Final synthesis workshop

Every exam question on complex variables is secretly the same question in disguise: given a contour integral $\oint_C f(z)\,dz$ , which of the unit's big theorems actually applies? This closing workshop of Unit 5 does not introduce new material — it wires together everything you already proved, from the Cauchy-Riemann equations through Cauchy's Integral Theorem and Formula, Taylor/Laurent series, and singularity classification, into one decision chain that ends at the Residue Theorem. You will practice reading a problem, checking analyticity, counting and classifying the singularities a specific contour encloses, and picking the fastest correct tool — the exact judgment call examiners reward and half-prepared students skip.

The core insight

Analyticity forces $\oint_C f\,dz=0$ (Cauchy's Theorem), a single well-placed singularity collapses that integral to a derivative value (Cauchy's Formula), and any finite set of isolated singularities reduces to $\oint_C f\,dz = 2\pi i\sum\operatorname{Res}$ (the Residue Theorem) — the whole unit is one chain, and exam skill is choosing the right link.

The core idea

The unit's logical chain runs in a single direction, and each arrow is a genuine dependency:

\text{analyticity (C-R eqns)} \;\Rightarrow\; \oint_C f\,dz = 0 \;\Rightarrow\; f(a) = \frac{1}{2\pi i}\oint_C \frac{f(z)}{z-a}dz \;\Rightarrow\; f(z) = \sum c_n(z-a)^n \;\Rightarrow\; \operatorname{Res} = c_{-1} \;\Rightarrow\; \oint_C f\,dz = 2\pi i\sum \operatorname{Res}.

where: $C$ is a simple closed contour traversed counter-clockwise, $a$ is a point inside $C$ , $c_n$ are the Taylor or Laurent coefficients of $f$ about $a$ , and $\operatorname{Res}$ (the residue) is the coefficient $c_{-1}$ of the $(z-a)^{-1}$ term in the Laurent expansion about each singularity enclosed by $C$ . The single most exam-relevant habit is running this decision tree before touching any formula. First ask whether $f$ is analytic everywhere inside and on $C$ — if so, Cauchy's Integral Theorem alone gives $0$ and you stop. If not, ask whether $f$ has exactly one singularity inside $C$ of the clean form $g(z)/(z-a)^{n+1}$ with $g$ analytic inside $C$ — Cauchy's Integral Formula then hands you $\oint_C f\,dz = \frac{2\pi i}{n!}g^{(n)}(a)$ directly, no series work needed. Only when the singularities are more general do you fall back to the full machinery: classify each isolated singularity (removable, pole of order $m$ , or essential) from its Laurent series, compute each residue, and sum via the Residue Theorem. A real-integral target is folded in by converting to a contour integral first (unit-circle substitution or a semicircular arc) and then applying the same three-way check. Two connections tie this back to earlier units. The Cauchy-Riemann equations force $u$ and $v$ to be harmonic conjugates, so Unit 2's Laplace equation $f_{xx}+f_{yy}=0$ resurfaces here as a consequence of analyticity; and the radius-of-convergence tests for Taylor and Laurent series are the same ratio/root tests from Unit 4, now applied with $|z-a|$ in place of $|x-c|$ .

Figure

A flowchart showing the decision tree for evaluating a contour integral, starting from is f analytic inside C, branching to Cauchys theorem giving zero, or to one clean singularity giving Cauchys formula, or to several isolated singularities giving the residue theorem

Derivation AWhy the residue at a pole of order m has that formula

The goal — proving that the residue of $f$ at an isolated pole of order $m$ equals $\operatorname{Res}_{z=a}f(z) = \frac{1}{(m-1)!}\lim_{z\to a}\frac{d^{m-1}}{dz^{m-1}}\left[(z-a)^m f(z)\right]$ — matters because it converts the abstract statement "residue equals $c_{-1}$ " into a concrete recipe you can execute on any pole without writing out a full Laurent series each time. Since a pole of order $m$ means $f$ has finitely many negative-power terms, we can isolate the coefficient we need with ordinary differentiation.

Step 1. Write the Laurent expansion of $f$ about the pole $z=a$ , valid on a punctured disc: $f(z) = \dfrac{c_{-m}}{(z-a)^m} + \cdots + \dfrac{c_{-1}}{z-a} + c_0 + c_1(z-a) + \cdots$ . Because the pole has order exactly $m$ , the most negative power present is $(z-a)^{-m}$ , so there are only finitely many singular terms.

Step 2. Multiply through by $(z-a)^m$ to clear every negative power at once: $(z-a)^m f(z) = c_{-m} + c_{-m+1}(z-a) + \cdots + c_{-1}(z-a)^{m-1} + c_0(z-a)^m + \cdots$ . This new function is analytic at $z=a$ (all powers are now non-negative), so it is a legitimate Taylor series and we may differentiate it term by term.

Step 3. The coefficient we want, $c_{-1}$ , now sits on the $(z-a)^{m-1}$ term of this Taylor series. Standard Taylor-coefficient extraction says the coefficient of $(z-a)^{m-1}$ in a Taylor series equals $\frac{1}{(m-1)!}$ times the $(m-1)$ -th derivative evaluated at $a$ . Differentiating $(z-a)^m f(z)$ exactly $m-1$ times and evaluating the limit as $z\to a$ isolates precisely that coefficient, since every term of lower or higher degree either vanishes or drops out under differentiation and the limit.

Step 4. Collecting the result: $\operatorname{Res}_{z=a}f(z) = c_{-1} = \dfrac{1}{(m-1)!}\lim_{z\to a}\dfrac{d^{m-1}}{dz^{m-1}}\left[(z-a)^m f(z)\right]$ . For a simple pole ( $m=1$ ) the derivative disappears and this collapses to the familiar shortcut $\operatorname{Res}_{z=a}f(z) = \lim_{z\to a}(z-a)f(z)$ .

Worked example

Problem. Evaluate $\displaystyle\oint_C \frac{z}{(z-1)(z+3)^2}\,dz$ for the contour $C: |z|=5$ , traversed counter-clockwise.

Step 1. Locate and classify the singularities. The integrand has a simple pole at $z=1$ (since $z-1$ appears to the first power) and a pole of order $m=2$ at $z=-3$ (since $z+3$ appears squared). Both lie inside $|z|=5$ , because $|1|=1<5$ and $|-3|=3<5$ — this rules out Cauchy's Theorem (not analytic inside $C$ ) and rules out the direct Integral Formula (more than one singularity enclosed), so the Residue Theorem is the correct tool.

Step 2. Residue at the simple pole $z=1$ , using the $m=1$ shortcut from the derivation above: $\operatorname{Res}_{z=1}\dfrac{z}{(z-1)(z+3)^2} = \lim_{z\to1}(z-1)\cdot\dfrac{z}{(z-1)(z+3)^2} = \dfrac{1}{(1+3)^2} = \dfrac{1}{16}$ .

Step 3. Residue at the order-2 pole $z=-3$ , using the $m=2$ formula (one derivative, then the limit): $\operatorname{Res}_{z=-3}f(z) = \lim_{z\to-3}\dfrac{d}{dz}\left[\dfrac{z}{z-1}\right] = \lim_{z\to-3}\dfrac{(z-1)-z}{(z-1)^2} = \dfrac{-1}{(-3-1)^2} = -\dfrac{1}{16}$ .

Step 4. Sum the residues and apply the Residue Theorem: $\tfrac{1}{16} + \left(-\tfrac{1}{16}\right) = 0$ , so $\oint_C f\,dz = 2\pi i \cdot 0 = \mathbf{0}$ .

Competitive stretch

Shrink the contour to $C:|z|=2$ instead. Now only $z=1$ is enclosed ( $|-3|=3>2$ lies outside), so the sum has a single term: $\oint_C f\,dz = 2\pi i \cdot \tfrac{1}{16} = \dfrac{\pi i}{8}$ . The pole locations never moved — only which ones a smaller contour happens to enclose changed, and that alone flipped the answer from $\frac{\pi i}{8}$ to $0$ .

Common traps

Reaching for the Residue Theorem on every problem. It always works, but recognizing that $f$ is analytic (answer $0$ ) or that only one clean singularity is enclosed (direct Integral Formula) is often faster and less error-prone on a timed exam.
Mixing up which theorem needs which hypothesis. Cauchy's Theorem requires analyticity inside and on $C$ ; the Integral Formula requires exactly one singularity of the specific $g(z)/(z-a)^{n+1}$ form; the Residue Theorem is the general fallback for any finite set of isolated singularities.
Forgetting an unresolved constant. Harmonic-conjugate problems end in an unknown additive constant that a boundary condition like $v(0,0)=0$ must pin down; skipping that step leaves the answer incomplete.
Not rechecking which singularities the current contour encloses. As the worked example shows, the same integrand on $|z|=2$ versus $|z|=5$ gives different answers because a different subset of poles counts.

Ask the tutor: For a mixed review session that jumps between C-R verification, contour-integral decision-making, and residue computation — mimicking exam conditions — ask the AI tutor to build you a combined practice set covering the whole Complex Variables unit.

How do I decide fast, under exam pressure, which theorem to use?

Run the three checks in order and stop at the first yes. Is $f$ analytic everywhere inside and on $C$ ? Then the answer is $0$ by Cauchy's Theorem — done. If not, is there exactly one singularity inside $C$ and does the integrand have the clean form $g(z)/(z-a)^{n+1}$ with $g$ analytic? Then use the Integral Formula directly. Only if neither shortcut applies do you classify every enclosed singularity and sum residues. Most exam problems are built to reward whichever of the first two checks you spot quickest.

Why did shrinking the contour from radius 5 to radius 2 change the answer?

The poles of $\dfrac{z}{(z-1)(z+3)^2}$ sit at fixed locations, $z=1$ and $z=-3$ , and never move — but the Residue Theorem only sums residues of singularities the contour actually encloses. At $|z|=5$ both poles are inside, and their residues happen to cancel to $0$ . At $|z|=2$ only $z=1$ is inside, so only its residue counts, giving $\pi i/8$ . Always redraw or re-check which poles satisfy $|a| <$ contour radius before summing.

What is the practical link between harmonic conjugates and this unit's contour-integral work?

The Cauchy-Riemann equations are the bridge: if $u$ and $v$ satisfy them, $f=u+iv$ is analytic, and analyticity is exactly the hypothesis every downstream theorem in this chain — Cauchy's Theorem, Cauchy's Formula, Taylor/Laurent expansion, residues — depends on. Finding a harmonic conjugate is really just verifying, from the real and imaginary parts, that an analytic $f(z)$ exists at all before you attempt to integrate or expand it.

Final synthesis workshop

The core idea

Worked example

Competitive stretch

Common traps

Before this