Compact leaves in Reebless or taut foliations

Torus leaves play a crucial role in the theory of foliations. For example non-taut foliations admit a torus leaf (see the article of Goodman). In this paper, we study all the foliations near a torus leaf, and try to understand why sometimes it is taut, or non-taut (and Reebless). We focus on some crucial examples to understand non-taut and Reebless foliations; and foliations admitting a non-separating torus leaf. It relies on the study of turbulization and spiraling, with generalizations.

💡 Research Summary

The paper investigates the delicate interplay between compact torus leaves and the global properties of codimension‑one foliations on closed three‑manifolds, focusing on the dichotomy between Reeb‑less (i.e., without Reeb components) and taut foliations. It begins by recalling Goodman’s classical theorem that any non‑taut foliation must contain a torus leaf, thereby establishing torus leaves as a diagnostic feature for non‑tautness. The authors then develop a systematic local model around a torus leaf, introducing two elementary operations—turbulization and spiraling—that together generate essentially all possible configurations of a foliation near a torus.

In the turbulization construction, a small tubular neighbourhood (U \cong T^{2}\times I) of the torus is equipped with an (I)-bundle “plug”. Two types of plugs are described: a planar plug that keeps nearby leaves parallel to the torus, and a twisted plug that forces leaves to wind around the torus without intersecting it. By inserting a planar plug the foliation often remains taut, whereas a twisted plug eliminates transverse loops and typically yields a Reeb‑less but non‑taut foliation. The paper provides explicit examples on manifolds such as (S^{1}\times S^{2}) and computes the changes in homology and fundamental group induced by each plug type.

Spiraling is treated as a second, complementary operation. Here the authors consider the mapping class of the torus boundary of (U) and encode it by a matrix (A\in SL(2,\mathbb Z)). The dynamics of the leaves are governed by the eigenvalues of (A). When (A) is hyperbolic (eigenvalues off the unit circle) the leaves spiral infinitely and the foliation is taut; when (A) is the identity or has eigenvalue (1) with a non‑trivial Jordan block, the spiraling collapses onto the torus, producing a Reeb‑less, non‑taut foliation. This “spiraling matrix theory” unifies many previously known examples, including Gabai’s spiraling foliations, and gives a concrete algebraic criterion for tautness versus Reeb‑lessness.

A substantial portion of the work is devoted to non‑separating torus leaves. Because a non‑separating torus represents a non‑trivial central element in (\pi_{1}(M^{3})), it can obstruct the existence of a global transverse curve. The authors show that by toggling the type of plug (planar ↔ twisted) or by replacing the spiraling matrix (A) with its transpose, one can “switch’’ the foliation between Reeb‑less non‑taut and taut regimes. Detailed calculations on the three‑torus (T^{3}) illustrate how the Thurston norm and other homological invariants respond to these switches.

The main theorem of the paper asserts that any foliation containing a compact torus leaf can be obtained, up to isotopy, by a finite sequence of turbulization and spiraling operations. Moreover, the taut versus Reeb‑less character of the resulting foliation is completely determined by (i) the spectral data of the spiraling matrix (A) and (ii) the choice of plug inserted during turbulization. This provides a precise, quantitative classification that extends beyond the earlier qualitative descriptions in the literature.

Finally, the authors discuss broader implications. The techniques suggest a pathway to study compact leaf phenomena in higher‑dimensional manifolds, to relate foliation dynamics to mapping‑class group actions, and to explore connections with contact topology via the Giroux correspondence. Open problems include characterizing the space of all possible spiraling matrices for a given manifold and understanding how these operations interact with measured lamination structures. The paper thus offers both a concrete toolkit for constructing and analyzing foliations with torus leaves and a conceptual framework that clarifies why certain torus‑leaf foliations are taut while others are merely Reeb‑less.