I know that the volume form on $S^1$ is $\omega= ydx-xdy$. But how I can derive that? The only things that I know are the definition of differential q-form, and the fact that the vector field $v= y \frac{\partial}{\partial x}-x\frac{\partial}{\partial y}$ never vanishes on $S^1$.
-
@Martín-BlasPérezPinilla I know this..ok, the form that I have written isn't correct...there must be $\omega=\omega_I dx \wedge dy$...but you don't answer my qeustion. – andreasvr Jan 28 '14 at 10:47
-
$ω_Idx\wedge dy$ is a 2-form, isn't? – Martín-Blas Pérez Pinilla Jan 28 '14 at 10:54
-
Yes is a 2 form – andreasvr Jan 28 '14 at 10:56
-
@Martín-BlasPérezPinilla I'm sorry, I say a nonsense thing. The $S^1$ is one dimensional, so i have to find a form like the fist that i have wrote in the top of the post... – andreasvr Jan 28 '14 at 11:02
2 Answers
You speak of "the" volume form. There are many volume forms on any orientable smooth manifold. Given $\omega$ a volume form, and a non-vanishing smooth function $f$, then $f\omega$ is again a volume form.
The task of finding "a" volume form is to find a non-vanishing top degree form on the manifold in question. To verify that a one form $\omega$ is a volume form on $S^1$ you of course need to check that for any tangent vector $X$ of $S^1$, the contraction $\omega(X) \neq 0$. Now, any one form on $\mathbb{R}^2$ can be written as $\omega_x \mathrm{d}x + \omega_y \mathrm{d}y$. Since you already know that the tangent space of $S^1$ is spanned by $y \partial_x - x\partial_y$, it suffices to find any pair of real functions $\omega_x, \omega_y$ such that $$ \omega(X) = \omega_x y - \omega_y x \neq 0 $$ for any $(x,y)\in S^1$. The choice that $\omega_y = -x$ and $\omega_x = y$ is just one of many possible choices.
Because of the freedom described above, there isn't one method to derive your given one form as the volume form on $S^1$ ... unless you add additional requirements. For example, on the unit circle the volume form you wrote down is the "natural" one (up to choice of orientation) given by the induced Riemannian metric. And perhaps this is what you are seeking, in the end:
The natural volume form on $\mathbb{R}^2$ is $\mathrm{d}x\mathrm{d}y$. Using the Riemannian metric on $\mathbb{R}^2$ we can write down the unit normal vector field to the unit circle, and this happens to be $x\partial_x + y\partial_y$. Since $\mathbb{R}^2$ is orientable and $S^1$ has a unit normal field, it is also orientable. And a volume form (indeed the one for the induced Riemannian metric) can be given by $$ \omega = (\mathrm{d}x\mathrm{d}y)(x \partial_x + y\partial_y, \cdot) $$
This method is not restricted to Riemannian manifolds. Suppose $\Sigma$ is a hypersurface in a smooth manifold $M$. Assume that $M$ is orientable and so $\omega$ is a volume form for $M$. Assume also that $\Sigma$ admits a field of "normal vectors", by which I mean that there exists a vector field $n$ defined along $\Sigma$ that is never in the tangent space to $\Sigma$. Then a volume form for $\Sigma$ can be found by contracting (taking the interior product) $\iota_n\omega$.
- 73,139
-
1Thank you very much. Your explanation is clear and answered at all my doubts. You have centered the problem and now, finally, i have understand! Again, thank you! – andreasvr Jan 28 '14 at 14:56
See the proof of Proposition 12.6 in http://www.math.toronto.edu/mat1300/orientation.11.pdf.
EDIT: Wikipedia gives the following reference for the deduction of the generalization of your formula: Flanders, Harley (1989). Differential forms with applications to the physical sciences.
EDIT2: $$x=\cos\theta, y=\sin\theta$$ $$xdy-ydx=\cos\theta\cos\theta d\theta - \sin\theta(-\sin\theta)d\theta=d\theta$$ and now, go back.
- 41,546
- 4
- 46
- 89
-
Maybe I don't explain well my problem. I'll try again. I want to know how I can derive a formula for a volume form on S^1. The proposition you suggest me to see solve a different problem. I don't want to know if a volume form exist, but I want to know how to write it. Clear? – andreasvr Jan 28 '14 at 10:53
-
OK. The proposition 12.6 in this case gives the 1-form $d\theta$, I think. – Martín-Blas Pérez Pinilla Jan 28 '14 at 11:11
-
-
2What is your definition of volume form? Is it unique? Can you imagine how to find it by parametrising $S^1$? – Siminore Jan 28 '14 at 12:48
-
@Siminore my definition are: a volume form on M is a form $\omega \in \Omega^{dim M}(M)$ such that $\forall p \in M$, $\omega_p$ is never 0. – andreasvr Jan 28 '14 at 13:50
-
@Siminore No, I don't know how I can find it. I know that i can parametrise $S^1$ with an angular variable...but this don't solve my problem.. – andreasvr Jan 28 '14 at 13:55
-
I think that your definition of volume form is too abstract to be actually useful. You should study some further equivalent definitions that turn out to be more useful to make computations. – Siminore Jan 28 '14 at 15:17