Schur Quartic: Canonical & Surface Forms

• Schur Quartic is defined in dual contexts: as the Cayley–Schur canonical form for binary quartics and as a unique quartic K3 surface in projective space.
• In classical invariant theory, binary quartics reduce to the form x⁴ + 6λx²y² + y⁴, which admits exactly six mixed-power decompositions.
• The Schur quartic surface is distinguished by hosting 64 lines, including a (24₄,32₃)-configuration, and is pivotal in studies of logarithmic geometry and invariant structures.

The term Schur quartic has two principal meanings in the current literature. In classical invariant theory, it denotes the Cayley–Schur canonical form for a general binary quartic under $GL_2(\mathbb{C})$-change of variables,

$X^4+6\lambda X^2Y^2+Y^4,$

with $\lambda$ determined by invariants (Reznick, 2012). In algebraic geometry, Schur’s quartic denotes the smooth quartic surface

$$X=\{x_0^4-x_0x_1^3-x_2^4+x_2x_3^3=0\}\subset \mathbb{P}^3,$$

a K3 surface which is recalled as the unique smooth complex quartic with $64$ lines, up to projective equivalence (Naskręcki et al., 8 Jul 2026). These usages are historically related by Schur’s name but mathematically distinct.

1. Classical binary-quartic meaning

For binary quartics, the Schur terminology belongs to the nineteenth-century theory of canonical forms. A general binary quartic can be reduced, after a linear change of variables, to

$x^4+6\lambda x^2y^2+y^4,$

which is described as the classical Cayley–Schur canonical form for a general binary quartic under $GL_2(\mathbb{C})$-change of variables (Reznick, 2012). In that context, $\lambda$ is determined by invariants, and the possible values of $\lambda$ for such normal forms are related by Möbius transformations; classically, one can arrange

$\lambda\in \{\pm 1,\pm \sqrt{3},\pm 1/\sqrt{3}\}$

after suitable linear changes of variables (Reznick, 2012).

Here “general” means outside a proper algebraic subset of $X^4+6\lambda X^2Y^2+Y^4,$0. For binary quartics this excludes degenerate quartics with multiple roots and special configurations where certain discriminants or invariants vanish and the algorithm degenerates (Reznick, 2012). In this sense, the Schur quartic is not a single quartic polynomial but a normal form representing the generic $X^4+6\lambda X^2Y^2+Y^4,$1-orbit.

This invariant-theoretic meaning is central because it separates two layers of structure. First, one performs $X^4+6\lambda X^2Y^2+Y^4,$2-reduction to the Schur form $X^4+6\lambda X^2Y^2+Y^4,$3. Second, one may seek a more structured decomposition in fixed coordinates. That second step is the subject of the mixed-power canonical forms developed by Reznick (Reznick, 2012).

2. Reznick’s refinement: $X^4+6\lambda X^2Y^2+Y^4,$4

A general binary quartic $X^4+6\lambda X^2Y^2+Y^4,$5 can be written as

$X^4+6\lambda X^2Y^2+Y^4,$6

where $X^4+6\lambda X^2Y^2+Y^4,$7 is a quadratic form and $X^4+6\lambda X^2Y^2+Y^4,$8 is a linear form (Reznick, 2012). In normalized coordinates,

$X^4+6\lambda X^2Y^2+Y^4,$9

the decomposition

$\lambda$0

exists in six different ways (Reznick, 2012).

This places the Schur quartic inside a broader family of canonical or normal forms for quartic forms. In Reznick’s language, the binary quartic case is a specialization of a general mixed-power scheme in which a binary form of degree $\lambda$1 is represented as a sum of $\lambda$2-th powers of linear forms together with powers of higher-degree forms, subject to the constant-count condition

$\lambda$3

and divisibility conditions $\lambda$4 (Reznick, 2012). For $\lambda$5, taking $\lambda$6, $\lambda$7, and $\lambda$8 yields precisely

$\lambda$9

The representation is non-unique but finite in number for general quartics. For fixed normalization and fixed linear data there are finitely many decompositions; in the normalized Cayley–Schur form there are exactly six (Reznick, 2012). If arbitrary linear changes of variables are allowed, the six decompositions transform accordingly, and equivalence also includes $$X=\{x_0^4-x_0x_1^3-x_2^4+x_2x_3^3=0\}\subset \mathbb{P}^3,$$0 with $$X=\{x_0^4-x_0x_1^3-x_2^4+x_2x_3^3=0\}\subset \mathbb{P}^3,$$1 and the sign change $$X=\{x_0^4-x_0x_1^3-x_2^4+x_2x_3^3=0\}\subset \mathbb{P}^3,$$2 (Reznick, 2012).

Proofs are organized by the Jacobian criterion and by apolarity. The standard technique is to show that at some parameter point the partial derivatives of the coefficient map span $$X=\{x_0^4-x_0x_1^3-x_2^4+x_2x_3^3=0\}\subset \mathbb{P}^3,$$3, or equivalently that there is no nonzero form apolar to all those partial derivatives (Reznick, 2012). This connects the Schur quartic to classical invariant theory, Waring-type decomposition, and apolar methods.

3. Extension to ternary quartics and canonical-form theory

The same framework extends beyond binary quartics. A general ternary quartic $$X=\{x_0^4-x_0x_1^3-x_2^4+x_2x_3^3=0\}\subset \mathbb{P}^3,$$4 can be written as

$$X=\{x_0^4-x_0x_1^3-x_2^4+x_2x_3^3=0\}\subset \mathbb{P}^3,$$5

where $$X=\{x_0^4-x_0x_1^3-x_2^4+x_2x_3^3=0\}\subset \mathbb{P}^3,$$6 and $$X=\{x_0^4-x_0x_1^3-x_2^4+x_2x_3^3=0\}\subset \mathbb{P}^3,$$7 (Reznick, 2012). This is a mixed-degree canonical form: one quadratic squared and three linear forms to the fourth power.

Reznick emphasizes that this differs from the classical result that a general ternary quartic is the sum of three squares of quadratics (Reznick, 2012). The mixed-power decomposition uses fewer quadratics at the expense of adding fourth powers of linear forms. As in the binary case, “general” again means outside a proper Zariski-closed subset of $$X=\{x_0^4-x_0x_1^3-x_2^4+x_2x_3^3=0\}\subset \mathbb{P}^3,$$8 (Reznick, 2012).

The ternary proof is again by the Jacobian criterion. One parametrizes a general quadratic with six parameters and three linear forms with nine parameters, giving $$X=\{x_0^4-x_0x_1^3-x_2^4+x_2x_3^3=0\}\subset \mathbb{P}^3,$$9, and then checks that the $64$0 resulting quartic partial derivatives span $64$1 at a carefully chosen point (Reznick, 2012). This is a Schur-type normal form only in a broad sense: not by monomial elimination, but by providing a canonical mixed-power decomposition of the generic orbit.

This broader theory clarifies the place of the binary Schur quartic. The classical Schur form is a normal form modulo $64$2, whereas Reznick’s decomposition gives a structured representation in fixed coordinates. A plausible implication is that “Schur quartic” names the invariant-theoretic reduction, while the later mixed-power theory explains how that reduced quartic can be decomposed into simpler algebraic constituents.

4. Schur’s quartic surface in $64$3

In algebraic geometry, Schur’s quartic is the surface

$64$4

(Naskręcki et al., 8 Jul 2026). As a smooth quartic hypersurface in $64$5, $64$6 is a K3 surface: $64$7 (Naskręcki et al., 8 Jul 2026).

Smoothness is verified from the partial derivatives

$64$8

Over characteristic $64$9, these vanish simultaneously only if $x^4+6\lambda x^2y^2+y^4,$0 and $x^4+6\lambda x^2y^2+y^4,$1, impossible in projective space, so $x^4+6\lambda x^2y^2+y^4,$2 is smooth in characteristic zero and over $x^4+6\lambda x^2y^2+y^4,$3 (Naskręcki et al., 8 Jul 2026).

The surface is distinguished by its lines. It contains exactly 64 lines, and it is recalled as the unique smooth complex quartic with 64 lines, up to projective equivalence, by the classification of Degtyarev–Itenberg–Sertöz (Naskręcki et al., 8 Jul 2026). Among those 64 lines, 48 are lines of the second kind in the sense of Rams–Schütt; their weak combinatorics are

$x^4+6\lambda x^2y^2+y^4,$4

and the associated logarithmic Chern slope is

$x^4+6\lambda x^2y^2+y^4,$5

(Naskręcki et al., 8 Jul 2026).

This surface meaning of Schur quartic is therefore rigid and singular in the moduli-theoretic sense recorded by the paper: one specific quartic K3 surface, unique up to projective equivalence among smooth complex quartics by the property of carrying $x^4+6\lambda x^2y^2+y^4,$6 lines (Naskręcki et al., 8 Jul 2026).

5. The $x^4+6\lambda x^2y^2+y^4,$7-configuration on Schur’s quartic

A recent development isolates a special half of the $x^4+6\lambda x^2y^2+y^4,$8 second-kind lines on Schur’s quartic (Naskręcki et al., 8 Jul 2026). Over

$x^4+6\lambda x^2y^2+y^4,$9

the paper lists 24 explicit lines in graph form

$GL_2(\mathbb{C})$0

with all parameters in $GL_2(\mathbb{C})$1 (Naskręcki et al., 8 Jul 2026). Substituting this parametrization into the quartic equation yields five homogeneous coefficients in $GL_2(\mathbb{C})$2; for the 24 chosen lines all $GL_2(\mathbb{C})$3 equalities vanish (Naskręcki et al., 8 Jul 2026).

Their incidence structure is exceptionally regular. For pairs of lines, one tests intersection by the determinant

$GL_2(\mathbb{C})$4

Among the $GL_2(\mathbb{C})$5 determinants, exactly 96 vanish and 180 are nonzero (Naskręcki et al., 8 Jul 2026). Appendix B lists 32 distinct points in $GL_2(\mathbb{C})$6, each incident with precisely three of the 24 lines. Each line appears in exactly four triples. Combinatorially, this is a

$GL_2(\mathbb{C})$7

with 24 lines, each line containing 4 triple points, and 32 triple points, each triple point lying on exactly 3 lines (Naskręcki et al., 8 Jul 2026). There are no double points and no quadruple points.

The reduced divisor

$GL_2(\mathbb{C})$8

is connected, and all intersections are ordinary triple points (Naskręcki et al., 8 Jul 2026). If $GL_2(\mathbb{C})$9 denotes the hyperplane class on $\lambda$0, then $\lambda$1, every line has self-intersection $\lambda$2, and the intersection calculations give

$\lambda$3

(Naskręcki et al., 8 Jul 2026). Thus the union of 24 lines is linearly equivalent to six times a plane section of $\lambda$4.

The configuration is one half of the $\lambda$5 lines of the second kind. The projective automorphism

$\lambda$6

preserves $\lambda$7, sends the 24-line divisor $\lambda$8 to another 24-line divisor $\lambda$9, and interchanges the two halves of the second-kind configuration (Naskręcki et al., 8 Jul 2026). The 64 triple points of the 48-line arrangement split into two disjoint subsets of 32 triple points each, one for $\lambda$0 and one for $\lambda$1; the 144 double points are exactly the intersections between a line in $\lambda$2 and a line in $\lambda$3 (Naskręcki et al., 8 Jul 2026).

6. Logarithmic geometry and adjacent quartic theories

Blowing up the $\lambda$4 triple points of the divisor $\lambda$5 produces a pair $\lambda$6 with reduced boundary

$\lambda$7

where $\lambda$8 is the strict transform and $\lambda$9 are the exceptional curves (Naskręcki et al., 8 Jul 2026). Since $\lambda\in \{\pm 1,\pm \sqrt{3},\pm 1/\sqrt{3}\}$0 is K3,

$\lambda\in \{\pm 1,\pm \sqrt{3},\pm 1/\sqrt{3}\}$1

The logarithmic Chern numbers are

$\lambda\in \{\pm 1,\pm \sqrt{3},\pm 1/\sqrt{3}\}$2

(Naskręcki et al., 8 Jul 2026). This gives a negative answer to the K3-surface specialization of the proposed $\lambda\in \{\pm 1,\pm \sqrt{3},\pm 1/\sqrt{3}\}$3 bound for transversal arrangements of rational curves (Naskręcki et al., 8 Jul 2026).

The paper also states that $\lambda\in \{\pm 1,\pm \sqrt{3},\pm 1/\sqrt{3}\}$4 is semistable in Sakai’s sense, $\lambda\in \{\pm 1,\pm \sqrt{3},\pm 1/\sqrt{3}\}$5 is big, and $\lambda\in \{\pm 1,\pm \sqrt{3},\pm 1/\sqrt{3}\}$6 is of log general type, so Sakai’s logarithmic Bogomolov–Miyaoka–Yau inequality applies (Naskręcki et al., 8 Jul 2026). In this formulation, Schur’s quartic becomes a testbed for the geography of log surfaces rather than only a line-rich quartic surface.

Adjacent literature uses Schur’s name or Schur-type constructions in different quartic settings. The paper on four-variable quartic inequalities studies the cones $\lambda\in \{\pm 1,\pm \sqrt{3},\pm 1/\sqrt{3}\}$7 and $\lambda\in \{\pm 1,\pm \sqrt{3},\pm 1/\sqrt{3}\}$8 and identifies extremal symmetric quartics such as $\lambda\in \{\pm 1,\pm \sqrt{3},\pm 1/\sqrt{3}\}$9, $X^4+6\lambda X^2Y^2+Y^4,$00, $X^4+6\lambda X^2Y^2+Y^4,$01, and the families $X^4+6\lambda X^2Y^2+Y^4,$02, $X^4+6\lambda X^2Y^2+Y^4,$03, and $X^4+6\lambda X^2Y^2+Y^4,$04 (Ando, 2022). A different invariant-theoretic direction analyzes $X^4+6\lambda X^2Y^2+Y^4,$05-invariant quartics and proves that the cones of invariant sums of squares and nonnegative forms are equal if and only if the number of variables is at most $X^4+6\lambda X^2Y^2+Y^4,$06 or odd (Debus et al., 2024). This suggests that the phrase Schur quartic is historically layered: in strict usage it denotes either the Cayley–Schur binary normal form or Schur’s quartic surface, while related quartic theories continue to organize themselves around Schur-type symmetry, invariant structure, and extremality.