The first minimal-hypersurface volume above an equator

← All digests

Family 349 · Differential geometry
Importance 5/5 · Provisional editorial rating
Sources consulted 2026-10-26

1. The question

An equator is the simplest closed minimal hypersurface in a round sphere. It has the least possible volume. What is the next possible volume level if the image is not an equator?

There is a natural family of candidates: products of two smaller round spheres sitting inside the unit sphere. Their radii can be chosen so that the positive curvatures from one factor balance the negative curvatures from the other. These are the minimal Clifford products.

The Solomon–Yau least-volume conjecture predicts that every non-equatorial closed minimal hypersurface has at least the volume of the smallest such product. Family 349 claims this sharp bound in every hypersurface dimension at least two, including smooth immersions with overlapping sheets.

The question is not whether there is some positive volume gap above the equator; earlier work already established a dimension-dependent gap. It is whether the gap ends at precisely the Clifford threshold, without extra topology, symmetry, or curvature hypotheses.

2. Background and essential definitions

Minimal hypersurfaces in a sphere

A hypersurface has one dimension less than its ambient manifold. A smooth immersion $F:M^m\to S^{m+1}$ is locally an embedded hypersurface but may identify different domain points globally. The induced metric $F^*g_{\mathrm{round}}$ measures lengths and volumes on the domain.

The hypersurface is minimal if its mean curvature vanishes. At a point, mean curvature is the average of the principal curvatures, which describe bending in independent tangent directions. Minimality means zero first variation of volume under small deformations; it does not mean that every deformation increases volume.

The image is totally geodesic when its second fundamental form vanishes. For a connected closed hypersurface in the round sphere, the relevant model is an equator, a great $m$-sphere. The theorem excludes a totally geodesic image, even if the domain parametrization has multiplicity.

The Clifford comparison family

For $1\leq k<m$, define

$$C_{k,m-k} =S^k\!\left(\sqrt{\frac{k}{m}}\right) \times S^{m-k}\!\left(\sqrt{\frac{m-k}{m}}\right) \subset S^{m+1}.$$

The squared radii add to one, so the product lies in the unit sphere. The chosen radii make its weighted sum of principal curvatures zero, hence it is minimal.

Let $\sigma_j$ be the volume of the unit $j$-sphere. The product volume is

$$\operatorname{Vol}(C_{k,m-k}) =\sigma_k\sigma_{m-k} \left(\frac{k}{m}\right)^{k/2} \left(\frac{m-k}{m}\right)^{(m-k)/2}.$$

Write $a_m$ for the minimum over the finitely many choices of $k$. This definition is enough to make the claimed threshold explicit without needing to classify which index minimizes it in every dimension.

For $m=2$, the product is the Clifford torus, with both circle radii $1/\sqrt2$. Its area is $2\pi^2$, while the equatorial sphere has area $4\pi$. Thus the predicted first non-equatorial level is larger than the equator by a factor $\pi/2$.

Why domain volume matters

For an immersion, the quantity being bounded is

$$\operatorname{Vol}(M,F^*g_{\mathrm{round}}).$$

It counts every sheet. If two domain points map to the same image point, their contributions are not merged into one. Measuring only the area of the image as a set would be a different problem.

3. Development history and previous partial results

Simons’s identities for minimal submanifolds supplied basic curvature rigidity and explained the special role of hypersurfaces with parallel second fundamental form, including Clifford products. Cheng–Li–Yau later obtained a dimension-dependent volume gap above the equator using heat-kernel comparison. A nonzero gap, however, does not identify the sharp next level. Simons, Cheng–Li–Yau.

In dimension two, Marques–Neves’s Willmore theorem yields the sharp area bound and Clifford-torus rigidity. Their proof uses a canonical family of surfaces and min–max theory: topology prevents the whole family from being deformed below a certain area threshold. This variational strategy is an important model for the new manuscript. Marques–Neves.

Higher-dimensional partial results retained geometric assumptions. Cheng–Wei–Zeng treated rotational cases. Viana obtained the sharp Clifford bound when both connected complementary regions are antipodally invariant. Related lower bounds for minimal cones, including work of Bernstein–Wang, connect volume questions to the density of conical singularities under additional topological conditions. Rotational results, Viana.

Nguyen and Ge–Li provided a useful reduction for immersions: an image point with $q$ preimages forces domain volume at least $q\sigma_m$. Since the Clifford threshold is less than $2\sigma_m$, any hypothetical counterexample would already be embedded. Ge–Li’s more recent general embedded bounds still leave the sharp threshold as the target. Nguyen, Ge–Li.

The new paper also proves an average-curvature estimate and uses an index characterization of Clifford products. It explicitly credits Nifa’s recent preprint with those conclusions through a stronger Schrödinger quadratic-form inequality. Those intermediate statements are therefore not presented here as originating solely in the new volume argument. Nifa’s preprint.

4. What the new manuscript claims

For every integer $m\geq2$, every closed connected smooth manifold $M^m$, and every smooth minimal immersion into the unit round sphere,

$$F:M^m\longrightarrow S^{m+1},$$

the manuscript asserts

$$F(M)\text{ not totally geodesic} \quad\Longrightarrow\quad \operatorname{Vol}(M,F^*g_{\mathrm{round}})\geq a_m.$$

Closed means compact without boundary. Embeddedness, orientability of the initial domain, symmetry, and prescribed topology are not additional hypotheses. Covering multiplicities are included in the volume.

The threshold is sharp because a Clifford product realizing the finite minimum has volume $a_m$. The stated main theorem is the inequality and its attainment by these examples; it should not automatically be expanded into a full classification of every equality case in arbitrary dimension.

For the embedded part of the argument, the manuscript also obtains

$$\int_\Sigma |A|^2\,d\mu\geq m\operatorname{Vol}(\Sigma),$$

where $A$ is the second fundamental form and $\Sigma$ is non-equatorial. Combined with spectral estimates, it identifies an embedded minimal hypersurface of Morse index $m+3$ as a Clifford product. The Morse index counts independent normal deformation directions that decrease area to second order.

These intermediate conclusions are used to show that a hypothetical smaller-volume counterexample has an extra negative direction beyond those already controlled by the variational family.

5. Why the advance matters

Importance: 5/5. If correct, the theorem identifies the sharp first nontrivial volume level for minimal hypersurfaces in every round sphere, resolving a longstanding geometric extremal question.

The scope is broad: the statement does not assume a genus, a symmetry group, or embeddedness. The immersion reduction explains why this breadth is possible while keeping multiplicity visible.

The claimed all-dimensional method is also important. In high dimensions, minimizing sequences and min–max constructions can develop singularities. A successful sharp-volume argument must control those limits rather than rely only on the smooth geometry of its starting hypersurfaces.

The family combines a geometric comparison, a topological detection mechanism, and an induction preventing singular subthreshold limits. The score reflects that claimed reach, not independent proof verification.

6. The main idea

Reduce a counterexample to an embedding

Form the Euclidean cone over the immersed hypersurface. Its density at the vertex is the domain volume divided by $\sigma_m$. If an image point has $q$ preimages, the corresponding nonzero point of the cone has density at least $q$, including tangent or coincident sheets.

Monotonicity then gives volume at least $q\sigma_m$. The manuscript shows $a_m<2\sigma_m$, so a counterexample below $a_m$ cannot have any image point with two preimages. It must be embedded.

This reduces the hard variational argument to embedded hypersurfaces without weakening the final immersion statement.

Use curvature to find an additional decreasing direction

A two-point comparison looks at the normal direction at one point relative to all other points. The paper treats both signs of principal curvature and derives the integrated estimate for $|A|^2$.

A conformal and spectral argument then controls the index. If the index is exactly $m+3$, the hypersurface is a Clifford product. A counterexample with volume below every Clifford product must consequently have index greater than $m+3$.

That extra negative direction is the resource used to lower the maximum area in a carefully constructed family. The index statement alone does not produce a smaller minimal hypersurface; the topological family is needed for that step.

Build a family whose topology cannot disappear

Take the signed distance to the Euclidean cone over the hypersurface and view its graph in a space with an additional time coordinate. Lorentz transformations preserve the difference between squared spatial length and squared time. After transforming and intersecting with the unit spatial sphere, level sets provide a family of hypersurfaces.

At closest-point contacts, minimality fixes the trace of a nonnegative Jacobian matrix. A determinant inequality controls the areas of the resulting levels, even when closest points are not unique.

The parameters lead to a projective family of dimension $m+3$. Along its nontrivial parameter loop, filling regions change to their complements. A degree-one cohomology class therefore has a nonzero $(m+3)$-rd power. This detects topology of the family without imposing a separate topological assumption on the original hypersurface.

The original hypersurface is the unique area maximum, with a uniform gap away from the central parameter. Its extra index direction allows that maximum to be lowered strictly while retaining the detecting topology and preventing concentration of area into tiny regions.

Keep the min–max output above the equator

The detecting condition forces the family to contain a boundary enclosing half the sphere’s volume with zero centroid. A hemisphere has half-volume but does not have zero centroid, so the spherical isoperimetric equality case is excluded. The paper obtains a uniform threshold

$$\gamma_m>\sigma_m$$

for these boundaries. This keeps the variational construction from collapsing back to an equator.

Min–max theory finds a stationary limiting hypersurface whose area lies between that lower threshold and the family’s maximum. Such an output is initially a varifold, a generalized surface allowing multiplicity and singularities. Smoothness is not automatic in all dimensions.

Rule out singular limits below the Clifford threshold

The proof proceeds by induction on hypersurface dimension. At a nonzero point of a Euclidean cone, a tangent cone splits off the radial line. The remaining geometry has lower cone dimension, allowing the earlier-dimensional volume statements to constrain it.

The manuscript uses this radial splitting to establish regularity and multiplicity one for the specific subthreshold limits arising in the argument. The induction does not apply the still-unproved dimension-$m$ theorem to a dimension-$m$ link. The base case $m=2$ requires no lower-dimensional volume hypothesis.

If counterexamples existed, compactness would then produce one of least area $H$, with

$$\sigma_m<\gamma_m\leq H<a_m.$$

The extra index direction lowers its detecting family to a maximum strictly below $H$. Min–max, followed by the same regularity argument, produces a smooth connected non-equatorial minimal hypersurface with area between $\gamma_m$ and $H$. That is a smaller counterexample, contradicting the definition of $H$.

The claimed proof therefore needs both ingredients: topology keeps the new surface above the equator, while cone induction makes it smooth enough to contradict leastness.

7. Scope, limitations, and open questions

The ambient space is a unit round sphere, and the submanifold is a closed minimal hypersurface. The theorem does not cover arbitrary ambient metrics, higher-codimension minimal submanifolds, or hypersurfaces with boundary.

It supplies a sharp lower volume threshold, not a complete volume spectrum or a classification of all minimal hypersurfaces. The existence of a gap above the equator does not imply that subsequent volume levels are discrete.

For immersions, replacing domain volume by the measure of the image would remove multiplicity and change the statement. For equality, the paper’s main theorem provides the sharp examples but does not state a universal equality-classification theorem in the form quoted here.

Finally, the manuscript credits prior average-curvature and index progress while supplying its own variational family and all-dimensional compactness argument. This digest reports that organization without verifying the estimates, the topological detection, or the cone induction.

8. Sources and further reading

The linked primary works explain the earlier volume gap, the two-dimensional sharp result, special higher-dimensional cases, and the curvature input. Optional background appears under Clifford torus, minimal surface.

Source manuscripts

• M1. OpenAI, The Solomon–Yau least-volume theorem (September 23, 2026).

Sources consulted on 26 October 2026 at repository revision 1581d4c. Download the manuscript citations.