WORKSHEETS for MATH 522

Advanced Calculus II
Instructor: Michael Taylor

Contents
1. §1.1, One variable calculus
2. §1.2, Euclidean spaces
3. §1.3, Vector spaces and linear transformations
4. §1.4, Determinants
A. Supplementary worksheet – more on determinants
5. §2.1, The derivative in several variables
6. §2.1, The derivative in several variables II
7. §2.1, Higher derivatives and power series
8. §2.1, Higher derivatives II, critical points
9. §2.2, Inverse function theorem
10. §2.2, Implicit function theorem
11. §2.3, Systems of differential equations and vector fields
12. §3.1, The Riemann integral in n variables
13. §3.1, The Riemann integral in n variables II, iterated integrals
14. §3.1, The Riemann integral in n variables III, change of variable formulas
15. §3.2, Surfaces and surface integrals
16. §3.2, Surfaces and surface integrals II

1
2

Introduction

These worksheets serve to guide the student through the text for Math 522
Introduction to Analysis in Several Variables - Advanced Calculus, by M. Taylor.
Each worksheet deals with material in a designated section of the text, and the idea
is that a student can do the exercises in a worksheet in consultation with the text,
and in that manner master the material in the text.
 3

Worksheet 1
§1.1, One variable calculus

1. Let I = [a, b] be a closed, bounded interval in R. Read the definition of a

partition P of I into intervals Jk = [xk , xk+1 ], associated to a collection of points
{xj }, satisfying
a = x0 < x1 < · · · < xN < xN +1 = b.
Write down the definitions of maxsize(P), ℓ(Jk ), and Q ≻ P.

2. Let f : I → R be a bounded function. Read (1.1.1)–(1.1.7), and write down

formulas for
I P (f ), I P (f ), I(f ), I(f ).

3. Take f as in #2. Note that

I(f ) ≤ I(f ).
We say that f is Riemann integrable, and write f ∈ R(I), provided

I(f ) = I(f ).

Then we write Z Z
b
f (x) dx = f dx = I(f ) = I(f ).
a
I

4. Proposition 1.1.1 says that f, g ∈ R(I) ⇒ f + g ∈ R(I) and

Z Z Z
(f + g) dx = f dx + g dx.
I I I

Read its proof.

5. Proposition 1.2.2 says that if f : I → R is continuous (we write f ∈ C(I)), then

f ∈ R(I). Read its proof.

6. Write down Darboux’s theorem, Theorem 1.1.4. One implication is that, if

f ∈ R(I), and if we have a sequence

Pν = {Jνk : 1 ≤ k ≤ ν}

of partitions of I, satisfying maxsize(Pν ) → 0, then

Z b X
ν
f (x) dx = lim f (ξνk ) ℓ(Jνk ),
a ν→∞
k=1

where we take arbitrary ξνk ∈ Jνk . These sums are called Riemann sums.
4

7. State the Fundamental Theorem of Calculus, given in Theorems 1.1.6 and 1.1.7,
and follow the proof. Note the use of the Mean Value Theorem in the proof of
Theorem 1.1.7.

8. Take a set S ⊂ I. Write down the definitions of upper content and outer measure,

cont+ (S), and m∗ (S),

given in (1.1.21) and (1.1.22).

9. Write down the sufficient conditions for a bounded function f : I → R to be

Riemann integrable, given in Proposition 1.1.11 and in Proposition 1.1.12. These
involve two different evaluations of the “size” of the set S of points in I at which f
is not continuous, namely cont+ (S) and m∗ (S).

10. Note the example of a bounded function that is not Riemann integrable, given in
(1.1.16). Note the examples of bounded, discontinuous functions that are Riemann
integrable, given in Exmples 1 and 2, after Proposition 1.1.11.

11. Do exercise 4, at the end of §1.1.

 5

Worksheet 2
§1.2, Euclidean spaces

1. Define the vector operations on Rn .

2. Given x, y ∈ Rn , define the dot product x · y.

3. Given x ∈ Rn , we define the norm |x| ∈ [0, ∞) by

√
|x| = x · x.

Consult Proposition 1.2.1 and show that the triangle inequality

|x + y| ≤ |x| + |y|

follows from Cauchy’s inequality

|x · y| ≤ |x| |y|.

4. Consult Proposition 1.2.2 for the proof of Cauchy’s inequality.

5. Given pj ∈ Rn , define what it means to say

pj converges to p as j → ∞.
(pj ) is Cauchy.

6. Given S ⊂ Rn , define what it means to say

S is closed,
S is open.

7. Given x, y ∈ Rn , we say

x ⊥ y ⇐⇒ x · y = 0.

Show that
|x + y|2 = |x|2 + |y|2 ⇐⇒ x ⊥ y.
8. Check out the notion of compactness, the Bolzano-Weierstrass theorem, and the
Heine-Borel theorem.
6

Worksheet 3
§1.3, Vector spaces and linear transformations

1. Define the concept of a vector space (over F = R or C). Note that Rn is a vector
space over R and Cn is a vector space over C. (Fn is a vector space over F.)

2. Let S = {v1 , . . . , vk } ⊂ V , a vector space. Define what it means to say

S spans V ,
S is linearly independent,
S is a basis of V .

3. Study Lemma 1.3.1 and Proposition 1.3.2, whose content is that

If V has a basis {v1 , . . . , vk } and if {w1 , . . . , wℓ } ⊂ V is linearly independent, then
ℓ ≤ k.
Show that this leads to Corollary 1.3.3:
If V is finite dimensional, then any two bases of V have the same number of
elements.
In such a case, dim V denotes the number of elements in a basis of V .

4. State Propositions 1.3.4 and 1.3.5.

5. State Proposition 1.3.6, the Fundamental Theorem of Linear Algebra, and show
how this follows from Propositions 1.3.4 and 1.3.5.

6. Deduce from the Fundamental Theorem of Linear Algebra that if V is finite

dimensional and A : V → V is linear, then

A injective ⇔ A surjective ⇔ A isomorphism.

7. State Proposition 1.3.9, characterizing when a matrix A ∈ M (n, F) is invertible,

in terms of the behavior of its columns.
 7

Worksheet 4
§1.4, Determinants

1. Given  
a b
A= , we set det A = ad − bc.
c d
Show that A : F2 → F2 is invertible if and only if det A ̸= 0. If this holds,
 
−1 1 d −b
A = .
det A −c a

2. Consult Proposition 1.4.1 and define det A for A ∈ M (n, F).

3. Show that the formula (1.4.30) for det A implies

det A = det At .

4. Read the proof of Proposition 1.4.3, that if A, B ∈ M (n, F),

det(AB) = (det A)(det B).

Show that this implies Corollary 1.4.4, i.e.,

A invertible =⇒ det A ̸= 0.

5. Read the proof of Proposition 1.4.6, that if A ∈ M (n, F),

A invertible ⇐⇒ det A ̸= 0.

See how this completes the result of Exercise 4.

6. Study Exercises 1–3 at the end of §1.4, treating the expansion of det A by minors
down the kth column, given A ∈ M (n, F).

7. Use an expansion by minors to evaluate det A for

 
1 1 1

A= 2 0 2 .
3 0 −1
8

Supplementary worksheet
§1.4, More on determinants

1. Verify the following method of computing 3 × 3 determinants. Given

 
a11 a12 a13
A =  a21 a22 a23  ,
a31 a32 a33

form a 3 × 5 rectangular matrix by copying the first two columns of A to the

right. The products in (1.4.16) with plus signs are the products of each of the three
downward sloping diagonals marked in bold below.
 
a11 a12 a13 a11 a12
 a21 a22 a23 a21 a22  .
a31 a32 a33 a31 a32

The products in (1.4.16) with minus signs are the products of each of the three
upward sloping diagonals marked in bold below.
 
a11 a12 a13 a11 a12
 a21 a22 a23 a21 a22  .
a31 a32 a33 a31 a32

2. Use the method described above to compute the determinants of

     
1 2 3 1 2 3 1 1 1
2 1 2, 1 2 3, 2 0 2 .
3 2 1 2 1 3 3 0 −1

3. Given A = (0 1 2), compute det At A and det AAt .

4. Compute the determinant of

 
1 2 3 4
0 2 3 4
 .
0 0 3 4
0 0 0 4
 9

Worksheet 5
§2.1, The derivative in several variables

As defined in §2.1, if O ⊂ Rn is open, a function F : O → Rm is differentiable at

x ∈ O, with derivative DF (x) ∈ L(Rn , Rm ), if and only if

(1) F (x + y) = F (x) + DF (x)y + R(x, y), R(x, y) = o(∥y∥).

One compares this with the partial derivative,

∂F 1 
(2) (x) = lim F (x + hej ) − F (x) ,
∂xj h→0 h

where {e1 , . . . , en } is the standard basis of Rn .

1. Verify the identity (2.1.4) connecting (1) and (2), when F is differentiable.

2. Study the argument in (2.1.12)–(2.1.14) that, if

S : M (n, R) −→ M (n, R), S(X) = X 2 ,

then
DS(X)Y = XY + Y X.
Then look at Exercise 2, at the end of §2.1. Show that

P3 (X) = X 3 =⇒ DP3 (X)Y = Y X 2 + XY X + X 2 Y.

3. Study the argument in (2.1.16)–(2.1.21) that, if

Φ : Gl(n, R) −→ M (n, R), Φ(X) = X −1 ,

then
DΦ(I)Y = −Y.
Note the use of the evaluation of the infinite series
∞
X
−1
(I + Y ) = (−1)k Y k , for ∥Y ∥ < 1.
k=0

Going further, as indicated in Exercise 3 at the end of §2.1, show that

DΦ(X)Y = −X −1 Y X −1 .
10

Worksheet 6
§2.1, The derivative in several variables II

1. Proposition 2.1.1 says that if F is of class C 1 (i.e., ∂F/∂xj is continuous on O

for each j) then F is differentiable at each x ∈ O. Study its proof.

2. The chain rule is given in (2.1.25). If F : O → U and G : U → Rk are

differentiable, then G ◦ F is differentiable, and

D(G ◦ F )(x) = DG(F (x))DF (x).

Study its proof.

3. Look at Exercise 0 at the end of §2.1, dealing with a strengthening of Proposition

4.1.1.

4. Look at Exercise 14 at the end of §2.1, describing a function that is not differ-
entiable at (0, 0) ∈ R2 , despite the fact that both its partial derivatives exist (but
they are not continuous).

5. Proposition 2.1.2 says that if O ⊂ Rn is open and F : O → Rm is of class C 2 ,

then

(1) ∂j ∂k F (x) = ∂k ∂j F (x), x ∈ O, j, k ∈ {1, . . . , n}.

Study the proof. It involves difference quotients and the mean value theorem.

6. Check out Exercise 15, at the end of §2.1, regarding a function g ∈ C 1 (R2 ) for
which ∂x ∂y g and ∂y ∂x g exist at each point of R2 , but

∂x ∂y g(0, 0) ̸= ∂y ∂x g(0, 0),

in contrast to (1) above. (In this case, g ∈

/ C 2 .)
 11

Worksheet 7
§2.1, Higher derivatives and power series

1. Formulas (2.1.35)–(2.1.36) present two different multi-index notations for deriva-

tives of order k:

f (J) (x) = ∂jk · · · ∂j1 f (x), J = (j1 , . . . , jk ), |J| = k,

(1)
f (α) (x) = ∂1α1 · · · ∂nαn f (x), α = (α1 , . . . , αn ), |α| = α1 + · · · + αn = k.

Become familiar with these notations.

2. Review the treatment of power series for functions of one variable in Chapter 1,
§1.1, Exercises 9–10. The formula (1.1.50) can be written

X
k
f (j) (0)
(2) f (t) = tj + Rk (t), t ∈ I = (−R, R),
j=0
j!

and two integral formulas are derived for the remainder Rk (t), depending on whether
f ∈ C k+1 (I) or f ∈ C k (I). See (1.1.51) and (1.1.54).
Note also the Cauchy and Lagrange formulas for the remainder, given in Appen-
dix A.4.

3. Study formulas (2.1.39)–(2.1.45), leading to the power series formula

X 1 (J)
(3) F (x) = F (0)xJ + Rk (x),
|J|!
|J|≤k

with the remainder Rk (x) given by (2.1.46).

4. Study formulas (2.1.47)–(2.1.52), leading to the power series formula

X 1
(4) F (x) = F (α) (0)xα + Rk (x),
α!
|α|≤k

with Rk (x) given by (2.1.53).

5. The results (2.1.46) and (2.1.53) require F to be of class C k+1 . Study Proposition
2.1.5, which establishes (4), with Rk (x) given by (2.1.56), for F of class C k .
12

Worksheet 8
§2.1, Higher derivatives II, critical points

1. Study formulas (2.1.58)–(2.1.62), in which it is established that, if O ⊂ Rn is

open and F : O → R is C 2 , then, for y ∈ O,

1
(1) F (x) = F (y) + DF (y)(x − y) + (x − y) · D2 F (y)(x − y) + R2 (x, y),
2

where D2 F (y) is the n × n Hessian matrix, given by (2.1.59), and

(2) R2 (x, y) = o(|x − y|2 ).

2. In the setting of Exercise 1, we say x0 ∈ O is a critical point of F if DF (x0 ) = 0.

Proposition 2.1.6 says that if F : O → R is C 2 and x0 ∈ O is a critical point, then
(i) D2 F (x0 ) positive definite ⇒ F has a local min at x0 ,
(ii) D2 F (x0 ) negative definite ⇒ F has a local max at x0 ,
(iii) D2 F (x0 ) strongly indefinite ⇒ x0 is a saddle point for F .
Show how this result follows from (1)–(2) above.

3. Study Proposition 2.1.7, characterizing when a matrix A = At ∈ M (n, R) is

positive definite, in terms of the behavior of the determinants of all the ℓ × ℓ upper
left submatrices, 1 ≤ ℓ ≤ n.

4. Returning to the setting of Exercise 1 above, check out the remainder formula
in (2.1.72).

5. Check out Proposition 2.1.10, regarding smoothness and derivatives of functions

given by convergent power series on a domain Re ⊂ Rn .
 13

Worksheet 9
§2.2, Inverse function theorem

1. Theorem 2.2.1 says that if Ω ⊂ Rn is open,

F : Ω → Rn is C 1 ,

p0 ∈ Ω, and DF (p0 ) is invertible, then F maps some open neighborhood U of p0

one-one and onto a neighborhood V of q0 = F (p0 ), and the inverse map

F −1 : V −→ U is C 1 .

Study its proof.

2. Proposition 2.2.2 gives a condition that guarantees that a C 1 map f : Ω → Rn

be one-one. State it and write down its proof. Describe the role it plays in the
proof of Theorem 2.2.1.

3. Theorem 2.2.3 is called the contraction mapping theorem. State it and write
down its proof. Describe its role in the proof of Theorem 2.2.1.

4. The map F : (0, ∞) × R → R2 in (2.2.21) defines polar coordinates. Study how

it illustrates the inverse function theorem.

5. Compare the iterative method (2.2.19) for solving F (x) = y for x with that
given in Exercise 1 at the end of §2.2 (Newton’s method).

6. Do Exercise 2 at the end of §2.2.

14

Worksheet 10
§2.2, Implicit function theorem

1. Explain how
x2 + y 2 = 1
defines y implicitly as a smooth function of x, in two ways, for x ∈ (−1, 1).

2. Theorem 2.2.5 is the implicit function theorem. It says that if x0 ∈ U , open in

Rm , y0 ∈ V , open in Rℓ , and

(1) F : U × V −→ Rℓ is C k , F (x0 , y0 ) = u0 ,

and if

(2) Dy F (x0 , y0 ) is invertible

(as an ℓ × ℓ matrix), then the equation

(3) F (x, y) = u0

defines

(4) y = g(x, u0 ),

for x near x0 (satisfying g(x0 , u0 ) = y0 ), and g is a C k map.

Study the proof given in (2.2.39)–(2.2.45), which brings in

   
x x0
(5) H :U ×V →R ×R ,m ℓ
H(x, y) = , H(x0 , y0 ) = .
F (x, y) u0

See from
 
I 0
(6) DH =
Dx F Dy F

that the hypothesis (2) implies DH(x0 , y0 ) is invertible. See how the inverse func-
tion theorem (Theorem 2.2.1) yields a smooth inverse

(7) G : O −→ U × V

to H, where O is a neighborhood of (x0 , u0 ) in Rm × Rℓ , and that G(x, u) has the

form
 
x
(8) G(x, u) = ,
g(x, u)
 15

yielding the identity

(9) F (x, g(x, u)) = u,

and hence satisfying (3)–(4).

3. Check out Proposition 2.2.6, which treats a C k map

F : Ω −→ R, F (x0 ) = u0 , Ω ⊂ Rn open,

under the hypothesis that

∇F (x0 ) ̸= 0.
It shows that if in particular
∂n F (x0 ) ̸= 0,
then you can solve F (x) = u0 for

xn = g(x1 , . . . , xn−1 ),

with (x10 , . . . , xn−1,0 , xn0 ) = x0 , for a C k function g.

4. Check out the relevance of the material of #3 to the example introduced in #1.

5. Do Exercise 7, at the end of §2.2.

16

Worksheet 11
§2.3, Systems of differential equations and vector fields

Material to be provided later.

 17

Worksheet 12
§3.1, The Riemann integral in n variables

1. Let R = I1 × · · · × In be a cell in Rn , where each Iν = [aν , bν ] is a closed,

bounded interval in R. Study the definition of a partition P of R, into cells Rα ,
α = (α1 , . . . , αn ), given in the beginning of §3.1. Write down the definitions of
maxsize(P), V (Rα ), and Q ≻ P.

2. Let f : R → R be a bounded function. Read (3.1.1)–(3.1.8), and write down

formulas for
I P (f ), I P (f ), I(f ), I(f ).

3. Take f as in #2. Note that

I(f ) ≤ I(f ).
We say f is Riemann integrable, and write f ∈ R(R), provided
I(f ) = I(f ).
Then we write Z
f (x) dV (x) = I(f ) = I(f ).
R

4. Proposition 3.1.1 is the multi-D Darboux theorem. One implication is that if

f ∈ R(R), and if we have a sequence
Pν = {Rνα : α ∈ Sν }
of partitions of R, satisfying maxsize(Pν ) → 0, then
Z X
f (x) dV (x) = lim f (ξνα ) V (Rνα ),
ν→0
R α∈Sν

where we take arbitrary ξνα ∈ Rνα . These sums are called Riemann sums. Compare
this result with Theorem 1.1.4. A key ingredient behind this result is that, whenever
g : R → R is bounded,
I Pν (g) → I(g), I Pν (g) → I(g).

5. Proposition 5.1.2 says that fj ∈ R(R), cj ∈ R ⇒ c1 f1 + c2 f2 ∈ R(R) and

Z Z Z
(c1 f1 + c2 f2 ) dV = c1 f1 dV + c2 f2 dV.
R R R

Compare this result with Proposition 1.1.1.

6. Proposition 3.1.3 says that if f : R → R is continuous, then f ∈ R(R). Read its

proof. Compare this with Proposition 1.1.2.
18

7. Take a set S ⊂ R, where R is a cell. Write down the definitions of upper content,
lower content, and volume,

cont+ (S), cont− (S), and V (S),

given in (3.1.13)–(3.1.14).

8. A set S ⊂ R is called contented provided cont+ (S) = cont− (S), in which case
the common value is denoted V (S). Proposition 3.1.4 says that a set S ⊂ R is
contented if and only if its boundary bS satisfies

cont+ (bS) = 0.

Read its proof.

9. Proposition 3.1.6 says that if f : R → R is bounded and if S is the set of its

points of discontinuity, then

cont+ (S) = 0 =⇒ f ∈ R(R).

Read its proof.

Note. A stronger result is established in Proposition 3.1.31, namely m∗ (S) = 0 ⇒
f ∈ R(R).

10. Propositions 3.1.7–3.1.8 give sufficient conditions that a set S ⊂ R have upper
content 0 (we then say S is a nil set). Write down these conditions.
 19

Worksheet 13
§3.1, The Riemann integral in n variables II
Iterated integrals

1. Theorem 3.1.9 takes a closed, bounded, contented set Σ ⊂ Rn−1

Ω = {(x, y) ∈ Rn : x ∈ Σ, g0 (x) ≤ y ≤ g1 (x)},

where gj are continuous on Σ, g0 < g1 . One is given f : Ω → R, continuous, and

Z g1 (x)
φ(x) = f (x, y) dy,
g0 (x)

which is continuous on Σ. The conclusion is that

Z Z
f dVn = φ dVn−1 .
Ω Σ

Study its proof. Note the role of Proposition 3.1.8. A corollary of Theorem 3.1.9 is
Z
V (Ω) = [g1 (x) − g0 (x)] dx.
Σ

2. Study the application of these results to

Z 1 p
A(D) = 2 1 − x2 dx = π.
−1

Here, D = {(x, y) ∈ R2 : x2 + y 2 ≤ 1}.

3. Study the application of results of # 1 to

Z p
n
V (B ) = 2 1 − |x|2 dx,
B n−1

where B n = {x ∈ Rn : |x| ≤ 1}, hence to

Z p
3
V (B ) = 2 1 − |x|2 dx
D
Z Z √
1 1−x2 p
=2 √ 1 − x2 − y 2 dy dx
−1 − 1−x2
4
= π.
3
20

4. Here is an extension of Theorem 3.1.9. We take Σ ⊂ Rk , closed, bounded, and

contented, gj : Σ → [0, ∞) continuous, g0 < g1 , and

Ω = {(x, y) ∈ Rn : x ∈ Σ, y ∈ Rn−k , g0 (x) ≤ |y| ≤ g1 (x)}.

We take f : Ω → R continuous. The conclusion is that

Z
φ(x) = f (x, y) dy
g0 (x)≤|y|≤g1 (x)

is continuous on Σ and Z Z
f dVn = φ dVk .
Ω Σ

Consider how the proof of Theorem 3.1.9 adapts to this situation.

5. Consider the application of #4 to solids of revolution

p
Ω = {(x, y, z) ∈ R3 : a ≤ x ≤ b, y 2 + z 2 ≤ g(x)},

including
Z b
V (Ω) = π g(x)2 dx,
a

and a derivation of the formula

Z 1
4
3
V (B ) = π (1 − x2 ) dx = π.
−1 3

6. Study the application of #4 to the recursive formula

Z 1
n
V (B ) = βn V (B n−1
), βn = (1 − x2 )(n−1)/2 dx,
−1

including
Z π/2
4 3
V (B ) = β4 V (B ), β4 = 2 cos4 t dt.
0

7. State the general Fubini theorem, Proposition 3.1.32. See how it implies Theorem
5.1.10 and the result of #4.
 21

Worksheet 14
§3.1, The Riemann integral in n variables III
Change of variable formulas

The central result of this part of §3.1, encompassing Propositions 3.1.10–3.1.14, is

Theorem 3.1.15, which says the following. Take

O, Ω ⊂ Rn open, G : O → Ω a C 1 diffeomorphism.

Assume f : Ω → R is supported on a compact subset of Ω and is Riemann integrable

(we say f ∈ Rc (Ω)). Then f ◦ G ∈ Rc (O), and
Z Z
(1) f (y) dV (y) = f (G(x)) | det DG(x)| dV (x).
Ω O

This is established in stages.

1. Proposition 3.1.10 says that if f : Rn → R is continuous and has compact

support (we say f ∈ Cc (Rn )), and if A ∈ Gℓ(n, R) (invertible matrix), then
Z Z
(2) f (x) dx = | det A| f (Ax) dx.

Study its proof. Note the role of Proposition 3.1.9.

2. Proposition 3.1.11 establishes other characterizations of I(f ) and I(f ), given

bounded f : R → R. In particular,
nZ o
I(f ) = inf g dV : g ∈ C(R), g ≥ f ,
R
(3) nZ o
I(f ) = sup g dV : g ∈ C(R), g ≤ f .
R

Study its proof.

3. See how (3) leads to the extension of (2) to all compactly supported Riemann
integrable functions f on Rn (we say f ∈ Rc (Rn )), in Proposition 3.1.12.
Note also Corollary 3.1.13:

(4) V (A(Σ)) = | det A| V (Σ),

when Σ ⊂ Rn is compact and contented, and A ∈ Gℓ(n, R).

4. Proposition 3.1.14 establishes Theorem 3.1.15, under the additional hypothesis

that f is continuous. Study its proof, following (3.1.48)–(3.1.56).
22

5. See how Theorem 3.1.15 is derived from Proposition 3.1.14, via (3).

6. Check out the use of polar coordinates to do double integrals, in (3.1.62)–(3.1.65),

including the formula
Z Z ρ Z 2π
(5) f (x, y) dx dy = f (r cos θ, r sin θ)r dθ dr,
0 0
Dρ

where Dρ = {(x, y) ∈ R2 : x2 + y 2 ≤ ρ2 }. Note that one can use polar coordinates

to establish
A(Dρ ) = πρ2 ,
and also to compute V (B 3 ),
Z p
3
V (B ) = 2 1 − x2 − y 2 dx dy
D
Z 2π Z 1 p
(6)
= 1 − r2 r dr dθ
0 0
4
= π.
3
Compare approaches described in #3 and #5 of Worksheet 13.

7. Study the use of spherical polar coordinates on R3 in Exercise 5 of §3.1, to obtain

Z Z 2π Z π Z 1
(7) f (x) dV (x) = f (G(ρ, θ, ψ))ρ2 sin θ dρ dθ dψ.
0 0 0
B3

See from this yet a fourth derivation of the formula (6).

8. Check out the extension of various results to integrals over all of Rn . In partic-
ular, see (3.1.71)–(3.1.74), deriving the identity
Z ∞ √
e−x dx = I =
2
(8) π
−∞

by representing I 2 as an integral over R2 and switching to polar coordinates. More

generally, show that
Z
e−|x| dx = I n = π n/2 .
2
(9)
Rn
 23

Worksheet 15
§3.2, Surfaces and surface integrals

1. A C k smooth m-dimensional surface M in Rn is covered by coordinate charts.

Each p ∈ M has a neighborhood U ⊂ M for which there is a C k map

(1) φ : O −→ U, one-to-one and onto,

with O ⊂ Rm open, such that

(2) Dφ(x) : Rm −→ Rn is injective, for all x ∈ O.

Given p ∈ U , we set

Tp M = Range Dφ(x0 ), φ(x0 ) = p,

(3)
Np M = ⊥ complement of Tp M in Rn .

If

(4) ψ : Ω −→ U, one-to-one and onto,

is another C k coordinate chart, we set

(5) F = ψ −1 ◦ φ : O −→ Ω.

Check out Lemma 3.2.1, saying F is a C k diffeomorphism, and study its proof.
Note the role of the Inverse Function Theorem. See that also

(6) Tp M = Range Dψ(x1 ), if ψ(x1 ) = p.

2. If M ⊂ Rn is a smooth m-dimensional surface, as in #1, we associate to each

coordinate chart φ, as in (1), an m × m matrix function of the form

(7) G(x) = Dφ(x)t Dφ(x),

called a metric tensor. See in (3.2.8) the connection with the inner product on Tp M
(p = φ(x)), inherited from the dot product on Rn . Note that

(7A) v · G(x)w = Dφ(x)w · Dφ(x)v,

for v, w ∈ Rm , and that

(7B) G(x) = gjk (x) , gjk (x) = ∂j φ(x) · ∂k φ(x).

Show via the chain rule that if we have another coordinate chart ψ (connected
with φ via (5)), with metric tensor

(8) H(y) = Dψ(y)t Dψ(y),

24

then

(9) G(x) = DF (x)t H(y) DF (x), for y = F (x).

3. If f : M → R is a continuous function supported in a coordinate patch U (as in

(1)), we define the surface integral of f as
Z Z p
(10) f dS = f (φ(x)) g(x) dx, g(x) = det G(x),
M O

with G(x) given by (7). Check out (3.2.14)–(3.2.15) to see that if ψ : Ω → U in (4)
is another coordinate chart, then (10) is equal to
Z p
(11) f (ψ(y)) h(y) dy, h(y) = det H(y),
Ω

with H(y) given by (8). Note the role of the change of variable formula, addressed
in Worksheet 14.

4. Check out (3.2.18)–(3.2.20), which says that if M ⊂ R3 is a 2D surface with

coordinate chart φ : O → U ⊂ M , and f is supported on U , then
Z Z
(12) f dS = f (φ(x)) |∂1 φ × ∂2 φ| dx1 dx2 .
M O

5. Check out (3.2.21)–(3.2.22), which says that if Ω ⊂ Rn−1 is open and M ⊂ Rn

is the graph of z = u(x), then

(13) φ(x) = (x, u(x))

provides a coordinate chart, in which the metric tensor formula (7B) becomes

∂u ∂u
(14) gjk (x) = δjk + ,
∂xj ∂xk

and in such a case

√
(15) g = (1 + |∇u|2 )1/2 .
 25

Worksheet 16
§3.2, Surfaces and surface integrals II

1. Show that the unit sphere S n−1 = {x ∈ Rn : |x| = 1} is a smooth surface. Show
that solving
x21 + · · · + x2k + · · · + x2n = 1
for xk yields coordinate charts

φ±
k :B
n−1
−→ Uk± ⊂ S n−1 ,

where
Uk± = {x ∈ S n−1 : ±x · ek > 0},
and {e1 , . . . , en } denotes the standard orthonormal basis of Rn .

2. Writing φ+
n in #1 as

p
φ+
n (x) = (x, u(x)) = (x, 1 − |x|2 ), x ∈ B n−1 ,

show that
|x|2
|∇u(x)| = , 2
1 − |x|2
and hence, from #5 of Worksheet 15,
p
g(x) = (1 − |x|2 )−1/2 ,

in this coordinate system. Deduce that the area An−1 of S n−1 is given by
Z
An−1 = 2 (1 − |x|2 )−1/2 dx.
B n−1

3. Deduce from #2 that

Z
A2 = 2 (1 − |x|2 )−1/2 dx
D
Z 2π Z 1
=2 (1 − r2 )−1/2 r dr dθ
0 0
= 4π.

4. Check out the argument in (3.2.24)–(3.2.28), yielding

Z Z hZ ∞ i
(4) f (x) dx = f (rω)rn−1 dr dS(ω),
0
Rn S n−1
26

for f integrable on Rn . If f is radial, i.e., f (x) = φ(|x|), deduce that

Z Z ∞
(5) φ(|x|) dx = An−1 φ(r)rn−1 dr,
0
Rn

where An−1 is the area of S n−1 .

5. Deduce from (5) that

1
(6) V (B n ) = An−1 .
n
In particular, V (B 3 ) = A2 /3. Compare the computations of V (B 3 ) in worksheets
20–21 and the computation of A2 in #3 above.

6. Combine (2), with n − 1 replaced by n, with (5) to show that

Z 1
(7) An = 2An−1 (1 − r2 )−1/2 rn−1 dr.
0

Use this with n = 2 to relate A2 to A1 = 2π. Compare the calculation in #3 above.

7. See #8 of Worksheet 14 for a derivation of the identity

Z
e−|x| dx = I n = π n/2 .
2
(8)
Rn

Take φ(r) = e−r in (5), and follow the arguments in (3.2.30)–(3.2.32) to see the
2

formula
2π n/2
(9) An−1 = ,
Γ(n/2)

where Γ(z) is Euler’s gamma function, defined for z > 0 by

Z ∞
(10) Γ(z) = e−s sz−1 ds.
0

8. Study the treatment of Γ(z) in (3.2.33)–(3.2.37), including Lemma 3.2.2, which

uses integration by parts to establish

(11) Γ(z + 1) = zΓ(z), z > 0,

and also the particular identities

1
(12) Γ(1) = 1, Γ = π 1/2 , Γ(k) = (k − 1)!
2
See also how these identities together with (9) yield the formulas

2π k 2π k
(13) A2k−1 = , A2k = .
(k − 1)! (k − 12 ) · · · ( 12 )
 27

9. Study Proposition 3.2.5, and its special case, arising in Exercise 10 at the end
of §3.2, which says that if Ω ⊂ Rn is open, c ∈ R, and

u : Ω −→ R is C k ,
(14) S = {x ∈ Ω : u(x) = c},
S ̸= ∅, and x ∈ S ⇒ ∇u(x) ̸= 0,

then S is a C k smooth, (n − 1)-dimensional surface. Also, for x ∈ S,

Nx S = span of ∇u(x),
(15)
Tx S = ⊥ complement of Nx S.

Apply this to

(16) S n−1 = {x ∈ Rn : u(x) = 1}, u(x) = |x|2 .

10. Apply #9 to
S = {(x′ , f (x′ )) : x′ ∈ Rn−1 },
which takes the form (14) with

u(x) = xn − f (x′ ), x′ = (x1 , . . . , xn−1 ), c = 0.