Differential Geometry
Differentiable Manifolds
Definition of topological manifold : It is a topological space (E, τ) so that
1. It is Hausdorff. 2. ∀x ∈ E there exists (U, ϕ) with U open and x ∈ U, such that ϕ : U → ϕ(U) is a homeomorphism. The pair (U, ϕ) is called chart and the real numbers (x1, . . . , xn) = ϕ(x) are called local coordinates. 3. (E, τ) has a countable basis of open sets.
Remark. Topological Manifolds are paracompact, i.e, every open cover has a locally finite refinement.
Remark. Paracompactness implies having Partition of unity.
Definition. A differentiable [C∞,Ck,Cω] structure on a topological ma- nifold M is a family of charts U = {(Uα, ϕα} so that S 1. Uα = M −1 2. If Uα ∩ Uβ 6= ∅ then ϕβ ◦ ϕα : ϕα(Uα ∩ Uβ) → ϕβ(Uα ∩ Uβ) are differentiable [C∞,Ck,Cω] with differentiable [C∞,Ck,Cω] inverse. In this case we say that (Uα, ϕα) and (Uβ, ϕβ) are compatible. 3. Completness property: If (V, ψ) is a chart which is compatible with every (Uα, ϕα) ∈ U then (V, ψ) ∈ U. Remark.It is not necessary to verify the third property because of the following proposition.
Proposition. Let M be Hausdorff with countable basis of open sets. Let {(Vα, ψα): α ∈ A} be a covering of M by C∞-compatible coordinate charts. Then ∃! C∞ structure containing the charts {(Vα, ψα): α ∈ A}. Examples of topological manifolds:
1. (Rn, Id) 2. If M is a C∞ manifold and N ⊂ M is an open subset then N is a C∞ manifold too.
1 3. By the previous example GL(n, R) is a manifold. 4. If M and N are C∞ manifolds so is M × N. 5. S2 is a Cω manifold.
X Definition of open equivalence relation: Let π : X → ∼ be the projection S −1 and [U] := x∈U {π (x)}. We will say that ∼ is open iff [U] is open for every U open.
Proposition.
1. ∼ is open iff π : X → X/ ∼ is open. 2. If ∼ is open and X has a countable basis of open sets, then so does X/ ∼.
Definition. Let ∼ be an equiv. relation on X. The graph R of ∼ is the subset of X × X defined by R = {(x, y): x ∼ y}.
Proposition. Let ∼ be an open eq. relation on a Hausdorff space. Then X/ ∼ is Hausdorff iff R is closed in X × X.
Applications.
n+1 1. Pn(R) is a manifold because ∼ is open and its graph is closed .(Recall, R −0 is endowed with the equivalence relation x ∼ y iff ∃t 6= 0 so that x = ty. n−1 R −0 Pn(R) := ∼ .) 2. G(n, k) is a C∞ manifold of dimension k(n−k). ("G(n, k) = {k-dim’l subspaces n F (n,k) kn of R }", actually G(n, k) := ∼ where F (k, n) is a manifold in R of k- frames in Rn).
Definition. Let M be a manifold and W ⊂ M open. We will say that f : W → R is a differentiable [C∞,Ck,Cω] map if ∀x ∈ W ∃(U, ϕ) coordinate chart (with x ∈ U) so that f ◦ ϕ−1 : ϕ(W ∩ U) → R is C∞.
Definition. Let A ⊂ M. We say that f : A → R is C∞ if it has a C∞ extension to an open set U ⊂ M such that A ⊂ U.
Definition. Let M and N be differentiable [C∞,Ck,Cω] manifolds. We will say that f : M → N is a differentiable [C∞,Ck,Cω] map if ∀x ∈ M ∃(U, ϕ) coordinate chart with x ∈ U and (V, ψ) coordinate with f(x) ∈ V so that f(U) ⊂ V and ψ ◦f ◦ϕ−1 : ϕ(W ∩U) → R is differentiable [C∞,Ck,Cω] as a map of euclidean spaces.
2 Definition. Let M and N be C∞ manifolds and f : M → N be a C∞ map. We define the rank of f as the rank of ψ ◦ f ◦ ϕ−1 where (U, ϕ) and (V, ψ) are charts as h ∂yi i −1 1 n above. That is, (rank f)x := rank ∂xj |ϕ(x) where ψ ◦ f ◦ ϕ = (y , . . . , y ). Definition. Let f, g be C∞ in a neighb. of x ∈ M. We will say that f ∼ g if ∃U open so that f(y) = g(y) ∀y ∈ U. The class [f] is called germ of C∞ function at x. The set of germs at x will be denoted by C∞(x, R).
Definition. Let x ∈ M and X : C∞(x, R) → R. If for every chart (U, ϕ) about x we have that there exist a1, . . . , an ∈ R so that
n X ∂ −1 X([f]) = ai (f ◦ ϕ ) , ∂x ϕ(x) i=1 i we will say that X is a tangent vector at x . Observation. If the equation holds for some chart (U, ϕ) about x, then it holds for every C∞-compatible chart overlapping at x.
Properties of tangent vectors:
1. X([f] + [g]) = X([f]) + X([g]) 2. X([λf]) = λX([f]) ∀λ ∈ R 3. X([f][g]) = X([f])[g] + [f]X([g])
Definition. The tangent space to M at x is the vector space of tangent vectors based at x. We will denote it by TxM.
Remark. If n = dimM then dimTxM = n.
∂ Definition. Given a coordinate chart (U, ϕ) about x ∈ M the basis ∂xi x , 1 ≤ i ≤ n is called the natural basis of TxM associated to this chart. ∂ ∂f◦ϕ−1 Here, ∂xi |x [f] = ∂xi |ϕ(x). Definition. Given f : M → N a C∞ map and x ∈ M we define the push-forward of f at x as follows:
f∗ |x: TxM → Tf(x)N (f∗ |x X)([g]) := X([g ◦ f])
∂ −1 ∂ Remark. With this definition we have ∂xi |x= ϕ∗ ( ∂xi |ϕ(x)).
3 Remark. Let ϕ and ψ be charts about x. Let
ψ ◦ ϕ−1(x1, . . . , xn) = (y1(x1, . . . , xn), . . . , yn(x1, . . . , xn)).
Then ∂ X ∂yk ∂ = (ϕ(x)) . ∂xi x ∂xi ∂yk x k Definition. We define the cotangent bundle to x at M to be the dual space ∗ of TxM. We denote it by Tx M.
i ∂ Remark. Let {dx |x, 1 ≤ i ≤ n} be the dual basis of ∂xi |x, 1 ≤ i ≤ n , then we ∗ will say that it is the natural basis of Tx M. Remark. Let ϕ and ψ be charts about x. Let
ϕ ◦ ψ−1(y1, . . . , yn) = (x1(y1, . . . , yn), . . . , xn(y1, . . . , yn)).
Then X ∂xi dxi = (ψ(x))dyk . x ∂yk x k Definition. The tangent bundle of M, denoted by TM is defined as follows: a TM := TxM. x∈M
Proposition. Let π : TM → M be defined by π(X) = x if X ∈ TxM. Then there is a unique topology on TM so that for each coordinate chart (U, ϕ) on M the set U˜ = π−1(U) is open and the map ϕ˜ : U˜ → ϕ(U) × Rn defined by n ! X i ∂ 1 n ϕ˜ v := (ϕ(x), v , . . . , v ) ∂xi i=1 x is a homeomorphism. With this topology, TM is a topological manifold of dimension 2n and the coordi- nate charts (U,˜ ϕ˜) define a differentiable [C∞,Ck,Cω] structure on TM relative to which π is an open mapping.
Definition. The cotangent bundle of M, denoted by T ∗M is defined as follows:
∗ a ∗ T M := Tx M. x∈M
4 ∞ Definition. A C map f : M → N is said to be an inmersion if f∗ is injective at every point.
Definition. An inmersed submanifold of a manifold N is a pair (M, f) where f : M → N is an injective inmersion.
Definition. An embedded submanifold of a manifold N is an inmersed sub- manifold (M, f) such that f : M → f(M) is a homeomorphism.
Theorem [Inverse function Theorem]. Let M and N be C∞ manifolds of dimension ∞ n. Let f : M → N be a C map. Suppose that xo ∈ M is such that
f∗ |xo : Txo M → Tf(xo)N is an isomorphism. Then there exists an open neighb. U of xo so that 1. f | U is injective. 2. f(U) is open in N. 3. f −1 : f(U) → U is C∞.
Theorem [Implicit function Theorem]. Let M and N be C∞ manifolds with ∞ −1 dimM > dimN and let f : M → N be C . Let yo ∈ f(M) and let Mo = f (yo). Suppose that ∀x ∈ Mo the map f∗ |x is surjective. Then Mo can be endowed with ∞ a C structure relative to which the inclusion map ı : Mo → M is an embedding. Furthermore, dimMo = dimM − dimN. Applications.
1. Sn is a C∞ manifold of dimension n. 2. SL(n, R) = {X ∈ GL(n, R) : det X = 1} is a C∞ manifold of dimension n2 − 1. t ∞ 3. O(n, R) = {X ∈ GL(n, R): X X = Idn} is a C manifold of dimension n(n − 1)/2.
Differential forms
Definition. Let V be a real n-dimensional space and let V ∗ be its dual space. We define the space of alternating k-forms as follows:
k ∗ Λ (V ) = {ω : V ⊕ · · · ⊕ V (k times) → R : ω is linear and alternating}.
k ∗ n!k! Observe that dimΛ (V ) = (n−k)! .
5 Remark. ω is linear and alternating if ω(v1, . . . , vn) is linear in each argument and
π ω(vπ(1), . . . , vπ(k)) = (−1) ω(v1, . . . , vn).
Definition. Let Λ0(V ∗) := R. We define the grassman algebra of V ∗ as follows n M Λ∗(V ∗) := Λk(V ∗). k=0 Observe that with this definition we have dimΛ∗(V ∗) = 2n. Λ∗(V ∗) will be endowed with a graded product operation as follows:
∧ :Λk(V ∗) × Λl(V ∗) → Λk+l(V ∗)(ω, τ) 7→ ω ∧ τ, where 1 X ω ∧ τ(v , . . . , v ) := (−1)πω(v , . . . , v )τ(v , . . . , v ). 1 k+l (k + l)! π(1) π(k) π(k+1) π(k+l) π∈Sk+l
Remark. If A ∈ End(V ∗) we extend A to an endomorphism of Λ∗(V ∗) in the following way:
A˜ :Λ∗(V ∗) → Λ∗(V ∗) A˜(ω ∧ τ) := Aω ∧ Aµ, ω, τ ∈ V ∗ = Λ1(V ∗).
A˜ preserves degrees by construction. Observe that if ω ∈ Λn(V ∗) then A˜ must be a multiple of ω. Actually, Aω˜ = det(A)ω.
Remark. Let M be a manifold and x ∈ M. Now we are able to construct the following spaces:
∗ ∗ Ln k ∗ • Λ (Tx M) := k=0 Λ (Tx M). k ∗ ` k ∗ • Λ (T M) := x∈M Λ (Tx M). ∗ ∗ ` ∗ ∗ • Λ (T M) := x∈M Λ (Tx M). Definition. Choose a chart (U, ϕ) about x with local coordinates (x1, . . . , xn). An k ∗ element ω |x ∈ Λ (Tx M) is called k-form at x and can be written as
X i1 ik ω |x= ai1...ik dx |x ∧ · · · ∧ dx |x .
1≤i1<... k ∗ n!k! Λ (T M) is a manifold of dimension n + (n−k)! . 6 Definition. We define a k-form on M as a section of the bundle π :Λk(T ∗M) → ∞ k ∗ M. That is, a C map ω : M → Λ (T M) so that π ◦ ω = idM . We will denote the space of k-forms on M by Ωk(M). ∗ Ln k 0 ∞ Notation. Ω (M) := k=0 Ω (M) and Ω (M) = C (M, R). Remark. For every k such that k > dimM we have Ωk(M) = 0. Definition. We define the operator d :Ω0(M) → Ω1(M) by df |x (X) := X([f]) X ∈ TxM. With this definition X ∂ i df |x= [f] dx |x . ∂xi i x Theorem. There exists a unique R-linear map d :Ω∗(M) → Ω∗(M) called the exterior derivative so that 1. d :Ωk(M) → Ωk+1(M). 2. d(f) = df defined as above for f ∈ C∞(M, R). 3. d(ω ∧ τ) = dω ∧ τ + (−1)degωω ∧ τ. 4. d2 = 0. Definition. Let f : M → N be a C∞ map. We define the pull back of f as the map f ∗ :Ω∗(N) → Ω∗(M) so that 1. f ∗(g) = g ◦ f for f ∈ Ω0(N) = C∞(N, R). ∗ k 2. f ω |x (X1,...,Xk) = ω |f(x) (f∗X1, . . . , f∗Xk) for ω ∈ Ω (N) with k ≥ 1. Properties of the pull-back map. 1. f ∗(ω ∧ τ) = f ∗ω ∧ f ∗τ 2. f ∗(gω + hτ) = f ∗(g) f ∗ω ∧ f ∗(h) f ∗τ 3. (f ◦ g)∗ = g∗ ◦ f ∗ Proposition. Pull-backs and d commute: d(f ∗ω) = f ∗(dω). n Proposition. M is orientable iff ∃ ω ∈ Ω (M) such that ω |x6= 0 for all x ∈ M. 7 Definition. An orientation on M is a choice of an equivalence class [ω] ∈ dim M ∞ + Ω (M) where ω1 ∼ ω2 iff ω1 = fω2 for some f ≥ 0, f ∈ C (M, R ). We denote M endowed with an orientation by [M]. Definition.Integral of n-forms. Let M be an orientable manifold of dimen- sion n. 1. If ω ∈ Ωn(Rn) has compact support, and ω = fdx1 ∧ · · · ∧ dxn then Z Z ω := fdx1 . . . dxn. n n R R 2. If ω ∈ Ωn(M) we define Z Z Z X X ∗ −1 ω = ραω := (ϕα) (ραω) [M] α∈a Uα α∈a ϕ(Uα) where {(Uα, ϕα): α ∈ A} is a positively oriented atlas and {ρα : α ∈ A} is a partition of unity subordinate to {(Uα, ϕα): α ∈ A}. Definition. A C∞ manifold with boundary is a Hausdorff space with a countable basis of open sets and a differentiable structure {(Uα, ϕα): α ∈ A} such n that it has compatibility on overlaps and ϕα(Uα) is open in H . We will denote the boundary of M by ∂M. Convention. If Hn is given the orientation [dx1 ∧ · · · ∧ dxn], then ∂Hn is given the orientation [(−1)ndx1 ∧ · · · ∧ dxn]. n−1 Theorem[Stokes]. If ω ∈ Ωc (M) where M is an oriented manifold with oriented boundary ∂M then Z Z dω = ω. [M] [∂M] Corollary. Let M be a compact manifold without boundary and θ ∈ Ωn(M). Then R if θ is exact we have M θ = 0. Definition. An open cover {Uα : α ∈ a} of a manifold M is said to be a good n cover if all non-empty intersections Uα0 ∩ · · · ∩ Uαp are diffeomorphic to R . Proposition. Every manifold admits a good cover. Corollary. Every compact manifold admits a finite good cover. Definition. A Riemannian Manifold is a pair (M, g) where M is a C∞ mani- fold and g is a map that assigns to any x ∈ M a non-degenerate symmetric positive definite bilinear form gx : TxM × TxM → R such that for all X,Y smooth vector fields on M, the map x 7→ gx(X,Y ) is smooth. 1 n ∂ Notation. Let (U, ϕ) be a chart with local coordinates (x , . . . , x ), and let { ∂x1 |x ∂ ,..., ∂xn |x} be the natural basis of TxM. Then we will adopt the following notation: ∂ ∂ gij(x) := gx , . ∂xi x ∂xj x ij We will also denote by g the entrances of the inverse matrix of (gij)ij. Proposition. Let M and N be manifolds and let g be a Riemannian metric on N. Let f : M → N be a C∞ inmersion. Then the map f ∗ defined as below defines a Riemannian metric -f ∗g- on M. ∗ (f g)x(Xx,Yx) := gf(x)(f∗Xx, f∗Yx). Theorem. Every manifold carries a Riemannian metric. Examples of Riemannian manifolds. 1. (R, gstand) n ∗ 2. If M ⊂ R then (M, ι gstand) is a Riemannian manifold isometrically immersed in an euclidean space. Here, ι : M → Rn is the inclusion map. 3. Lie Groups. 4. If (M, g) and (N, h) are Riemannian manifolds, we can endow M × N with the product metric: If π1 : M × N → M and π2 : M × N → M are the projections then the product metric is defined as follows ∗ ∗ g ⊕ h(X,Y ) := π1(g)(X,Y ) + π2(h) (X,Y ). Definition. Let (M, g) be a Riemannian manifold. An Affine Connection is a map ∇ : Γ(TM) × Γ(TM) → Γ(TM)(X,Y ) 7→ ∇X Y satisfying the following conditions: 9 ∞ 1. ∇fX+gY Z = f∇X Z + g∇Y Z for f, g ∈ C (M, R). 2. ∇X (Y + Z) = ∇X Y + ∇X Z. 3. ∇X (fY ) = f∇X Y + X(f)Y. Theorem[Fundamental theorem of Riemannian Geometry] . ∇ exists and is uni- que. i Definition. We define the Christoffel symbols Γjk : M → R so that ∂ X k ∂ ∂ ∇ j = Γij k . ∂xi ∂x ∂x k Then, 1 X ∂gjl ∂gjk ∂gkl Γi = gil − + . jk 2 ∂xk ∂xl ∂xj l P i ∂ P i ∂ Remark. Let X = i a ∂xi and Y = i b ∂xi . Then, ! X ∂bj ∂ X ∂ ∇ Y = ai + bjΓk . X ∂xi ∂xj ij ∂xk i k,i This allows us to define ∇Xx Y |x∈ TxM where Xx ∈ TxM and Y ∈ Γ(TM) as follows: P i ∂ i If Xx = i α ∂xi |x where now α is a number for all i (instead of a function) and Y is as before, then ! ∂bj ∂ ∂ X i X j k ∇Xx Y |x= α + b (x)Γij(x) . ∂xi x ∂xj x ∂xk x i k,i Definition. Let γ :(a, b) → M. We say that V is a vector field along γ if it is a function that for every t assigns a tangent vector of the form X ∂ V (t) = vi(t) . ∂xi γ(t) i Sometimes we will write V |γ(t) instead of V (t). In particular, we define the vector field γ˙ along the curve γ as follows d γ˙ (t) := γ∗ . dt t 10 Observe that if ϕ ◦ γ = (x1, . . . , xn) we have that X dxi ∂ γ˙ (t) = (t) . dt ∂xi γ(t) i Remark. Because of the observations made in previous definition, given X ∈ Γ(TM) and a curve γ it has sense to consider a vector field along γ defined as follows ∇γ˙ X (t) = ∇γ˙ (t)X |γ(t) . D Theorem. There exists a unique operator dt on vector fields along a given curve γ so that D D D 1. dt (V + W ) = dt (V ) + dt (W ). D df D ∞ 2. dt (fV ) = dt V + f dt (V ) for f ∈ C (R). D 3. If X ∈ Γ(TM) is such that X |γ(t)= V (t) we have dt (V ) = ∇γ˙ X. P i ∂ Remark. If V (t) = i v (t) ∂xi |γ(t) then k i ! DV X X dv X k dx j ∂ (t) = (t) + Γij(γ(t)) (t)v (t) . dt dt dt ∂xk γ(t) k k i,j Definition. Let M be a differentiable manifold with affine connection ∇. A vector DV field V along a curve γ : I → M is called parallel when dt = 0, for all t ∈ I. Proposition. Let M be a differentiable manifold with an affine connection ∇. Let γ : I → M be a differentiable curve in M and let V0 be a vector tangent to M at γ(t0), t0 ∈ I (i.e. V0 ∈ Tγ(t0)M). Then there exists a unique parallel vector field V along γ, such that V (t0) = V0. Definition. Let M be a differentiable manifold with an affine connection ∇ and a Riemannian metric h, i. A connection is said to be compatible with the metric h, i, when for any smooth curve γ and any pair of parallel vector fields P and P 0 along γ, we have hP,P 0i = constant. Proposition. Let M be a Riemannian manifold. A connection ∇ on M is compati- ble with a metric if and only if for any vector fields V and W along the differentiable curve γ : I → M we have d DV DW hV,W i = h ,W i + hV, i, t ∈ I. dt dt dt 11 Corollary. A connection ∇ on a Riemannian manifold M is compatible with the metric if and only if X(hY,Zi) = h∇X Y,Zi + hY, ∇X Zi,X,Y,Z ∈ Γ(TM). Definition. An affine connection ∇ on a smooth manifold M is said to be symme- tric when ∇X Y − ∇Y X = [X,Y ] for all X,Y ∈ Γ(TM). Theorem, (Levi-Civita). Given a Riemannian manifold M, there exists a unique affine connection ∇ on M satisfying the conditions: 1. ∇ is symmetric. 2. ∇ is compatible with the Riemannian metric. n k ∂ Remark. For the Euclidean space R we have Γij = 0. As a consequence ∇ ∂ ∂x = ∂xi j n Dv P dvk ∂ P ∂ 0. Thus, in , = and ∇X Y = X(yk) . R dt k dt ∂xk k ∂xk Dγ˙ Definition. γ :(a, b) → R is a geodesic if dt = 0 . P dxi ∂ 1 n Observe that since γ˙ (t) = i dt (t) ∂xi |γ(t) (where ϕ ◦ γ = (x , . . . , x )) then 2 k i j ! Dγ˙ X X d x X k dx dx ∂ (t) = (t) + Γij(γ(t)) (t) (t) . dt dt2 dt dt ∂xk γ(t) k k i,j dγ Remark. If γ :(a, b) → R is a geodesic, | dt | is constant. Lemma. there exists a unique vector field G on TM whose trajectories are of the form t → (γ(t), γ0(t)), where γ is a geodesic on M. Definition. The vector field G defined above is called the geodesic field on TM and its flow is called the geodesic flow on TM. Lemma. If the geodesic γ(t, q, v) is defined on the interval (−δ, δ), then the geodesic γ(t, q, av), a ∈ R, a > 0, is defined on the interval (−δ/a, δ/a) and γ(t, q, av) = γ(at, q, v). Proposition. Given p ∈ M, there exist a neighborhood V of p in M, a number ε > 0 and a C∞ mapping γ :(−2, 2) × U → M, U = {(q, w) ∈ TM; q ∈ V, w ∈ TqM, |w| < ε} , 12 such that t → γ(t, q, w), t ∈ (−2, 2), is the unique geodesic of M which, at the instant t = 0, passes through q with velocity w, for every q ∈ V and for every w ∈ TqM, with |w| < ε. Definition. We define the exponential map exp : U → M by v exp(q, v) = γ(1, q, v) = γ |v|, q, , (q, v) ∈ U. |v| In most applications we shall utilize the restriction of exp to an open subset of the tangent space TqM, that is, expq : Bε(0) ⊂ TqM → M, expq(v) = exp(q, v). Proposition. Given q ∈ M, there exists an ε > 0 such that expq : Bε(0) ⊂ TqM → M is a diffeomorphism of Bε(0) onto an open subset of M. Definition. A segment of the geodesic γ :[a, b] → M is called minimizing if `(γ)) ≤ `(c), where c is an arbitrary piecewise differentiable curve joining γ(a) to γ(b). Definition. A parametrized surface in M is a differentiable mapping s : A ⊂ R2 → M.A vector field along s is a mapping which associates to each q ∈ A a vector V (q) ∈ Ts(q)M, and which is differentiable in the following sense: if f is a differentiale function on M, then the mapping q → V (q)f is differentiable. If V is a vector field along s : A → M, let us define the covariant derivatives DV DV DV ∂u and ∂v in the following way. ∂u (u, v0) is the covariant derivative along the curve u → s(u, v0) of the restriction of V to this curve. Lemma(symmetry). If M is a differentiable manifold with a symmetric connec- tion ans s : A → M is a parametrized surface then: D ∂s D ∂s = . ∂v ∂u ∂u ∂v Lemma(Gauss). Let p ∈ M and let v ∈ TpM such that expq v is defined. Let w ∈ TpM ∼ Tv(TpM). Then hdv expp(v), dv expp(w)i = hv, wi. 13 Definition. If expp is a diffeomorphism of a neighborhood V of the origin in TpM, expp V = U is called a normal neighborhood of p. If Bε(0) is such that Bε(0) ⊂ V , we call expp B(0) = Bε(p) the normal ball, or geodesic ball. Proposition. Let p ∈ M, U a normal neighborhood of p, and B ⊂ U a normal ball of center p. Let γ : [0, 1] → B be a geodesic segment with γ(0) = p. If c : [0, 1] → M is any piecewise differentiable curve joining γ(0) to γ(1) then `(γ) ≤ `(c) and if equality holds then γ([0, 1]) = c([0, 1]). Theorem. For any p ∈ M there exist a neighborhood W of p and a number δ > 0, such that, for every q ∈ W , expq is a diffeomorphism on Bδ(0) ⊂ TqM and W ⊂ expq(Bδ(0)), that is, W is a normal neighborhood of each of its points. Definition. The curvature R of a Riemannian manifold M is a correspondence that associates to every pair X,Y ∈ Γ(TM) a mapping R(X,Y ) : Γ(TM) → Γ(TM), R(X,Y )Z = ∇Y ∇X Z − ∇X ∇Y Z + ∇[X,Y ]Z,Z ∈ Γ(TM), where ∇ is the Riemannian connection of M. h i Remark. In particular, since ∂ , ∂ = 0, we obtain ∂xi ∂xj ∂ ∂ ∂ ∂ R , = ∇ ∂ ∇ ∂ − ∇ ∂ ∇ ∂ . ∂xi ∂xj ∂xk ∂xj ∂xi ∂xi ∂xj ∂xk That is, the curvature measures the non-commutativity of the covariant derivative. Proposition. The curvature R of a Riemannian manifold has the following pro- perties: 1. R is bilinear in Γ(TM) × Γ(TM). 2. For any X,Y ∈ Γ(TM), the curvature operator R(X,Y ) : Γ(TM) → Γ(TM) is linear. Proposition (Bianchi Identity). R(X,Y )Z + R(Y,Z)X + R(Z,X)Y = 0. 14 Proposition. Let σ ⊂ TpM be a two-dimensional subspace of the tangent space TpM and let x, y ∈ σ be two linearly independent vectors. Then (x, y, x, y) K(x, y) = |x ∧ y|2 does not depend on the choice of the vectores x, y ∈ σ. Definition. Given a point p ∈ M and a two-dimensional subspace σ ⊂ TpM, the real number K(x, y) = K(σ), where {x, y} is any basis of σ, is called the sectional curvature of σ at p. Definition. Let x = zn be a unit vector in TpM; we take an orthonormal basis {z1, . . . , zn−1} of the hyperplane in TpM orthogonal to x and consider the following averages: 1. Ricci curvature: 1 X Ric (x) = hR(x, z )x, z i, i = 1, . . . , n − 1. p n − 1 i i i 2. Scalar curvature: 1 X 1 X R(p) = Ric (z ) = hR(z , z )z , z i, j = 1, . . . , n − 1. n p j n(n − 1) i j i j j ij Lemma. D D D D ∂f ∂f V − V = R , V. ∂t ∂s ∂s ∂t ∂s ∂t Definition. Let γ : [0, a] → M be a geodesic in M. A vector field J along γ is said to be a Jacobi Field if it satisfies the Jacobi equation D2J (t) + R(γ0(t),J(t))γ0(t) = 0, for allt ∈ [0, a]. dt2 Example. Let M be a Riemannian manifold of constant sectional curvature K, and let γ : [0, `] → M be a normalized geodesic on M. Then the Jacobi equation ca be written as D2J + KJ = 0. dt2 15 Let w(t) be a parallel field along γ with hγ0(t), w(t)i = 0 and |w(t)| = 1. Then √ sin(t K) √ w(t), ifK > 0, K J(t) = tw(t), ifK = 0, √ sinh(t K) √ w(t), ifK < 0, K is a solution of the Jacobi equation with initial conditions J(0) = 0, J 0(0) = w(0). Theorem. Let γ : [0, a] → M be a geodesic. Then a Jacobi field J along γ with J(0) = 0 is given by 0 J(t) = Dtγ0(0) expp(tJ (0)), t ∈ [0, a]. Definition. Let γ : [0, a] → M be a geodesic. the point γ(t0) is said to be con- jugate to γ(0) along γ, if there exists a Jacobi field J along γ, not identically zero, with J(0) = 0 = J(t0). The maximum number of such linearly independent fields is called the multiplicity of the conjugate point γ(t0). The set of (first) conjugate points to the point p ∈ M, for all the geodesics that start at p, is called the conjugate locus of p and is denoted by C(p). Proposition. Let γ : [0, a] → M be a geodesic and put γ(0) = p. The point 0 q = γ(t0), t0 ∈ (0, a], is conjugate to p along γ if and only if v0 = t0γ (0) is a critical point of expp. In addition, the multiplicity of q as a conjugate of p is equal to the dimension of the kernel of the linear map Dv0 expp. Corollary. Suppose that J(0) = 0. Then hJ 0(0), γ0(0)i = 0 if and only if hJ, γ0i(t) = 0; in particular, the space of Jacobi fields J with J(0) = 0 and hJ, γ0i = 0 has dimension equal to n − 1. Definition. A Riemannian manifold is said to be extendible if there exists a riemannian manifold M 0 such that M is isometric to a proper open subset of M 0. Definition. A Riemannian manifold M is geodesically complete if for all p ∈ M, the exponential map expp is defined for all v ∈ TpM, i.e., if any geodesic γ(t) starting from p is defined for all values of t ∈ R. Proposition. If M is geodesically complete then M is non-extendible. Definition. The distance d(p, q) is defined by d(p, q) = infimum of the lengths of all curves fp,q where fp,q is a piecewise differentiable curve joining p to q. Proposition. With the distance d, M is a metric space. 16 Proposition. The topology induced by d on M coincides with the original topology on M. Corollary. If p0 ∈ M, the function f : M → R given by f(p) = d(p, p0) is continuous. Theorem (Hopf-Rinow). Let M be a Riemannian manifold and let p ∈ M. the following assertions are equivalent: 1. expp is defined on all of TpM. 2. The closed and bounded sets of M are compact. 3. M is complete as a metric space. 4. M is geodesically complete. 5. There exists a sequence of compact subsets Kn ⊂ M, Kn ⊂ Kn+1 and ∪nKn = M, such that if qn ∈/ Kn then d(p, qn) → ∞. In addition, any of the statements above implies that for any q ∈ M there exists a geodesic γ joining p to q with `(γ) = d(p, q). Corollary. If M is compact then M is complete. Corollary. A closed submanifold of a complete Riemannian manifold is complete in the induced metric; in particular, the closed submanifolds of Euclidean space are complete. Lemma. Let M be a complete Riemannian manifold with K(p, σ) ≤ 0, for all p ∈ M and for all σ ⊂ TpM. Then for all p ∈ M, the conjugate locus C(p) =; in particular the exponential map expp : TpM → M is a local diffeomorphism. Lemma. Let M be a complete Riemannian manifold and let f : M → N be a local diffeomorphism onto a Riemannian manifold N which has the following property: for all p ∈ M and for all v ∈ TpM, we have |Dpf(v)| ≥ |v|. Then f is a covering map. Theorem (Hadamard). Let M be a complete Riemannian manifold with K(p, σ) ≤ n 0, for all p ∈ M and for all σ ⊂ TpM. Then M is diffeomorphic to R , n = dimM; more precisely expp : TpM → M is a diffeomorphism. 17