🔢 mathematics

Schottky versus super Schottky in genus 4

This paper resolves the final open case of the relationship between Schottky and super Schottky ideals by proving that for genus 4, the smallest power $d$ such that the $d$ -th power of the Schottky ideal is contained in the super Schottky ideal is $d=4$ , achieved through a rank analysis of the codifferential of the super period map involving degeneration to a vanishing theta-null and first-order deformation calculations.

Original authors: Ron Donagi, Simone Noja

Published 2026-09-16

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Ron Donagi, Simone Noja

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the vast landscape of modern mathematics, there is a branch dedicated to understanding the shapes of curved surfaces, specifically those that exist in complex, multi-dimensional spaces. Imagine a surface that can twist and turn in ways a simple sheet of paper cannot, possessing a specific number of "holes" or handles that define its fundamental character. Mathematicians call this number the "genus." For a long time, researchers have been trying to map out all possible shapes of these surfaces and understand how they relate to a different, more rigid kind of geometric object known as an abelian variety. This relationship is captured by a famous tool called the period map, which acts like a translator, converting the flexible geometry of a surface into the rigid coordinates of a grid.

However, in recent decades, physicists and mathematicians have discovered that these surfaces can be "super" surfaces. These are not just ordinary shapes; they carry an extra layer of hidden information, a kind of shadow dimension that behaves differently from the visible parts. This concept, central to theories about the fundamental structure of the universe, requires a new version of the period map to translate these super surfaces. The big question has been: how different are the rules for these super surfaces compared to the ordinary ones? Specifically, mathematicians wanted to know if the mathematical "rules" that define ordinary surfaces are enough to describe the super versions, or if the super versions require entirely new, deeper rules that ordinary ones do not see.

For many years, experts knew the answer for most sizes of these surfaces. They found that for surfaces with a certain number of holes, the ordinary rules were almost enough, but not quite. There was a specific power, a mathematical exponent, that determined how many times you had to apply the ordinary rules before they finally matched the super rules. For surfaces with five or more holes, and for those with six or more holes, this number was known to be equal to the number of holes itself. But there was one stubborn case left over: surfaces with exactly four holes. For this specific size, it was unclear whether the ordinary rules needed to be applied four times or just three times to match the super rules. This gap remained the last open door in a long series of investigations.

In this new work, a team of mathematicians has finally closed that door. They have proven that for surfaces with four holes, the ordinary rules must indeed be applied four times to fully capture the super rules. This means the answer is the same for this case as it is for all the larger cases studied before. The researchers did not just guess this; they performed a rigorous, step-by-step calculation to demonstrate it beyond any doubt. Their proof relies on a clever strategy of looking at a specific, slightly broken version of the surface where the math is easier to handle, and then carefully watching how the properties change as the surface is smoothed back out to its perfect form.

The journey began by focusing on a special type of four-holed surface that has a "vanishing theta-null." In plain language, this is a surface where a particular mathematical quantity, which usually measures a kind of hidden balance, drops to zero. On these special surfaces, the usual tools for translation break down and become infinite, making them difficult to study directly. To get around this, the researchers used a technique called regularization. They multiplied the breaking tools by a small factor that went to zero at the same rate, effectively canceling out the infinity and leaving behind a finite, manageable number. This allowed them to see what the translation looked like right at the moment the surface was in this special, broken state.

When they looked at this moment, they found that the translation tool produced a result that was not fully flexible. It could only move in four specific directions out of a possible six. In the language of the field, the "rank" of this tool was four, not six. If the rank had stayed at four even after the surface was smoothed out, it would have meant that the ordinary rules applied only three times were enough, and the answer would have been three. But the researchers knew that the surface they were studying was just a snapshot, and the real question was about the smooth surfaces nearby.

To find out what happened on the smooth surfaces, the team had to calculate how the translation tool changed as they moved away from the special, broken state. They chose a very specific way to smooth the surface, one that moved in a direction perpendicular to the special state, ensuring they were seeing the most direct change possible. They then calculated the first tiny step of this change. This involved a complex but precise accounting of how the hidden dimensions of the surface shifted. They tracked how the "Gaussian map," a tool that measures how the surface curves, changed during this shift. By combining this with the behavior of the translation tool, they were able to isolate the new information that appeared.

The calculation revealed a crucial detail: the tiny change was not zero. It produced a new, non-zero value that pushed the translation tool out of its four-directional limit. This meant that as soon as the surface was smoothed out, even by a tiny amount, the translation tool gained the ability to move in all six possible directions. The rank jumped from four to six. This jump is the smoking gun. It proves that on a general, smooth surface with four holes, the translation tool is fully flexible and requires the full power of the fourth application of the ordinary rules to be contained within the super rules.

The researchers confirmed this result by checking the numbers with extreme care. They verified that the specific type of surface they used for the calculation was not a fluke, but representative of the general case. Because the property of having a fully flexible translation tool is stable, finding it on one specific smooth surface proves it exists for almost all smooth surfaces of this type. The proof is complete and absolute. The smallest power of the ordinary rules needed to contain the super rules for a four-holed surface is four.

This finding resolves a question that had lingered for years, aligning the case of four holes with the behavior of all other known cases. It confirms that the complexity of these super surfaces grows in a predictable, uniform way as the number of holes increases. The work required deep expertise in the geometry of curves and the behavior of their hidden dimensions, but the result is a clean, definitive answer. The gap is closed, and the map of these mathematical landscapes is now complete for this important family of shapes.

Technical Summary: Schottky versus Super Schottky in Genus 4

Problem Statement
The paper addresses a specific question in the theory of super Riemann surfaces and their moduli spaces, originally posed by Witten and Felder, Kazhdan, and Polishchuk. The central problem is to determine the smallest integer $d$ such that the $d$ -th power of the classical Schottky ideal $I_g$ is contained within the super Schottky ideal $\tilde{I}_g$ .

The Schottky ideal $I_g$ defines the locus of Jacobians within the moduli space of principally polarized abelian varieties $A_g$ . The super Schottky ideal $\tilde{I}_g$ defines the schematic image of the super period map $\tilde{p}: M_{g}^{+,0} \to A_g$ , where $M_{g}^{+,0}$ is the open substack of even supermoduli with vanishing zero-modes ( $h^0(C, L) = 0$ ).

Previous results established that:

For odd genus $g \geq 5$ , $d = g$ .
For even genus $g \geq 6$ , $d = g$ (proven by Y. Shen).
For even genus $g \geq 4$ , it was known that $d \in \{g-1, g\}$ .

The case $g=4$ remained the only open instance where the minimal power $d$ had not been determined. The authors aim to resolve this gap by proving that for genus 4, $d = 4$ .

Methodology
The proof relies on analyzing the codifferential of the super period map, $d^*\tilde{p}$ . The condition $(I_g)^{g-1} \not\subset \tilde{I}_g$ is equivalent to showing that the codifferential, when restricted to the conormal space of the Jacobian locus, has maximal rank $2g-2$ (which is 6 for $g=4$ ).

The methodology proceeds through the following steps:

Sheaf-Theoretic Framework: The authors utilize the sheaves $\mathcal{O}(a, b, c)$ on $C \times C$ to control analytic kernels and obstruction classes. Specifically, they interpret the codifferential as multiplication by the Szegő kernel $s$ . For a conormal vector $q$ (a canonical quadric), the image $s_1(q)$ is a bivector in $\bigwedge^2 H^0(K_C^{3/2})$ . The rank of this bivector determines the value of $d$ .
Degeneration to Vanishing Theta-Nulls: Since the Szegő kernel is undefined when $h^0(C, L) > 0$ (vanishing theta-null), the authors work with a regularized version. They consider a family of curves degenerating to a curve $C_0$ with a vanishing theta-null ( $h^0(C_0, L_0) = 2$ ).
- At the central fiber $C_0$ , the limiting bivector $\omega_0$ is computed explicitly. For a non-hyperelliptic genus 4 curve, the space of canonical quadrics is 1-dimensional. The authors show that the rank of $\omega_0$ at the theta-null is 4, not the generic 6.
First-Order Variation and Rank Jump: To prove the generic rank is 6, the authors must show that the rank jumps from 4 to 6 under a transverse deformation.
- They select a specific cyclic trigonal curve with a vanishing theta-null and a transverse deformation where the cubic equation is fixed.
- They compute the first variation of the regularized conormal bivector $\dot{\omega}_0$ . This involves differentiating a global Gaussian-map identity.
- The variation is projected onto the normal quotient $\bigwedge^2 R^\vee$ , where $R$ is the 2-dimensional null space of the limiting rank-4 matrix.
- Using Serre duality and explicit calculations of the Kodaira–Spencer class and the obstruction map $B_\xi$ , they determine the polar coefficient of the Szegő kernel.
- They demonstrate that the first normal symbol $\nu$ is non-zero.
Rank Jump Lemma: A linear algebra argument (Lemma 7.1) shows that if the first-order variation of the bivector is non-zero on the null space of the limiting matrix, the rank of the matrix jumps to the maximal value (6) for nearby fibers.

Key Contributions and Results

Resolution of the Genus 4 Case: The paper proves Theorem 1.1: In genus 4, the smallest $d$ such that $(I_4)^d \subset \tilde{I}_4$ is $d=4$ . This confirms that the $(g-1)$ -st power of the Schottky ideal is not contained in the super Schottky ideal for $g=4$ .
Explicit Computation of the Codifferential: The authors provide an explicit calculation of the codifferential of the super period map at a vanishing theta-null, showing it has rank 4.
Non-vanishing of the First Variation: The core technical contribution is the proof that the first variation of the regularized conormal bivector is non-zero on the null space of the limiting matrix. This is achieved through a detailed analysis of cyclic trigonal curves, the Gaussian map, and the interplay between the Szegő kernel and the obstruction to lifting sections of the spin bundle.
Independence from Higher Genus Results: The argument for $g=4$ is self-contained and does not rely on the results for $g \geq 6$ established by Shen.

Significance
The paper closes the final gap in the classification of the containment of powers of the Schottky ideal within the super Schottky ideal for all genera $g \geq 4$ . By establishing that $d=g$ for genus 4, it confirms a uniform behavior across all even and odd genera (where $d=g$ for $g \geq 5$ and $g \geq 6$ , and now $g=4$ ).

The work clarifies the geometry of the super period map near the theta-null divisor, demonstrating that while the map degenerates (rank drops) at the divisor, the generic behavior is determined by a non-vanishing first-order term. This result is significant for the understanding of the structure of supermoduli spaces and the relationship between super Riemann surfaces and their bosonic reductions. The authors note that their proof relies on standard exact sequences, degeneration techniques, and explicit kernel calculations, providing a concrete verification of the theoretical expectations set by previous work in the field.

Technical Summary: Schottky versus Super Schottky in Genus 4

More like this