$\gamma _1$ and $\gamma _2$ are two closed curves in $\Bbb C\setminus\{0\}$ both with parameter interval $[a,b]$. We define the curve $\gamma(t)=\gamma_1(t)\gamma_2(t)$.

I've just shown that $\gamma$ is closed in $\Bbb C\setminus\{0\}$. Now I have to show that $w(\gamma ,0)=w(\gamma _1 ,0)+w(\gamma _2 ,0)$. I first used a theorem saying that under the given circumstances(or so I thought), $w(\gamma,0)=(1/2\pi i)\int_\gamma (dz/z)$, which easily solved the problem. It turned out, however, that for this to be true, $\gamma$ has to be piecewise $C^1$, which we don't know if it is, so I can't use that theorem. Any ideas what to do instead?

EDIT: My book defines winding number as $w(\gamma,0)=\frac{1}{2\pi}argvar(\gamma)$, and $argvar(\gamma)=\theta(b)-\theta(a)$, where $\theta$ is an arbitrary continuous argument function along $\gamma$.

  • $\begingroup$ what exactly do you call $w$ ? $\endgroup$ – Glougloubarbaki May 30 '13 at 20:44
  • 2
    $\begingroup$ How are you defining the winding number and which textbook are you using? You are right that it is a simple consequence of the integral formula for winding number if the curve is piecewise smooth. However, for more general curves, you need to use a covering space argument. More context would allow me to provide a more appropriate answer for you situation. $\endgroup$ – Potato May 30 '13 at 20:50
  • $\begingroup$ Sorry, I've defines the winding number now. $\endgroup$ – MBrown May 30 '13 at 21:27
  • $\begingroup$ Which book? The point of defining winding number is that you can't define a single valued continuous argument function on $\mathbb C\backslash \{0\}$, so that definition needs some tweaking. $\endgroup$ – Potato May 30 '13 at 21:32

Here's a more geometric description of what's going on.

Since all the curves avoid the origin, you can parametrize them in polar coordinates using Euler's formula (treating $\mathbb{C} \setminus 0 \cong \mathbb{R}^2 \setminus \{0,0\}$). $$ \gamma(t) = \begin{pmatrix} r(t) \cos \theta(t) \\ r(t) \sin \theta(t)\end{pmatrix}, $$ where $r$ and $\theta$ are continuous functions from $[a, b] \to \mathbb{R}$, and $r(t) > 0$ for all $t$. Now, normalize by projecting onto the unit circle: $$ \begin{array}{rccc} \epsilon = \frac{\gamma}{||\gamma||}: &[a,b] &\longrightarrow &S^1 \\ & t &\longmapsto &\begin{pmatrix} \cos \theta(t) \\ \sin \theta(t)\end{pmatrix} \end{array} $$ This map lifts to a map $\tilde{\epsilon}: [a, b] \to \mathbb{R}$ that is unique once you specify $\tilde{\epsilon}(a)$ (or the image of any one point, for that matter). In other words, $$ \epsilon = p \circ \tilde{\epsilon}, $$ where $p$ is the projection $\mathbb{R} \to S^1$, $u \to (\cos u, \sin u)$.

If $\gamma$ is a closed curve, then $\epsilon$ is too, so $$ \tilde{\epsilon}(b) - \tilde{\epsilon}(a) = 2\pi w $$ for some $w \in \mathbb{Z}$. This $w$ is the winding number of $\gamma$ around $0$.

Now, to address your actual question, in which you multiply the curves in the complex plane, first parametrize them as maps $[a, b] \to \mathbb{R}^2 \setminus \{0, 0\}$: $$ \gamma_1(t) = \begin{pmatrix} r_1(t) \cos \theta_1(t) \\ r_1(t) \sin \theta_1(t)\end{pmatrix} $$ and $$ \gamma_2(t) = \begin{pmatrix} r_2(t) \cos \theta_2(t) \\ r_2(t) \sin \theta_2(t)\end{pmatrix}. $$ Now, $$ \begin{align} \gamma(t) &= \gamma_1(t) \gamma_2(t) \\ &= r_1(t) r_2(t) \begin{pmatrix} \cos \theta_1(t) \cos \theta_2(t) - \sin \theta_1(t) \sin \theta_2(t) \\ \sin \theta_1(t) \cos \theta_2(t) - \cos \theta_1(t) \sin \theta_2(t) \end{pmatrix} \\ &= r_1(t) r_2(t) \begin{pmatrix} \cos \big(\theta_1(t) + \theta_2(t) \big) \\ \sin \big(\theta_1(t) + \theta_2(t) \big) \end{pmatrix}. \end{align} $$ From this you can see that the winding number is additive.


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.