MATH2023 Multivariable Calculus Problem Set 6

L3 (Fall 2019)

Solution to Extra Problem Set 6

1. (a) The gradient field of the air pressure function is

∇𝑃(𝑥, 𝑦) = 〈𝑃𝑥 (𝑥, 𝑦), 𝑃𝑦 (𝑥, 𝑦)〉 = 〈−12𝑥, −4𝑦〉.
The direction of the wind at (−1, 3) is the direction of steepest descent of 𝑃, which is given by
−∇𝑃(−1, 3) = −〈−12(−1), −4(3)〉 = 〈−12,12〉.
A unit vector indicating this direction is
1 1 1
〈−12,12〉 = 〈− , 〉.
√(−12)2 + 122 √2 √2
(b) The path traced by the smoke particle is a field line of the vector field −∇𝑃. On solving the differential
equation
𝑑𝑥 𝑑𝑦
= ,
12𝑥 4𝑦
1 1
we obtain ln|𝑥| = ln|𝑦| + 𝐶0 , i.e. 𝑥 = 𝐶𝑦 3 , where 𝐶 is a non-zero constant. Since the smoke particle is
12 4
emitted at (−1, 3), the required field line has to pass through the point (−1, 3), and so we have −1 = 𝐶(33 ),
1
i.e. 𝐶 = − . Therefore the required path has equation
27

1 3
𝑥=− 𝑦 ,
27
where 𝑦 ≥ 3.

2. Let
𝐅(𝑥, 𝑦, 𝑧) = 〈𝑢(𝑥, 𝑦, 𝑧), 𝑣(𝑥, 𝑦, 𝑧), 𝑤(𝑥, 𝑦, 𝑧)〉.
Then 𝑓𝐅 = 〈𝑓𝑢, 𝑓𝑣, 𝑓𝑤〉, and so its curl and divergence are as follows:
(a)
𝜕 𝜕 𝜕 𝜕𝑓 𝜕𝑢 𝜕𝑓 𝜕𝑣 𝜕𝑓 𝜕𝑤
∇ ⋅ (𝑓𝐅) = (𝑓𝑢) + (𝑓𝑣) + (𝑓𝑤) = ( 𝑢 + 𝑓 ) + ( 𝑣 + 𝑓 ) + ( 𝑤 + 𝑓 )
𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑧 𝜕𝑧

𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑢 𝜕𝑣 𝜕𝑤
=( 𝑢+ 𝑣+ 𝑤) + (𝑓) ( + + ) = (∇𝑓) ⋅ 𝐅 + (𝑓)(∇ ⋅ 𝐅).
𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑦 𝜕𝑧

(b)
𝐢 𝐣 𝐤
𝜕 𝜕 𝜕 𝜕 𝜕 𝜕 𝜕 𝜕 𝜕
∇ × (𝑓𝐅) = | | = 〈 (𝑓𝑤) − (𝑓𝑣), (𝑓𝑢) − (𝑓𝑤), (𝑓𝑣) − (𝑓𝑢)〉
𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑦 𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑦
𝑓𝑢 𝑓𝑣 𝑓𝑤
𝜕𝑓 𝜕𝑤 𝜕𝑓 𝜕𝑣 𝜕𝑓 𝜕𝑢 𝜕𝑓 𝜕𝑤 𝜕𝑓 𝜕𝑣 𝜕𝑓 𝜕𝑢
=〈 𝑤+𝑓 − 𝑣−𝑓 , 𝑢+𝑓 − 𝑤−𝑓 , 𝑣+𝑓 − 𝑢−𝑓 〉
𝜕𝑦 𝜕𝑦 𝜕𝑧 𝜕𝑧 𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦

𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑤 𝜕𝑣 𝜕𝑢 𝜕𝑤 𝜕𝑣 𝜕𝑢
=〈 𝑤− 𝑣, 𝑢 − 𝑤, 𝑣 − 𝑢〉 + (𝑓) 〈 − , − , − 〉
𝜕𝑦 𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑦

𝐢 𝐣 𝐤 𝐢 𝐣 𝐤
𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕 𝜕 𝜕
=| | + (𝑓) | | = (∇𝑓) × 𝐅 + (𝑓)(∇ × 𝐅).
𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑦 𝜕𝑧
𝑢 𝑣 𝑤 𝑢 𝑣 𝑤

Page 1 of 10
MATH2023 Multivariable Calculus Problem Set 6
L3 (Fall 2019)

3. (a) Let 𝐜 ≔ 〈𝑎, 𝑏, 𝑐〉. Then 𝐜 × 𝐫 = 〈𝑏𝑧 − 𝑐𝑦, 𝑐𝑥 − 𝑎𝑧, 𝑎𝑦 − 𝑏𝑥〉, so

𝜕 𝜕 𝜕
∇ ⋅ (𝐜 × 𝐫) = (𝑏𝑧 − 𝑐𝑦) + (𝑐𝑥 − 𝑎𝑧) + (𝑎𝑦 − 𝑏𝑥) = 0
𝜕𝑥 𝜕𝑦 𝜕𝑧

and
𝐢 𝐣 𝐤
𝜕 𝜕 𝜕
∇ × (𝐜 × 𝐫) = | | = 〈2𝑎, 2𝑏, 2𝑐〉 = 2𝐜.
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝑏𝑧 − 𝑐𝑦 𝑐𝑥 − 𝑎𝑧 𝑎𝑦 − 𝑏𝑥

(b) We have 𝑓(𝑟)𝐫 = 𝑓(√𝑥 2 + 𝑦 2 + 𝑧 2 )〈𝑥, 𝑦, 𝑧〉, so

𝜕 𝜕 𝜕
∇ ⋅ (𝑓(𝑟)𝐫) = (𝑥𝑓 (√𝑥 2 + 𝑦 2 + 𝑧 2 )) + (𝑦𝑓 (√𝑥 2 + 𝑦 2 + 𝑧 2 )) + (𝑧𝑓 (√𝑥 2 + 𝑦 2 + 𝑧 2 ))
𝜕𝑥 𝜕𝑦 𝜕𝑧

𝑥 2 𝑓 ′ (√𝑥 2 + 𝑦 2 + 𝑧 2 ) 𝑦 2 𝑓 ′ (√𝑥 2 + 𝑦 2 + 𝑧 2 )
= 𝑓 (√𝑥 2 + 𝑦 2 + 𝑧 2 ) + + 𝑓 (√𝑥 2 + 𝑦 2 + 𝑧 2 ) +
√𝑥 2 + 𝑦 2 + 𝑧 2 √𝑥 2 + 𝑦 2 + 𝑧 2
𝑧 2 𝑓 ′ (√𝑥 2 + 𝑦 2 + 𝑧 2 )
+ 𝑓 (√𝑥 2 + 𝑦 2 + 𝑧 2 ) +
√𝑥 2 + 𝑦 2 + 𝑧 2
= √𝑥 2 + 𝑦 2 + 𝑧 2 𝑓 ′ (√𝑥 2 + 𝑦 2 + 𝑧 2 ) + 3𝑓 (√𝑥 2 + 𝑦 2 + 𝑧 2 ) = 𝑟𝑓 ′ (𝑟) + 3𝑓(𝑟).
Hence, if the vector field 𝑓(𝑟)𝐫 has zero divergence on ℝ3 ∖ {(0, 0, 0)}, then
𝑟𝑓 ′ (𝑟) + 3𝑓(𝑟) = 0.
This differential equation can be rewritten as
𝑓 ′ (𝑟) 3
=− ,
𝑓(𝑟) 𝑟
so ln 𝑓(𝑟) = −3 ln 𝑟 + 𝐶0 for some constant 𝐶0 . Taking exponentials on both sides, we have
𝑓(𝑟) = 𝐶𝑟 −3
for some positive constant 𝐶.

4. Since 𝑓: ℝ3 → ℝ is a potential function of the conservative vector field 𝐅: ℝ3 → ℝ3 , we have 𝐅 = ∇𝑓 on ℝ3 .

Now writing
𝐫(𝑡) = 〈𝑥(𝑡), 𝑦(𝑡), 𝑧(𝑡)〉
and observing that ‖𝐫 ′ (𝑡)‖2 = 𝐫 ′ (𝑡) ⋅ 𝐫 ′ (𝑡), we have

𝑑 1
𝐸 ′ (𝑡) = [ 𝑚𝐫 ′ (𝑡) ⋅ 𝐫 ′ (𝑡) − 𝑓(𝑥(𝑡), 𝑦(𝑡), 𝑧(𝑡))]
𝑑𝑡 2
1 ′′ (𝑡)
𝜕𝑓 𝑑𝑥 𝜕𝑓 𝑑𝑦 𝜕𝑓 𝑑𝑧
= 𝑚 (𝐫
⏟ ⋅ 𝐫 ′ (𝑡) + 𝐫 ′ (𝑡) ⋅ 𝐫 ′′ (𝑡)) − ( + + )
2 ⏟𝜕𝑥 𝑑𝑡 𝜕𝑦 𝑑𝑡 𝜕𝑧 𝑑𝑡
product rule
chain rule
= 𝑚𝐫 ′′ (𝑡) ⋅ 𝐫 ′ (𝑡) − ∇𝑓(𝑥(𝑡), 𝑦(𝑡), 𝑧(𝑡)) ⋅ 𝐫 ′ (𝑡)
= 𝐅(𝐫(𝑡)) ⋅ 𝐫 ′ (𝑡) − 𝐅(𝐫(𝑡)) ⋅ 𝐫 ′ (𝑡)
=0
for every 𝑡, and so 𝐸(𝑡) is a constant function. ∎

Page 2 of 10
MATH2023 Multivariable Calculus Problem Set 6
L3 (Fall 2019)

5. Note that the divergence of 〈𝑥 cos 𝑦 , − sin 𝑦 , 𝑧 − 𝑒 𝑥𝑦 〉 is

𝜕 𝜕 𝜕
∇ ⋅ 〈𝑥 cos 𝑦 , − sin 𝑦 , 𝑧 − 𝑒 𝑥𝑦 〉 = (𝑥 cos 𝑦) + (− sin 𝑦) + (𝑧 − 𝑒 𝑥𝑦 )
𝜕𝑥 𝜕𝑦 𝜕𝑧

= (cos 𝑦) + (− cos 𝑦) + 1 = 1 ≢ 0
on ℝ3 , so it cannot be the curl of any vector field 𝐅: ℝ3 → ℝ3 . In other words, there does not exist any vector
3 3
field 𝐅: ℝ → ℝ such that
∇ × 𝐅 = 〈𝑥 cos 𝑦 , − sin 𝑦 , 𝑧 − 𝑒 𝑥𝑦 〉.

6. (a) Let 𝐫: [𝑎, 𝑏] → ℝ2 be a parametrization of the curve 𝐶, so that 𝐫 ′ (𝑡) is a tangent vector to 𝐶 at the point
𝐫(𝑡). According to the diagram, at every point 𝐫(𝑡) on the curve 𝐶, we have
𝐅(𝐫(𝑡)) ⋅ 𝐫 ′ (𝑡) < 0
because the angle between the assigned vector 𝐅(𝐫(𝑡)) and the tangent vector 𝐫 ′ (𝑡) is always obtuse.
Therefore
𝑏
∫𝐅 ⋅ 𝑑𝐫 = ∫ 𝐅(𝐫(𝑡)) ⋅ 𝐫 ′ (𝑡)𝑑𝑡 < 0.
𝐶 𝑎

(b) Let 𝐅(𝑥, 𝑦) = 〈𝑓(𝑥, 𝑦), 𝑔(𝑥, 𝑦)〉. We focus on the 𝑥- and 𝑦-components of those vectors assigned to points
near 𝑃. From the diagram we observe that as one moves from left to right through 𝑃 (i.e. as 𝑥 increases),
𝜕𝑓
the horizontal components 𝑓(𝑥, 𝑦) of the assigned vectors are decreasing, so (𝑃) < 0:
𝜕𝑥
𝑦

𝑃 𝑥

On the other hand, as one moves from bottom to top through 𝑃 (i.e. as 𝑦 increases), the vertical

𝜕𝑔
components 𝑔(𝑥, 𝑦) of the assigned vectors are also decreasing, so (𝑃) < 0:
𝜕𝑦
𝑦

𝑃 𝑥

Therefore
𝜕𝑓 𝜕𝑔
(∇ ⋅ 𝐅)(𝑃) = (𝑃) + (𝑃) < 0.
𝜕𝑥 𝜕𝑦

Page 3 of 10
MATH2023 Multivariable Calculus Problem Set 6
L3 (Fall 2019)
𝑦
7. (a) The three edges of the triangle 𝐶 can be parametrized by
𝐫1 (𝑡) = 〈𝑡, 0〉, 𝐫2 (𝑡) = 〈1 − 𝑡, 𝑡〉 and 𝐫3 (𝑡) = 〈0, 1 − 𝑡〉 1
𝐶
respectively, where 𝑡 ∈ [0, 1] in each of these parametrizations. So
𝐫1 ′ (𝑡) = 〈1, 0〉, 𝐫2 ′ (𝑡) = 〈−1, 1〉 and 𝐫3 ′ (𝑡) = 〈0, −1〉
𝑥
for every 𝑡 ∈ (0, 1), and so 0 1

1 1 1
∫(𝑥 + 𝑦)𝑑𝑠 = ∫ (𝑡 + 0)√12 + 02 𝑑𝑡 + ∫ (1 − 𝑡 + 𝑡)√(−1)2 + 12 𝑑𝑡 + ∫ (0 + 1 − 𝑡)√02 + (−1)2 𝑑𝑡
𝐶 0 0 0

1
= ∫ (1 + √2)𝑑𝑡 = 1 + √2.
0

(b) The curve 𝐶 has a given parametrization 𝐫(𝑡) = 〈𝑡 cos 𝑡 , 𝑡 sin 𝑡 , 𝑡〉 for 𝑡 ∈ [0, 2𝜋], so we have
𝐫 ′ (𝑡) = 〈cos 𝑡 − 𝑡 sin 𝑡 , sin 𝑡 + 𝑡 cos 𝑡 , 1〉
for every 𝑡 ∈ (0, 2𝜋), and so
2𝜋
∫𝑧 𝑑𝑠 = ∫ 𝑡√(cos 𝑡 − 𝑡 sin 𝑡)2 + (sin 𝑡 + 𝑡 cos 𝑡)2 + 12 𝑑𝑡
𝐶 0

2𝜋
1 3 2𝜋 2√2 3
= ∫ 𝑡√2 + 𝑡 2 𝑑𝑡 = [ (2 + 𝑡 2 )2 ] = [(1 + 2𝜋 2 )2 − 1].
0 3 0 3

𝑦
(c) The curve 𝐶 can be decomposed into two parts. The first part is the line
4
segment from (2, 4) to (−2, 4), which can be parametrized by
𝐫1 (𝑡) = 𝑡〈−2, 4〉 + (1 − 𝑡)〈2, 4〉 = 〈2 − 4𝑡, 4〉 𝐶

for 𝑡 ∈ [0, 1]. The second part is the portion of the parabola 𝑦 = 𝑥 2 from
(−2, 4) to (2, 4), which can be parametrized by 𝑥
−2 0 2
𝐫2 (𝑡) = 〈𝑡, 𝑡 2 〉
for 𝑡 ∈ [−2, 2]. Now we have 𝐫1 ′ (𝑡) = 〈−4, 0〉 for 𝑡 ∈ (0, 1) and 𝐫2 ′ (𝑡) = 〈1, 2𝑡〉 for 𝑡 ∈ (−2, 2), so
1 2
∮𝐅 ⋅ 𝑑𝐫 = ∫ 〈(2 − 4𝑡)2 (4), (2 − 4𝑡)3 (4)2 〉 ⋅ 〈−4, 0〉𝑑𝑡 + ∫ 〈(𝑡)2 (𝑡 2 ), (𝑡)3 (𝑡 2 )2 〉 ⋅ 〈1, 2𝑡〉𝑑𝑡
𝐶 0 −2

1 2
9856
= ∫ −64(1 − 2𝑡)2 𝑑𝑡 + ∫ (𝑡 4 + 2𝑡 8 )𝑑𝑡 = .
0 −2 45 𝑦
Alternative solution: 4
2
𝐶 is a simple closed curve in ℝ whose interior is the region defined by 𝐶
𝑅
𝑅 = {(𝑥, 𝑦) ∈ ℝ2 : −2 ≤ 𝑥 ≤ 2 and 𝑥 2 ≤ 𝑦 ≤ 4},
so by Green’s Theorem,
𝑥
−2 0 2
𝜕 𝜕 2
∮𝐅 ⋅ 𝑑𝐫 = ∬ [ (𝑥 3 𝑦 2 ) − (𝑥 𝑦)] 𝑑𝐴 = ∬ (3𝑥 2 𝑦 2 − 𝑥 2 )𝑑𝐴
𝐶 𝑅 𝜕𝑥 𝜕𝑦 𝑅

2 4 2
= ∫ ∫ (3𝑥 2 𝑦 2 − 𝑥 2 )𝑑𝑦 𝑑𝑥 = ∫ [𝑥 2 (43 − (𝑥 2 )3 ) − 𝑥 2 (4 − 𝑥 2 )]𝑑𝑥
−2 𝑥 2 −2

2
9856
= ∫ (60𝑥 2 + 𝑥 4 − 𝑥 8 )𝑑𝑥 = .
−2 45

Page 4 of 10
MATH2023 Multivariable Calculus Problem Set 6
L3 (Fall 2019)

(d) The curve 𝐶 can be parametrized by

𝐫(𝑡) = 〈𝑡, 𝑡 2 , 𝑡 3 〉
for 𝑡 ∈ [0, 1]. Now we have 𝐫 ′ (𝑡) = 〈1, 2𝑡, 3𝑡 2 〉 for 𝑡 ∈ (0, 1), so
1
∫𝐅 ⋅ 𝑑𝐫 = ∫ 〈(𝑡 2 )(𝑡 3 ), (𝑡)2 , (𝑡)(𝑡 3 )〉 ⋅ 〈1, 2𝑡, 3𝑡 2 〉𝑑𝑡
𝐶 0

1
23
= ∫ (𝑡 5 + 2𝑡 3 + 3𝑡 6 )𝑑𝑡 = .
0 21

(e) The curve 𝐶 can be parametrized by

2
𝐫(𝑡) = 〈𝑡, 𝑡 3 〉
1
2
for 𝑡 ∈ [0, 1]. Now we have 𝐫 ′ (𝑡) = 〈1, 𝑡 −3 〉 for 𝑡 ∈ (0, 1), so
3

1 1
2 2 1 5 5 5
∫𝑥𝑦𝑑𝑥 + 𝑥 2 𝑑𝑦 = ∫ [(𝑡) (𝑡 3 ) (1) + (𝑡)2 ( 𝑡 −3 )] 𝑑𝑡 = ∫ 𝑡 3 𝑑𝑡 = .
𝐶 0 3 0 3 8

(f) We consider a parameter 𝑢 = 𝑥. Then according to the equation of the given plane we have
𝑦 = 2 − 𝑥 = 2 − 𝑢.
Now substituting these into the equation (𝑥 − 1)2 + (𝑦 − 1)2 + 𝑧 2 = 2 of the given sphere, we obtain
(𝑢 − 1)2 + (2 − 𝑢 − 1)2 + 𝑧 2 = 2,

𝑧2
i.e. (𝑢 − 1)2 + = 1. Thus we may now take a parameter 𝑡 ∈ [0, 2𝜋] such that
2

𝑧 𝑧
𝑦
𝑢 − 1 = cos 𝑡 and = − sin 𝑡,
√2
(the negative sign before sin 𝑡 is due to the clockwise orientation when viewed 2
from the origin), so the curve of intersection 𝐶 can be parametrized by

𝑥 = 𝑢 = 1 + cos 𝑡 0
2 𝑥
{𝑦 = 2 − 𝑢 = 1 − cos 𝑡
𝑧 = −√2 sin 𝑡 𝐶

or equivalently,
𝐫(𝑡) = 〈1 + cos 𝑡 , 1 − cos 𝑡 , −√2 sin 𝑡〉
for 𝑡 ∈ [0, 2𝜋]. Now we have 𝐫 ′ (𝑡) = 〈− sin 𝑡 , sin 𝑡 , −√2 cos 𝑡〉 for 𝑡 ∈ (0, 2𝜋), so
2𝜋
∫𝑦 𝑑𝑥 + 𝑧 𝑑𝑦 + 𝑥 𝑑𝑧 = ∫ [(1 − cos 𝑡)(− sin 𝑡) + (−√2 sin 𝑡)(sin 𝑡) + (1 + cos 𝑡)(−√2 cos 𝑡)]𝑑𝑡
𝐶 0

2𝜋
= ∫ (− sin 𝑡 − √2 cos 𝑡 + sin 𝑡 cos 𝑡 − √2)𝑑𝑡
0

= −2𝜋√2.

Page 5 of 10
MATH2023 Multivariable Calculus Problem Set 6
L3 (Fall 2019)

8. We check whether there exists a function 𝑓: 𝐷 → ℝ such that

𝑓𝑥 (𝑥, 𝑦, 𝑧) = 2𝑥𝑦
𝑓𝑦 (𝑥, 𝑦, 𝑧) = 𝑥 2 + ln 𝑧
.
𝑦+2
𝑓
{𝑧 (𝑥, 𝑦, 𝑧) =
𝑧
The first equation of this system requires that

𝑓(𝑥, 𝑦, 𝑧) = ∫ 2𝑥𝑦 𝑑𝑥 = 𝑥 2 𝑦 + 𝑔(𝑦, 𝑧)

for some function 𝑔 depending on 𝑦 and 𝑧 only. Taking partial derivatives with respect to 𝑦 we get
𝑓𝑦 (𝑥, 𝑦, 𝑧) = 𝑥 2 + 𝑔𝑦 (𝑦, 𝑧),
so comparing this with the second equation of the system we must have 𝑔𝑦 (𝑦, 𝑧) = ln 𝑧, i.e.

𝑔(𝑦, 𝑧) = ∫ ln 𝑧 𝑑𝑦 = 𝑦 ln 𝑧 + ℎ(𝑧),

and so 𝑓(𝑥, 𝑦, 𝑧) = 𝑥 2 𝑦 + 𝑦 ln 𝑧 + ℎ(𝑧) for some function ℎ depending on 𝑧 only. Taking partial derivatives
with respect to 𝑧 we get
𝑦
𝑓𝑧 (𝑥, 𝑦, 𝑧) = + ℎ′ (𝑧),
𝑧
2
so comparing this with the third equation of the system we must have ℎ′ (𝑧) = , i.e.
𝑧

2
ℎ(𝑧) = ∫ 𝑑𝑧 = 2 ln 𝑧 + 𝑘
𝑧
for some constant 𝑘. Thus if we choose 𝑘 = 0 and let
𝑓(𝑥, 𝑦, 𝑧) = 𝑥 2 𝑦 + (𝑦 + 2) ln 𝑧 ,
then 𝐅 = ∇𝑓 on 𝐷. This shows that 𝐅 is conservative in 𝐷. Finally 𝐶 is a curve in 𝐷 from the point
(−1, −2, 2) to the point (3, 0, 𝑒 2 ), so

∫𝐅 ⋅ 𝑑𝐫 = ∫∇𝑓 ⋅ 𝑑𝐫 = 𝑓(3, 0, 𝑒 2 ) − 𝑓(−1, −2, 2) = (2 ln 𝑒 2 ) − (−2) = 6.

𝐶 𝐶

9. The curve 𝐶 can be parametrized by 𝐫(𝑡) = 〈cos 𝑡 , sin 𝑡〉 for 𝑡 ∈ [0, 2𝜋]. Now we have 𝐫 ′ (𝑡) = 〈− sin 𝑡 , cos 𝑡〉,
so
2𝜋
−𝑦 𝑥 − sin 𝑡 cos 𝑡
∮ 2 2
𝑑𝑥 + 2 2
𝑑𝑦 = ∫ [ 2 2
(− sin 𝑡) + (cos 𝑡)] 𝑑𝑡
𝐶 𝑥 + 4𝑦 𝑥 + 4𝑦 0 (cos 𝑡) + 4(sin 𝑡) (cos 𝑡) + 4(sin 𝑡)2
2

𝜋 3𝜋
2𝜋
1 2 sec 2 𝑡 2 sec 2 𝑡
=∫ 𝑑𝑡 = ∫ 𝑑𝑡 + ∫ 𝑑𝑡
0 cos 2 𝑡 + 4 sin2 𝑡 − 1 + 4 tan 𝑡
𝜋 2 𝜋 1 + 4 tan2 𝑡
2 2
∞ ∞
1 1
=∫ 2
𝑑𝑢 + ∫ 2
𝑑𝑢 = [arctan 2𝑢]∞
−∞ = 𝜋,
−∞ 1 + 4𝑢 −∞ 1 + 4𝑢
𝜋 𝜋 𝜋 3𝜋
where we have used the substitution 𝑢 = tan 𝑡 for 𝑡 ∈ (− , ) and for 𝑡 ∈ ( , ).
2 2 2 2

−𝑦 𝑥
Note: Since the vector field 𝐅(𝑥, 𝑦) = 〈 , 〉 is undefined at (0, 0) and (0, 0) lies in the interior
𝑥 2 +4𝑦 2 𝑥 2 +4𝑦 2

of the simple closed curve 𝐶, we cannot apply Green’s theorem to compute this line integral.

Page 6 of 10
MATH2023 Multivariable Calculus Problem Set 6
L3 (Fall 2019)

10. The curve 𝐶 is a simple closed curve in ℝ2 whose interior 𝑅 is a trapezoidal region with area
[8 − (−2)](8 − 5) [8 − (−2)][5 − (−1)]
∬ 𝑑𝐴 = + = 45.
𝑅
⏟ 2 ⏟ 2
Area of the triangular region Area of the triangular region
with vertices (8,5), (4,8) and (−2,5) with vertices (8,5), (−4,−1) and (−2,5)

By Green’s Theorem, we have

𝜕 𝜕
∮(2𝑥 cos 𝑥 2 + 3𝑦)𝑑𝑥 + (7𝑥 + 𝑦 3 𝑒 𝑦 )𝑑𝑦 = ∬ [ (7𝑥 + 𝑦 3 𝑒 𝑦 ) − (2𝑥 cos 𝑥 2 + 3𝑦)] 𝑑𝐴
𝐶 𝑅 𝜕𝑥 𝜕𝑦

= ∬ 4 𝑑𝐴 = 4(45) = 180.
𝑅

11. With the given parametrization of the simple closed curve 𝐶 in ℝ2 , the area of the interior of 𝐶 is given by
1 1 2𝜋
∮−𝑦 𝑑𝑥 + 𝑥 𝑑𝑦 = ∫ [(−𝑎 sin3 𝑡)(−3𝑎 cos 2 𝑡 sin 𝑡) + (𝑎 cos 3 𝑡)(3𝑎 sin2 𝑡 cos 𝑡)]𝑑𝑡
2 𝐶 2 0

3𝑎2 2𝜋 3𝑎2 2𝜋 2
= ∫ [sin4 𝑡 cos 2 𝑡 + cos 4 𝑡 sin2 𝑡]𝑑𝑡 = ∫ sin 𝑡 cos 2 𝑡 𝑑𝑡
2 0 2 0

3𝑎2 2𝜋 2 3𝑎2 2𝜋 1 − cos 4𝑡 3𝜋𝑎2

= ∫ sin 2𝑡 𝑑𝑡 = ∫ 𝑑𝑡 = .
8 0 8 0 2 8

12. For each 𝑘 ∈ {1, 2, … , 𝑛}, let 𝐿𝑘 be the line segment from (𝑥𝑘−1 , 𝑦𝑘−1 ) to (𝑥𝑘 , 𝑦𝑘 ) which is parametrized by
𝐫𝑘 (𝑡) = 〈𝑥𝑘−1 , 𝑦𝑘−1 〉 + 𝑡〈𝑥𝑘 − 𝑥𝑘−1 , 𝑦𝑘 − 𝑦𝑘−1 〉
= 〈𝑥𝑘−1 + 𝑡(𝑥𝑘 − 𝑥𝑘−1 ), 𝑦𝑘−1 + 𝑡(𝑦𝑘 − 𝑦𝑘−1 )〉
for 𝑡 ∈ [0, 1]. Then 𝑃 is simply 𝐿1 + 𝐿2 + ⋯ + 𝐿𝑛 , the area bounded by the polygon 𝑃 is
1 1 1
∮−𝑦 𝑑𝑥 + 𝑥 𝑑𝑦 = ( ∫ −𝑦 𝑑𝑥 + 𝑥 𝑑𝑦 ) + ⋯ + ( ∫ −𝑦 𝑑𝑥 + 𝑥 𝑑𝑦 )
2 𝑃 2 𝐿1 2 𝐿𝑛
𝑛
1
1
= ∑ ∫ [−(𝑦𝑘−1 + 𝑡(𝑦𝑘 − 𝑦𝑘−1 ))(𝑥𝑘 − 𝑥𝑘−1 ) + (𝑥𝑘−1 + 𝑡(𝑥𝑘 − 𝑥𝑘−1 ))(𝑦𝑘 − 𝑦𝑘−1 )]𝑑𝑡
2 0 𝑘=1
𝑛
1 𝑦𝑘 − 𝑦𝑘−1 2 𝑥𝑘 − 𝑥𝑘−1 2 𝑡=1
= ∑ [−(𝑥𝑘 − 𝑥𝑘−1 ) (𝑦𝑘−1 𝑡 + 𝑡 ) + (𝑦𝑘 − 𝑦𝑘−1 ) (𝑥𝑘−1 𝑡 + 𝑡 )]
2 2 2 𝑡=0
𝑘=1
𝑛
1 𝑦𝑘−1 + 𝑦𝑘 𝑥𝑘−1 + 𝑥𝑘
= ∑ [−(𝑥𝑘 − 𝑥𝑘−1 ) ( ) + (𝑦𝑘 − 𝑦𝑘−1 ) ( )]
2 2 2
𝑘=1
𝑛
1
= ∑(𝑥𝑘−1 𝑦𝑘 − 𝑥𝑘 𝑦𝑘−1 ).
2
𝑘=1

13. ℝ2 ∖ 𝑋 is simply connected only in (c); ℝ2 ∖ 𝑋 is not simply connected in (a), (b), (d) and (e). ℝ2 ∖ 𝑋 is not even
connected in (b) and (e).

Page 7 of 10
MATH2023 Multivariable Calculus Problem Set 6
L3 (Fall 2019)

14. (a) Applying the 2-D Divergence Theorem (Corollary 6.65) to the vector field
𝑓∇𝑔 − 𝑔∇𝑓 = 𝑓〈𝑔𝑥 , 𝑔𝑦 〉 − 𝑔〈𝑓𝑥 , 𝑓𝑦 〉 = 〈𝑓𝑔𝑥 − 𝑔𝑓𝑥 , 𝑓𝑔𝑦 − 𝑔𝑓𝑦 〉
and the simple closed curve 𝐶 in ℝ2 , we have
𝜕 𝜕
̂ 𝑑𝑠 = ∬ [
∮(𝑓∇𝑔 − 𝑔∇𝑓) ⋅ 𝐧 (𝑓𝑔𝑥 − 𝑔𝑓𝑥 ) + (𝑓𝑔𝑦 − 𝑔𝑓𝑦 )] 𝑑𝐴
𝐶 𝑅 𝜕𝑥 𝜕𝑦

= ∬ [(𝑓𝑥 𝑔𝑥 + 𝑓𝑔𝑥𝑥 − 𝑔𝑥 𝑓𝑥 − 𝑔𝑓𝑥𝑥 ) + (𝑓𝑦 𝑔𝑦 + 𝑓𝑔𝑦𝑦 − 𝑔𝑦 𝑓𝑦 − 𝑔𝑓𝑦𝑦 )]𝑑𝐴

𝑅

= ∬ [𝑓(𝑔𝑥𝑥 + 𝑔𝑦𝑦 ) − 𝑔(𝑓𝑥𝑥 + 𝑓𝑦𝑦 )]𝑑𝐴

𝑅

= ∬ [𝑓(∇ ⋅ ∇𝑔) − 𝑔(∇ ⋅ ∇𝑓)]𝑑𝐴 .

𝑅

(b) Let 𝑔(𝑥, 𝑦) ≡ −1. Then ∇𝑔 = 𝟎 and ∇ ⋅ ∇𝑔 = 0. Now if 𝑓 is a function satisfying 𝑓𝑥𝑥 + 𝑓𝑦𝑦 ≡ 0, then
applying (a) to these functions 𝑓 and 𝑔 we have

∮(𝑓𝟎 − (−1)∇𝑓) ⋅ 𝐧
̂ 𝑑𝑠 = ∬ [𝑓0 − (−1)(𝑓𝑥𝑥 + 𝑓𝑦𝑦 )]𝑑𝐴 ,
𝐶 𝑅

i.e.

̂ 𝑑𝑠 = ∬ (𝑓𝑥𝑥 + 𝑓𝑦𝑦 )𝑑𝐴 .

∮∇𝑓 ⋅ 𝐧
𝐶 𝑅

15. (a) The statement is true.

If 𝐅 = 〈𝑓, 𝑔, ℎ〉 and 𝐆 = 〈𝑢, 𝑣, 𝑤〉, then 𝐅 + 𝐆 = 〈𝑓 + 𝑢, 𝑔 + 𝑣, ℎ + 𝑤〉, so
∇ × (𝐅 + 𝐆) = 〈(ℎ + 𝑤)𝑦 − (𝑔 + 𝑣)𝑧 , (𝑓 + 𝑢)𝑧 − (ℎ + 𝑤)𝑥 , (𝑔 + 𝑣)𝑥 − (𝑓 + 𝑢)𝑦 〉
= 〈(ℎ𝑦 + 𝑤𝑦 ) − (𝑔𝑧 + 𝑣𝑧 ), (𝑓𝑧 + 𝑢𝑧 ) − (ℎ𝑥 + 𝑤𝑥 ), (𝑔𝑥 + 𝑣𝑥 ) − (𝑓𝑦 + 𝑢𝑦 )〉
= 〈ℎ𝑦 − 𝑔𝑧 , 𝑓𝑧 − ℎ𝑥 , 𝑔𝑥 − 𝑓𝑦 〉 + 〈𝑤𝑦 − 𝑣𝑧 , 𝑢𝑧 − 𝑤𝑥 , 𝑣𝑥 − 𝑢𝑦 〉
= (∇ × 𝐅) + (∇ × 𝐆).

(b) The statement is false.

The expression ∇ × (𝐅 ⋅ 𝐆) on the left-hand side is not even well-defined. 𝐅 ⋅ 𝐆 is a scalar-valued function
of three real variables but is not a vector field in three dimensions, so the curl of 𝐅 ⋅ 𝐆 is not defined.

(c) The statement is false.

Although the expressions on both sides are well-defined, they are not equal in general. To see this we
consider the vector fields 𝐅 = 𝑥𝐢 and 𝐆 = 𝐣. Then 𝐅 × 𝐆 = 𝑥𝐢 × 𝐣 = 𝑥𝐤 and so
∇ × (𝐅 × 𝐆) = ∇ × (𝑥𝐤) = −𝐣.
But since ∇ × 𝐅 = ∇ × 𝐆 = 𝟎, we have (∇ × 𝐅) × (∇ × 𝐆) = 𝟎 ≠ ∇ × (𝐅 × 𝐆).

Page 8 of 10
MATH2023 Multivariable Calculus Problem Set 6
L3 (Fall 2019)

(d) The statement is true.

Let 𝐫: [𝑎, 𝑏] → ℝ𝑛 be a smooth parametrization of 𝐶. Then 𝐬: [𝑎, 𝑏] → ℝ𝑛 defined by
𝐬(𝑡) = 𝐫(𝑎 + 𝑏 − 𝑡)
is a parametrization of −𝐶, so with the substitution 𝑢 = 𝑎 + 𝑏 − 𝑡 we have
𝑏 𝑏
∫ 𝐅 ⋅ 𝑑𝐫 = ∫ 𝐅(𝐬(𝑡)) ⋅ 𝐬 ′ (𝑡)𝑑𝑡 = ∫ 𝐅(𝐫(𝑎 + 𝑏 − 𝑡)) ⋅ [−𝐫 ′ (𝑎 + 𝑏 − 𝑡)]𝑑𝑡
−𝐶 𝑎 𝑎

𝑎 𝑏
= ∫ 𝐅(𝐫(𝑢)) ⋅ 𝐫 ′ (𝑢)𝑑𝑢 = − ∫ 𝐅(𝐫(𝑢)) ⋅ 𝐫 ′ (𝑢)𝑑𝑢 = − ∫𝐅 ⋅ 𝑑𝐫.
𝑏 𝑎 𝐶

[The key idea is that when the orientation of 𝐶 is reversed, each tangent vector 𝐫 ′ (𝑡) is also reversed, so an
extra negative sign is introduced to the dot product 𝐅(𝐫(𝑡)) ⋅ 𝐫 ′ (𝑡).]

(e) The statement is false.

Let 𝐫: [𝑎, 𝑏] → ℝ𝑛 be a smooth parametrization of 𝐶. Then 𝐬: [𝑎, 𝑏] → ℝ𝑛 defined by
𝐬(𝑡) = 𝐫(𝑎 + 𝑏 − 𝑡)
is a parametrization of −𝐶, so with the substitution 𝑢 = 𝑎 + 𝑏 − 𝑡 we have
𝑏 𝑏
∫ 𝑓𝑑𝑠 = ∫ 𝑓(𝐬(𝑡))‖𝐬 ′ (𝑡)‖𝑑𝑡 = ∫ 𝑓(𝐫(𝑎 + 𝑏 − 𝑡))‖𝐫 ′ (𝑎 + 𝑏 − 𝑡)‖𝑑𝑡
−𝐶 𝑎 𝑎

𝑎 𝑏
= ∫ 𝑓(𝐫(𝑢))‖𝐫 ′ (𝑢)‖(−𝑑𝑢) = ∫ 𝑓(𝐫(𝑢))‖𝐫 ′ (𝑢)‖𝑑𝑢 = ∫𝑓𝑑𝑠 .
𝑏 𝑎 𝐶

[The key idea is that when the orientation of 𝐶 is reversed, although each tangent vector 𝐫 ′ (𝑡) is also
reversed, its length ‖𝐫 ′ (𝑡)‖ is still the same, so the quantity 𝑓(𝐫(𝑡))‖𝐫 ′ (𝑡)‖ is not affected.]

(f) The statement is true.

Let 𝐅(𝑥, 𝑦) = 〈𝑓(𝑥, 𝑦), 𝑔(𝑥, 𝑦)〉. Then

∫ 𝑓 𝑑𝑥 + 𝑔 𝑑𝑦 = ∫ 𝐅 ⋅ 𝑑𝐫 = − ∫𝐅 ⋅ 𝑑𝐫 = − (∫𝑓 𝑑𝑥 + 𝑔 𝑑𝑦 )
−𝐶 −𝐶 𝐶 𝐶

because of the statement in (d).

(g) The statement is false.

If a force field 𝐅 is not conservative in a region, then the work done by 𝐅 in moving a particle around a
closed curve in that region may not be zero in general.

(h) The statement is true.

Let 𝑅 denote the interior of the given curve 𝐶 and let 𝐅(𝑥, 𝑦) = 〈0, 𝑥〉. Then by Green’s Theorem,
𝜕 𝜕
∮𝑥 𝑑𝑦 = ∮𝐅 ⋅ 𝑑𝐫 = ∬ [ (𝑥) − (0)] 𝑑𝐴 = ∬ 1 𝑑𝐴 = Area of 𝑅.
𝐶 𝐶 𝑅 𝜕𝑥 𝜕𝑦 𝑅

Page 9 of 10
MATH2023 Multivariable Calculus Problem Set 6
L3 (Fall 2019)

(i) The statement is true.

Let 𝑅 denote the interior of the given curve 𝐶 and let 𝐅(𝑥, 𝑦) = 〈𝑓(𝑥), 𝑔(𝑦)〉. Then by Green’s Theorem,
𝜕 𝜕
∮𝑓(𝑥)𝑑𝑥 + 𝑔(𝑦)𝑑𝑦 = ∮𝐅 ⋅ 𝑑𝐫 = ∬ [ (𝑔(𝑦)) − (𝑓(𝑥))] 𝑑𝐴 = ∬ 0 𝑑𝐴 = 0.
𝐶 𝐶 𝑅 𝜕𝑥 𝜕𝑦 𝑅

(j) The statement is true.

By the statement (i) we see that the line integrals of 𝐅 are path-independent on ℝ2 , so 𝐅 is conservative in
ℝ2 . Alternatively, since
𝜕 𝜕
[𝑔(𝑦)] − [𝑓(𝑥)] = 0
𝜕𝑥 𝜕𝑦

for every (𝑥, 𝑦) ∈ ℝ2 and since ℝ2 is simply connected, we can conclude that 𝐅(𝑥, 𝑦) = 〈𝑓(𝑥), 𝑔(𝑦)〉 is
conservative in ℝ2 .

(k) The statement is false.

If the region 𝑅 is not simply connected, then 𝐅 may not be conservative in 𝑅 in general. To see this, we
2
let 𝑅 = ℝ ∖ {(0, 0)}, and let 𝑓, 𝑔: 𝑅 → ℝ be the functions defined by
𝑦 𝑥
𝑓(𝑥, 𝑦) = − and 𝑔(𝑥, 𝑦) = .
𝑥 2 + 𝑦2 𝑥2 + 𝑦2

𝑦 2 −𝑥 2
Then even though 𝑔𝑥 (𝑥, 𝑦) = (𝑥 2 = 𝑓𝑦 (𝑥, 𝑦) for every (𝑥, 𝑦) ∈ 𝑅, the vector field 𝐅 = 〈𝑓, 𝑔〉 is still not
+𝑦 2 )2

conservative in 𝑅 because for the unit circle 𝐶 centered at the origin oriented counterclockwise,

∮𝐅 ⋅ 𝑑𝐫 = 2𝜋 ≠ 0.
𝐶

Page 10 of 10