Differential Geometry II - Smooth Manifolds

Winter Term 2024/2025

Lecturer: Dr. N. Tsakanikas
Assistant: L. E. Rösler

Exercise Sheet 11 – Solutions

Exercise 1: Show that there is a smooth vector field on S2 which vanishes at exactly one
point.
[Hint: Use the stereographic projection [Exercise Sheet 2, Exercise 5] and consider one of
the coordinate vector fields.]

Solution: We view (u, v), resp. (e

u, ve), as the component functions of σ, resp. σ
e, where
   
x y x y
σ(x, y, z) = , = (u, v) and σ e(x, y, z) = , = (eu, ve),
1−z 1−z 1+z 1+z

so that
u
e ve
u = u(e
u, ve) = and v = v(e
u, ve) = ,
e2
u + ve2 e2
u + ve2
resp.
u v
u
e=u
e(u, v) = and v = v (u, v) = .
u2 + v 2 u2 + v 2
e e

Note that σ e ◦ σ −1 is given essentially by the same formula as σ ◦ σ e−1 (with the roles of
(u, v) and (e u, ve) reversed), and thus its Jacobian is essentially the same matrix as the one
in [Exercise Sheet 10, Exercise 1(d)] (where everything is now expressed in terms of (u, v)
instead of (e u, ve)); see [Exercise Sheet 2, Exercise 5].
∂
We now
consider the first coordinate vector field X := ∂u associated with the chart
2 2 ∂ ∂
S \{N }, σ for S . It follows from Proposition 7.2 that X = 1 ∂u +0 ∂v is a smooth vector
2
field on S \ {N }, since its component functions with respect to the smooth coordinate
∂ ∂
frame ∂u , ∂v are constant, and it is obvious that X does not vanish on S2 \ {N }. We
claim that X extends to a smooth vector field on the whole S2 and that it vanishes
precisely at the north pole N ∈ S2 . Indeed, on S2 \ {N, S} we have

∂e
u ∂ ∂e
v ∂ v 2 − u2 ∂ −2uv ∂
X= + = 2 2 2
+ 2 2 2
∂u ∂e
u ∂u ∂e v (u + v ) ∂e u (u + v ) ∂ev

∂
∂
= ve2 − u
e2 + −2e
uve .
∂eu ∂e
v

1
Since N = (0, 0, 1) ∈ S2 corresponds under σ u, ve) = (0, 0) ∈ R2 , we infer that X can
e to (e
2
be extended to a vector field on S by defining its value at N to be zero; namely,
(
∂
, if p ̸= N,
X : S2 → T S2 , p 7→ ∂u p
0, if p = N.

The above expression for X  ∂also∂ shows that its component functions with
respect to the
2
smooth coordinate frame ∂ ue , ∂ev associated with the chart S \ {S}, σ e are smooth,
2
and hence X is smooth (also) on S \ {S} by Proposition 7.2. Therefore, X is a smooth
vector field on S2 which vanishes only at the north pole N of S2 , as claimed.

Exercise 2:

(a) Let F : M → N be a smooth map. Let X ∈ X(M ) and Y ∈ X(N ). Show that X and
Y are F -related if and only if for every smooth real-valued function f defined on an
open subset of N , we have
X(f ◦ F ) = (Y f ) ◦ F.

(b) Consider the smooth map

F : R → R2 , t 7→ (cos t, sin t)

and the smooth vector fields

d ∂ ∂
X= ∈ X(R) and Y =x −y ∈ X(R2 ).
dt ∂y ∂x
Show that X and Y are F -related.

(c) Let F : M → N be a diffeomorphism and let X ∈ X(M ). Prove that there exists
a unique smooth vector field Y on N that is F -related to X. The vector field Y is
denoted by F∗ X and is called the pushforward of X by F .

(d) Consider the open submanifolds

M := (x, y) ∈ R2 | y > 0 and x + y > 0 ⊆ R2



and
N := (u, v) ∈ R2 | u > 0 and v > 0 ⊆ R2


and the map

x

F : M → N, (x, y) 7→ x + y, +1 .
y
(i) Show that F is a diffeomorpism and compute its inverse F −1 .
(ii) Compute the pushforward F∗ X of the following smooth vector field X on M :

∂
X(x,y) = y 2 .
∂x (x,y)

2
(e) Naturality of integral curves: Let F : M → N be a smooth map. Show that X ∈ X(M )
and Y ∈ X(N ) are F -related if and only if F takes integral curves of X to integral
curves of Y .
Solution:
(a) For any point p ∈ M and any smooth real-valued function f defined on an open
neighborhood of F (p) we have
X(f ◦ F )(p) = Xp (f ◦ F ) = dFp (Xp )(f )
and


(Y f ) ◦ F (p) = (Y f ) F (p) = YF (p) f.
Therefore, X and Y are F -related (i.e., dFp (Xp ) = YF (p) for every p ∈ M ) if and only
if for every smooth real-valued function f defined on an open subset of N it holds that
X(f ◦ F ) = (Y f ) ◦ F .
(b) 1st way: We prove the claim using the definition of F -related vector fields. To this
end, recall that the differential of F at an arbitrary point t ∈ R is represented (with
respect to the bases {d/dt t } for Tt R ∼
= R and {∂/∂x F (t) , ∂/∂y F (t) } for TF (t) R2 ∼
= R2 )
by the Jacobian of F at t, which is the 2 × 1-matrix
 
− sin(t)
.
cos(t)
Hence,
   
− sin(t) − sin(t) ∂ ∂
dFt (Xt ) = · (1) = = − sin(t) + cos(t) = YF (t)
cos(t) cos(t) ∂x F (t) ∂y F (t)

for any t ∈ R, which shows that X and Y are F -related.

2nd way: We may alternatively prove the assertion using (a) as follows: For every
smooth real-valued function f = f (x, y) defined on an open subset of R2 and for any
t ∈ R we have
d
X(f ◦ F )(t) = Xt (f ◦ F ) = (f ◦ F )
dt t
 
∂f
∂f
T
F (t) · F1′ (t), F2′ (t)


= F (t) ,
∂x ∂y
∂f
∂f

= − sin(t) F (t) + cos(t) F (t)
∂x ∂y
and



(Y f ) ◦ F (t) = (Y f ) F (t) = YF (t) f
!
∂ ∂
= cos(t) − sin(t) f
∂y F (t) ∂x F (t)

∂f
∂f

= cos(t) F (t) − sin(t) F (t) .
∂y ∂x

3
It follows from part (a) that X and Y are F -related.
(c) Since F is a diffeomorphism, we may define the following rough vector field on N :


Y : N → T N, q 7→ dFF −1 (q) XF −1 (q) .

It is clear that this is the unique (rough) vector field on N that is F -related to X. We
now observe that Y is the composition of the following smooth maps (see also [Exercise
Sheet 5, Exercise 4(a)]):
F −1 X dF
N −→ M −→ T M −→ T N,
so it is smooth by [Exercise Sheet 3, Exercise 3(e)].
Remark. Given a diffeomorphism F : M → N , the pushforward of any X ∈ X(M ) by F
is defined explicitly by the formula


(F∗ X)q = dFF −1 (q) XF −1 (q) ,

as already demonstrated in the proof of (c) above. As long as the inverse map F −1 of F
can be computed explicitly, the pushforward of a smooth vector field can be computed
directly from this formula. This observation will be applied in (d) below.
(d) It is straightforward to check that the inverse of F is given by the formula

−1
 u u
F (u, v) = u − , .
v v
The differential of F at an arbitrary point (x, y) ∈ M is represented by the Jacobian of
F at (x, y), given by  
1 1
DF (x, y) =  1 x ,
− 2
y y
and thus dFF −1 (u,v) is represented by the matrix
 
1 1
 u u 
DF u − , = v
 v − v2  .
v v
u u
For any (u, v) ∈ N we have

u2 ∂
XF −1 (u,v) = X(u− u , u ) = .
v v v 2 ∂x (u− u , u )
v v

Therefore, we obtain
u2 ∂ u ∂
(F∗ X)(u,v) = 2 + .
v ∂u (u,v) v ∂v (u,v)

(e) Assume first that X and Y are F -related. Let γ be an integral curve of X. By
definition and by [Exercise Sheet 4, Exercise 5(b)] we obtain

(F ◦ γ)′ (t) = dFγ(t) γ ′ (t) = dFγ(t) Xγ(t) = YF (γ(t)) = Y(F ◦γ)(t) ,

4
which shows that F ◦ γ is an integral curve of Y .
Assume now that F takes integral curves of X to integral curves of Y . Let p ∈ M
and let γ : (−ε, ε) → M be an integral curve of X starting at p. Then γ(0) = p and
γ ′ (0) = Xp . Moreover, by assumption, F ◦ γ : (−ε, ε) → N is an integral curve of Y
starting at F (p), so Y(F ◦γ)(0) = (F ◦ γ)′ (0). Therefore, by [Exercise Sheet 4, Exercise 5(b)]
we obtain
YF (p) = (F ◦ γ)′ (0) = dFp γ ′ (0) = dFp Xp .

Since p ∈ M was arbitrary, we conclude that X and Y are F -related.

Exercise 3: Let M be a smooth manifold and let X and Y be two smooth vector fields
on M . Show that the Lie bracket [X, Y ] of X and Y , defined by

[X, Y ] : C ∞ (M ) → C ∞ (M ), f 7→ XY f − Y Xf,

is also a smooth vector field on X.

Solution: The R-linearity of [X, Y ] follows immediately from the R-linearity of both X
and Y . Let us now verify the product rule:

[X, Y ](f g) = XY (f g) − Y X(f g)

= X f Y g + gY f − Y f Xg + gXf


= Xf
Yg + f XY g + Xg Y f + g XY f −


Y
 f Xg
 + f Y Xg + Y g Xf + g Y Xf



= f XY g − Y Xg + g XY f − Y Xf
= f [X, Y ]g + g [X, Y ]f.

In conclusion, [X, Y ] is a smooth vector field on M by Proposition 7.5.

Exercise 4: Let M be a smooth n-manifold and let X, Y ∈ X(M ).

(a) Coordinate formula for the Lie bracket: Let

n n
X
i ∂ X ∂
X= X and Y = Yj
i=1
∂xi j=1
∂xj

be the coordinate expressions for X and Y , respectively, in terms of some smooth

local coordinates (xi ) for M . Show that the Lie bracket [X, Y ] has the following
coordinate expression:
n X n  j j

i ∂Y i ∂X ∂
X
[X, Y ] = X i
−Y i
.
j=1 i=1
∂x ∂x ∂xj

 ∂ ∂ 
(b) Compute the Lie brackets

,
∂xi ∂xj
of the coordinate vector fields ∂/∂xi in any
smooth chart U, (xi ) for M .

5
(c) Assume now that M = R3 ,
∂ ∂ ∂ ∂ ∂
X=x + + x(y + 1) and Y = +y ,
∂x ∂y ∂z ∂x ∂z
and compute the Lie bracket [X, Y ].
Solution:
(a) Denote by U ⊆ M the coordinate domain. For any f ∈ C ∞ (U ) we have
! !
X ∂f X ∂f
[X, Y ](f ) = XY (f ) − Y X(f ) = X Yj j −Y Xi i
j
∂x i
∂x
     
j ∂f ∂f i ∂f ∂f
X X
j i
= X(Y ) j + Y X j
− Y (X ) i + X Y
j
∂x ∂x i
∂x ∂xi
X  ∂Y j ∂f   X  i
 
i j i ∂ ∂f j ∂X ∂f i j ∂ ∂f
= X +Y X − Y +X Y
j,i
∂xi ∂xj ∂xi ∂xj i,j
∂xj ∂xi ∂xj ∂xi
X  ∂Y j j
     
i ∂X ∂f ∂ ∂f ∂ ∂f
X
i i j
= X i
−Y i j
− (X Y ) i j
− j
i,j
∂x ∂x ∂x i,j
∂x ∂x ∂x ∂xi
| {z }
=0
X  ∂Y j j

i i ∂X ∂f
= X i
− Y i j
.
i,j
∂x ∂x ∂x

(b) Recall that the component functions of each coordinate vector
field ∂/∂xj in the
coordinate frame (∂/∂xi ) associated with the smooth chart U, (xi ) are constant, so it
follows immediately from (a) that
 
∂ ∂
, = 0.
∂xi ∂xj
(c) By part (a) we obtain

∂
[X, Y ] = (x · 0 − 1 · 1) + (1 · 0 − 0 · 0) + (x(y + 1) · 0 − 1 · 0)
∂x

∂
+ (x · 0 − 1 · 0) + (x · 0 − 0 · 0) + (x(y + 1) · 0 − y · 0)
∂y

∂
+ (x · 0 − 1 · (y + 1)) + (1 · 1 − 0 · x) + (x(y + 1) · 0 − y · 0)
∂z
∂ ∂
=− −y .
∂x ∂z

Exercise 5 (Properties of the Lie bracket): Let M be a smooth manifold. Show that the
Lie bracket satisfies the following identities for all X, Y, Z ∈ X(M ):
(a) Bilinearity: For all a, b ∈ R we have
[aX + bY, Z] = a[X, Z] + b[Y, Z],
[Z, aX + bY ] = a[Z, X] + b[Z, Y ].

6
(b) Antisymmetry:
[X, Y ] = −[Y, X].

(c) Jacobi identity:      

X, [Y, Z] + Y, [Z, X] + Z, [X, Y ] = 0.

(d) For all f, g ∈ C ∞ (M ) we have

[f X, gY ] = f g[X, Y ] + (f Xg)Y − (gY f )X.

Solution:
(a) We first make the following observation: given λ, µ ∈ R and U, V, W ∈ X(M ), for any
f ∈ C ∞ (M ) it holds that

(λV + µW ) U f = λV U f + µW U f and U (λV + µW )f = λU V f + µU W f.

Indeed, for any p ∈ M we have

(λV + µW ) U f (p) = (λV + µW )p (U f ) = (λVp + µWp )(U f )
= λVp (U f ) + µWp (U f ) = λV (U f )(p) + µW (U f )(p)


= λV U f + µW U f (p),

which yields the first equality above, while the second one is obtained analogously.
Now, given a, b ∈ R, using the previous observation, for any f ∈ C ∞ (M ) we have



[aX + bY, Z](f ) = aX + bY Zf − Z aX + bY f
= aXZf + bY Zf − aZXf − bZY f



= a XZf − ZXf + b Y Zf − ZY f
= a[X, Z](f ) + b[Y, Z](f )


= a[X, Z] + b[Y, Z] (f ),

which yields the first part of the statement, while the second one is obtained similarly.
(b) For any f ∈ C ∞ (M ) we have


[X, Y ](f ) = XY f − Y Xf = − Y Xf − XY f = −[Y, X](f ),

which yields the statement.

(c) By expanding all the brackets and using linearity we obtain
     
X, [Y, Z] + Y, [Z, X] + Z, [X, Y ] =
= X[Y, Z] − [Y, Z]X + Y [Z, X] − [Z, X]Y + Z[X, Y ] − [X, Y ]Z
= XY Z − XZY − Y ZX + ZY X + Y ZX − Y XZ − ZXY + XZY +
+ ZXY − ZY X − XY Z + Y XZ
= 0.

7
(d) We first make the following observation: if V ∈ X(M ) and s, t ∈ C ∞ (M ), then

(sV ) h = s (V h) (as smooth functions on M ),

since for any p ∈ M we have

(sV ) h (p) = (sV )p h = s(p)Vp h = s(p)Vp h = s(p)(V h)(p) = s (V h) (p).

Now, fix f, g ∈ C ∞ (M ). Using the previous observation and the fact that smooth
vector fields are derivations of C ∞ (M ) by Proposition 7.5, for any h ∈ C ∞ (M ) we have

[f X, gY ](h) = (f X)(gY )(h) − (gY )(f X)(h)

= (f X) g(Y h) − (gY ) f (Xh)
= g(f X)(Y h) + (Y h)(f X)(g) − f (gY )(Xh) − (Xh)(gY )(f )



= gf X(Y h) + f (Xg)(Y h) − f g Y (Xh) − g(Y f )(Xh)


= f g (XY − Y X)(h) + (f Xg)Y (h) − (gY f )X(h)


= f g[X, Y ] + (f Xg)Y − (gY f )X (h),

whence the desired relation.

Remark. A Lie algebra (over R) is an R-vector space g endowed with a map g × g →

g, called the Lie bracket and usually denoted by (X, Y ) 7→ [X, Y ], which satisfies the
following properties for all X, Y, Z ∈ g:
(a) Bilinearity: For all a, b ∈ R we have

[aX + bY, Z] = a[X, Z] + b[Y, Z],

[Z, aX + bY ] = a[Z, X] + b[Z, Y ].

(b) Antisymmetry:
[X, Y ] = −[Y, X].

(c) Jacobi identity:      

X, [Y, Z] + Y, [Z, X] + Z, [X, Y ] = 0.

According to Exercise 4, the infinite-dimensional R-vector space X(M ) of all smooth

vector fields on a smooth manifold M is a Lie algebra under the Lie bracket. Here are
two more examples of Lie algebras:
(1) The R-vector space Mn (R) of real n × n matrices equipped with the commutator
bracket [A, B] := AB − BA becomes an n2 -dimensional Lie algebra, which is denoted
by gl(n, R).

(2) If V is an R-vector space, then the R-vector space of all linear maps from V to itself
equipped with the commutator bracket [L1 , L2 ] := L1 ◦ L2 − L2 ◦ L1 becomes a Lie
algebra, which is denoted by gl(V ).
Under our usual identification of n × n matrices with linear maps from Rn to itself,
gl(Rn ) is the same as gl(n, R).

8
.

Exercise 6: Let F : M → N be a smooth map.

(a) Naturality of the Lie bracket: Let X1 , X2 ∈ X(M ) and Y1 , Y2 ∈ X(N ) be smooth
vector fields such that Xi is F -related to Yi for i ∈ {1, 2}. Show that [X1 , X2 ] is
F -related to [Y1 , Y2 ].

(b) Pushforwards of Lie brackets: Assume that F is a diffeomorphism and consider

X1 , X2 ∈ X(M ). Show that F∗ [X1 , X2 ] = [F∗ X1 , F∗ X2 ].
Solution:
(a) Since Xi is F -related to Yi for i ∈ {1, 2}, by Exercise 2(a) we infer that for every
smooth real-valued function f defined on an open subset of N we have

X1 (f ◦ F ) = (Y1 f ) ◦ F and X2 (f ◦ F ) = (Y2 f ) ◦ F.

Therefore,

[X1 , X2 ](f ◦ F ) = X1 X2 (f ◦ F ) − X2 X1 (f ◦ F )



= X1 (Y2 f ) ◦ F − X2 (Y1 f ) ◦ F



= Y1 (Y2 f ) ◦ F − Y2 (Y1 f ) ◦ F


= [Y1 , Y2 ](f ) ◦ F,

and thus Exercise 1(a) implies that [X1 , X2 ] is F -related to [Y1 , Y2 ] .

(b) Follows immediately from part (a) and from Exercise 2(c).

Exercise 7:
(a) Restricting smooth vector fields to submanifolds: Let M be a smooth manifold, let S
be an immersed submanifold of M , and let ι : S ,→ M be the inclusion map. Prove
the following assertions:

(i) If Y ∈ X(M ) and if there is X ∈ X(S) that is ι-related to Y , then Y ∈ X(M ) is

tangent to S.
(ii) If Y ∈ X(M ) is tangent to S, then there is a unique smooth vector field on S,
denoted by Y |S , which is ι-related to Y .
[Hint: Determine first the candidate vector field on S and then use Theorem 5.6
and Proposition 5.17 to show that it is smooth.]

(b) Lie brackets of smooth vector fields tangent to submanifolds: Let M be a smooth
manifold and let S be an immersed submanifold of M . If Y1 and Y2 are smooth vector
fields on M that are tangent to S, then show that their Lie bracket [Y1 , Y2 ] is also
tangent to S.
Solution:
(a)(i) Since X is ι-related to Y , it holds that Yp = dιp (Xp ) for all p ∈ S, which means
that Yp ∈ Tp S for all p ∈ S, i.e., Y is tangent to S.

9
(a)(ii) Since by hypothesis we have Yp ∈ Tp S ∼ = dιp (Tp S) for all p ∈ S, we may define a
rough vector field X : S → T S by requiring that, for any p ∈ S, its value Xp ∈ Tp S is
the unique vector such that dιp (Xp ) = Yp . By the injectivity of dιp , it is clear that X is
unique, and that it is ι-related to Y , so it remains to show that X is smooth. To this
end, let p ∈ S be arbitrary. By Proposition 5.17 there is an open neighborhood V of p in

S such that V is embedded in M . By Theorem 5.6 there exists a smooth chart U, (xi )
for M such that V ∩ U is a k-slice in U – we may assume that V ∩ U is the k-slice given
by xk+1 = . . . = xn = 0 – and (x1 , . . . , xk ) are local coordinates for S in V ∩ U . Consider
the coordinate representation
n
X ∂
Y = Yi
i=1
∂xi

of Y on U . Since Y is tangent to S, by Proposition 7.8 (evaluating the above expression

at the coordinate function xi with i > k) we infer that Y k+1 = . . . = Y n = 0 on V ∩ U .
Therefore,
X ∂
X= Y i |U ∩V i
1≤i≤k
∂x U ∩V

is the coordinate representation of X on V ∩ U , and each component Y i |U ∩V is smooth

by [Exercise Sheet 8, Exercise 5(a)], so X is smooth on U ∩ V , and we are done.
(b) Since Y1 , Y2 ∈ X(M ) are tangent to S, by part (a)(ii) there exist (unique) smooth
vector fields W1 , W2 on S such that Wj is ι-related to Yj for each j ∈ {1, 2}, where
ι : S ,→ M is the inclusion map. But now Exercise 6(a) implies that [W1 , W2 ] ∈ X(S) is
ι-related to [Y1 , Y2 ] ∈ X(M ), and thus [Y1 , Y2 ] is tangent to S by part (a)(i).

10