Generalized Fibonacci numbers and automorphisms of K3 surfaces with Picard number 2

Kwangwoo Lee

REVIEW 3 major objections 5 minor 10 references

Generalized Fibonacci numbers and automorphisms of K3 surfaces with Picard number 2

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper claims that for K3 surfaces with Picard lattice $L_m(a)$, the automorphism group is generated by the automorphism attached to the minimal $n$ with $m \mid a_n$, and that the same dictionary proves generalized Fibonacci numbers…

desk verdict Theorem 1.1 is likely true, but the proof leans on a misapplied proposition and Section 5 has an exponent error that needs fixing. read the letter →

arxiv 2411.13038 v1 pith:ESAXWDAZ submitted 2024-11-20 math.AG

Kwangwoo Lee This is my paper · ORCID

classification math.AG MSC 11B3914C0514J2814J50

keywords GeneralizedFibonaccinumberK3surfaceAutomorphismSalempolynomialResultantPicard2Symplectic

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper establishes a two-way bridge between generalized Fibonacci numbers and the automorphism groups of certain K3 surfaces with Picard number 2. For a surface whose Picard lattice has intersection matrix $L_m(a)=m\begin{bmatrix}2&a\\a&-2\end{bmatrix}$ with $m\ge 2$, it claims that whenever $m$ divides the $n$-th generalized Fibonacci number $a_n$, the isometry $(AB)^n$ on the Néron-Severi lattice lifts to an automorphism: symplectic for $n$ even, anti-symplectic for $n$ odd. When $n$ is the smallest index with $m\mid a_n$ and $5\nmid n$, that automorphism generates the whole automorphism group, giving its full characteristic polynomial on $H^2$. The same geometric input proves a converse number-theoretic test, namely that an integer $n$ is a generalized Fibonacci number exactly when $(a^2+4)n^2\pm 4$ is a perfect square with the sign fixed by parity, and the divisibility rule $a_k\mid a_q$ if and only if $k\mid q$.

What carries the argument

The machinery is a pair of matrices $A=\begin{bmatrix}1&0\\a&-1\end{bmatrix}$ and $B=\begin{bmatrix}1&a\\0&-1\end{bmatrix}$, which generate the isometry group $O(L_m(a))\cong \mathbb{Z}/2*\mathbb{Z}/2$. Their product has the Fibonacci form $(AB)^n=\begin{bmatrix}a_{2n-1}&a_{2n}\\a_{2n}&a_{2n+1}\end{bmatrix}$, so the trace is $(a^2+4)a_n^2+(-1)^n2$. The proof uses the lemma that an isometry of the Néron-Severi lattice acts trivially on the discriminant group exactly when $((AB)^n\mp I)L_m(a)^{-1}$ is an integer matrix, and that matrix integrality collapses to the single condition $m\mid a_n$. Once the isometry is integral, the Torelli theorem extends it to a K3 automorphism, and the classification of possible cyclotomic factors of the characteristic polynomial together with the cyclicity of $\mathrm{Aut}(X)$ identify the minimal $n$ as the generator's exponent.

What would settle it

For $a=1$ and $m=61$, the theorem says the generator is the anti-symplectic automorphism with Néron-Severi action $(AB)^{15}$; finding an automorphism of that K3 surface whose Néron-Severi action is $(AB)^6$, whose trace $322$ is a valid Salem trace, would falsify Theorem 1.1 unless the paper's resultant obstruction is invalid.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is Theorem 1.1: for a K3 surface $X$ with Picard lattice $L_m(a)$ and $m\ge 2$, the integrality condition that $(AB)^n$ act by $\pm\mathrm{id}$ on the discriminant group is exactly the divisibility $m\mid a_n$, so each such $n$ produces a symplectic or anti-symplectic automorphism with $g^*|_{NS(X)}=(AB)^n$. If $5\nmid n$ and $n$ is minimal with that divisibility, then $\mathrm{Aut}(X)=\langle g\rangle$, and the characteristic polynomial of $g^*$ on $H^2(X,\mathbb{Z})$ is $S(x)(x-1)^{20}$ for even $n$ and $S(x)(x+1)^{20}$ for odd $n$, where $S(x)=x^2-((a^2+4)a_n^2+(-1)^n2)x+1$ is the Salem polynomial. The paper then reverses the flow of argument: applying these automorphisms to previously known Salem-trace results proves Theorem 1.3, characterizing generalized Fibonacci numbers by a Pell-type square condition, and Theorem 1.5, that $a_k$ divides $a_q$ exactly when $k$ divides $q$. In short, the paper claims that the minimal Fibonacci index of $m$ is the single arithmetic invariant that determines $\mathrm{Aut}(X)$.

Load-bearing premise

The conclusion relies on having already proved that $\mathrm{Aut}(X)$ is infinite cyclic for these surfaces; if $\mathrm{Aut}(X)$ had extra generators, the minimal Fibonacci index would not identify the generator.

Editorial extensions

If this is right

For any $m\ge 2$ and even $n$ with $m\mid a_n$, the K3 surface $X_{L_m(a)}$ admits a symplectic automorphism acting as $(AB)^n$ on its Néron-Severi lattice; for odd $n$, it admits an anti-symplectic one.
When $5\nmid n$ and $n$ is minimal, $\mathrm{Aut}(X)$ is infinite cyclic with a single generator, so every automorphism of that surface is a power of $g$, and the topological entropy is the logarithm of the Salem number of $S(x)$.
An integer $n$ is the $k$-th generalized Fibonacci number $a_k$ exactly when $(a^2+4)n^2+4$ is a perfect square for even $k$, or $(a^2+4)n^2-4$ is a perfect square for odd $k$; this generalizes Whitney's classical Fibonacci test.
The generalized Fibonacci sequence is a divisibility sequence in the strong sense that $a_k\mid a_q$ if and only if $k\mid q$, with the auxiliary facts $\gcd(a_k,a_{k+1})=1$ and $a_k\mid a_q\Rightarrow a_k\mid a_{q-k}$.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Because the divisibility rule $a_k\mid a_q$ iff $k\mid q$ is a statement about the recurrence alone, a purely arithmetic proof independent of K3 geometry should exist; finding it would separate the number theory from the Torelli machinery.
The theorem suggests an arithmetic recipe for constructing automorphisms of these K3 surfaces: iterate the recurrence to find the minimal $n$ with $m\mid a_n$, then write down $(AB)^n$; the same recipe predicts the Salem trace and entropy, which could be checked numerically for small $a$ and $m$.
The cases with $5\mid n$ are left open by the generator conclusion, but the resultant computations in Section 5 already rule out many cyclotomic combinations and provide a template for treating those cases.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Kwangwoo Lee's paper has a real new result hiding inside a proof that needs repair. Theorem 1.1, identifying the exponent of the automorphism generator with the rank of apparition of m in the generalized Fibonacci sequence, is genuinely new and basically correct. The gluing computation in Section 3.1 is solid: for even n with m|a_n, ((AB)^n - I)L^{-1} is integral, so (AB)^n acts trivially on the discriminant; gluing with id on the transcendental lattice gives a symplectic automorphism. The odd case works the same way with -id, giving an anti-symplectic automorphism. The minimality argument then correctly identifies the generator once the existence of these automorphisms is granted.

The soft spots are real. The sentence "by Proposition 2.6" to conclude symplectic is not justified. Takada's Proposition 2.6 gives necessary conditions for l=1,2,5,10,25,50, and for a=1, l=5 satisfies those conditions for any even n. So the cited proposition cannot rule it out. The fix is to drop that reference and rely on the gluing, which already forces the action on the 2-form to be trivial because phi|T_X=id. Alternatively, add a resultant argument for general a; Section 5 tries to do this for a=1, but that section has a worse problem.

Section 5's Lemma 5.1 computes resultants with (x-lambda^n)(x-lambda^{-n}), but the characteristic polynomial of (AB)^n is (x-lambda^{2n})(x-lambda^{-2n}). The trace formula in Lemma 2.2 confirms this. So the resultants, and the examples built on them (5.3-5.5), are off by a factor of 2 in the exponent. The 5|m case is therefore not handled correctly as written.

Theorem 1.5(1) has a spurious proof; the claim is true by the standard Euclidean algorithm, and the K3 argument with consecutive powers doesn't lead to a contradiction. Theorems 1.3 and 1.5 are standard Lucas sequence facts and should cite the general theory, not rederive them via K3 automorphisms.

Overall: the main theorem is likely correct, but the written proof has a gap in the symplectic step, and Section 5 contains a genuine error. The paper deserves a referee, but heavy revision is needed. I wouldn't cite the current version in my own work until the proof is cleaned up.

Referee Report

3 major / 5 minor

Summary. The paper studies K3 surfaces X whose Picard lattice has intersection matrix L_m(a) = m[[2,a],[a,-2]] with m ≥ 2, and relates the automorphism group Aut(X) to generalized Fibonacci numbers a_n defined by a_0=0, a_1=1, a_{n+2}=a a_{n+1}+a_n. The main theorem (Theorem 1.1) asserts that if m | a_n and n is even (resp. odd), then X admits a symplectic (resp. anti-symplectic) automorphism g with g^*|NS(X) = (AB)^n, where A,B are the involutions generating O(L_m(a)); moreover, if 5 ∤ n and n is minimal, then this g is the generator of Aut(X) ≅ Z. The paper also proves a criterion for an integer to be a generalized Fibonacci number (Theorem 1.3) and a divisibility criterion a_k | a_q iff k | q (Theorem 1.5). The arguments combine discriminant-group computations, the Torelli theorem, and resultant calculations, building on prior work of the author and of Takada.

Significance. If the main theorem is correct, it gives an explicit description of the automorphism group of a family of Picard-number-2 K3 surfaces in terms of the minimal index of m in a generalized Fibonacci sequence, which is a concrete and valuable result. The converse number-theoretic criterion in Theorem 1.3 is a natural generalization of Whitney's Fibonacci test and is interesting in its own right. The paper also demonstrates a useful technique of using resultants of Salem polynomials and cyclotomic factors to rule out certain actions on the transcendental lattice. However, the proof of the central claim has substantial gaps, and the divisibility theorem is not rigorously established as written.

major comments (3)

[Section 3.1, Theorem 1.1] The assertion in the proof that 'if 5 ∤ n, then g*|NS(X) = (AB)^n is a symplectic automorphism by Proposition 2.6' is not justified. The automorphism g constructed by gluing (AB)^n on NS(X) with the identity on T_X is symplectic for every even n with m | a_n, because it acts trivially on H^{2,0}; Proposition 2.6 is irrelevant to that conclusion. The actual gap is in the 'Moreover' claim: to conclude Aut(X) = ⟨g⟩ from minimality of n, the proof must exclude the possibility that the generator of Aut(X) is a mixed automorphism with l ∈ {5,10,25,50} and index k | n. Proposition 2.6 does not exclude this; for a = 1 and even n, the necessary condition 5(τ − 2) = 25 f_n^2 is automatically a square, so l = 5 satisfies the numerical test of Proposition 2.6. The resultant computations of Section 5 (Lemma 5.1) are the tool needed to rule out such cases, but they are not invoked in the proof of Theorem 1.1 and are stated only for a = 1. As written, the conclusion that the generator index equals the minimal n with m | a_n does not follow.
[Section 5, Lemma 5.1] There is a mismatch between the polynomial in Lemma 5.1 and the trace of (AB)^n. The resultant is computed for (x − λ^n)(x − λ^{−n}), whose trace is λ^n + λ^{−n}. For a = 1, however, trace((AB)^n) = f_{2n−1} + f_{2n+1} = λ^{2n} + λ^{−2n}; the two are equal only for n = 0. The proof in (5.1) sets µ = λ^n and then substitutes µ + µ^{−1} = trace((AB)^n), which is false. The formulas stated in Lemma 5.1 are the correct ones for the polynomial (x − λ^{2n})(x − λ^{−2n}), so the lemma can be repaired by replacing λ^n with λ^{2n} throughout, but as written the proof is inconsistent. This matters because Section 5 is the intended mechanism for ruling out mixed automorphisms, and the error obscures the correction.
[Section 3.3, Theorem 1.5(1)] The proof of Theorem 1.5(1) does not contain a valid contradiction. After deriving N = 1 for the generator index, the observation that one of g^k and g^{k+1} is symplectic and the other anti-symplectic is automatic from parity and is consistent with every k; it does not force k = 1. The claimed contradiction is spurious. The statement gcd(a_k, a_{k+1}) = 1 is true and follows directly from Lemma 2.1(2), so the proof should be replaced by that one-line argument. Since Theorem 1.5(2) and the reverse implication in Theorem 1.5(3) both use Theorem 1.5(1), this gap affects the divisibility criterion as well.

minor comments (5)

[Section 1, Introduction] There are several typos and awkward phrasings, for example 'it induce s' and 'Cayley-Oguiso' should be 'Cayley–Oguiso' with the en-dash.
[Section 3.3, Theorem 1.5(3)] The concern that the case 5 | k is unhandled in Theorem 1.5(3) does not land: the proof of (3) does not use the condition 5 ∤ n, and the forward direction relies only on the existence part of Theorem 1.1, which has no 5-condition. The proof is, however, incomplete for k = 1 because Theorem 1.1 requires m ≥ 2, although that case is trivial.
[Section 5, formulas] The notation '52f4 n' and similar strings in Lemma 5.1 should be typeset as 5^2 f_n^4; as printed they are easily misread as '52 f^{4n}' and the ambiguity is confusing.
[Example 4.3] The line 'if l = 2, then h is anti-symplectic, hence k = 1 or 5 by again Remark 3.1' is hard to follow; the connection to m ∤ f_50 should be stated more explicitly.
[Section 2.4] The sentence 'A Salem trace is an algebraic integer τ > 2 whose other conjugates lie in (−2, 2)' should read 'all conjugates other than τ itself lie in (−2, 2)', since the definition is about the conjugates of τ.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the main index computation is a direct lattice/Torelli argument, and reliance on [6] and [4] is prior published work rather than a self-referential reduction.

full rationale

The claimed derivation is not circular. The construction of the automorphism in Theorem 1.1 is a direct calculation: from m | a_n, equation (3.1) shows ((AB)^n - I)L_m(a)^{-1} is integral, Lemma 2.4 gives the discriminant action, and gluing with id on the transcendental lattice gives an isometry that the Torelli theorem promotes to an automorphism. The generator claim uses the cyclicity theorem from [6] as a black box, but [6] is a published paper with its own proof and is not itself derived from the current theorems; citing it is standard prior-work dependence, not a circular reduction. The converse direction of Theorem 1.3 applies the automorphism existence to obtain the square criterion; it does not assume the criterion to prove the automorphism. Theorem 1.5 is derived from Theorem 1.1, not presupposed. The skeptical note that Proposition 2.6 does not rule out l = 5 for even n is a proof gap or possible error in the symplectic-generator step, but it is not circularity: the conclusion is not equivalent to an input by construction, and no fitted parameter is renamed as a prediction. The self-citations to [4] and [6] are load-bearing but externally published; under the stated rules they count as real evidence, so the circularity score is 0.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The central claim rests on the Torelli theorem, the existence result for K3 surfaces with prescribed Picard lattice, and the structure theorems of [4] and [6]. No new physical or mathematical entities are postulated.

assumptions (5)

standard math Global Torelli theorem for K3 surfaces
Used in Section 3.1 to extend isometries of H^2 to automorphisms, and in Section 3.2 for the converse construction.
domain assumption Existence of a projective K3 surface with specified Picard lattice
Proposition 2.5, cited from [4], is used throughout to create surfaces X_{L_m(a)} for any even lattice of signature (1,1).
domain assumption O(L_m(a)) ≅ Z_2 * Z_2 and Aut(X) ≅ Z for m≥2
Imported from [6, Theorem 1]; the paper's main theorem assumes the generator acts as (AB)^n.
domain assumption Classification of Salem traces for Picard number 2 K3 surfaces ([4, Main Theorem] and [4, Lemma 3])
Used in Theorem 1.3 to connect traces to squares and to construct automorphisms from arithmetic conditions.
domain assumption Proposition 2.6 from [8] on possible cyclotomic factors
Used in Sections 3.1 and 5 to restrict the cyclotomic factor and the symplectic or anti-symplectic type.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Generalized Fibonacci numbers and automorphisms of K3 surfaces with Picard number 2." pith.science (2026). https://pith.science/paper/ESAXWDAZ

@misc{pith2026241113038,
  author       = {Pith},
  title        = {Pith review of: Generalized Fibonacci numbers and automorphisms of K3 surfaces with Picard number 2},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ESAXWDAZ}},
  note         = {Machine review of arXiv:2411.13038}
}

read the original abstract

Using the properties of generalized Fibonacci numbers, we determine the automorphism groups of some K3 surfaces with Picard number 2. Conversely, using the automorphisms of K3 surfaces with Picard number 2, we prove the criterion for a given integer n is to be a generalized Fibonacci number. Moreover, we show that the generalized k-th Fibonacci number divides the generalized q-th Fibonacci number if and only if k divides q.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

10 extracted references · 10 canonical work pages

[1]

Festi, D., Garbagnati, A., van Geemen, B., van Luijk, R.: The Cayley–Oguiso automorphism of positive entropy on a K3 surface. J. Mod. Dyn. 7 (1), 75–97 (2013)

work page 2013
[2]

Gromov, M.: On the entropy of holomorphic maps. Enseign. Math. 49, 217–235 (2003)

work page 2003
[3]

Cambridge Unive rsity Press, Cambridge (2016)

Huybrechts D.: Lectures on K3 Surfaces. Cambridge Unive rsity Press, Cambridge (2016)

work page 2016
[4]

Hashimoto, K., Keum, J., Lee, K.: K3 surfaces with Picard number 2, Salem polynomials and Pell equation. J. Pure Appl. Algebra. 224, 432-443 (2020)

work page 2020
[5]

EMS Tracts Math., 32 EMS Publishing House, Berlin (2020)

Kondo, S.: K3 surfaces. EMS Tracts Math., 32 EMS Publishing House, Berlin (2020)

work page 2020
[6]

Lee, K.: Automorphisms of some K3 surfaces and Beauville involutions of their Hilbert schemes. Rend. Circ. Mat. Palermo, II. Ser. 72 (4), 2443-2457 (2023) 10 KW ANGWOO LEE

work page 2023
[7]

In: Algebraic geometry in east Asia-Taipei 2011

Oguiso, K.: Free Automorphisms of positive entropy on sm ooth K¨ ahler surfaces. In: Algebraic geometry in east Asia-Taipei 2011. Adv. Stud. Pure Math. 65, 187-199 (2015)

work page 2015
[8]

arXiv:2405.15195 (2024)

Takada, Y .: Dynamical degrees of automorphisms of K3 sur faces with Picard number 2. arXiv:2405.15195 (2024)

work page arXiv 2024

Show all 10 references

[9]

The Fib onacci quarterly, 10 (4), 413-422 (1972)

Whitney, R.E.: Advanced problems and solutions. The Fib onacci quarterly, 10 (4), 413-422 (1972)

1972
[10]

Israel J

Y omdin, Y .: V olume growth and entropy. Israel J. Math.57, 285–300 (1987) DEPARTMENT OF MATHEMATICS , K OREA INSTITUTE FOR ADVANCED STUDY, 85 H OEGI - RO, D ONGDAEMUN -GU , S EOUL , R EPUBLIC OF KOREA Email address: klee0222@gmail.com

1987

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

Festi, D., Garbagnati, A., van Geemen, B., van Luijk, R.: The Cayley–Oguiso automorphism of positive entropy on a K3 surface. J. Mod. Dyn. 7 (1), 75–97 (2013)

work page 2013

[2] [2]

Gromov, M.: On the entropy of holomorphic maps. Enseign. Math. 49, 217–235 (2003)

work page 2003

[3] [3]

Cambridge Unive rsity Press, Cambridge (2016)

Huybrechts D.: Lectures on K3 Surfaces. Cambridge Unive rsity Press, Cambridge (2016)

work page 2016

[4] [4]

Hashimoto, K., Keum, J., Lee, K.: K3 surfaces with Picard number 2, Salem polynomials and Pell equation. J. Pure Appl. Algebra. 224, 432-443 (2020)

work page 2020

[5] [5]

EMS Tracts Math., 32 EMS Publishing House, Berlin (2020)

Kondo, S.: K3 surfaces. EMS Tracts Math., 32 EMS Publishing House, Berlin (2020)

work page 2020

[6] [6]

Lee, K.: Automorphisms of some K3 surfaces and Beauville involutions of their Hilbert schemes. Rend. Circ. Mat. Palermo, II. Ser. 72 (4), 2443-2457 (2023) 10 KW ANGWOO LEE

work page 2023

[7] [7]

In: Algebraic geometry in east Asia-Taipei 2011

Oguiso, K.: Free Automorphisms of positive entropy on sm ooth K¨ ahler surfaces. In: Algebraic geometry in east Asia-Taipei 2011. Adv. Stud. Pure Math. 65, 187-199 (2015)

work page 2015

[8] [8]

arXiv:2405.15195 (2024)

Takada, Y .: Dynamical degrees of automorphisms of K3 sur faces with Picard number 2. arXiv:2405.15195 (2024)

work page arXiv 2024