Milnor — *Morse Theory* (Fast Track 3.03) — Audit + Gap Plan

Book: John W. Milnor, Morse Theory, based on lecture notes by M. Spivak and R. Wells. Annals of Mathematics Studies 51, Princeton University Press, 1963. vi + 153 pp. Free PDF hosted on the Ranicki archive at https://webhomes.maths.ed.ac.uk/~v1ranick/papers/milnmors.pdf.

Fast Track entry: 3.03 (the canonical small monograph on critical-point theory for smooth manifolds, sitting between general differential topology and the Floer / symplectic chapter).

Purpose of this plan: lightweight audit-and-gap pass (P1-lite + P2 + P3-lite of the orchestration protocol). Output is a concrete punch-list of new units so that Morse Theory (MMT hereafter) is covered to the equivalence threshold (≥95% effective coverage of theorems, examples, exercises, notation, sequencing, intuition, applications — see docs/plans/FASTTRACK_EQUIVALENCE_PLAN.md §3.4).

This pass is intentionally not a full P1 audit. MMT is short (~150 pp.) and extraordinarily distilled — the editorial style is to deliver each theorem with minimal preamble — so the TOC alone already determines most of the audit content, and a line-number-level inventory would be redundant with the punch-list. The pass works from MMT's TOC (Parts I–IV + Appendix), the existing Codex symplectic / spin / K-theory units that already cite MMT, and the gap is the punch-list.

---

§1 What MMT is for

MMT is the canonical small monograph on Morse theory — the study of how the topology of a smooth manifold is determined by the critical-point data of a smooth function on it. Despite being ~150 pages, MMT delivers the entire classical theory and culminates in a complete proof of the Bott periodicity theorem for the orthogonal and unitary groups, obtained as an application of Morse theory to the path space of a symmetric space. The editorial style is severe: Milnor states the minimum necessary lemmas, gives short and exact proofs, and arrives at deep theorems with a velocity that no later textbook has matched.

Distinctive content, organised by the four parts of the book:

1. Part I — Non-degenerate smooth functions on a manifold (§§1–7). The foundations. Critical point of a smooth $f:M\to \mathbb{R}$, non-degenerate critical point, Hessian and Morse index, the Morse lemma giving local normal form $f = -x_1^2 - \cdots - x_\lambda^2 + x_{\lambda+1}^2 + \cdots + x_n^2$. Reeb's theorem: a compact manifold with exactly two non-degenerate critical points is homeomorphic to a sphere. The two foundational theorems linking critical data to topology: Theorem 3.1 (passing a non-critical level changes nothing up to diffeomorphism) and Theorem 3.2 (passing a level containing one non-degenerate critical point of index $\lambda$ attaches a $\lambda$-handle / $\lambda$-cell). The Morse inequalities for Betti numbers in terms of critical-point counts. Existence of non-degenerate functions via embedding in Euclidean space. The Lefschetz hyperplane theorem as a Morse-theoretic application to complex projective manifolds. Compare Audin-Damian Part I and Banyaga-Hurtubise Chapters 3–6 [ref: Audin-Damian Morse Theory and Floer Homology Part I; Banyaga-Hurtubise Lectures on Morse Homology Chs. 3–6].
2. Part II — A rapid course in Riemannian geometry (§§8–10). Just enough Riemannian geometry to do calculus of variations on geodesics: covariant differentiation $\nabla$, the curvature tensor $R$, geodesics, the exponential map, and completeness (Hopf-Rinow). Genuinely "rapid" — Milnor compresses a semester of Riemannian geometry into roughly 25 pages, keeping only what is load-bearing for Part III.
3. Part III — Calculus of variations applied to geodesics (§§11–19). The path space $\Omega (M;p,q)$, the energy functional $E$, the first variation, the Hessian $E_{\ast \ast }$ at a critical path (geodesic), Jacobi fields as the null-space of $E_{\ast \ast }$, conjugate points. The Morse Index Theorem: the index of $E_{\ast \ast }$ at a geodesic equals the number of conjugate points (with multiplicity) interior to the geodesic. Finite-dimensional approximation of $\Omega$ by broken geodesics — the technique that makes infinite-dimensional Morse theory rigorous in this book. Topology of the full path space (it has the homotopy type of a CW complex with one $\lambda$-cell per geodesic of index $\lambda$). Existence of non-conjugate points. Topology vs. curvature: Theorem 19.4 (Bonnet-Myers via Morse) and the Cartan-Hadamard theorem.
4. Part IV — Applications to Lie groups and symmetric spaces (§§20–24). Symmetric spaces, Lie groups as symmetric spaces. Whole manifolds of minimal geodesics: the key technical device for the periodicity proof — when the set of minimal geodesics from $p$ to $q$ is itself a manifold, one can do Morse theory by induction on dimension. The Bott periodicity theorem for $\mathrm{U}(n)$ (§23): $\Omega \mathrm{U}\simeq \mathbb{Z}\times \mathrm{BU}$, proved via Morse theory on the path space of $\mathrm{U}(2n)$ joining $I$ to $-I$. The periodicity theorem for $\mathrm{O}(n)$ (§24): the eight-fold periodicity, proved by iterating the minimal-geodesics construction through a chain $\mathrm{O}(16n) \supset \mathrm{O}(16n)/\Omega_1 \supset \cdots$ down to a one-point space. This is the original Morse-theoretic proof; the K-theoretic / Clifford-module reformulation (Atiyah-Bott-Shapiro
   1964. is logically independent but historically posterior.
5. Appendix — The homotopy type of a monotone union. A technical CW approximation lemma used in Part III to handle ascending unions of approximations.
6. Editorial signature. MMT is the gold-standard example of Milnor's "deliver the heart of the field in a small book" style. Compare Topology from the Differentiable Viewpoint (76 pp.) and Characteristic Classes (with Stasheff, 230 pp.). Subsequent textbooks on Morse theory (Schwarz 1993, Banyaga-Hurtubise 2004, Audin-Damian 2014, Nicolaescu 2007/2011) all explicitly cite MMT as the canonical classical reference and structure their classical-theory chapters around its Part I [ref: Schwarz Morse Homology Birkhäuser 1993 Introduction; Nicolaescu An Invitation to Morse Theory 2nd ed. 2011 Part I].
7. Relation to later programmes. The infinite-dimensional Morse theory of Floer (Hamiltonian / Lagrangian / instanton Floer homology) replaces Milnor's finite-dim approximation by a gradient flow in a Banach / Hilbert manifold setting, recovers Milnor-style index information from the trajectory moduli spaces, and packages the output as a Morse complex (Schwarz 1993). MMT is the conceptual prerequisite for that programme. Audin-Damian Part I is essentially "MMT Part I rewritten in the Morse-complex framework that will be needed for Floer in Part II" [ref: Audin-Damian Part I; Schwarz §§1–3].

MMT is not a first introduction to smooth manifolds. It assumes familiarity with manifolds, tangent bundles, vector fields, the implicit function theorem, and CW complexes (the latter only at the level of cell attachment). It is the canonical entry point to Morse theory if one wants the classical finite-dim treatment with the Bott-periodicity payoff. The modern Morse-complex / Morse-homology programme is a later development; MMT itself does not define Morse homology as a chain complex (that is Thom 1949 / Smale 1960s / Witten 1982 / Floer 1988 / Schwarz 1993).

---

§2 Coverage table (Codex vs MMT)

Cross-referenced against the current Codex corpus. ✓ = covered, △ = partial / different framing, ✗ = not covered.

MMT topic	Codex unit(s)	Status	Note
Critical point, Hessian, non-degeneracy (Morse condition)	02.05.05-multivariable-taylor-extrema.md	△	Classical multivariable extrema only; no manifold Hessian, no Morse-index definition. Touches the Euclidean shadow but not the geometry.
Morse lemma (local normal form)	—	✗	Gap. Foundational.
Morse index $\lambda (p)$ of a non-degenerate critical point	—	✗	Gap. Cited by name in Floer units (05.08.02, 05.08.04) without an anchor.
Reeb's theorem (sphere recognition by two critical points)	—	✗	Gap. Charming and short — natural Beginner-tier hook.
Theorem 3.1 (sublevel sets across non-critical levels are diffeomorphic)	—	✗	Gap. Load-bearing for handle attachment.
Theorem 3.2 / handle attachment ($\lambda$-cell attachment at index-$\lambda$ critical)	—	✗	Gap. This is the theorem linking analysis to topology in Morse theory.
Morse inequalities (weak + strong) for Betti numbers	—	✗	Gap.
Existence of Morse functions (generic via Whitney embedding)	—	✗	Gap.
Lefschetz hyperplane theorem (Morse-theoretic proof)	—	✗	Gap. Connects to Hodge theory / Hartshorne (04-algebraic-geometry/); currently unanchored in either chapter.
Covariant derivative $\nabla$	△	△	Covered functionally in 03.05.09-curvature.md and 03.05-bundles/ connection units, but as a connection on a vector bundle rather than the Levi-Civita-on-a-Riemannian-manifold framing MMT uses. Adequate for FT-equivalence; no new unit needed.
Curvature tensor $R$	△	△	Covered in 03.05.09-curvature.md at the bundle-curvature level. The Riemannian-curvature-tensor specialisation (sectional curvature, $R(X,Y)Z$ formula) is partial. Marginal gap.
Geodesic, exponential map, Hopf-Rinow completeness	△	△	Geodesic flow appears in 05.02.06-geodesic-flow-hamiltonian.md from the Hamiltonian-symplectic side; the Riemannian-geometry framing (Levi-Civita geodesic, ${\exp }_p$, Hopf-Rinow) has no dedicated unit. Marginal gap — one short unit closes it.
Path space $\Omega (M;p,q)$	—	✗	Gap. Foundational for the variational half of MMT and for loop-space arguments throughout topology.
Energy functional $E$, first variation formula	—	✗	Gap.
Hessian $E_{\ast \ast }$ at a critical geodesic	—	✗	Gap.
Jacobi field, Jacobi equation	—	✗	Gap. Cited in 03.09.18-berger-holonomy.md informally.
Conjugate point, conjugate-point multiplicity	—	✗	Gap.
Morse Index Theorem ($\mathrm{index}(E_{\ast \ast })=$ count of interior conjugate points)	—	✗	Gap (high priority — the central theorem of Part III).
Finite-dimensional approximation of $\Omega$ by broken geodesics	—	✗	Gap. Master-tier only — technique unit.
Path space has CW type with one $\lambda$-cell per geodesic of index $\lambda$	—	✗	Gap. The bridge from variational Morse theory to algebraic topology of loop spaces.
Bonnet-Myers theorem via Morse	—	✗	Gap. Compact-with-positive-curvature ⇒ finite fundamental group.
Cartan-Hadamard theorem (non-positive curvature ⇒ ${\exp }_p$ covering map)	—	✗	Gap.
Symmetric space; Lie group as symmetric space	△	△	03.03-lie/ Lie group units are present (Lie group, Lie algebra, Maurer-Cartan, etc.); a unit explicitly identifying $G$ with $(G\times G)/\Delta G$ as a symmetric space is absent. Marginal gap.
Whole manifolds of minimal geodesics (Bott's induction device)	—	✗	Gap. Master-tier; the technical heart of the periodicity proof.
Bott periodicity for $\mathrm{U}$ via Morse on path space of $\mathrm{U}(2n)$	03.08.07-bott-periodicity.md	△	The Bott-periodicity unit already cites MMT §23 by name as its Master anchor, but presents the statement via clutching / classifying-space arguments and only gestures at the Morse-theoretic proof in a single line (see 03.08.07 line 490, line 508, line 510). The Morse-theoretic proof is the anchor citation but has no anchor unit.
Bott periodicity for $\mathrm{O}$ via iterated minimal-geodesics chain	03.08.07-bott-periodicity.md (real case)	△	Same as above. The real eight-fold periodicity is stated and proved via Clifford modules (Lawson-Michelsohn route) but the Morse proof is referenced only obliquely.
Appendix: homotopy type of a monotone union (CW approximation)	—	✗	Gap (low priority — technical lemma).

Aggregate coverage estimate: ~5–10% of MMT has corresponding Codex units. Part I (the foundations) is essentially uncovered. Part II is mostly covered (the Codex has Riemannian curvature and connections, but not Levi-Civita-and-Hopf-Rinow framed). Part III is essentially uncovered. Part IV — the Bott periodicity culmination — has a downstream unit (03.08.07) but its Morse-theoretic proof is unanchored: the unit cites MMT §23 as its Master tier anchor while having no prerequisite chain that supplies the Morse-theoretic background.

Silent Morse dependencies in the symplectic chapter. The Floer sub-chapter (content/05-symplectic/floer/) consists of four shipped units (05.08.01 Arnold conjecture, 05.08.02 Floer homology, 05.08.03 Maslov index, 05.08.04 Conley-Zehnder index). All four cite Audin-Damian Morse Theory and Floer Homology as their primary anchor; Audin-Damian Part I is MMT Part I in modern notation, and all four units invoke Morse-theoretic concepts (index, gradient flow, critical-point counting, Morse complex) without any prerequisite unit that defines them. Four Floer units silently depend on Morse-theoretic content with no anchor.

---

§3 Gap punch-list (P3-lite — units to write, priority-ordered)

Priority 1 — load-bearing classical Morse theory (Part I + Part III core):

1. 03.02.02 Critical point, Hessian, Morse index, Morse lemma. Smooth $f:M\to \mathbb{R}$, the intrinsic Hessian at a critical point, the non-degeneracy condition, and the Morse-lemma normal form. MMT §2 anchor. Three-tier, ~1500 words. Foundational; unblocks every other Morse unit.
2. 03.02.03 Morse functions and handle attachment (Theorems 3.1, 3.2). Sublevel-set diffeomorphism across non-critical levels; cell attachment at non-degenerate critical points of index $\lambda$. MMT §3 anchor. Three-tier; Beginner gives the height-function-on-torus picture (MMT §1), Intermediate gives Theorems 3.1 and 3.2 with the gradient-flow proof sketch, Master gives the full proofs. ~2000 words.
3. 03.02.04 Morse inequalities and Reeb's theorem. Weak and strong Morse inequalities relating critical-point counts to Betti numbers; Reeb sphere-recognition theorem as a corollary. MMT §§4–5 anchor. ~1500 words. Three-tier.
4. 03.12.X1 Path space $\Omega (M;p,q)$, energy functional, first variation. Definition of the path space, energy vs length, first variation of $E$, critical paths = geodesics. MMT §§11–12 anchor; also covered in Bott-Tu §17 indirectly via loop spaces. ~1500 words. Slots into 03.12-homotopy/ next to existing loop-space-adjacent units.
5. 03.02.05 Jacobi fields, conjugate points, Morse Index Theorem. Jacobi equation as the second-variation EL equation, conjugate points as zeros of Jacobi fields, the Index Theorem ($\mathrm{index}(E_{**}) = \sum$ interior conjugate-point multiplicities). MMT §§13–15 anchor. Three-tier; ~2500 words; the longest unit in the punch-list because the Index Theorem proof is genuinely technical. Master tier includes the finite-dim-approximation argument from MMT §16.

Priority 2 — Part II patch + Part III topology + Bott connection:

6. 03.03.X1 Levi-Civita connection, exponential map, Hopf-Rinow. Patches the Part II gap: the existing 03.05.09-curvature.md covers bundle curvature but not Levi-Civita; Hopf-Rinow is absent. MMT §10 anchor; do Cross with do Carmo Riemannian Geometry as a secondary anchor. ~1500 words.
7. 03.12.X2 Path-space CW structure: cells from geodesics of index $\lambda$. The theorem that $\Omega (M;p,q)$ has the homotopy type of a CW complex with one $\lambda$-cell per geodesic of index $\lambda$. MMT §17 anchor. Two-tier (Intermediate + Master); ~1500 words. The bridge from Part III variational Morse to the loop-space algebraic topology that feeds Bott.
8. 03.02.06 Topology and curvature: Bonnet-Myers and Cartan-Hadamard. Both as Morse-theoretic applications of the Index Theorem. MMT §§19 + 22 anchor. ~1500 words. Three-tier.
9. 03.08.X1 Bott periodicity for $\mathrm{U}$: the Morse-theoretic proof. Currently 03.08.07-bott-periodicity.md cites MMT §23 as its Master anchor while only stating the result. Add a sibling unit (or expand the existing Master tier into a separate unit) giving the Morse-theoretic proof: path space of $\mathrm{U}(2n)$ from $I$ to $-I$, the set of minimal geodesics is ${\mathrm{Gr}}_n({\mathbb{C}}^{2n})$, induct on $n$. MMT §23 anchor; Milnor-Stasheff §24 secondary. ~2000 words; Master tier only. This is the unit that completes the citation chain in the K-theory chapter.

Priority 3 — Lefschetz hyperplane and the real-case Bott proof:

10. 03.02.07 Lefschetz hyperplane theorem via Morse. Andreotti-Frankel / Milnor Morse-theoretic proof on the affine complement. MMT §7 anchor; Andreotti-Frankel 1959 originator citation. ~1500 words. Connects the Morse units to 04-algebraic-geometry/ (where the Lefschetz hyperplane theorem is also a gap on the Hartshorne audit).
11. 03.08.X2 Bott periodicity for $\mathrm{O}$: iterated minimal geodesics. The eight-fold periodicity proved by the chain of symmetric spaces $\mathrm{O}(16n)\supset \mathrm{O}(16n)/\mathrm{U}(8n)\supset \cdots$. MMT §24 anchor. Master tier only; ~2000 words. Optional — the Clifford-module proof (Lawson-Michelsohn) is already covered in 03.09-spin-geometry/, so this unit is the historical Morse-theoretic proof, not the only proof.

Priority 4 — technical and survey units (optional, Master-only):

12. 03.02.X1 Finite-dimensional approximation of $\Omega$ by broken geodesics. MMT §16 technique unit. Master-only, ~1200 words. Useful as the conceptual bridge to Floer theory (where the analogous device is the analytic compactification of trajectory moduli spaces).
13. 03.02.X2 Symmetric space; Lie group as symmetric space. Closes the Part IV §§20–22 gap. Sits between 03.03-lie/ and the new Bott units. ~1500 words. Three-tier.
14. Pointer in 05.08.02 Floer homology to the new 03.02.0X Morse units. Single-paragraph weaving edit, not a new unit; recorded here so it is not forgotten in Pass-W.

---

§4 Implementation sketch (P3 → P4)

For a full MMT coverage pass, items 1–5 are the priority-1 minimum and also unblock the Floer units' silent Morse dependency. Realistic production estimate (mirroring earlier batches):

• ~3 hours per Morse unit. The corpus average is 2.5–3 hours; MMT units trend slightly above average because the Index Theorem proof and the Bott periodicity proof both have nontrivial Master tiers.
• Priority 1: 5 units × ~3 hours = ~15 hours.
• Priority 2: 4 units × ~3 hours = ~12 hours.
• Priority 3: 2 units × ~3.5 hours = ~7 hours.
• Priority 4 (optional): 2 units + 1 weaving edit ≈ ~6 hours.
• Total for full coverage: ~40 hours. Priority 1 alone closes the Floer-dependency gap and is the minimal useful chunk: ~15 hours, two to three focused days.

Originator-prose targets. Per docs/plans/FASTTRACK_EQUIVALENCE_PLAN.md §10, the following units should carry originator-prose citations:

• Marston Morse, "Relations between the critical points of a real function of $n$ independent variables," Trans. AMS 27 (1925) 345–396 — the foundational paper introducing the Morse inequalities. Cite in unit 03.02.04.
• Marston Morse, The Calculus of Variations in the Large (AMS Colloquium Publications XVIII, 1934) — book-length consolidation; cite in unit 03.02.05.
• Raoul Bott, "The stable homotopy of the classical groups," Proc. Nat. Acad. Sci. USA 43 (1957) 933–935; the long-form paper Annals 70 (1959) 313–337. Cite in 03.08.X1 and 03.08.X2.
• Stephen Smale, "On the structure of manifolds," Amer. J. Math. 84 (1962) 387–399, and the $h$-cobordism theorem (1962) — Morse-theoretic application. Cite as a "downstream" pointer in 03.02.03.
• John Milnor, Morse Theory (1963) — the canonical consolidation. Cite throughout.
• Andreotti, Frankel, "The Lefschetz theorem on hyperplane sections," Annals of Math. 69 (1959) 713–717 — cite in 03.02.07.

Notation crosswalk. MMT writes:

• $f^a=\{x\in M:f(x)\le a\}$ for sublevel sets.
• $\lambda$ for the Morse index.
• $\Omega (M;p,q)$ for the path space from $p$ to $q$, and $\Omega(M) = \Omega(M; p, p)$ for the based loop space when convenient.
• $E$ for the energy functional, $L$ for length.
• $E_{\ast \ast }$ for the Hessian of $E$ at a critical path.

Audin-Damian and Schwarz use slightly different notation: $\mathrm{Crit}(f)$, $\mathrm{ind}(p)$ for index, $W^s(p)$ / $W^u(p)$ for stable/unstable manifolds (these are absent from MMT, which does not introduce gradient-flow stable manifolds explicitly — that is a Smale 1960s reformulation). The Codex notation decision (per docs/specs/UNIT_SPEC.md §11) should: adopt MMT's $f^a$, $\lambda$, $\Omega (M;p,q)$, $E$, $L$ verbatim, and add the Smale-era $W^s$/$W^u$ in unit 03.02.03 as the modern reformulation, with a notation paragraph cross-referencing both. Record in §Notation of units 03.02.02 and 03.02.03.

---

§5 What this plan does NOT cover

• A line-number-level inventory of every named theorem in MMT (full P1 audit). Deferred; the TOC + the punch-list above are sufficient at the scale of a 150-page monograph.
• Schwarz, Morse Homology (Birkhäuser 1993) — Fast Track 3.04. Deferred to a separate audit. The Morse-complex / Morse-homology programme is logically downstream of MMT and substantively distinct (gradient flow, transversality, signs, the Morse-Smale-Witten complex); the Codex priority is to anchor classical MMT first and then build the complex on top.
• Audin-Damian Part II — Floer Homology. Already partially covered in 05.08-floer/ and will get its own audit pass alongside the Floer-specific texts (Donaldson Floer / FT 3.06; Hutchings ECH notes).
• Donaldson Floer homology and Yang-Mills Morse theory — FT 3.06. Deferred.
• The $h$-cobordism theorem of Smale 1962 as a substantive unit. Mentioned in the originator citations and as a downstream pointer in 03.02.03; a full unit on the $h$-cobordism theorem is a separate effort (one of the few classical Milnor results not in MMT itself; it is in Milnor's later Lectures on the h-Cobordism Theorem 1965).
• Witten's supersymmetric Morse theory (Witten 1982 "Supersymmetry and Morse theory") — the analytic/physical reformulation that motivates Floer. Deferred to a topological-QFT audit; relevant to the existing gauge-theory and Chatterjee QFT plans.

---

§6 Acceptance criteria for FT equivalence (MMT)

Per docs/plans/FASTTRACK_EQUIVALENCE_PLAN.md §3.4, the book is at equivalence-coverage when:

• ≥95% of MMT's named theorems map to Codex units. Current ~5–10%; after priority-1 units rises to ~55%; after priority-1+2 to ~80%; after priority-1+2+3 to ~95%; full ≥95% requires priorities 1–3 (priority 4 is optional deepening).
• ≥90% of MMT's worked examples (height function on torus / sphere, ${\mathbb{CP}}^n$ as a Morse example, $f(x)=\sum a_i{x}_i^2$ on $S^n$, the Bott periodicity calculation on $\mathrm{U}(2n)$) have either a direct unit or are referenced from a unit covering them.
• The Bott periodicity unit 03.08.07-bott-periodicity.md cites its new Morse-theoretic-proof sibling (item 9 in the punch-list) rather than citing MMT §23 directly without an in-Codex anchor.
• All four Floer units (05.08.01–05.08.04) cite the new Morse units as prerequisites and have Pass-W weaving paragraphs pointing back to them.
• Notation decisions are recorded in 03.02.02 and 03.02.03 (see §4).
• Originator-prose citations of Morse 1925, Morse 1934, Bott 1957/1959, Andreotti-Frankel 1959, and Smale 1962 are present in the relevant units.

The 5 priority-1 units close the Floer-dependency gap entirely and roughly half the absolute coverage gap. Priority-2 closes Part III and connects to Bott. Priority-3 closes Lefschetz hyperplane and the real Bott proof. Priority-4 is master-tier deepening.

---

§7 Sourcing

• Free. Ranicki archive hosts the full PDF at https://webhomes.maths.ed.ac.uk/~v1ranick/papers/milnmors.pdf (5.4 MB, ~160 pp. scanned typescript). This is a free educational mirror of the 1963 Annals of Mathematics Studies publication; cite as Milnor, J., Morse Theory, Annals of Mathematics Studies 51, Princeton University Press, 1963.
• Print. Princeton University Press has kept MMT in print continuously since 1963 (ISBN 0-691-08008-9). Used copies are easy to find.
• Local copy. Add to reference/fasttrack-texts/03-modern-geometry/ as Milnor-MorseTheory.pdf to mirror the pattern of other free FT texts (e.g. Bott already lives in this directory as Bott-LecturesOnMorseTheory.pdf per 03.08.07-bott-periodicity.md line 27).
• Companion peer texts (cited in §1):
  • M. Schwarz, Morse Homology, Progress in Mathematics 111, Birkhäuser
    1993. The Morse-complex / Floer-finite-dim reformulation. FT 3.04.
  • M. Audin, M. Damian, Morse Theory and Floer Homology, Universitext, Springer 2014 (English translation of the 2010 French original). The standard modern reference; Part I is the Morse-complex rewriting of MMT Part I.
  • A. Banyaga, D. Hurtubise, Lectures on Morse Homology, Texts in the Mathematical Sciences 29, Kluwer 2004. Self-contained graduate text structured around a complete proof of the Morse Homology Theorem.
  • L. Nicolaescu, An Invitation to Morse Theory, Universitext, Springer (2nd ed. 2011). The most accessible modern introduction; covers Morse-Smale flows, min-max theory, moment maps and equivariant cohomology, and complex Morse / Picard-Lefschetz theory in addition to the classical material.
• Originator-paper archive locations:
  • Morse 1925 Trans. AMS — JSTOR.
  • Bott 1957 PNAS — open access via pnas.org.
  • Bott 1959 Annals — JSTOR.
  • Andreotti-Frankel 1959 Annals — JSTOR.
  • Smale 1962 Amer. J. Math. — JSTOR.