Can a harmonic vector field $X$ on a Riemannian surface $(M,g)$ possess a limit cycle(An isolated periodic orbit)?

Note that the Laplacian of a vector field is defined via natural correspondence between the space of vector fields and the space of $1$-forms.(The natural correspondence arising from the Riemannian metric). If this correspondence is denoted by $i$ then the Laplacian of a vector field $X$ is defined as $\Delta X=i^{-1} \Delta (i(X))$. Where the latter Laplacian is the natural Laplacian on the space of differential forms.

In particular is there a quadratic vector field $$\begin{cases} x'= ax+by+\lambda(x^2-y^2)+txy \\ y'=cx+dy+\mu (x^2-y^2)+sxy\end{cases}$$

which has at least one limit cycle?

The second question: Assume that $X$ is a vector field on a Riemannian surface Assume that $\gamma$ is a periodic orbit of $X$. Is it true to say that there is a point $p\in \gamma$ such that $\Delta X$ is tangent to $\gamma$ at $p$?

  • $\begingroup$ I think the answers to both the questions depend on the de Rham cohomology of M - that is, whether or not the limit cycle bounds a simply connected domain. Could you clarify this point? As for the second question, the 1-form i(X) is closed and co-closed. Hence, the inner product of i(X) with its Laplacian ( as defined by the metric g) can be seen to be positive. $\endgroup$
    – Chaitanya
    Nov 23, 2017 at 12:38

1 Answer 1


Consider the infinite strip, $[0,1]\times\mathbb{R}$ with $[0,y]$ identified with $[1,y]$ for all $y\in\mathbb{R}$, and with the Riemannian metric $g=dxdx+dydy$. Consider $X=\partial_x +f(y)\partial_y$. The Laplacian is $\Delta X =f''(y) \partial_y$.

For the first question, let $f=-y$, so that the line $y=0$ is a limit cycle and $\Delta X=0$. For the second question, let $f=y^2$, so that $y=0$ is again a periodic orbit, but now, everywhere on the curve, $\Delta X=2\partial_y$ is orthogonal to the tangent $\partial_x$.

  • $\begingroup$ A limit cycle is in priori a compact set but the line $x=0$ is not a compact set on the cylinder. Moreover is your vector field , a well defind vector field on the cylinder? $\endgroup$ Jan 6, 2018 at 20:14
  • 1
    $\begingroup$ I had the roles of x and y reversed in the second paragraph. I've now fixed this. Is this now correct? $\endgroup$
    – user119324
    Jan 6, 2018 at 22:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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