Abstract
By refining the volume estimate of Heintze and Karcher [11], we obtain a sharp pinching estimate for the genus of a surface in \(\mathbb S^{3}\), which involves an integral of the norm of its traceless second fundamental form. More specifically, we show that if g is the genus of a closed orientable surface \(\Sigma \) in a 3-dimensional orientable Riemannian manifold M whose sectional curvature is bounded below by 1, then \(4 \pi ^{2} g(\Sigma ) \le 2\left( 2 \pi ^{2}-|M|\right) +\int _{\Sigma } f(|{\mathop {A}\limits ^{\circ }}|)\), where \( {\mathop {A}\limits ^{\circ }} \) is the traceless second fundamental form and f is an explicit function. As a result, the space of closed orientable embedded minimal surfaces \(\Sigma \) with uniformly bounded \(\Vert A\Vert _{L^3(\Sigma )}\) is compact in the \(C^k\) topology for any \(k\ge 2\).
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
Suppose \(\Sigma \) is an umbilical surface in the sphere \(\mathbb S^{3}\), i.e. the difference of the principal curvatures \(k_2-k_1\equiv 0\), then \(\Sigma \) is a geodesic sphere. Our result shows that in a 3-manifold with sectional curvature bounded below by 1, if the two principal curvatures \(k_1\), \(k_2\) are close enough in an integral sense, then it is homeomorphic to a sphere. More generally, we can show that if M is a closed Riemannian 3-manifold whose sectional curvature is bounded below by 1, then a certain integral involving the gap between the two principal curvatures of a closed surface controls the genus of \(\Sigma \). See Theorem 2.
For example, as a corollary of our result, we can give a sharp bound of the genus of a surface in \(\mathbb S^{3}\):
Theorem 1
(Corollary 4) For any closed embedded surface \(\Sigma \) of genus g in \(\mathbb S^{3}\),
Here \({\mathop {A}\limits ^{\circ }}\) is the traceless second fundamental form of \(\Sigma \) and \(f(t)=\sqrt{2} t+\left( t^{2}-2\right) \tan ^{-1}\left( \frac{t}{\sqrt{2}}\right) \).
The equality holds if and only if \(\Sigma \) is a geodesic sphere or a Clifford torus.
As a result of this pinching estimate, by using a compactness result of Choi-Schoen [5], we can prove the following compactness result:
Corollary 1
(Corollary 11) Let M be a closed orientable three-dimensional Riemannian manifold whose sectional curvature is bounded from below by 1 and \(C\ge 0\). Then the space of closed orientable embedded minimal surfaces \(\Sigma \) with \( \Vert A\Vert _{L^3(\Sigma )} \le C\) is compact in the \(C^k\) topology for any \(k\ge 2\).
As another corollary, we can prove the following \(L^3\) gap theorem for minimal surfaces in \(\mathbb S^3\).
Corollary 2
(Corollary 6) A closed embedded minimal surface \(\Sigma \) in \(\mathbb S^{3}\) with \(\int _\Sigma |A|^3 dS<3\sqrt{2}\pi ^2\) is an equator.
There is a number of existing results of pinching estimates and rigidity theorems for surfaces in \(\mathbb R^3\) or \(\mathbb R^n\) in the literature. For example, Pogorelov [17, p. 493] proved that if \(\Sigma \) is a closed convex surface and the ratio of its principal curvatures are uniformly close to 1, then \(\Sigma \) is close to a round sphere. In [20], a similar result is proved by replacing the pointwise condition by some integral condition. In [7, 8], the authors proved that if the \(L^2 \) norm of the traceless second fundamental form is small enough, then the surface is \(W^{2, 2}\) or \(C^0\) close to a sphere. See also [18, Ch. 6] for a survey on this matter.
There are works on the rigidity theorems in other space forms as well. Most noticably, there are pioneering rigidity theorems for minimal submanifolds in a sphere, due to Simons [19], Lawson [14] and Chern-do Carmo-Kobayashi [4], in which rigidity for oriented closed minimal submanifolds in \(\mathbb S^{n+p}\) with \( |A|^2\le \frac{n}{2-\frac{1}{p}}\) is obtained. For pinching estimates, in [3, Theorem 1.3] (cf. also [16]), for a closed hypersurface in the standard space forms, a pinching result of the variance of the mean curvature H is proved, which involves a bound on the \(L^2\) norm of the traceless second fundamental form. On the other hand, this estimate is modelled on the umbilical hypersurfaces (i.e. geodesic spheres) and no topological or geometric information is known if the \(L^2\) norm of the traceless second fundamental form is large, whereas our result can be used to give some topological information about the surface given a bound on a certain integral of the norm of the traceless second fundamental form.
The organization of the remaining sections in this paper is as follows. In Sect. 2, we present a proof of an estimate for the genus of a surface embedded in a three-manifold with curvature bounded below by 1. Section 3 establishes the connection between this estimate and the norm of the traceless second fundamental form, establishing the main result. Furthermore, we explore some applications, including a compactness result regarding the space of minimal surfaces in \(\mathbb S^3\). In the Appendix, Sect. 4, we conduct a comparative analysis between our estimate and the classical estimate involving the area and the \(L^2\) norm of the traceless second fundamental form. By providing numerical examples, we demonstrate that our estimate cannot be derived from the classical estimate and, in fact, performs consistently better in all the computed cases.
2 An Estimate of the Genus
To prove the main result, we will need the following proposition, which results from a refinement of a volume estimate by Heintze and Karcher [11].
Proposition 1
Let \(\Sigma \) be a closed orientable embedded surface in a closed orientable three-dimensional Riemannian manifold M. Assume the sectional curvature of M is bounded from below by 1. Let \(k_1\le k_2\) be the principal curvatures of \(\Sigma \) (w.r.t. some unit normal), then
Here \(g(\Sigma )\) is the genus of \(\Sigma \) and |M| is the volume of M.
The following lemma is sketched in [6] and is well-known to experts. For reader’s convenience, we provide slightly more details here.
Lemma 1
With the assumption in Proposition 1, \(\Sigma \) separates M into two components.
Proof
This is equivalent to showing that \(H_{0}(M \backslash \Sigma ) \cong \mathbb Z \oplus \mathbb Z\).
As the Ricci curvature of M is positive, its first Betti number \(b_1(M)\) is zero. By deRham’s theorem, \(H^1(M, \mathbb R)=0\), and hence by universal coefficient theorem, \(\textrm{rank}(H_1(M))=0\) as a finitely generated Abelian group. By Poincare duality, the cohomology group \(H^{2}(M; \mathbb Z)\) has rank zero. Since M is compact, the cohomology groups are all finitely generated Abelian groups and consequently we have that every element of \(H^{2}(M ; \mathbb Z)\) sequence ( [10, p. 200])
Since \(\mathbb Z\) has no non-zero torsion element, we have \(\textrm{im}(\phi )=0\). Using the first isomorphism theorem and exactness we deduce that \(\textrm{ker}(g)=\textrm{im}(f) \cong \mathbb Z / \textrm{ker}(f) =\mathbb Z / \textrm{im}(\phi )\cong \mathbb Z\). Applying the first isomorphism theorem to g, we get the isomorphism \(H^{3}(M, \Sigma ; \mathbb Z) / \mathbb Z \cong \textrm{im}(g)\cong \mathbb Z\), from which it follows that \(H^{3}(M, \Sigma ; \mathbb Z) \cong \mathbb Z \oplus \mathbb Z\). By Poincare-Alexander-Lefschetz duality ( [1, Theorem 8.3]), we then get the isomorphism
\(\square \)
Proof of Proposition 1
By Lemma 1, \(\Sigma \) separates M into two components, \(\Omega _1\) and \(\Omega _2\).
It is clear that our inequality does not depend on the choice of the normal, so we can without loss of generality assume \(\{k_i\}_{i=1}^{2}\) (with \(k_1\le k_2\)) are the principal curvatures of \(\Sigma \) w.r.t. the unit outward normal of \(\Omega _1\). By [11, Corollary 3.3.1] (cf. also [15, (10”), (11”)]), the volume of \(\Omega _1\) is bounded from above:
where c(p) is the cut distance of \(\Sigma \) w.r.t. \(\Omega _1\). In [11], the integrand is estimated by using the AM-GM inequality. In order to obtain a more precise estimate, we keep the integrand unchanged. The cut distance at each point is bounded by above by the focal length, which is not longer than \(\cot ^{-1}(k_2)\) by Jacobi field comparison, as we have a lower bound on the sectional curvature. So direct integration gives
Similarly, with respect to the outward unit normal of \(\Omega _2\), the principal curvatures are \(-k_2\le -k_1\), so we have
By adding the two inequalities and using the identity \(\tan ^{-1} x=\frac{\pi }{2}-\cot ^{-1} x\),
By the Gauss-Bonnet theorem, the result follows. Note also that \(|M|\le |\mathbb S^3|=2\pi ^2\) by the volume comparison theorem. \(\square \)
3 Applications
Our aim is to draw some conclusions from Proposition 1.
Lemma 2
For any \(k_2\ge k_1\),
The equality holds if and only if \(k_1=\pm k_2\).
Proof
Let \(k_{2}-k_{1}=2 t \) and \( k_{2}+k_{1} =2s \). It is equivalent to show that \(f (t, s)=2\left( t^{2}-1\right) \tan ^{-1} t+\left( 1+ s^{2}-t^{2} \right) \left[ \tan ^{-1}(s +t)-\tan ^{-1}(s -t)\right] \ge 0\) for \(t \ge 0\). By symmetry we can also assume \(s \ge 0\).
This is an equality when \(s =0\), so it suffices to show that \(\frac{\partial f}{\partial s }\ge 0\). We compute
As \(\frac{\partial f}{\partial s }(0, s)=0\), it in turn suffices to show that \(\frac{\partial }{\partial t}\frac{\partial f}{\partial s }\ge 0\). By direct computation,
Therefore the inequality holds. From the above, the equality holds if and only if \(s =0\) or \(t=0\), which is equivalent to \(k_1=\pm k_2\). \(\square \)
Theorem 2
With the same assumption and notation in Proposition 1, we have
If the equality holds, then every point on \(\Sigma \) is either umbilical or of zero mean curvature. Moreover, the cut distance (from both sides of \(\Sigma \)) at each point of \(\Sigma \) must equal to the focal length.
Proof
Let \(k_1\le k_2\) be the principal curvatures, then we have \( |{\mathop {A}\limits ^{\circ }}|=\frac{k_2-k_1}{\sqrt{2}}\). By Proposition 1
So by Lemma 2,
We remark that \(\sqrt{2}|{\mathop {A}\limits ^{\circ }}|+\left( |{\mathop {A}\limits ^{\circ }}|^{2}-2\right) \tan ^{-1}\left( \frac{ | {\mathop {A}\limits ^{\circ }}|}{\sqrt{2}}\right) \) is strictly increasing in \(|{\mathop {A}\limits ^{\circ }}|\) by Lemma 3. In particular, the integrand on the RHS is always non-negative.
We now assume the equality holds. By [11, 3.4.8], all planes of M containing a tangent vector to a geodesic segment emanating orthogonally from \(\Sigma \) have sectional curvature 1, up to its focal length. (2.1) also shows that \(R^M_{1221}=1\) and so the sectional curvature along \(\Sigma \) must be constantly equal to 1. \(\square \)
In view of Theorem 2, in the following, we define the function
Lemma 3
The function f is a strictly increasing convex function on \([0,\infty )\), with \(f(0)=0\).
Proof
Direct computation gives \(f^{\prime }(t)=\frac{2 \sqrt{2} t^2}{2+t^2}+2 t \tan ^{-1}\left( \frac{t}{\sqrt{2}}\right) >0\) and \(f''(t)=\frac{2 \sqrt{2} t\left( 6+t^2\right) }{\left( 2+t^2\right) ^2} +2 \tan ^{-1} \left( \frac{t}{\sqrt{2}}\right) >0 \) for \(t>0\). \(\square \)
Remark 1
The inequality in Theorem 2 can also be written in a form such that the integrand is expressed as a power series of \(|{\mathop {A}\limits ^{\circ }}|\):
Observe that the Taylor series above is alternating and starts from the third order term. Indeed we have the following lemma.
Lemma 4
For \(t\ge 0\), \(f(t) \le \frac{2 \sqrt{2} t^{3}}{3}\) and the inequality is strict unless \(t=0\).
Proof
For \(t>0\), we have
\(\square \)
By Lemma 4, instead of the bound in Theorem 2, we can bound the genus by a weaker but simpler expression:
Corollary 3
With the same assumption and notation in Proposition 1, we have
If the equality holds, then \(\Sigma \) is umbilical.
Corollary 4
For any closed embedded surface \(\Sigma \) in \(\mathbb S^{3}\), we have
-
(1)
$$\begin{aligned} 4 \pi ^{2} g(\Sigma ) \le \int _{\Sigma }\left[ \sqrt{2}|{\mathop {A}\limits ^{\circ }}|+\left( |{\mathop {A}\limits ^{\circ }}|^{2}-2\right) \tan ^{-1}\left( \frac{|{\mathop {A}\limits ^{\circ }}|}{\sqrt{2}}\right) \right] dS. \end{aligned}$$
The equality holds if and only if \(\Sigma \) is a geodesic sphere or is a Clifford torus.
-
(2)
$$\begin{aligned} 2 \pi ^{2} g(\Sigma ) \le \frac{ \sqrt{2}}{3} \int _{\Sigma }|{\mathop {A}\limits ^{\circ }}|^{3} dS. \end{aligned}$$
The equality holds if and only if \(\Sigma \) is a geodesic sphere.
Proof
First note that a closed embedded surface in \(\mathbb S^3\) is orientable. By Theorem 2 and Corollary 3, we only have to consider the equality case. By Lemma 4, it suffices to prove (1).
By Lemma 2, at each point in \(\Sigma \), we have either \(k_1=k_2\) or \(k_1+k_2=0\). Suppose there is an umbilical point p such that the principal curvatures \(k_i\ne 0\). Then by continuity, \(k_1=k_2\) in a neighborhood of p. By the usual proof of the Schur’s lemma by using the Codazzi equation, we conclude that on this neighborhood, \(k_i=\text {constant}\ne 0\). By a standard connectedness argument we then conclude that \(\Sigma \) has constant principal curvatures and is a geodesic sphere. The remaining possibility is of course \(k_1+k_2\equiv 0\), i.e. \(\Sigma \) is a minimal surface.
Suppose now \(\Sigma \) is a minimal surface. From the proof of Proposition 1, the cut distance (from both sides of \(\Sigma \)) at each point must also equal to the focal length \(\cot ^{-1}k_2\). Assume there exists a point such that \(k_1=k_2=0\), then the focal point of \(\Sigma \) inside \(\Omega _1\) is also a cut point, and so its distance from \(\Sigma \) is \(\cot ^{-1}(0)=\frac{\pi }{2}\). By [13, Theorem A(2)], \(\Omega _1\) is the hemisphere and so \(\Sigma \) is the equator. Therefore, if \(\Sigma \) is not the equator, then \(k_2>0\) on \(\Sigma \). By [14, Proposition 1.5], \(\Sigma \) must be a torus. By Brendle’s proof of the Lawson’s conjecture [2], \(\Sigma \) is a Clifford torus.
Finally, if \(\Sigma \) is a Clifford torus or a geodesic sphere, then either \(k_1=k_2\) or \(k_1=-k_2\), i.e. the equality in Lemma 2 holds. Therefore (1) becomes an equality. \(\square \)
For any surface, we have \(|{\mathop {A}\limits ^{\circ }}|^{2}=|A|^{2}-2\,H^{2}\), where \(H=\frac{1}{2}(k_1+k_2)\) is the normalized mean curvature. In particular, \(|{\mathop {A}\limits ^{\circ }}|^{2}=|A|^{2}\) on a minimal surface.
Corollary 5
For any closed embedded minimal surface \(\Sigma \) in \(\mathbb S^{3}\), we have
The equality holds if and only if \(\Sigma \) is an equator or is a Clifford torus.
This result combined with Lemma 4 gives a gap theorem:
Corollary 6
A closed embedded minimal surface \(\Sigma \) in \(\mathbb S^{3}\) with \(\int _\Sigma |A|^3 dS<3\sqrt{2}\pi ^2\) is an equator.
Corollary 7
If \(\Sigma \) is a closed embedded surface in \(\mathbb S^3\) such that \(\displaystyle |{\mathop {A}\limits ^{\circ }}|<\sqrt{2} \beta \), where \(\beta \) is the unique solution to
for some \( g_0\in \mathbb N\), then the genus of \(\Sigma \) is less than \(g_0\). In particular, if \(\displaystyle \beta +(\beta ^2-1)\tan ^{-1}\beta = \frac{2\pi ^2}{|\Sigma |}\) and \(\displaystyle |{\mathop {A}\limits ^{\circ }}|<\sqrt{2} \beta \), then \(\Sigma \) is homeomorphic to a sphere.
Remark 2
The condition in Corollary 7 cannot be weakened to \(k_1-k_1\le 2\beta \). Indeed, for the Clifford torus \(\Sigma \), we have \(k_1=-1, k_2=1\) and \(|\Sigma |=2\pi ^2\). So in this case we have \(2\beta =2\), \(\chi (\Sigma )=0\) and \(k_2-k_1=2\).
Yang and Yau [21] proved the following eigenvalue estimate: Let \(\Sigma \) be a compact Riemannian surface of genus g, then for any metric on \(\Sigma \),
where \(\lambda _1\) is the first Laplacian eigenvalue of \(\Sigma \). It was observed by El Soufi and Ilias [9, Remark 2.3] that the same proof gives the improved bound
where \(\lfloor \cdot \rfloor \) is the floor function. See also [12] for a further improvement.
Using (3.3) and Theorem 2, we have
Corollary 8
With the same assumption and notation in Proposition 1, we have
where \(f(t)=\sqrt{2} t+\left( t^{2}-2\right) \tan ^{-1}\left( \frac{t}{\sqrt{2}}\right) \).
When M is the 3-sphere, we have the following
Corollary 9
For a closed embedded surface in \(\mathbb {S}^{3}\), we have
The equality holds if and only if \(\Sigma \) is a geodesic sphere or is a Clifford torus.
The following theorem is due to Choi and Schoen.
Theorem 3
[5, Theorem 2] Assume M is a closed 3-dimensional manifold with positive Ricci curvature. There exists a constant C depending only on M and an integer g such that if \(\Sigma \) is a closed embedded minimal surface of genus g in M, then
It is interesting to compare it with the following result:
Corollary 10
Assume M is a closed orientable 3-dimensional manifold whose sectional curvature is bounded below by 1 and \(\Sigma \) is a closed embedded orientable minimal surface of genus \(g\ge 1\) in M, then we have
where f is given by (3.2). In paricular, \(\displaystyle \max _{\Sigma }|A| \ge C>0\), where C depends on the genus g and the volume of M only.
Proof
For a minimal surface, \(|A|=|{\mathop {A}\limits ^{\circ }}|\). As remarked above, f is strictly increasing, and so by Theorem 2,
By the proof of the area estimate in [6, Proposition 4] and the improvement (3.4), we have \(\textrm{Area}(\Sigma )\le 8 \pi \left\lfloor \frac{g+3}{2}\right\rfloor \). Therefore
In particular, if \(M=\mathbb S^3\), then
\(\square \)
Corollary 11
Let M be a closed orientable three-dimensional Riemannian manifold whose sectional curvature of M is bounded from below by 1 and \(C\ge 0\). Then
-
(1)
The space of closed orientable embedded minimal surfaces \(\Sigma \) with \(\int _\Sigma f(|A|)\le C\) is compact in the \(C^k\) topology for any \(k\ge 2\).
-
(2)
The space of closed orientable embedded minimal surfaces \(\Sigma \) with \(\int _\Sigma |A|^3 \le C\) is compact in the \(C^k\) topology for any \(k\ge 2\).
Proof
By Theorem 2 and Corollary 3, both spaces are contained in the space of closed orientable embedded minimal surfaces with an upper bound of the genus and are closed in the \(C^k\) topology, \(k\ge 2\). Therefore they are both compact by [5, Theorem 1]. \(\square \)
Data Availability
Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.
Notes
The computation file is available upon request.
References
Bredon, G.E.: Topology and geometry, vol. 139. Springer Science & Business Media, Berlin (2013)
Brendle, S.: Embedded minimal tori in \(S^3\) and the Lawson conjecture. Acta Math. 211(2), 177–190 (2013)
Cheng, X., Zhou, D.: Rigidity for nearly umbilical hypersurfaces in space forms (2012). arXiv preprint arXiv:1208.1786
Chern, S.-S., do Carmo, M., Kobayashi, S.: Minimal submanifolds of a sphere with second fundamental form of constant length. In: Functional analysis and related fields, pages 59–75. Springer (1970)
Choi, H.I., Schoen, R.: The space of minimal embeddings of a surface into a three-dimensional manifold of positive Ricci curvature. Invent. Math. 81(3), 387–394 (1985)
Choi, H.I., Wang, A.N.: A first eigenvalue estimate for minimal hypersurfaces. J. Differ. Geomet. 18(3), 559–562 (1983)
De Lellis, C., Müller, S.: Optimal rigidity estimates for nearly umbilical surfaces. J. Differ. Geomet. 69(1), 075–110 (2005)
De Lellis, C., Müller, S.: A \(C^0\) estimate for nearly umbilical surfaces. Calc. Var. Partial Differ. Equ. 26(3), 283–296 (2006)
El Soufi, A., Ilias, S.: Le volume conforme et ses applications d’après Li et Yau. Séminaire de théorie spectrale et géométrie 2, 1–15 (1983)
Hatcher, A.: Algebraic topology. Cambridge University Press, Cambridge (2005)
Heintze, E., Karcher, H.: A general comparison theorem with applications to volume estimates for submanifolds. Ann. Sci. Ecole Norm. Sup 11(4), 451–470 (1978)
Karpukhin, M., Vinokurov, D.: An improved Yang-Yau inequality for the first Laplace eigenvalue (2021). arXiv preprint arXiv:2106.00627
Kasue, A.: Ricci curvature, geodesics and some geometric properties of Riemannian manifolds with boundary. J. Math. Soc. Japan 35(1), 117–131 (1983)
Lawson Jr, H.B.: Complete minimal surfaces in \(S^3\). Annal. Math. pages 335–374 (1970)
Montiel, S., Ros, A.: Compact hypersurfaces: the Alexandrov theorem for higher order mean curvatures. Differ. Geomet. Pitman Monogr. Surv. Pure Appl. Math. 52, 279–296 (1991)
Perez, D.: On nearly umbilical hypersurfaces. Thesis (2011)
Pogorelov, A.V.: Extrinsic geometry of convex surfaces. Mathematical Society, Amer (1973)
Reshetnyak, Y.G.: Stability theorems in geometry and analysis, vol. 304. Springer Science & Business Media, Berlin (2013)
Simons, J.: Minimal varieties in Riemannian manifolds. Ann. Math., pages 62–105 (1968)
Vodopyanov, S.K.: Estimates of the deviation of quasi-umbilical surfaces from a sphere. Sib. Math. J. 11(5), 971–987 (1970)
Yang, P., Yau, S.T.: Eigenvalues of the Laplacian of compact Riemann surfaces and minimal submanifolds. Ann. Scuola Norm. Sup. Pisa Cl. Sci. (4) 7(1), 55–63 (1980)
Funding
Open Access funding enabled and organized by CAUL and its Member Institutions
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix: Comparison With the Classical Estimate
Appendix: Comparison With the Classical Estimate
In this appendix, we present a comparison between our estimate and a more well-known estimate involving the \(L^2\) norm of \( {\mathop {A}\limits ^{\circ }} \).
It is well-known that the genus of \(\Sigma \) can also be bounded in terms of \(\int _\Sigma |{\mathop {A}\limits ^{\circ }}|^2 \) when M has sectional curvature at least 1. To see this, note that the Gaussian curvature \(K_{\Sigma }\) of \(\Sigma \) satisfies \(K_{\Sigma } \ge 1+k_1 k_2\) by the Gauss equation and the assumption that the sectional curvature of M is at least 1. As \(-k_1 k_2 \le \frac{1}{4}(k_1-k_2)^2=\frac{1}{2}|{\mathop {A}\limits ^{\circ }}|^2\), by integrating this inequality and using the Gauss-Bonnet formula, we have
It appears that this estimate is simpler compared to (3.1), and the \(L^2\) norm of \( {\mathop {A}\limits ^{\circ }} \) in the estimate seems to be preferable to the \(L^3\) norm of \( {\mathop {A}\limits ^{\circ }} \) as seen in Corollary 3.
However, it is important to note that this inequality does not imply our estimate (3.1). In fact, we have verified this through numerous numerical examples, which indicate that (3.1) is better than (4.1). This observation demonstrates that (i) our estimate is novel and (ii) it is not possible to derive (3.1) from (4.1). Note also that we do not impose any assumption on the area in Theorem 2.
To compare the two estimates, we have computed examples using “ellipsoids” in \(\mathbb S^3\). Upon examining the proof of (4.1), it is evident that it cannot be sharp for “small” surfaces that are nearly umbilical, which have genus 0. (However, it can still be sharp after taking the floor on the right-hand side.)
To aid our computation, we observe that the surface integral \(\int _{\Sigma }|{\mathop {A}\limits ^{\circ }}|^2 dS\) is conformally invariant. Specifically, the spherical metric can be expressed as \(g_{\mathbb S^3}=\frac{4}{(1+|x|^2)^2}{\overline{g}}\) under the stereographical projection, where \({\overline{g}}\) represents the flat metric on \(\mathbb R^3\). Therefore, we can calculate
where \(r=|x|\) and \({\overline{A}}\) denotes the second fundamental form with respect to \({\overline{g}}\). From this, we see that \(|{\mathop {{A}}\limits ^{\circ }}|= \frac{1+r^2}{2} |{\mathop {{\overline{A}}}\limits ^{\circ }}|\) and \(\int _{\Sigma }|{\mathop {A}\limits ^{\circ }}|^2 dS_{g_{\mathbb S^3}}= \int _{\Sigma }|{\mathop {{\overline{A}}}\limits ^{\circ }}|^2 dS_{{\overline{g}}} \).
We now consider the family of ellipsoids \(\{\frac{{x_1}^2}{a^2}+\frac{{x_2}^2}{b^2}+\frac{{x_3}^2}{c^2}=1\}\), where the genus \(g=0\). We compare the differences between the two estimates \(E_2=\lfloor 1-\frac{1}{4\pi }|\Sigma |_{g_{\mathbb S^3}}+\frac{1}{8\pi }\int _\Sigma |{\mathop {{\overline{A}}}\limits ^{\circ }}|^2 dS_{{\overline{g}}}\rfloor \) and \(E_f=\lfloor \frac{1}{4\pi ^2}\int _\Sigma f\left( |{\mathop { A}\limits ^{\circ }}|\right) dS_{ g_{\mathbb S^3}} \rfloor \), where \(\lfloor \cdot \rfloor \) is the floor function.
Using Mathematica,Footnote 1 we have computed the differences \(E_2-E_f\) for all a, b, c ranging between \({1, \cdots , 10}\). Our experiment demonstrates that for all a, b, c in this range, \(E_2-E_f\ge 0\), and the differences can be as large as 16 (occurs when \((a, b, c)=(1, 10, 10)\)). Consequently, our estimate (3.1) is better than the classical estimate (4.1).
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Kwong, KK. A Sharp Estimate for the Genus of Embedded Surfaces in the 3-Sphere. J Geom Anal 34, 251 (2024). https://doi.org/10.1007/s12220-024-01689-4
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s12220-024-01689-4