x, y) = (x2 + y 2 )1/2 .

In cylindrical coords, base is

0 ≤ r ≤ 2 sin θ
0≤ θ ≤π

The top is z = r.

(a)
� π � 2 sin θ � r � π � 2 sin θ
V = 1dz rdr dθ = r · rdr dθ
0 0 0 0 0
� π � π
8
= r3 /3|20 sin θ dθ = sin3 θdθ
0 3 0
� π/2
16
= sin3 θ dθ
3 0
with the last step by symmetry. Then use table 113, n = 3 to get
16 2 32
= · = .
3 3 9

1
(b)

1 π 2 sin θ r
� � � � � �
1
z̄ = zdV = zdz rdr dθ
V G V 0 0 0
1 π 2 sin θ � 2 �r
� �
= z /2 0 r dr dθ
V 0 0
1 π 2 sin θ 3 2 π 4
� � �
= r /2 dr dθ = sin θ dθ
V 0 0 V 0
4 π/2 4
�
4 13π 3π
= sin θ dθ = = .
V 0 V 242 4V

Now V = 32/9 so z̄ = 27π

128
≈ 0.6627 and this is approximately a third of the
maximum height (=2) of G.

2. Looking at the original integral, we see that G is the region

√
0 ≤ z ≤ 4 − x2 , 0 ≤ y ≤ 2x, 0 ≤ x ≤ 2

(a)

(b) We may re-describe G as

√
0 ≤ y ≤ 2x, 0≤x≤ 4 − z2, 0≤z≤2

This gives √
� 2 � 4−z 2 � 2x
f (x, y, z)dy dx dz
0 0 0

(c) We may redescribe G as

√ �
y/2 ≤ x ≤ 4 − z 2 , 0≤z≤ 4 − (y/2)2 , 0≤y≤4

2
This gives
� 4 � √4−(y/z)2 � √
4−z 2
f (x, y, z)dx dz dy.
0 0 y/2

(d) f = 1. Then the original integral is

√
� 2 � 2x � 4−x2 � 2 � 2x � 2
2 1/2
dz dy dx = (4 − x ) dy dx = 2x(4 − x2 )1/2 dx
0 0 0 0 0 0
�2
�
2 2 3/2 �
= − (4 − x ) � = 16/3.
3 0

The integral from (b) is

√ √
� 2 � 4−z 2 � 2x � 2 � 4−z 2
dy dx dz = 2xdx dz
0 0 0 0 0
� 2 � 2 �(4−z2 )1/2
� 2 �2
(4 − z 2 )dz = 4z − z 3 /3 0 = 16/3.
�
= x 0 dz =
0 0

The integral from (c) is

� 4 � √4−(y/z)2 � √4−z2 � 4 � √4−(y/2)2 √
dx dz dy = ( 4 − z 2 − y/2) dz dy =
0 0 y/2 0 0
� 4 � √4−(y/2)2 √ � 4 � √4−(y/2)2
= 4 − z 2 dz dy – y/2 dz dy
0 0 0 0
Trig sub’s and/or tables would be needed in order to get these to work out
in exact form.

3. If V is the volume of the full vessel, we have that

� � �
Energy = dE
V

where
dE = g z dm = 103 g z dV
since the density of water is 103 . Therefore we compute:
� � � � 100 � 9 � 81
3
Energy = 10 g z dV = 103 g z dz dx dy = . . . = 5.14·1010 joules.
1 4
V 0 −9 81
x

3
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 18.02 Problem Set 8
At MIT problem sets are referred to as ’psets’. You will see this term used occasionally
within the problems sets.
The 18.02 psets are split into two parts ’part I’ and ’part II’. The part I are all taken
from the supplementary problems. You will find a link to the supplementary problems
and solutions on this website. The intention is that these help the student develop
some fluency with concepts and techniques. Students have access to the solutions
while they do the problems, so they can check their work or get a little help as they
do the problems. After you finish the problems go back and redo the ones for which
you needed help from the solutions.
The part II problems are more involved. At MIT the students do not have access
to the solutions while they work on the problems. They are encouraged to work
together, but they have to write their solutions independently.

Part I (12 points)

At MIT the underlined problems must be done and turned in for grading.
The ‘Others’ are some suggested choices for more practice.
A listing like ’§1B : 2, 5b, 10’ means do the indicated problems from supplementary
problems section 1B.
1 Vector fields and Line Integrals in the plane.
§4A: 1d, 2bc, 3bd, 4 ; Others: 1b, 2a, 3ac
§4B: 1ab, 2b, 3; Others: 1cd, 2a

2 Fundamental Theorem for line integrals, path independence

§4C: 1, 3; Others: 2

3 Gradient fields, potential functions.

§4C: 5a(by V2 Method 1), 5b(by V2 Method 2), 6ab(by both methods); Others: 2
 Part II (23 points)

Problem 1 (3 :1,2)
a) Write down in xy-coordinates the vector field F whose vector at P = (x, y) runs
in the vertical direction from P to the line L : x + y = 1.
�
b) Show without calculation that for this field, C F · dr �= 0, if C is any positively
oriented right triangle with legs parallel to the axes and hypotenuse on L.

Problem 2 (5 :2,2,1)
xi + yj
Consider the vector field F = −c . (This is the gravitational field of an in­
x2 + y 2
finitely long uniform wire along the z-axis.). Take c = 1 and find by direct calculation
(i.e., using a parametrization of C) the work done by the gravitational field in moving
a unit mass along each of the following paths C in the xy-plane.
a) C is the half-line y = 1, x ≥ 0.
b) C is the circle of radius a with center at the origin, traced counterclockwise.
c) C is the line from (0,1) to (1,0).

Problem 3 (4 :2,2)
a) Let F = −�(ln r). Show that F is the field in problem 2, when c = 1.
�
b) Show that for any path C joining two points P1 and P2 the values of C F · dr
depends only on the ratio r2 /r1 , where ri is the distance of Pi to the origin.

Problem 4 (6 :1,2,1,2)
Let F = �(x2 y + 2xy 2 ). Let C be the quarter ellipse 9x2 + 4y 2 = 1 running from the
positive x-axis to the positive y-axis.
� �
a) Write the integral C F · dr in the form C M dx + N dy.
�
b) By parametrizing the curve C write the integral C F · dr as an explicit definitie
integral in t. (Do not evaluate.)
�
c) Use the Fundamental Theorem to (easily) compute C F · dr.

�d) Use path independence to choose a different path and (again easily) compute
C
F · dr.

Problem 5 (5 :1,2,2)
Let F = xyi + x3 j.
a) Show F is not conservative.
b) Try to find a potential function for F using method 1 from the reading (section
V2). What goes wrong?
c) Answer the same question for method 2.
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
 18.02 Problem Set 8, Part II Solutions
Problem 1
a) F = (1 − x − y)j. (See diagram (a) below.)
y y y
C2
(x, 1 − x)
F C1 C3 C1
C3
x C2 x x

diagram a diagram b(i) diagram b(ii)

b) The two possibilities for C are shown in diagrams b(i) and b(ii).
�
On C3 : (in both cases) F = 0 ⇒ C3 F · dr = 0.
�
On C2 : (in both cases) F · T = 0 ⇒ C2 F · dr = 0.
� �
On C1 : case b(i) F · T < 0 ⇒ C1 F · dr = C1 F · T ds < 0.
� � y
case b(i) F · T > 0 ⇒ C1 F · dr = C1 F · T ds > 0. 1
� � �
Therefore in both cases C F · dr = C1 +C2 +C3 F · dr = C1 F · dr =
� 0.

Problem 2 � �
x y
a) Work = F · dr = − dx − dy x
C C x2 + y 2 x2 + y 2
Path: x = x, y = 1, where 0 ≤ x < ∞ Path for problem 2a
� ∞ �∞
x 1 2
�
⇒ work = − 2 dx = − ln(x + 1)�� = −∞.
0 x +1 2 0 y
b) Path:x = a cos t, y = a sin t, 0 ≤ t ≤ 2π
� 2π � 2π
a cos t 1 sin t
⇒ work = − 2 (−a sin t) dt − (a cos t) dt = 0 dt = x
0 a a2 0 a
0.
c) Path:x = t, y = 1 − t, 0 ≤ t ≤ 1
� 1
t 1−t
⇒ work = − 2 dt − 2 (−dt) Path for problem 2b
0 2t − 2t + 1 2t − 2t + 1
� 1 �1
2t − 1 1 2
� y
=− 2 − 2t + 1
dt = − ln(2t − 2t + 1) � = 0.
0 2t 2 �
0

Problem 3 x
� ∂r x x ∂r y
a) r = x2 + y 2 ⇒ =� = , likewise = .
∂x x2 + y 2 r ∂y r Path for problem 2c
∂ ln r ∂ ln r 1 x 1 y x y
⇒ −� ln r = −� , � = −� · , · � = −� 2 , 2 �. QED
∂x ∂y r r r r x + y x + y2
2
� P2
r2
b) F · dr = − ln r|PP21 = −(ln r2 − ln r1 ) = − ln . QED
P1 r1
Problem 4 � �
a) F = �2xy + 2y 2 , x2 + 4xy� ⇒ c F · dr = c (2xy + 2y 2 ) dx + (x2 + 4xy) dy.
b) Path: x = 13 cos t, y = 12 sin t, 0 ≤ t ≤ π/2.
� � π/2 � �� � y
2 2 2 1 1/2
⇒ F · dr = cos t sin t + sin t − sin t dt
C 0 6 4 3
� �� �
1 2 4 1
+ cos t + cos t sin t cos t dt x
9 6 2 1/3
� π/2 � �
1 2 1 3 1 3 1 2
= − cos t sin t − sin t + cos t + cos t sin t dt.
0 9 6 18 3
� (0,1/2) Path for problem 4
�(0,1/2)
c) F · dr = x2 y + 2xy 2 �(1/3,0) = 0.
(1/3,0)
� � �
d) By path independence: C
F · dr = C1
F · dr + C2
F · dr. y
2 2
1/2
On C1 :y = 0, dy = 0 ⇒ (2xy + 2y ) dx + (x + 4xy) dy = 0.
C2
On C2 :x = 0, dx = 0 ⇒ (2xy + 2y 2 ) dx + (x2 + 4xy) dy = 0.
x
C1 1/3
�
⇒ each of the integrals is 0 ⇒ C F · dr = 0.

Problem 5 Path for problem 4d

a) curl F = (Nx − My )k = (3x2 − x)k �= 0 ⇒ F is not conservative.
b) We pretend F is conservative and look for a potential function f using method 1.
Since F is not conservative this will run into trouble.
Method 1. We � use the path�shown. y
3
f (x1 , y1 ) = F · dr = xy dx + x dy. (x1 , y1 )
C1 +C2 C1 +C2
On C1 : y = 0, dy = 0 ⇒ xy dx + x3 dy = 0. C2
On C2 : x = x1 , dx = 0 ⇒ xy dx + x3 dy = x31 dy.
� y1 x
⇒ f (x1 , y1 ) = x31 dy = x31 y1 ⇒ f (x, y) = x3 y. C1
0
So far so good, the trouble is �f = �3x2 y, x3 � =
� F. Path for problem 5b
c) Again pretending there is a potential function f .
x2 y
Method 2: fx = xy ⇒ f = + g(y).
2
x2 x2
fy = + g � (y) = x3 ⇒ g � (y) = x3 − .
2 2
The trouble here is that no possible function g(y) can satisfy this equation.
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.