We define the exponential map at
p
∈
M
{\displaystyle p\in M}
by
exp
p
:
T
p
M
⊃
B
ϵ
(
0
)
⟶
M
,
v
t
⟼
γ
p
,
v
(
t
)
,
{\displaystyle \exp _{p}:T_{p}M\supset B_{\epsilon }(0)\longrightarrow M,\quad vt\longmapsto \gamma _{p,v}(t),}
where
γ
p
,
v
{\displaystyle \gamma _{p,v}}
is the unique geodesic with
γ
p
,
v
(
0
)
=
p
{\displaystyle \gamma _{p,v}(0)=p}
and tangent
γ
p
,
v
′
(
0
)
=
v
∈
T
p
M
{\displaystyle \gamma _{p,v}'(0)=v\in T_{p}M}
and
ϵ
{\displaystyle \epsilon }
is chosen small enough so that for every
t
∈
[
0
,
1
]
,
v
t
∈
B
ϵ
(
0
)
⊂
T
p
M
{\displaystyle t\in [0,1],vt\in B_{\epsilon }(0)\subset T_{p}M}
the geodesic
γ
p
,
v
(
t
)
{\displaystyle \gamma _{p,v}(t)}
is defined. So, if
M
{\displaystyle M}
is complete, then, by the Hopf–Rinow theorem ,
exp
p
{\displaystyle \exp _{p}}
is defined on the whole tangent space.
Let
α
:
I
→
T
p
M
{\displaystyle \alpha :I\rightarrow T_{p}M}
be a curve differentiable in
T
p
M
{\displaystyle T_{p}M}
such that
α
(
0
)
:=
0
{\displaystyle \alpha (0):=0}
and
α
′
(
0
)
:=
v
{\displaystyle \alpha '(0):=v}
. Since
T
p
M
≅
R
n
{\displaystyle T_{p}M\cong \mathbb {R} ^{n}}
, it is clear that we can choose
α
(
t
)
:=
v
t
{\displaystyle \alpha (t):=vt}
. In this case, by the definition of the differential of the exponential in
0
{\displaystyle 0}
applied over
v
{\displaystyle v}
, we obtain:
T
0
exp
p
(
v
)
=
d
d
t
(
exp
p
∘
α
(
t
)
)
|
t
=
0
=
d
d
t
(
exp
p
(
v
t
)
)
|
t
=
0
=
d
d
t
(
γ
p
,
v
(
t
)
)
|
t
=
0
=
γ
p
,
v
′
(
0
)
=
v
.
{\displaystyle T_{0}\exp _{p}(v)={\frac {\mathrm {d} }{\mathrm {d} t}}{\Bigl (}\exp _{p}\circ \alpha (t){\Bigr )}{\Big \vert }_{t=0}={\frac {\mathrm {d} }{\mathrm {d} t}}{\Bigl (}\exp _{p}(vt){\Bigr )}{\Big \vert }_{t=0}={\frac {\mathrm {d} }{\mathrm {d} t}}{\Bigl (}\gamma _{p,v}(t){\Bigr )}{\Big \vert }_{t=0}=\gamma _{p,v}'(0)=v.}
So (with the right identification
T
0
T
p
M
≅
T
p
M
{\displaystyle T_{0}T_{p}M\cong T_{p}M}
) the differential of
exp
p
{\displaystyle \exp _{p}}
is the identity. By the implicit function theorem,
exp
p
{\displaystyle \exp _{p}}
is a diffeomorphism on a neighborhood of
0
∈
T
p
M
{\displaystyle 0\in T_{p}M}
. The Gauss Lemma now tells that
exp
p
{\displaystyle \exp _{p}}
is also a radial isometry.
The exponential map is a radial isometry
edit
Let
p
∈
M
{\displaystyle p\in M}
. In what follows, we make the identification
T
v
T
p
M
≅
T
p
M
≅
R
n
{\displaystyle T_{v}T_{p}M\cong T_{p}M\cong \mathbb {R} ^{n}}
.
Gauss's Lemma states:
Let
v
,
w
∈
B
ϵ
(
0
)
⊂
T
v
T
p
M
≅
T
p
M
{\displaystyle v,w\in B_{\epsilon }(0)\subset T_{v}T_{p}M\cong T_{p}M}
and
M
∋
q
:=
exp
p
(
v
)
{\displaystyle M\ni q:=\exp _{p}(v)}
. Then,
⟨
T
v
exp
p
(
v
)
,
T
v
exp
p
(
w
)
⟩
q
=
⟨
v
,
w
⟩
p
.
{\displaystyle \langle T_{v}\exp _{p}(v),T_{v}\exp _{p}(w)\rangle _{q}=\langle v,w\rangle _{p}.}
For
p
∈
M
{\displaystyle p\in M}
, this lemma means that
exp
p
{\displaystyle \exp _{p}}
is a radial isometry in the following sense: let
v
∈
B
ϵ
(
0
)
{\displaystyle v\in B_{\epsilon }(0)}
, i.e. such that
exp
p
{\displaystyle \exp _{p}}
is well defined.
And let
q
:=
exp
p
(
v
)
∈
M
{\displaystyle q:=\exp _{p}(v)\in M}
. Then the exponential
exp
p
{\displaystyle \exp _{p}}
remains an isometry in
q
{\displaystyle q}
, and, more generally, all along the geodesic
γ
{\displaystyle \gamma }
(in so far as
γ
p
,
v
(
1
)
=
exp
p
(
v
)
{\displaystyle \gamma _{p,v}(1)=\exp _{p}(v)}
is well defined)! Then, radially, in all the directions permitted by the domain of definition of
exp
p
{\displaystyle \exp _{p}}
, it remains an isometry.
The exponential map as a radial isometry
Recall that
T
v
exp
p
:
T
p
M
≅
T
v
T
p
M
⊃
T
v
B
ϵ
(
0
)
⟶
T
exp
p
(
v
)
M
.
{\displaystyle T_{v}\exp _{p}\colon T_{p}M\cong T_{v}T_{p}M\supset T_{v}B_{\epsilon }(0)\longrightarrow T_{\exp _{p}(v)}M.}
We proceed in three steps:
T
v
exp
p
(
v
)
=
v
{\displaystyle T_{v}\exp _{p}(v)=v}
: let us construct a curve
α
:
R
⊃
I
→
T
p
M
{\displaystyle \alpha :\mathbb {R} \supset I\rightarrow T_{p}M}
such that
α
(
0
)
:=
v
∈
T
p
M
{\displaystyle \alpha (0):=v\in T_{p}M}
and
α
′
(
0
)
:=
v
∈
T
v
T
p
M
≅
T
p
M
{\displaystyle \alpha '(0):=v\in T_{v}T_{p}M\cong T_{p}M}
. Since
T
v
T
p
M
≅
T
p
M
≅
R
n
{\displaystyle T_{v}T_{p}M\cong T_{p}M\cong \mathbb {R} ^{n}}
, we can put
α
(
t
)
:=
v
(
t
+
1
)
{\displaystyle \alpha (t):=v(t+1)}
.
Therefore,
T
v
exp
p
(
v
)
=
d
d
t
(
exp
p
∘
α
(
t
)
)
|
t
=
0
=
d
d
t
(
exp
p
(
t
v
)
)
|
t
=
1
=
Γ
(
γ
)
p
exp
p
(
v
)
v
=
v
,
{\displaystyle T_{v}\exp _{p}(v)={\frac {\mathrm {d} }{\mathrm {d} t}}{\Bigl (}\exp _{p}\circ \alpha (t){\Bigr )}{\Big \vert }_{t=0}={\frac {\mathrm {d} }{\mathrm {d} t}}{\Bigl (}\exp _{p}(tv){\Bigr )}{\Big \vert }_{t=1}=\Gamma (\gamma )_{p}^{\exp _{p}(v)}v=v,}
where
Γ
{\displaystyle \Gamma }
is the parallel transport operator and
γ
(
t
)
=
exp
p
(
t
v
)
{\displaystyle \gamma (t)=\exp _{p}(tv)}
. The last equality is true because
γ
{\displaystyle \gamma }
is a geodesic, therefore
γ
′
{\displaystyle \gamma '}
is parallel.
Now let us calculate the scalar product
⟨
T
v
exp
p
(
v
)
,
T
v
exp
p
(
w
)
⟩
{\displaystyle \langle T_{v}\exp _{p}(v),T_{v}\exp _{p}(w)\rangle }
.
We separate
w
{\displaystyle w}
into a component
w
T
{\displaystyle w_{T}}
parallel to
v
{\displaystyle v}
and a component
w
N
{\displaystyle w_{N}}
normal to
v
{\displaystyle v}
. In particular, we put
w
T
:=
a
v
{\displaystyle w_{T}:=av}
,
a
∈
R
{\displaystyle a\in \mathbb {R} }
.
The preceding step implies directly:
⟨
T
v
exp
p
(
v
)
,
T
v
exp
p
(
w
)
⟩
=
⟨
T
v
exp
p
(
v
)
,
T
v
exp
p
(
w
T
)
⟩
+
⟨
T
v
exp
p
(
v
)
,
T
v
exp
p
(
w
N
)
⟩
{\displaystyle \langle T_{v}\exp _{p}(v),T_{v}\exp _{p}(w)\rangle =\langle T_{v}\exp _{p}(v),T_{v}\exp _{p}(w_{T})\rangle +\langle T_{v}\exp _{p}(v),T_{v}\exp _{p}(w_{N})\rangle }
=
a
⟨
T
v
exp
p
(
v
)
,
T
v
exp
p
(
v
)
⟩
+
⟨
T
v
exp
p
(
v
)
,
T
v
exp
p
(
w
N
)
⟩
=
⟨
v
,
w
T
⟩
+
⟨
T
v
exp
p
(
v
)
,
T
v
exp
p
(
w
N
)
⟩
.
{\displaystyle =a\langle T_{v}\exp _{p}(v),T_{v}\exp _{p}(v)\rangle +\langle T_{v}\exp _{p}(v),T_{v}\exp _{p}(w_{N})\rangle =\langle v,w_{T}\rangle +\langle T_{v}\exp _{p}(v),T_{v}\exp _{p}(w_{N})\rangle .}
We must therefore show that the second term is null, because, according to Gauss's Lemma, we must have:
⟨
T
v
exp
p
(
v
)
,
T
v
exp
p
(
w
N
)
⟩
=
⟨
v
,
w
N
⟩
=
0.
{\displaystyle \langle T_{v}\exp _{p}(v),T_{v}\exp _{p}(w_{N})\rangle =\langle v,w_{N}\rangle =0.}
⟨
T
v
exp
p
(
v
)
,
T
v
exp
p
(
w
N
)
⟩
=
0
{\displaystyle \langle T_{v}\exp _{p}(v),T_{v}\exp _{p}(w_{N})\rangle =0}
:
The curve chosen to prove lemma
Let us define the curve
α
:
[
−
ϵ
,
ϵ
]
×
[
0
,
1
]
⟶
T
p
M
,
(
s
,
t
)
⟼
t
v
+
t
s
w
N
.
{\displaystyle \alpha \colon [-\epsilon ,\epsilon ]\times [0,1]\longrightarrow T_{p}M,\qquad (s,t)\longmapsto tv+tsw_{N}.}
Note that
α
(
0
,
1
)
=
v
,
∂
α
∂
t
(
s
,
t
)
=
v
+
s
w
N
,
∂
α
∂
s
(
0
,
t
)
=
t
w
N
.
{\displaystyle \alpha (0,1)=v,\qquad {\frac {\partial \alpha }{\partial t}}(s,t)=v+sw_{N},\qquad {\frac {\partial \alpha }{\partial s}}(0,t)=tw_{N}.}
Let us put:
f
:
[
−
ϵ
,
ϵ
]
×
[
0
,
1
]
⟶
M
,
(
s
,
t
)
⟼
exp
p
(
t
v
+
t
s
w
N
)
,
{\displaystyle f\colon [-\epsilon ,\epsilon ]\times [0,1]\longrightarrow M,\qquad (s,t)\longmapsto \exp _{p}(tv+tsw_{N}),}
and we calculate:
T
v
exp
p
(
v
)
=
T
α
(
0
,
1
)
exp
p
(
∂
α
∂
t
(
0
,
1
)
)
=
∂
∂
t
(
exp
p
∘
α
(
s
,
t
)
)
|
t
=
1
,
s
=
0
=
∂
f
∂
t
(
0
,
1
)
{\displaystyle T_{v}\exp _{p}(v)=T_{\alpha (0,1)}\exp _{p}\left({\frac {\partial \alpha }{\partial t}}(0,1)\right)={\frac {\partial }{\partial t}}{\Bigl (}\exp _{p}\circ \alpha (s,t){\Bigr )}{\Big \vert }_{t=1,s=0}={\frac {\partial f}{\partial t}}(0,1)}
and
T
v
exp
p
(
w
N
)
=
T
α
(
0
,
1
)
exp
p
(
∂
α
∂
s
(
0
,
1
)
)
=
∂
∂
s
(
exp
p
∘
α
(
s
,
t
)
)
|
t
=
1
,
s
=
0
=
∂
f
∂
s
(
0
,
1
)
.
{\displaystyle T_{v}\exp _{p}(w_{N})=T_{\alpha (0,1)}\exp _{p}\left({\frac {\partial \alpha }{\partial s}}(0,1)\right)={\frac {\partial }{\partial s}}{\Bigl (}\exp _{p}\circ \alpha (s,t){\Bigr )}{\Big \vert }_{t=1,s=0}={\frac {\partial f}{\partial s}}(0,1).}
Hence
⟨
T
v
exp
p
(
v
)
,
T
v
exp
p
(
w
N
)
⟩
=
⟨
∂
f
∂
t
,
∂
f
∂
s
⟩
(
0
,
1
)
.
{\displaystyle \langle T_{v}\exp _{p}(v),T_{v}\exp _{p}(w_{N})\rangle =\left\langle {\frac {\partial f}{\partial t}},{\frac {\partial f}{\partial s}}\right\rangle (0,1).}
We can now verify that this scalar product is actually independent of the variable
t
{\displaystyle t}
, and therefore that, for example:
⟨
∂
f
∂
t
,
∂
f
∂
s
⟩
(
0
,
1
)
=
⟨
∂
f
∂
t
,
∂
f
∂
s
⟩
(
0
,
0
)
=
0
,
{\displaystyle \left\langle {\frac {\partial f}{\partial t}},{\frac {\partial f}{\partial s}}\right\rangle (0,1)=\left\langle {\frac {\partial f}{\partial t}},{\frac {\partial f}{\partial s}}\right\rangle (0,0)=0,}
because, according to what has been given above:
lim
t
→
0
∂
f
∂
s
(
0
,
t
)
=
lim
t
→
0
T
t
v
exp
p
(
t
w
N
)
=
0
{\displaystyle \lim _{t\rightarrow 0}{\frac {\partial f}{\partial s}}(0,t)=\lim _{t\rightarrow 0}T_{tv}\exp _{p}(tw_{N})=0}
being given that the differential is a linear map. This will therefore prove the lemma.
We verify that
∂
∂
t
⟨
∂
f
∂
t
,
∂
f
∂
s
⟩
=
0
{\displaystyle {\frac {\partial }{\partial t}}\left\langle {\frac {\partial f}{\partial t}},{\frac {\partial f}{\partial s}}\right\rangle =0}
: this is a direct calculation. Since the maps
t
↦
f
(
s
,
t
)
{\displaystyle t\mapsto f(s,t)}
are geodesics,
∂
∂
t
⟨
∂
f
∂
t
,
∂
f
∂
s
⟩
=
⟨
D
∂
t
∂
f
∂
t
⏟
=
0
,
∂
f
∂
s
⟩
+
⟨
∂
f
∂
t
,
D
∂
t
∂
f
∂
s
⟩
=
⟨
∂
f
∂
t
,
D
∂
s
∂
f
∂
t
⟩
=
1
2
∂
∂
s
⟨
∂
f
∂
t
,
∂
f
∂
t
⟩
.
{\displaystyle {\frac {\partial }{\partial t}}\left\langle {\frac {\partial f}{\partial t}},{\frac {\partial f}{\partial s}}\right\rangle =\left\langle \underbrace {{\frac {D}{\partial t}}{\frac {\partial f}{\partial t}}} _{=0},{\frac {\partial f}{\partial s}}\right\rangle +\left\langle {\frac {\partial f}{\partial t}},{\frac {D}{\partial t}}{\frac {\partial f}{\partial s}}\right\rangle =\left\langle {\frac {\partial f}{\partial t}},{\frac {D}{\partial s}}{\frac {\partial f}{\partial t}}\right\rangle ={\frac {1}{2}}{\frac {\partial }{\partial s}}\left\langle {\frac {\partial f}{\partial t}},{\frac {\partial f}{\partial t}}\right\rangle .}
Since the maps
t
↦
f
(
s
,
t
)
{\displaystyle t\mapsto f(s,t)}
are geodesics,
the function
t
↦
⟨
∂
f
∂
t
,
∂
f
∂
t
⟩
{\displaystyle t\mapsto \left\langle {\frac {\partial f}{\partial t}},{\frac {\partial f}{\partial t}}\right\rangle }
is constant. Thus,
∂
∂
s
⟨
∂
f
∂
t
,
∂
f
∂
t
⟩
=
∂
∂
s
⟨
v
+
s
w
N
,
v
+
s
w
N
⟩
=
2
⟨
v
,
w
N
⟩
=
0.
{\displaystyle {\frac {\partial }{\partial s}}\left\langle {\frac {\partial f}{\partial t}},{\frac {\partial f}{\partial t}}\right\rangle ={\frac {\partial }{\partial s}}\left\langle v+sw_{N},v+sw_{N}\right\rangle =2\left\langle v,w_{N}\right\rangle =0.}