Let $F$ be an orientable surface (possibly with boundary) with a foliation $\cal F$ with $k$-prong saddle singularities only for $k\geq 3,$ (as in figure borrowed from Farb-Margalit book). Suppose that $\alpha, \beta$ are transversal multi-curves which are isotopic in $F$. Are they necessarily transversally isotopic? No, if $\alpha, \beta$ are separated by a Reeb component. But what about a Reebless foliation? If still not, is there a broad, natural condition on $\cal F$ which guarantees that to be true?

If we allow for additional singularities in $\partial F$, then we need to assume that the endpoints of $\alpha$ and $\beta$ coincide as a transversal isotopy cannot move an endpoint through a singularity. 3- and 4-prong singularities


1 Answer 1


"Reebless" is not strong enough for this to be true. One should also assume that there is no "half-Reeb" annulus, i.e. for every circle leaf $C$ and for each side of $C$ the nearby leaves on that side are not allowed to spiral into $C$.

With that assumption, let me describe the proof for the closed surface case; it is easily adapted to cover the bounded case.

By perturbing $\alpha,\beta$ through transversal curves, we may assume that $\alpha,\beta$ are transverse to each other. Their intersection $\alpha \cap \beta$ consists of $n$ transverse intersection points for some $n \ge 0$. We proceed by induction.

If $n=0$ then $\alpha\cup\beta$ bounds an annulus $A$ with a singular foliation transverse to the boundary. By the index formula, the foliation restricted to $A$ has no singularities. Also, there cannot be any circle leaf in $A$, because such a circle would be isotopic to a core curve of $A$, and it would follow that $A$ contains a half-Reeb annulus. Thus $A$ is homeomorphic to $S^1 \times [0,1]$ with the foliation into arcs $x \times [0,1]$. Now we can transversely isotope $\alpha$ to $\beta$ through $A$.

If $n > 1$ then there exist arcs $a \subset \alpha$, $b \subset \beta$ such that $a \cap b = \partial a \cap \partial b = \{x,y\}$, and the simple closed curve $a \cup b$ bounds a disc $D \subset S$ whose interior is disjoint from $\alpha \cup \beta$. Furthermore, at the two endpoints $x,y$ the foliation is externally tangent to $D$. By the index formula, the foliation restricted to $D$ has no singularity, and has neither a closed leaf nor a leaf which has both endpoints on $a$ or on $b$. It follows that $D$ is foliated like the vertical arcs in $\mathbb{R}^2$ restricted to $$\{(x,y) \mid x^2 + (y-1)^2 \le 4, x^2 + (y+1)^2 \le 4\} $$ Now we can transversely isotope $\alpha$ across $D$ so as to remove two points of intersection.

  • $\begingroup$ Thanks, Lee! This is exactly what I was looking for. And now it makes me curious: Measured foliations have no half-Reeb annuli. Would you know if the opposite holds too: No half-Reeb annuli implies existence of a regular invariant measure? $\endgroup$
    – Adam
    Feb 28, 2018 at 20:00
  • $\begingroup$ If by "regular" you mean "full support" then no, there's generally not any such measure, because of the Denjoy construction. $\endgroup$
    – Lee Mosher
    Feb 28, 2018 at 20:58

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.