Differential Geometry II - Smooth Manifolds
Winter Term 2023/2024
Lecturer: Dr. N. Tsakanikas
Assistant: L. E. Rösler
Exercise Sheet 5 – Solutions
Exercise 1:
(a) Let (x, y) denote the standard coordinates on R2 . Verify that (e
x, ye) are smooth global
2
coordinates on R , where
e = x and ye = y + x3 .
x
Let p be the point (1, 0) ∈ R2 (in standard coordinates), and show that
∂ ∂
̸= ,
∂x p ∂e
x p
even though the coordinate functions x and x
e are identically equal.
(This shows that each coordinate vector ∂/∂xi |p depends on the entire coordinate
system, not just on the single coordinate function xi .)
(b) Polar coordinates on R2 : Consider the map
Φ : W := (0, +∞) × (−π, π) → R2
(r, θ) 7→ (r cos θ, r sin θ).
(i) Show that Φ is a diffeomorphism onto its image U := Φ(W ).
(Therefore, Φ−1 can be considered as a smooth chart on R2 , and it is common
to call its component functions the polar coordinates (r, θ) on R2 .)
(ii) Let p be a point in R2 whose polar coordinate representation is (r, θ) = (2, π/2),
and let v ∈ Tp R2 be the tangent vector whose polar coordinate representation is
∂ ∂
v=3 − .
∂r p ∂θ p
Compute the coordinate representation of v in terms of the standard coordinate
vectors
∂ ∂
, .
∂x p ∂y p
1
(c) Spherical coordinates on R3 : Consider the map
Ψ : W := (0, +∞) × (−π, π) × (0, π) → R3
(r, φ, θ) 7→ (r cos φ sin θ, r sin φ sin θ, r cos θ).
(i) Show that Ψ is a diffeomorphism onto its image U := Ψ(W ).
(Therefore, Ψ−1 can be considered as a smooth chart on R3 , and it is common
to call its component functions the spherical coordinates (r, φ, θ) on R3 .)
(ii) Express the coordinate vectors
∂ ∂ ∂
, ,
∂r p ∂φ p ∂θ p
of this chart at some point p ∈ U in terms of the standard coordinate vectors
∂ ∂ ∂
, , .
∂x p ∂y p ∂z p
Solution:
(a) Consider the function
ψ : R2 → R2 , (x, y) 7→ (x, y + x3 ).
Observe that ψ is smooth and bijective with inverse function
ψ −1 : R2 → R2 , (e e3 ,
x, ye) 7→ x
e, ye − x
which is also smooth. Hence, ψ is a global smooth coordinate chart on R2 ; in other words,
its components (ex, ye) are smooth global coordinates on R2 .
We have
∂e
x ∂e
y
(x, y) = 1 and (x, y) = 3x2 ,
∂x ∂x
and hence
∂ ∂ ∂ ∂
=1· +3· ̸= .
∂x p ∂e
x p ∂e
y p ∂e x p
(b) We deal with (i) and (ii) separately.
(i) Geometrically, r ∈ (0, +∞) is the distance from the origin, and θ ∈ (−π, π) is
the angle from the negative x-axis. Observe now that the image of Φ is the plane
without the non-positive x-axis, that is,
U = Φ(W ) = (x, y) ∈ R2 | x > 0 .
Note also that Φ : W → U is bijective with inverse
p y
Φ−1 : U → W, (x, y) 7→ x2 + y 2 , arctan .
x
Since both Φ and Φ−1 are clearly smooth, we conclude that Φ : W → U is a
diffeomorphism.
2
(ii) We have
∂ π ∂ π ∂ ∂
= cos + sin =
∂r p 2 ∂x p 2 ∂y p ∂y p
and
∂ π ∂ π ∂ ∂
= −2 sin + 2 cos = −2 ,
∂θ p 2 ∂x p 2 ∂y p ∂x p
so v has the following coordinate representation in standard coordinates:
∂ ∂
v=2 +3 .
∂x p ∂y p
(c) We deal with (i) and (ii) separately.
(i) Geometrically, r ∈ (0, +∞) is the distance from the origin, φ ∈ (−π, π) is the
angle from the x < 0 half of the (x, z)-plane, and θ ∈ (0, π) is the angle from
the positive z-axis. Observe now that the image of Ψ is the 3-dimensional space
without the z-axis and the non-positive x-axis, that is,
3 3 3
U = Ψ(W ) = R \ (0, 0, z) ∈ R | z ∈ R ∪ (x, 0, 0) ∈ R | x ≤ 0 ,
and also that Ψ : W → U is bijective. Furthermore, Ψ is clearly smooth with
Jacobian matrix
cos φ sin θ −r sin φ sin θ r cos φ cos θ
JΨ = sin φ sin θ r cos φ sin θ r sin φ cos θ
cos θ 0 −r sin θ
and Jacobian determinant
det JΨ = −r2 sin θ,
which does not vanish for any (r, θ) ∈ (0, +∞) × (0, π). Hence, Ψ is a local
diffeomorphism by the inverse function theorem, and since it is bijective, it is
actually a diffeomorphism.
(ii) Since
r = (x2 + y 2 + z 2 )1/2
and
r sin θ = (x2 + y 2 )1/2 ,
we have
∂ ∂x ∂ ∂y ∂ ∂z ∂
= + +
∂r p ∂r ∂x p ∂r ∂y p ∂r ∂z p
∂ ∂ ∂
= cos φ sin θ + sin φ sin θ + cos θ
∂x p ∂y p ∂z p
!
1 ∂ ∂ ∂
= 2 x +y +z ,
(x + y + z 2 )1/2
2 ∂x p ∂y p ∂z p
3
∂ ∂x ∂ ∂y ∂ ∂z ∂
= + +
∂φ p ∂φ ∂x p ∂φ ∂y p ∂φ ∂z p
∂ ∂
= −r sin φ sin θ + r cos φ sin θ
∂x p ∂y p
∂ ∂
= −y +x ,
∂x p ∂y p
and
∂ ∂x ∂ ∂y ∂ ∂z ∂
= + +
∂θ p ∂θ ∂x p ∂θ ∂y p ∂θ ∂z p
∂ ∂ ∂
= r cos φ cos θ + r sin φ cos θ − r sin θ
∂x p ∂y p ∂z p
xz ∂ yz ∂ ∂
= + − (x2 + y 2 )1/2 .
(x2 2 1/2
+ y ) ∂x p (x2 2 1/2
+ y ) ∂y p ∂z p
Exercise 2: Consider the inclusion ι : S2 ,→ R3 , where both S2 and R3 are endowed with
the standard smooth structure. Let p = (p1 , p2 , p3 ) ∈ S2 with p3 > 0. What is the image
of the differential dιp : Tp S2 → Tp R3 ?
Solution: Observe that the given point p ∈ S2 is contained in the domain of the smooth
chart (U3+ , φ+ 2
3 ) for S , where
U3+ = (x1 , x2 , x3 ) ∈ R3 | x3 > 0
and
φ+ + 2 2 1 2 3 1 2
3 : U3 ∩ S → B , (x , x , x ) 7→ (x , x )
with coordinate functions φ1 and φ2 (defined in the obvious manner), and recall that the
inverse of φ+
3 is the map
p
+ −1 2 + 2 1 2 1 2 1 2 2 2
(φ3 ) : B → U3 ∩ S , (u , u ) 7→ u , u , 1 − (u ) − (u ) ,
see Example 1.3 (2) and Example 1.8 (2). Therefore, the coordinate representation b ι of
2 3 + + 3
ι : S ,→ R with respect to the charts (U3 , φ3 ) and (R , IdR3 ) is the function
p
−1
ι (u1 , u2 ) = IdR3 ◦ ι ◦ (φ+
1 2 1 2
) (u , u ) = u , u , 1 − (u 1 )2 − (u2 )2 ,
b 3
and the coordinate representation pb of p ∈ S2 is pb = φ(p) = (p1 , p2 ). Since the Jacobian
matrix of b
ι, given by
1 0
J(u1 , u2 ) =
0 1 ,
1 2
−√ u
1 2 2 2
− √ u
1 2 2 2
1−(u ) −(u ) 1−(u ) −(u )
represents dιp : Tp S2 → Tp R3 in the coordinate bases
( ) ( )
∂ ∂ ∂ ∂ ∂
1
, 2
⊆ Tp S2 and 1
, 2
, 3
⊆ Tp R3 ,
∂φ p ∂φ p ∂x p ∂x p ∂x p
4
we deduce that
!
∂ ∂ ∂ p1 ∂
dιp =1· +0· −p · 3
∂φ1 p ∂x1 p ∂x2 p 1 − (p1 )2 + (p2 )2 ∂x p
1
∂ p ∂
= −
∂x1 p p3 ∂x3 p
and
!
∂ ∂ ∂ p2 ∂
dιp =0· +1· −p · 3
∂φ2 p ∂x1 p ∂x2 p 1 − (p1 )2 + (p2 )2 ∂x p
2
∂ p ∂
= − .
∂x2 p p3 ∂x3 p
Thus, the image of dιp is the R-vector space spanned by the above two vectors, which
1 2
can be identified with the vectors (1, 0, − pp3 ) and (0, 1, − pp3 ), respectively, in R3 . It is now
easy to check that this 2-dimensional R-vector space is the orthogonal complement of ⟨p⟩;
namely,
dιp Tp S2 = ⟨p⟩⊥ ∼ = v ∈ R3 | ⟨v, p⟩ = 0 .
Exercise 3: Let M1 , . . . , Mk be smooth manifolds. Show that T (M1 × . . . × Mk ) is
diffeomorphic to T (M1 ) × . . . × T (Mk ).
Solution: For each 1 ≤ i ≤ k, denote by
πi : M1 × . . . × Mk → Mi
the projection onto the i-the factor. This map is smooth by Exercise 4, Sheet 3, and
hence its global differential
d(πi ) : T (M1 × . . . × Mk ) → T Mi
is also a smooth map by Exercise 4(a). Again by Exercise 4, Sheet 3 we thus obtain a
smooth map
α : T (M1 × . . . × Mk ) → T M1 × . . . × T Mk
given by α = d(π1 ), . . . , d(πk ) . Note that if p = (p1 , . . . , pk ) ∈ M1 × . . . × Mk , then α
restricted to the fiber Tp (M1 × . . . × Mk ) is just the map defined in Exercise 3, Sheet 4,
so it is in particular an isomorphism. Therefore, α is bijective. It remains to show that α
is a diffeomorphism.
To this end, for every 1 ≤ i ≤ k, let (Ui , (xji i )ji ) be a smooth chart for Mi , and
denote by pri : T Mi → Mi the projection. By construction of the tangent bundle,
ji ji ji
pr−1
i Ui , (xi )ji , (vi )ji is a smooth coordinate chart, where (vi )ji are the coordinates
∂
of a point (p, v) ∈ T Mi (with p ∈ Ui ) in terms of the basis j of Tp Mi . This yields
∂xi i
p ji
the chart
−1 j1 j1 jk jk
pr−1
1 U1 × . . . × prk Uk , ((x1 )j1 , (v1 )j1 ), . . . , ((xk )jk , (vk )jk )
5
for T (M1 ) × . . . × T (Mk ). On the other hand, if we denote by
pr : T (M1 × . . . × Mk ) → M1 × . . . × Mk
the projection, then this also yields the chart
pr−1 (U1 × . . . × Uk ), (xji i )iji , (viji )iji
for T (M1 × . . . × Mk ). In terms of these charts, the map α is just given by
(xji i )iji , (viji )iji 7→ (xj11 )j1 , (v1j1 )j1 , . . . , (xjkk )jk , (vkjk )jk
which is clearly a diffeomorphism. Hence α is locally a diffeomorphism, and as it is
bijective, it is actually a diffeomorphism; see part (f) of Exercise 4, Sheet 6.
Exercise 4:
(a) Let F : M → N be a smooth map. Show that its global differential dF : T M → T N
(which is just the map whose restriction to each tangent space Tp M ⊆ T M is dFp ) is
also a smooth map.
(b) Let F : M → N and G : N → P be smooth maps. Prove the following assertions:
(i) d(G ◦ F ) = dG ◦ dF : T M → T P .
(ii) d(IdM ) = IdT M : T M → T M .
(iii) If F is a diffeomorphism, then dF : T M → T N is also a diffeomorphism, and it
holds that (dF )−1 = d(F −1 ).
Solution:
(a) Using the local expression for dFp in coordinates,
!
∂ ∂ Fbj ∂
dFp i
= i
(b
p) j ,
∂x p ∂x ∂y F (p)
we see that dF has the following coordinate representation in terms of natural coordinates
for T M and T N :
!
∂
e−1 (x1 , . . . , xn , v 1 , . . . , v n ) = ψe ◦ dF vi i
ψe ◦ dF ◦ φ
∂x φ−1 (x)
!
∂ bj ∂
F
i
= ψe v
∂xi ∂y j F ◦φ−1 (x)
!
∂ b1
F ∂ bn
F
1 n i i
= Fb (x), . . . , Fb (x), (x)v , . . . , (x)v .
∂xi ∂xi
Since F is smooth, and thus its coordinate representation Fb = ψ ◦ F ◦ φ−1 is smooth, the
above coordinate representation of dF is smooth, and hence dF is smooth, as claimed.
(b) All assertions follow immediately from Exercise 1, Sheet 4.
6
Exercise 5:
(a) Let f : X → S be a map from a topological space X to a set S. Show that if X is
connected and if f is locally constant, i.e., for every x ∈ X there exists a neighborhood
U of x in X such that f |U : U → S is constant, then f is constant.
[Hint: Show that f is continuous when S is endowed with the discrete topology.]
(b) Let M and N be smooth manifolds and let F : M → N be a smooth map. Assume
that M is connected. Show that dFp : Tp M → TF (p) N is the zero map for each p ∈ M
if and only if F is constant.
[Hint: Use (a). You may also use (without proof) the fact that any topological
manifold is locally (path) connected.]
Solution:
(a) We endow S with the discrete topology, and we claim that f : X → S is continuous.
Since then the singletons in S are open, to prove the claim, it suffices to show that the
fibers of f are open subsets of X. Fix s ∈ S and pick x ∈ f −1 (s). Since f is locally
constant, there exists an open neighborhood U of x in X such that f |U : U → S is
constant, so for every u ∈ U we have f (u) = f (x) = s, and hence u ∈ f −1 (s). Therefore,
the open neighborhood U of x is contained in the fiber f −1 (s), i.e., x ∈ U ⊆ f −1 (s). Since
x ∈ f −1 (s) was arbitrary, f −1 (s) is an open subset of X, and since s ∈ S was arbitrary,
we conclude that f is continuous.
Since S is endowed with the discrete topology, every singleton in S is also closed, and
thus every fiber of f is also closed, since f is continuous. In other words, the fibers of f
are both closed and open subsets of X, which is a connected space by assumption, and
hence each one of them is either empty or the whole space X. It follows that f is constant.
(b) Assume first that F is constant and let p ∈ M . For every f ∈ C ∞ (N ), the composite
map f ◦ F : M → R is constant, and hence for every v ∈ Tp M , by Lemma 3.5 we have
dFp (v)(f ) = v(f ◦ F ) = 0. In conclusion, dFp is the zero linear transformation for every
p ∈ M.
Assume now that dFp is the zero map for each p ∈ M . By assumption and by (a),
to prove that F is constant, it suffices to show that F is locally constant. Fix p ∈ M .
Since F is smooth, there are smooth charts (U, φ) for M containing p and (V, ψ) for
N containing F (p) such that F (U ) ⊆ V and the composite map Fb = ψ ◦ F ◦ φ−1 is
smooth. By shrinking U if necessary, we may assume that U is connected, and thus φ(U )
is also connected. Now, for each q ∈ U we know that the differential dFq is represented
in coordinate bases by the Jacobian matrix of Fb. Since dFq = O for every q ∈ U by
assumption, we infer that
∂ Fbj
qb = 0 for every i, every j, and every qb = φ(q) ∈ φ(U ),
∂xi
and hence Fb is constant on φ(U ). It follows that F = φ ◦ Fb ◦ ψ −1 is constant on U . Since
p ∈ M was arbitrary, we conclude that F is locally constant, as desired.