\( \newcommand{\cross}{\times} \newcommand{\isom}{\cong} \newcommand{\RP}{\mathbb{RP}} \newcommand{\hatC}{\hat{\CC}} \newcommand{\CC}{\mathbb{C}} \newcommand{\DD}{\mathbb{D}} \newcommand{\EE}{\mathbb{E}} \newcommand{\HH}{\mathbb{H}} \newcommand{\RR}{\mathbb{R}} \newcommand{\ZZ}{\mathbb{Z}} \newcommand{\orb}{\mbox{orb}} \newcommand{\Isom}{\operatorname{Isom}} \newcommand{\GL}{\operatorname{GL}} \newcommand{\Sim}{\operatorname{Sim}} \newcommand{\Aut}{\operatorname{Aut}} \newcommand{\Exp}{\mathsc{Exp}} \newcommand{\Log}{\mathsc{Log}} \)

MA3B8 Complex analysis
Term I 2026-2027

Schedule

Week Date of Monday Topics Lecture notes Example sheet Comments
1 Oct. 5 Tuesday: Admin. Some highlights of the module. Standard notations. Arguments, multi-valued functions. Domains. Branches. The complex square root is a multi-valued function. Riemann surfaces (only mentioned in passing and not examinable). Wednesday: Contours, simple, closed. Contour integrals. Linearity in the integrand and in the domain of integration. The integral of \(dz/z\) about the unit circle is \(2 \pi i\). Laurent polynomials. Integral of a Laurent polynomial about the unit circle. Element of arc length, arc-length functional, scaling properties. Friday: Reparametrisation. ML-inequality. Element of argument, variation of argument functional, scaling properties. Branches of argument. Element of radius, element of logarithm of radius. "Better" definition of complex square root. Begin discussion of complex logarithm. Tuesday Wednesday Friday One Tuesday: I defined a multi-valued function as having domain "some set" and codomain "some powerset". However, I should also have mentioned that the image of a point can never be the empty set. Wednesday: A student asked if contours are continuous. The answer is "yes". Piecewise continuous implies continuous, by the "gluing lemma". Another student asked if we allow contours to be defined on (say) all of \(\RR\). The answer is "no". Perhaps we would call that an "improper contour", to match "improper integral"? Friday: I got confused during the proof of the lifting lemma - what I called uniqueness is actually a proof that \(\theta(t)\) is an argument of \(\gamma(t)\). The actual uniqueness proof is the last paragraph of Proposition 3.4 in the typed notes.
2 Oct. 12 Tuesday: Element of complex logarithm. Holomorphicity, holomorphic implies continuous, algebra of holomorphic functions. Examples and non-examples. Error, the (real) total derivative, conformality. A sketch of the squaring function. Wednesday: Power series, radius of convergence, standard properties, power series expansions. Complex exponential and trigonometric functions via power series, standard properties, zeros of complex cosine and sine. A sketch of the complex exponential. Complex logarithm, relation to the argument, branches. Segments, convexity, hulls. Friday: Triangles, positive, negative, degenerate. Integrals about triangles. Midpoint subdivision, properties of smaller triangles. Goursat's lemma and its proof. Tuesday Wednesday Friday Two Friday: A student asks "Where do we use Lemma 4.8 other than in the proof of Goursat's lemma?" It is also used in the proofs of Lemma 4.10 (computing the total derivative) and Theorem 12.1 (Cauchy's reproduction formula).
3 Oct. 19 Tuesday: Holomorphic functions integrate to zero about rectangles. Affine segments, piecewise-linear contours. Polygons. Orientation convention. Holomorphic functions integrate to zero about polygons. Decomposing polygons into triangles. Wednesday: Annuli. Holomorphic functions integrate to zero about annuli via far-too-clever proof. Regions, holomorphic functions integrate to zero about regions (Cauchy's theorem for regions). Uniformisation of Riemann surfaces (only mentioned in passing and not examinable). Primitives, holomorphic functions have primitives in discs. Friday: Has primitive in \(U\) iff integrates to zero about all closed contours in \(U\). Standard simplex, faces. Singular simplices, chains, boundary operator. Cycles and boundaries. Tuesday Wednesday Friday Three Tuesday: A student asks "Why do we need the interior of a polygon to be dense in the polygon?" Because we want polygons to be two-dimensional, not one-dimensional. A student asks "Can a polygon have infinitely many sides?" No, because contours have only finitely many pieces. A student asks "Why are the pieces (after cutting \(P\)) convex?" Because we cut by entire lines, not just line segments. Wednesday: A student asks "Why did the orientations (of the rectangle and annulus) work out? What do we do with an example where the orientations don't work out?" The orientations will always work out. This is because holomorphic maps preserve signed angles so they preserve the relation "to the left of". Friday: A student asks "Is \(C_n(X)\) the free abelian group generated by singular \(n\)-simplices?" Yes.
4 Oct. 26 Tuesday: Cycles, boundaries, and homology. First homology of convex domain is trivial. Boundary squared is zero. Coning operator. Non-example - the punctured plane has non-trivial homology (proof deferred). Midpoint subdivision operator in dimensions one and two. Midpoint subdivision commutes with taking boundaries. Wednesday: Subdivision shrinks summands. Straightening, straightening commutes with taking boundaries. Cauchy's theorem for one-boundaries, its proof, and a more common statement. Friday: Domains with non-trivial first homology. Annuli, homeomorphisms. Regions, accessibility. Tuesday Wednesday Friday Four Tuesday: A student asks "Does a domain have trivial first homology if and only if it is convex?" No. Another student asks "What does it mean for one group to be less than another?" That is our notation for the relation "is a subgroup of". Wednesday: A student asks "How do we know that the straightened one-simplex is disjoint from the original (summand of the subdivided) singular one-simplex?" We don't and we don't need to - we only need that the difference (original - straightened) has the same integral as a closed contour. We do not need simplicity.
5 Nov. 2 Tuesday: Positive orientation of the boundary of a region. Cauchy's theorem for regions. The boundary of a region is a one-boundary. The reproduction formula. Holomorphic functions satisfy the reproduction formula. Wednesday: The (complex) mean value theorem. Analytic functions. Analytic functions are holomorphic. Holomorphic functions are analytic, reproduction of coefficients, Cauchy estimate. Friday: Morera's theorem. The fundamental theorem of complex analysis. Elementary computation of a contour integral. Tuesday Wednesday Friday Five Wednesday: A student asks "Why is the complex exponential analytic?" I sketched three proofs. (1) By Abel's rearrangement theorem, power series are analytic inside their disc of convergence. (2) The addition law implies \( \Exp(z) = \Exp(z - z_0 + z_0) = \Exp(z - z_0) \Exp(z_0) \); this gives a power series expansion of \(\Exp(z)\) about \(z_0\). (3) Functions given by power series are holomorphic (inside their disc of convergence) so are analytic by the version of Cauchy's theorem we are proving today. A student asks "How much of Sections 11 and 12 (Regions) is examinable?" You are responsible for the statements (definitions, lemmas, propositions, and theorems) and the exercises. You are not responsible for the proofs. I would be very grateful for any feedback on this material (as it is new). Friday: A student asks "Why can't we just use the indefinite integral \(\int dz/z = \Log(z) + c\)?" Because \(\Log\) is not a function - it is a multivalued function. Said another way, we need to know which branch of \(\Log\) to use. We will return to this topic.
6 Nov. 9 Six
7 Nov. 16
8 Nov. 23
9 Nov. 30
10 Dec. 7