Formal definition of the pack [Intermediate]
Bott-Tu Chapter I covers the de Rham complex, Mayer-Vietoris, integration on manifolds, the Poincaré lemma, the MV-induction computation of H ∗ ( S n ) , the Thom isomorphism via differential forms with the global angular form, the non-orientable case, and the sphere-bundle Euler class with the Hopf index theorem. Several of its exercises do not anchor to a single Codex unit — they cross-cut the Mayer-Vietoris machinery, the good-cover induction, the Thom-isomorphism construction, and the Hopf-index calculation simultaneously.
This pack collects fourteen such exercises — four easy, six medium, four hard — each with a hint and a full solution. It is meant to be read alongside its prerequisite units rather than as a standalone development. The exercises are loosely grouped by Bott-Tu section: degree theory and the homotopy operator (easy), MV-induction computations on spheres and punctured spaces (medium), and Hopf-index-flavored sphere-bundle calculations and Stokes-on-boundary applications (hard).
The conventions throughout are Bott-Tu's: the global angular form ψ on a sphere bundle satisfies d ψ = − π ∗ e ( E ) with the originator-text negative sign; the Thom class is normalised so that ∫ fiber Φ = 1 .
Key theorem with full solution [Intermediate]
Before the pack proper, we work one exercise in full as an exemplar of the format. The remaining thirteen follow the same structure (problem, hint, full answer in <details> blocks).
Lead exercise. Compute $H^ _{\mathrm{dR}}(S^n)f or e v er y n\geq 1$ by Mayer-Vietoris induction.*
Solution. Cover S n by two open hemispheres U + = S n ∖ { south pole } and U − = S n ∖ { north pole } . Each hemisphere is diffeomorphic to R n , so contractible; the intersection U + ∩ U − deformation-retracts to the equator S n − 1 .
The Mayer-Vietoris long exact sequence reads
⋯ → H k − 1 ( U + ∩ U − ) δ H k ( S n ) → H k ( U + ) ⊕ H k ( U − ) → H k ( U + ∩ U − ) → ⋯
By contractibility of U ± , H k ( U ± ) = 0 for k ≥ 1 and H 0 ( U ± ) = R . By the inductive hypothesis on S n − 1 , H k ( U + ∩ U − ) = H k ( S n − 1 ) .
For n = 1 : S 0 is two points, so H 0 ( S 0 ) = R 2 and H k = 0 for k ≥ 1 . The MV sequence gives H 0 ( S 1 ) = R and H 1 ( S 1 ) = R (from the cokernel of H 0 ( U + ) ⊕ H 0 ( U − ) → H 0 ( U + ∩ U − ) = R 2 , which is one-dimensional).
For n ≥ 2 : the MV sequence and the inductive hypothesis H ∗ ( S n − 1 ) = ( R , 0 , … , 0 , R ) in degrees zero and n − 1 give
Degree zero: H 0 ( S n ) = R (connected).
Degrees 1 ≤ k ≤ n − 1 : H k ( S n ) = 0 .
Degree n : the connecting map δ : H n − 1 ( S n − 1 ) → H n ( S n ) is an isomorphism, so H n ( S n ) = R .
Final: H dR ∗ ( S n ) = ( R , 0 , … , 0 , R ) in degrees zero and n , zero otherwise. □
This is the canonical computation. The same template — cover by two contractible pieces with intersection the lower sphere — recovers every S n from S 0 . Degree-theory and Hopf-index calculations downstream all rely on the explicit generator of H n ( S n ) produced by this argument.
Exercises [Intermediate]
Exercise 1 (easy, numeric). Cohomology of R n ∖ { 0 } .
Compute H dR ∗ ( R n ∖ { 0 }) for n ≥ 1 .
Hint
R n ∖ { 0 } deformation-retracts to S n − 1 . Cohomology is a homotopy invariant.
Answer
R n ∖ { 0 } deformation-retracts to the unit sphere S n − 1 via x ↦ x /∣ x ∣ . By homotopy invariance,
H dR ∗ ( R n ∖ { 0 }) = H dR ∗ ( S n − 1 ) = ( R , 0 , … , 0 , R )
in degrees zero and n − 1 , zero otherwise.
For n = 1 : R ∖ { 0 } has two components, so H 0 = R 2 . For n = 2 : punctured plane has H 0 = H 1 = R , with H 1 generated by d θ .
Exercise 2 (easy, numeric). Cohomology of R 2 ∖ { n points } .
Compute H dR ∗ ( R 2 ∖ { p 1 , … , p n }) for n distinct points in the plane.
Hint
Mayer-Vietoris on a cover by punctured disks around each point and an open set containing the rest. Or homotopy-equivalence to a wedge of n circles.
Answer
R 2 ∖ { n points } is homotopy-equivalent to a wedge of n circles ⋁ i = 1 n S i 1 . By Mayer-Vietoris on a wedge,
H k ( i ⋁ S i 1 ) = i ⨁ H ~ k ( S i 1 ) ⊕ H k ( pt ) .
So H 0 = R (connected), H 1 = R n (one generator per puncture, the angular form d θ i around p i ), H k = 0 for k ≥ 2 .
The n generators are precisely the angular forms d θ i around each of the n punctures. This is Bott-Tu Worked Example 7 in §5.
Exercise 3 (easy, numeric). Punctured torus.
Compute H dR ∗ ( T 2 ∖ { p }) for the two-torus minus a point.
Hint
The punctured torus deformation-retracts to a wedge of two circles.
Answer
The punctured torus deformation-retracts to its 1-skeleton, which is a wedge of two circles S 1 ∨ S 1 (the two generating loops of the torus, joined at one vertex after removing the 2-cell containing p ).
H 0 = R , H 1 = R 2 , H k = 0 for k ≥ 2 .
The Euler characteristic is 1 − 2 + 0 = − 1 , consistent with χ ( T 2 ) − χ ( pt ) = 0 − 1 = − 1 via additivity of χ .
Exercise 4 (easy, symbolic). Pullback of d θ under the squaring map.
Let f : S 1 → S 1 be the map z ↦ z 2 (in complex coordinates). Compute f ∗ [ d θ ] as a multiple of [ d θ ] in H 1 ( S 1 ) .
Hint
In angular coordinates, f sends θ ↦ 2 θ , so f ∗ ( d θ ) = 2 d θ .
Answer
In angular coordinates θ ∈ [ 0 , 2 π ) on S 1 , the squaring map z ↦ z 2 corresponds to θ ↦ 2 θ . So f ∗ ( d θ ) = d ( 2 θ ) = 2 d θ .
On cohomology, f ∗ : H 1 ( S 1 ) → H 1 ( S 1 ) is multiplication by 2 . The integer 2 is the degree of f , computed cohomologically.
In general, the map z ↦ z n has degree n , and f ∗ [ d θ ] = n [ d θ ] .
Exercise 5 (medium, proof). Degree of the antipodal map.
Show that the antipodal map a : S n → S n , a ( x ) = − x , has degree ( − 1 ) n + 1 .
Hint
a is the composition of n + 1 reflections, each reversing one coordinate. Each reflection has degree − 1 .
Answer
Write a as the composition of n + 1 reflections r i : ( x 1 , … , x n + 1 ) ↦ ( x 1 , … , − x i , … , x n + 1 ) . Each r i is a smooth orientation-reversing diffeomorphism of S n , hence has degree − 1 (since r i ∗ vol S n = − vol S n , integration against the volume form changes sign).
By multiplicativity of degree under composition,
deg ( a ) = i = 1 ∏ n + 1 deg ( r i ) = ( − 1 ) n + 1 .
Consequence: for n even, a has degree − 1 (orientation-reversing); for n odd, a has degree + 1 (orientation-preserving). This explains why R P n is orientable iff n is odd. Bott-Tu §4 Worked Example 6.
Exercise 6 (medium, proof). Polynomial map C → C has degree n .
Let p : C → C be a degree-n complex polynomial. Show that the induced map p : S 2 → S 2 on the one-point compactification has topological degree n .
Hint
The map p is properly homotopic to z n . Use that homotopic maps have the same degree.
Answer
Write p ( z ) = a n z n + a n − 1 z n − 1 + ⋯ + a 0 with a n = 0 . The straight-line homotopy
p t ( z ) = ( 1 − t ) p ( z ) + t ⋅ a n z n = a n z n + ( 1 − t ) ( a n − 1 z n − 1 + ⋯ + a 0 )
is proper as a map C × [ 0 , 1 ] → C (it tends to ∞ uniformly on ∣ z ∣ → ∞ ), so it extends to a homotopy p t : S 2 → S 2 between p and a n z n .
The map a n z n has degree n : away from the origin and infinity, a n z n wraps each circle ∣ z ∣ = r around a n ⋅ r n ⋅ S 1 exactly n times. By homotopy invariance of degree, deg ( p ) = n .
Consequence: every degree-n polynomial has n roots counted with multiplicity (the fundamental theorem of algebra), since the equation p ( z ) = w 0 has n preimages for generic w 0 — exactly the degree.
Exercise 7 (medium, proof). Möbius bundle is nontrivial as a vector bundle.
The Möbius bundle M → S 1 is the rank-one real vector bundle whose total space is the Möbius strip. Show it is not isomorphic to the constant bundle S 1 × R → S 1 .
Hint
The constant bundle has a nowhere-zero section (the constant 1). The Möbius bundle does not.
Answer
The constant rank-one bundle S 1 × R → S 1 has the global section s ( x ) = ( x , 1 ) , which is nowhere zero.
The Möbius bundle has no nowhere-zero global section. Suppose σ : S 1 → M were a nowhere-zero section. Choosing a continuous trivialisation along an open arc U ⊂ S 1 , σ becomes a function U → R ∖ { 0 } . Continuing around the loop, the trivialisation flips sign (the Möbius monodromy), so σ as a function returns to its negative. By the intermediate value theorem applied to the function as it crosses zero, σ must vanish somewhere — contradiction.
Therefore the Möbius bundle is not isomorphic to the constant bundle. Its first Stiefel-Whitney class w 1 ∈ H 1 ( S 1 ; Z /2 ) = Z /2 is the nonzero element. Bott-Tu §6 Exercise 6.10.
Exercise 8 (medium, symbolic). Euler class of T S 2 .
Compute the Euler class e ( T S 2 ) ∈ H 2 ( S 2 ; R ) by integrating a representative form (e.g., 4 π 1 sin ϕ d ϕ ∧ d θ in spherical coordinates).
Hint
The Euler class evaluated on the fundamental class gives the Euler characteristic. For S 2 , χ = 2 .
Answer
A representative of e ( T S 2 ) is the curvature of any connection on T S 2 . The standard round-metric Levi-Civita connection has curvature 4 π 1 sin ϕ d ϕ ∧ d θ (the Gauss-Bonnet integrand normalised). Integrating:
∫ S 2 4 π 1 sin ϕ d ϕ ∧ d θ = 4 π 1 ∫ 0 π sin ϕ d ϕ ∫ 0 2 π d θ = 4 π 1 ⋅ 2 ⋅ 2 π = 1.
Wait — this integrates to 1 , not 2 . The correct normalisation places the Euler class as 2 π 1 K vol (Gauss-Bonnet), with K = 1 on the unit sphere, giving
∫ S 2 2 π 1 sin ϕ d ϕ ∧ d θ = 2 π 1 ⋅ 2 ⋅ 2 π = 2 = χ ( S 2 ) .
So e ( T S 2 ) = 2 [ pt ] ∈ H 2 ( S 2 ; R ) ≅ R , where [ pt ] is the generator dual to the fundamental class. Equivalently, ∫ S 2 e ( T S 2 ) = χ ( S 2 ) = 2 . This is Bott-Tu §6 Exercise 6.14, related to the Hopf index theorem.
Exercise 9 (medium, proof). Projection formula.
Let π : E → M be a rank-r oriented vector bundle and Φ ∈ Ω c v r ( E ) a representative of the Thom class. Show the projection formula : for any ω ∈ Ω ∗ ( M ) ,
π ∗ ( Φ ∧ π ∗ ω ) = ω
where π ∗ : Ω c v ∗ ( E ) → Ω ∗− r ( M ) is integration along the fiber.
Hint
By the Fubini-style splitting on each fiber, π ∗ ( Φ ∧ π ∗ ω ) = ω ⋅ π ∗ ( Φ ) . Use π ∗ ( Φ ) = 1 by the Thom-class normalisation.
Answer
Write a local product chart E ∣ U ≅ U × R r with coordinates ( x , y ) , x ∈ U and y ∈ R r . Express Φ = ϕ ( x , y ) d y 1 ∧ ⋯ ∧ d y r + ( lower-vertical-degree terms ) and π ∗ ω a pure base form (no d y factors). Their wedge is ϕ ( x , y ) ω ( x ) ∧ d y 1 ∧ ⋯ ∧ d y r in top vertical degree.
Integration along the fiber 03.04.09 integrates the d y -coefficient over R r for fixed x :
π ∗ ( Φ ∧ π ∗ ω ) = ω ( x ) ⋅ ∫ R r ϕ ( x , y ) d y 1 ∧ ⋯ ∧ d y r = ω ( x ) ⋅ π ∗ ( Φ ) ( x ) .
By the Thom-class normalisation ∫ fiber Φ = 1 , π ∗ ( Φ ) = 1 on M . Therefore π ∗ ( Φ ∧ π ∗ ω ) = ω .
The projection formula is the operational identity of the Thom isomorphism: it says integration-along-the-fiber composed with multiplication-by-the-Thom-class is the identity on Ω ∗ ( M ) . Bott-Tu §6 Exercise 6.2.
Exercise 10 (medium, proof). Twisted Poincaré on R P 2 .
Use the orientation local system or M on R P 2 to compute H k ( R P 2 ; or M ) for each k .
Hint
Twisted Poincaré duality on a closed non-orientable manifold says H k ( M ; R ) ≅ H n − k ( M ; or M ) . Compute the LHS first.
Answer
The cohomology of R P 2 with constant real coefficients is H 0 = R , H 1 = 0 , H 2 = 0 (the integer cohomology has Z /2 -torsion in degrees one and two, which vanishes when tensored with R ).
Twisted Poincaré duality on the closed non-orientable surface R P 2 :
H k ( R P 2 ; R ) ≅ H 2 − k ( R P 2 ; or M ) .
So H 0 ( or M ) = H 2 ( R ) = 0 , H 1 ( or M ) = H 1 ( R ) = 0 , H 2 ( or M ) = H 0 ( R ) = R .
The orientation-twisted top cohomology H 2 ( R P 2 ; or M ) = R supplies the orientation-twisted fundamental class — the class that integrates orientation-twisted top forms to ± 1 . This is the Bott-Tu §7 reformulation of orientation as a local system. The non-twisted top cohomology vanishes because R P 2 is non-orientable.
Exercise 11 (hard, proof). Hopf invariant of the Hopf map.
The Hopf map h : S 3 → S 2 is the map ( z 1 , z 2 ) ↦ [ z 1 : z 2 ] ∈ C P 1 ≅ S 2 . Compute its Hopf invariant H ( h ) ∈ Z .
Hint
Choose a generator ω ∈ H 2 ( S 2 ) with ∫ S 2 ω = 1 . Pull back by h to get h ∗ ω ∈ H 2 ( S 3 ) = 0 , so h ∗ ω = d η for some η ∈ Ω 1 ( S 3 ) . The Hopf invariant is ∫ S 3 η ∧ h ∗ ω .
Answer
Take ω the area form on S 2 normalised to ∫ S 2 ω = 1 . Pull back by h : h ∗ ω ∈ Ω 2 ( S 3 ) is closed and represents the class h ∗ [ ω ] ∈ H 2 ( S 3 ) = 0 , so h ∗ ω = d η for some η ∈ Ω 1 ( S 3 ) .
A natural choice of η is the contact form on S 3 : η = 2 1 ( x 1 d y 1 − y 1 d x 1 + x 2 d y 2 − y 2 d x 2 ) , where ( x i , y i ) are the real coordinates on S 3 = {( z 1 , z 2 ) : ∣ z 1 ∣ 2 + ∣ z 2 ∣ 2 = 1 } . This is the Boothby-Wang connection form on the Hopf bundle.
The Hopf invariant is
H ( h ) := ∫ S 3 η ∧ h ∗ ω = ∫ S 3 η ∧ d η .
Computing in coordinates: η ∧ d η = 2 1 d vol S 3 when ω is normalised so ∫ S 2 ω = 1 . The integral ∫ S 3 d vol S 3 = vol ( S 3 ) = 2 π 2 for the unit S 3 , but the normalisation of ω and η matches up so that H ( h ) = 1 .
The Hopf map has Hopf invariant 1 — it is the generator of π 3 ( S 2 ) = Z , with the Hopf invariant providing the isomorphism. Bott-Tu §11 Exercise 11.19 (specialised to h ).
Exercise 12 (hard, proof). Hopf index theorem for T 2 → S 2 .
A vector field on T 2 has total index (sum of local indices at zeros) equal to χ ( T 2 ) = 0 . Construct an explicit vector field on T 2 exhibiting this — say, with two simple zeros of opposite index.
Hint
A vector field on T 2 is a section of T ( T 2 ) , which is a trivialisable bundle. The trivialisation gives a constant non-vanishing vector field; perturb to introduce isolated zeros.
Answer
The tangent bundle T ( T 2 ) is trivialisable (the torus is parallelisable: T ( T 2 ) ≅ T 2 × R 2 via the angular coordinates). The constant vector field ∂ θ 1 has no zeros — total index zero, consistent with χ ( T 2 ) = 0 .
For an explicit field with two zeros of opposite index, take the vector field V = sin θ 1 ∂ θ 1 + cos θ 2 ∂ θ 2 . Zeros: sin θ 1 = 0 , cos θ 2 = 0 , i.e., θ 1 ∈ { 0 , π } and θ 2 ∈ { π /2 , 3 π /2 } , giving four zeros. The local index at each is ± 1 depending on the sign of the Jacobian of V at that point: at ( 0 , π /2 ) and ( π , 3 π /2 ) the Jacobian is positive (index + 1 ); at ( 0 , 3 π /2 ) and ( π , π /2 ) the Jacobian is negative (index − 1 ). Total index: ( + 1 ) + ( + 1 ) + ( − 1 ) + ( − 1 ) = 0 , consistent with χ ( T 2 ) = 0 .
A simpler example with two zeros: V = sin θ 1 ∂ θ 1 + sin θ 2 ∂ θ 2 has zeros at θ 1 , θ 2 ∈ { 0 , π } , giving four points; the local index alternates as before.
The key principle (Hopf index theorem): the sum of local indices of any vector field with isolated zeros on a compact oriented manifold equals the Euler characteristic. Bott-Tu §11 Exercise 11.26.
Exercise 13 (hard, proof). Stokes' theorem on a manifold with boundary — applied.
Let M be a compact oriented n -manifold with boundary ∂ M . Show that for every closed ( n − 1 ) -form ω on M , ∫ ∂ M ω = 0 .
Hint
Stokes' theorem: ∫ ∂ M ω = ∫ M d ω . Since ω is closed, d ω = 0 .
Answer
Stokes' theorem on M with boundary ∂ M states ∫ M d ω = ∫ ∂ M ω for any compactly-supported ( n − 1 ) -form ω . If ω is closed, then d ω = 0 , so ∫ M d ω = 0 , and therefore ∫ ∂ M ω = 0 .
Consequence (no-go for cobordism by exact form). If ω = d η is exact, both sides are zero — boundary integration only sees the de Rham cohomology class [ ω ] ∈ H dR n − 1 ( M ) , not the form itself. This is the start of cobordism theory: the integral of a closed form over a closed cobordant pair ( N 1 , N 2 = ∂ M ) is invariant under the cobordism, since ∫ N 1 ω − ∫ N 2 ω = ∫ ∂ M ω = 0 .
Consequence (Pontryagin numbers cobordism-invariant). For closed manifolds, Pontryagin and Chern numbers are integrals of closed forms (representatives of characteristic classes). By Stokes-on-boundary, two cobordant manifolds have the same Pontryagin numbers — the foundational fact of cobordism theory.
Exercise 14 (hard, proof). Mayer-Vietoris induction with general good cover.
Show that on a compact manifold M with finite good cover { U 1 , … , U n } , the de Rham cohomology H dR ∗ ( M ) is finite-dimensional.
Hint
Induct on the number of opens. The base case is M contractible. The inductive step uses the Mayer-Vietoris long exact sequence and the fact that direct sums and quotients of finite-dimensional spaces are finite-dimensional.
Answer
Induct on k = number of opens needed to cover M .
Base (k = 1 ). M = U 1 is contractible, so H dR ∗ ( M ) = R in degree zero, zero in higher degrees. Finite-dimensional.
Inductive step. Suppose M is covered by k ≥ 2 opens. Set V 1 = U 1 and V 2 = U 2 ∪ ⋯ ∪ U k (the union of the remaining k − 1 ). By the inductive hypothesis, H ∗ ( V 1 ) , H ∗ ( V 2 ) , and H ∗ ( V 1 ∩ V 2 ) are all finite-dimensional — the first because V 1 is contractible, the second by inductive hypothesis on k − 1 opens, the third because V 1 ∩ V 2 = ( U 1 ∩ U 2 ) ∪ ⋯ ∪ ( U 1 ∩ U k ) is itself covered by k − 1 opens (each contractible by the good-cover hypothesis on intersections).
The Mayer-Vietoris long exact sequence
⋯ → H k ( M ) → H k ( V 1 ) ⊕ H k ( V 2 ) → H k ( V 1 ∩ V 2 ) → H k + 1 ( M ) → ⋯
squeezes H k ( M ) between two finite-dimensional spaces. Specifically, H k ( M ) is an extension of a subspace of H k ( V 1 ) ⊕ H k ( V 2 ) by a quotient of H k − 1 ( V 1 ∩ V 2 ) . Both are finite-dimensional, so H k ( M ) is finite-dimensional.
This completes the induction. Corollary: every compact smooth manifold has finite-dimensional de Rham cohomology — the finite-type theorem , Bott-Tu §5. The proof here is the canonical MV-induction argument.
Pass 4 Agent B exercise pack EP1. Bott-Tu Chapter I supplement: Mayer-Vietoris and degree theory across §1.4, §3, §4, §5, §6, §7, §11.