I was looking for proofs of the generalized Cauchy integral theorem:

Theorem (Generalized Cauchy integral theorem): Let $\Omega\subset\mathbb{C}$ be an open set and $\gamma:[a,b]\to\Omega$ be a piecewise $C^1$ closed curve such that $\text{ind}_\gamma(z)=0$ for all $z\in\mathbb{C}\backslash\Omega$. Let $f:\Omega\to\mathbb{C}$ be an holomorphic function. Then $\int_\gamma f(z)dz=0$.

Now all the proofs that I found not using topological arguments were based on proving Cauchy's generalized integral formula using Liouville's theorem. However, after we have shown that a function holomorphic on an open set is analytic on that set, isn't it possible to prove Cauchy's theorem by dividing $\mathbb{C}$ into its connected components induced by $\gamma$, and then using green's theorem on each simple closed piecewise $C^1$ component of $\gamma$? The only problem that I see is that looking at the conditions for Green's theorem, I saw that it requires $\gamma$ to be piecewise smooth instead of piecewise $C^1$. So is there a way around this condition on the curve and do you think that otherwise the proof is valid?


1 Answer 1


Let $\Omega \subset \mathbb{R}^2$ be a bounded and open set, and let $f$ be holomorphic on $\Omega$ and $C^1$ on a neighborhood of $\overline{\Omega}$. The proof of Cauchy's theorem through Green's theorem uses $d(f\,dz) = 0$ to deduce $$\int_{\partial \Omega}f\,dz = \int_{\Omega}d(f\,dz) = 0.$$ The only question is how regular does $\partial \Omega$ need to be for Green's theorem to be true?

The standard direct proof of Green's theorem works for open sets that can be divided into finitely many type I regions and into finitely many type II regions. These sets can have pieces of their boundary that are merely continuous.

A version of Green's theorem that follows directly from the Stokes theorem for manifolds with corners is that if $\Omega$ is a bounded open set such that $\overline{\Omega}$ is a $C^2$ manifold with corners and $f, g$ are $C^1$ on a neighborhood of $\overline{\Omega}$, then $$\int_{\Omega}(g_x - f_y) \,dx\,dy = \int_{\partial \Omega}(f\,dx + g\,dy).$$

According to "Introduction to Analysis in Several Variables" here, the regularity assumption can be relaxed to $\overline{\Omega}$ only being $C^1$ with corners. It can even be relaxed to $\Omega$ having Lipschitz boundary. A direct proof of this version of Green's theorem can be found in appendix I of "Measure Theory and Integration" by Taylor. The book refers to "Geometric Measure Theory" by Federer for "substantially more sophisticated generalizations" of the Stoke's theorem.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .