I'm reading Conway's complex analysis book and on page 84 he says:

Quite often I see this version of the theorem:

Let $f$ be analytic in the simply connected domain $D$ and let $C$ be a simple closed positively oriented contour that lies in $D$. If $z_0$ is a point that lies interior to $C$, then $$f(z_0)=\frac{1}{2\pi i}\int_C\frac{f(z)}{z-z_0}$$ (From this book page 235)

The Conway's formulation seems more general. The problem is I can't understand why Conway's formulation implies in the second formulation. Is the winding number in the case of the second formulation one?

  • 1
    $\begingroup$ That the winding number is one, is exactly en.wikipedia.org/wiki/Jordan_curve_theorem (as $z_0$ is in the interior). And this is the reason why the first theorem implies the second $\endgroup$ – b00n heT Aug 29 '16 at 13:50
  • $\begingroup$ @b00nheT but the winding number is zero when $\gamma$ is homotopic to a constant, no? $\endgroup$ – user42912 Aug 29 '16 at 19:02
  • $\begingroup$ One can derive the first theorem without the knowledge of Jordan curve theorem & the theorem that says that the winding number around Jordan curve is $\pm 1$. But note: there is a different definition for "simply connected domain" that considers homotopic curves and not Jordan curves, i.e. simply connected domain is a region where we can deform any loop into a point. But if you know the Jordan curve theorem, then you can immediately define simple connected domain without any hocus pocus... $\endgroup$ – Hulkster Aug 29 '16 at 19:38
  • $\begingroup$ @Juho Please could you explain the relationship between my doubt and the Jordan Curve theorem. I didn't understand what this theorem has to do with everything else. Thank you $\endgroup$ – user42912 Aug 29 '16 at 22:28
  • $\begingroup$ Note also that you can't always deform loop $\gamma$ into a point without crossing the point $a$. $\endgroup$ – Hulkster Aug 29 '16 at 23:33

enter image description here

Topology - Munkres 2ed, p. 403


The winding number may in fact be 'defined' as $$ n(\gamma;a)= \frac{1}{2\pi i} \oint_\gamma \frac{1}{z-a} \; dz$$ and this is typically also used in the proof. In Conwey's version $G$ need not be simply connected, $\gamma$ may intersect itself, but $\gamma$ should not effectively 'wind around' any points in the complement of $G$. The winding number is indeed one in the second formulation.

  • $\begingroup$ But if $\gamma$ doesn't "wind around" then the winding number is zero, no? $\endgroup$ – user42912 Aug 29 '16 at 14:04
  • $\begingroup$ Indeed. When $\gamma$ is the boundary of an open domain $U$ and goes around counter-clockwise (this corresponds to a curve $C$ as in the second formulation) then $n(\gamma;a)=+1$ if $a\in U$ and zero otherwise. Conwey's formulation allows for the curve $\gamma$ to wind around and then unwind again. For example, if you have an annulus the geometric image of $\gamma$ may in fact encircle the hole in the middle! but still have winding number 0 around points in the hole (you should probably make a drawing for this). $\endgroup$ – H. H. Rugh Aug 29 '16 at 14:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.