JOVANA
Explore Library Glossary Getting Started Three Levels Fields How it works Mission
Join the mission
All guides

The Residue Theorem

The last guide handed you the residue and the recipes to compute it. Now watch a single theorem swallow every version of Cauchy's theorem you have met: a closed-loop integral equals 2 pi i times the sum of the residues it encircles — global integral, local sum.

One theorem to rule the integrals

In the previous guide you learned what a residue is — the single coefficient a_(-1) sitting in front of 1/(z - z_0) in the Laurent series — and you collected fast recipes to compute it without ever expanding the whole series. That number felt like a small bookkeeping detail. This guide reveals why it is the most important coefficient in all of complex analysis: it is the ONLY one that survives integration around a loop. Everything else integrates to zero, and the residue is what is left standing.

Recall the one integral that started it all. Take a circle around z_0 and integrate (z - z_0)^n term by term. Every power except n = -1 has an antiderivative that is single-valued around the loop, so it returns to its starting value and the integral is zero. The lone exception is (z - z_0)^(-1), whose 'antiderivative' is the multivalued log: going once around adds 2 pi i to the argument, and the integral comes out to exactly 2 pi i. So out of the entire two-sided Laurent series, integrating around the singularity keeps precisely one term — the a_(-1) term — and reports 2 pi i times a_(-1).

Write that as a clean rule. For every integer n other than -1, the integral of (z - z_0)^n dz around the circle is 0; for n = -1 alone it is 2 pi i. Integrating the Laurent series of f term by term therefore wipes out everything and leaves 2 pi i times a_(-1), which is exactly 2 pi i times the residue read off as that Laurent coefficient. That is the whole theorem in miniature, for a single singularity — the rest of this guide just lets the loop hold several at once.

Stating the theorem

Now stop looking at one singularity and let the loop be larger, enclosing several of them. The residue theorem says: if f is holomorphic inside and on a simple closed contour C, except at finitely many isolated singularities z_1, ..., z_m strictly inside C, then the integral of f(z) dz around C, taken counterclockwise, equals 2 pi i times the sum of the residues at those points. The integral does not care how complicated f is between the singularities — it reads off only the local residue at each one and adds them up.

integral over C  of  f(z) dz   =   2 pi i  *  sum over k=1..m  Res(f, z_k)

   C  -- simple closed contour, counterclockwise
   z_k -- the isolated singularities of f strictly inside C
   Res(f, z_k) -- the residue at z_k (the a_(-1) of its Laurent series)
The residue theorem: a global contour integral is a local sum of residues.

Why is the answer simply the sum? Because Cauchy's theorem lets you deform the one big loop C — without changing the integral — into a bouquet of tiny circles, one hugging each singularity, joined by hairline corridors that cancel in pairs. The region between the big loop and the little circles is singularity-free, so f is holomorphic there and that part contributes nothing. Each tiny circle then contributes its own 2 pi i times Res from the calculation above. Add them up and you have the theorem. The proof is contour deformation doing all the work.

Why it is the grand generalization of Cauchy

Look at what happens when there are no singularities inside C at all. The sum on the right is empty, so the integral is zero — and that is exactly the Cauchy integral theorem you proved a rung ago. The residue theorem contains it as the case 'sum of nothing.' Cauchy's theorem said holomorphic functions integrate to zero around loops; the residue theorem upgrades that to: they integrate to zero EXCEPT for an explicit, computable correction at each singularity.

It also swallows the Cauchy integral formula. Take the function g(z) = f(z)/(z - z_0) with f holomorphic. This has a simple pole at z_0, and its residue there — by the simple-pole recipe from the last guide — is just f(z_0). Feed that single residue into the theorem and the integral of f(z)/(z - z_0) dz comes out to 2 pi i times f(z_0). That IS the Cauchy integral formula. So both of the great theorems of the previous rung are now two small special cases of one master statement.

A worked loop, end to end

Let us run the whole machine once on a concrete loop. Integrate f(z) = 1/(z^2 - 1) counterclockwise around the circle |z| = 2. The denominator factors as (z - 1)(z + 1), so there are two simple poles, at z = 1 and z = -1 — and both sit well inside the circle of radius 2. Everywhere else inside the loop f is holomorphic. The residue theorem says we only need the residue at each of these two points.

  1. Locate the singularities inside C. The poles are at z = 1 and z = -1; both have |z| = 1 < 2, so both are enclosed. (Nothing on the circle itself, which we will always need to check.)
  2. Residue at z = 1. Using the simple-pole recipe Res = lim (z - 1) f(z) = 1/(z + 1) evaluated at z = 1, which gives 1/2.
  3. Residue at z = -1. Likewise Res = lim (z + 1) f(z) = 1/(z - 1) evaluated at z = -1, which gives -1/2.
  4. Sum and multiply. The residues add to 1/2 + (-1/2) = 0, so the integral is 2 pi i times 0 = 0.

The integral vanishes — but notice it did NOT vanish because f is holomorphic inside (it is not; there are two poles), and not because we found an antiderivative. It vanished because the two residues happened to cancel. Shrink the loop to circle only z = 1 and the answer becomes 2 pi i times 1/2 = pi i instead. The residue theorem makes the loop's contents the whole story: which poles you enclose, and what their residues are.

Winding numbers and honest hypotheses

The version stated above assumes a simple loop traversed once counterclockwise, so each enclosed singularity is wrapped exactly once. The full theorem is sharper: if the contour winds around z_k some integer number of times — its winding number n(C, z_k), counting counterclockwise turns as positive and clockwise as negative — then each residue is weighted by that count. The general formula is the integral around C equals 2 pi i times the sum of n(C, z_k) Res(f, z_k). Go around twice and you double the contribution; go clockwise and you flip its sign.

With the theorem and these caveats in hand, you are now armed for the real prize. The next three guides spend a real-axis integral with no elementary antiderivative, close it off with an added arc into a contour, and read the answer straight from the residues inside — for trigonometric integrals over a period in guide 3, for improper integrals tamed by Jordan's lemma in guide 4, and for branch-cut and series-summing tricks in guide 5. But always remember the catch that those guides will keep insisting on: the theorem only gives the closed-loop integral, so you must separately prove the added arc contributes nothing before the real integral falls out.