Congruence properties of Lehmer-Euler numbers

Guo-Dong Liu; Takao Komatsu

REVIEW 2 major objections 3 minor 2 cited by

Congruence properties of Lehmer-Euler numbers

T0 review · 2 major / 3 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Lehmer-Euler numbers have periodic residues modulo powers of three

desk verdict Real new congruences for Lehmer-Euler numbers, but Theorem 4's proof needs a filled-in binomial congruence argument; the result survives. read the letter →

arxiv 2501.01178 v1 pith:6ESPLULQ submitted 2025-01-02 math.NT math.CO

Takao Komatsu , Guo-Dong Liu This is my paper · ORCID

classification math.NTmath.CO MSC 11B6811A0705A19

keywords Lehmer-EulernumbersEulerBernoullicongruencesLucas'stheorembinomialcoefficientsperiodicitycentralfactorial

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 studies Lehmer-Euler numbers, the class of numbers defined in 1935 through cubic roots of unity as a generalization of Euler and Bernoulli numbers, and proves that their nonzero entries obey strict periodic congruences modulo powers of three. The main proved result is that modulo 27, the nonzero sequence $W_{3n}$ cycles through the residues $1, -1, -8, -1, 1, 8$ with period 6. It also proves $W_{3n} \equiv (-1)^n \pmod{9}$, gives recurrence and explicit formulas for the numbers, and introduces incomplete and higher-order analogues. The paper conjectures the full rule: if $3n \equiv 3m \pmod{2 \cdot 3^k}$, then $W_{3n} \equiv W_{3m} \pmod{3^{k+1}}$.

What carries the argument

The load-bearing object is the Lehmer-Euler sequence itself, with generating function $3/(e^t + e^{\omega t} + e^{\omega^2 t}) = (\sum_{l \geq 0} t^{3l}/(3l)!)^{-1}$, so that $W_n$ is nonzero exactly when $3$ divides $n$. The congruence proofs run through the recurrence $W_{3n} = -\sum_{k=0}^{n-1} \binom{3n}{3k} W_{3k}$, which expresses each term as an integer linear combination of earlier terms. The proof evaluates the appearing binomial-coefficient sums via the identity $\sum \binom{3n}{3k+\ell} x^{3k+\ell} = \frac{1}{3}\sum_{j=0}^2 \omega^{\ell j}(1+\omega^j x)^{3n}$, turning them into explicit powers of $3$ times expressions in $\sqrt{-3}$. Lucas's theorem gives these sums modulo $3$, and the paper asserts a stronger unproved binomial congruence modulo arbitrary powers of $3$ to move the congruences up to modulo 27.

What would settle it

Direct computation of $W_{3n}$ for $n = 0, 1, \dots, 50$ using the recurrence $W_{3n} = -\sum_{k=0}^{n-1} \binom{3n}{3k} W_{3k}$ and reduction modulo 27 can settle the main theorem's claim: it must yield the sequence $1, -1, -8, -1, 1, 8$ (mod 27) with period 6. Additionally, testing the asserted binomial congruence for specific pairs, say $n = 10$ and each $k$, would show whether the gap in the proof is real.

Watch

Extended reading notes

Core claim

The central claim is Theorem 4: for every $n \geq 0$, the congruences $W_{9n} \equiv (-1)^n$, $W_{9n+3} \equiv (-1)^{n-1}$, and $W_{9n+6} \equiv 8(-1)^{n-1}$ hold modulo 27. From these, the paper derives the period-6 pattern of $W_{3n}$ modulo 27 and, by an analogous argument that it sketches, the corresponding residue patterns modulo $3^4$ and $3^5$. The paper also proposes Conjecture 1 as the general law, asserting that the residue of $W_{3n}$ modulo $3^{k+1}$ depends only on the class of $n$ modulo $2 \cdot 3^k$. In addition, the paper establishes a new identity expressing Euler numbers in terms of central factorial numbers through a polynomial sequence defined by a two-term recurrence, and it extends the recurrence, determinant, and explicit-formula toolkit to incomplete and higher-order Lehmer-Euler numbers.

Load-bearing premise

The proof of the modulo-27 congruences relies on the unproved assertion that $\binom{9n}{9k+3\ell} \equiv \binom{3n}{3k+\ell}$ modulo every power of three once $n$ is large enough, which is stronger than Lucas's theorem and is stated without proof.

Editorial extensions

If this is right

If Theorem 4 is correct, the residue of $W_{3n}$ modulo 27 is determined entirely by $n$ modulo 6, reproducing the explicit six-entry pattern in the paper.
If Conjecture 1 is true, the congruence class of $W_{3n}$ modulo any $3^{k+1}$ is a function only of $n$ modulo $2 \cdot 3^k$, mirroring the classical result for Euler numbers modulo powers of two but with period doubled at each level.
The identity with central factorial numbers provides a fresh explicit formula for Euler numbers in terms of the polynomial sequence $\Delta(x,k)$, which may be useful for studying divisibility of Euler numbers.
The incomplete and higher-order Lehmer-Euler numbers inherit the same recurrence and determinant structure, so the congruence machinery can be applied to those families as well.

Reading between the lines

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

The unproved binomial congruence $\binom{9n}{9k+3\ell} \equiv \binom{3n}{3k+\ell} \pmod{3^j}$ for all $j$ when $n$ is large is the delicate step; if it can be established by p-adic methods, the proof of Theorem 4 would be fully rigorous and the same method might prove Conjecture 1.
The observed period doubling suggests that $W_{3n}$ is 3-adically well-behaved: the residue modulo $3^{k+1}$ depends on $n$ modulo $2 \cdot 3^k$, which is consistent with a 3-adic analytic continuation whose Taylor coefficients encode the roots-of-unity sums.
Applying the same generating-function identity with $r$-th roots of unity to the higher-order numbers $W_{r,n}^{(\alpha)}$ defined in Section 5 may yield analogous periodic congruences modulo powers of $r$.

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

Quick take: this paper has genuine new results on Lehmer-Euler numbers, but the proof of the mod-27 theorem has a gap that needs fixing before publication. The main claims look true and are worth refereeing.

What's actually new: the mod 9 and mod 27 congruences (Theorems 3 and 4), the periodicity table in Proposition 2, and the polynomial Delta(x,k) identities (Theorems 5 and 6). The methods are standard — generating functions, Lucas's theorem, induction — but the results are legitimate extensions, not rehashes of earlier work. Theorem 3 is clean and complete. The induction proofs for Theorems 5 and 6 are solid, and they give a neat connection between Euler numbers and central factorial numbers.

The soft spots are in Section 3. In the proof of Theorem 4, the authors assert that binom(9n, 9k+3ℓ) ≡ binom(3n, 3k+ℓ) modulo 3^j for all j≥1 once n is large enough. That's false as stated: take n=5, k=0, ℓ=1; the difference is 14175, which is 81 mod 243, so the mod-3^5 version fails. What the proof actually needs is the mod-27 case, and that case is true, but the paper doesn't prove it. The stress-test note sketches a valid argument using the product formula for factorials; that argument should be inserted. Without it, the induction in Theorem 4 relies on an assertion that is both unproved and overstrong. Also, Proposition 2's proof explicitly omits most cases, saying 'Other identities are similarly shown.' For a table of periodicity at four different moduli, that's a lot of unverified content. I'd want the cases filled in or a code snippet verifying them.

Minor quibbles: the 'In fact' refinements in the Theorem 4 proof (e.g., the sharper congruences for n≥n0) are asserted without proof and are not needed for the main claims. The paper could drop them safely. On the positive side, the citations look appropriate — the authors cite the relevant literature, including their own prior work, without over-reliance.

Who this is for: number theorists working on Bernoulli/Euler-type numbers and their congruences. The impact is narrow but real. The gaps are fixable, and the stress-test's counterexample doesn't sink the main theorem. I'd send this to a competent referee and ask for a repaired proof of Theorem 4 and a fuller Proposition 2.

Referee Report

2 major / 3 minor

Summary. The paper studies Lehmer-Euler numbers W_n, defined in 1935 by Lehmer via cubic roots of unity. The authors prove recurrence, explicit, and determinant formulas (Theorems 1 and 2), and establish congruence properties modulo 9 and 27 (Theorems 3 and 4). They then tabulate periodic residue patterns modulo powers of 3 and propose Conjecture 1 generalizing the period. The paper also introduces incomplete Lehmer-Euler numbers (Section 4), higher-order Lehmer-Euler numbers (Section 5), and a new polynomial sequence Delta(x,k) whose identities with Euler numbers and central factorial numbers are proved in Theorems 5 and 6. The main claimed new arithmetic results are the modulo 27 congruences and the conjectured periodicity.

Significance. If the proof gaps are repaired, Theorems 3 and 4 give a clean congruence pattern for Lehmer-Euler numbers modulo powers of 3, and the paper collects useful recurrence and determinant identities. The new polynomial sequence and its identities with central factorial numbers (Theorems 5 and 6) are proven in full and appear correct. However, the main arithmetic theorem, Theorem 4, currently depends on an unproved and, as stated, false binomial congruence assertion, so the central claim is not yet established in the written form. The paper contains reproducible definitions and several complete induction proofs, which is a strength.

major comments (2)

[Section 3, proof of Theorem 4] After quoting Lucas's theorem, the proof states: 'In fact, these congruences hold for (mod 3^l) (forall l >= 1) if n is enough large.' This assertion is not proved and is false as stated. For n = 5, k = 0, and ell = 1, the difference binom(45,3) - binom(15,1) = 14190 - 15 = 14175 is not divisible by 3^5 = 243. The subsequent derivation of the congruences for W_{9n}, W_{9n+3}, and W_{9n+6} modulo 27 relies on the vanishing of the mixed sums S1 and S2 modulo 27, which depends on the mod-27 case of this binomial congruence. Since this is the key input to Theorem 4, the proof is incomplete. The mod-27 case may be true and provable, for instance by writing (3m)! = 3^m m! prod_{i=1}^m (3i-1)(3i-2), but that argument is not supplied in the paper.
[Section 3.1, proof of Proposition 2] The proof of Proposition 2, which asserts periodic residue patterns modulo 3^4 and 3^5, explicitly says 'Other identities are similarly shown and their proofs are omitted.' The displayed congruences for W_{27n+r} modulo 81 and the longer pattern modulo 243 are not derived from Theorem 4 or from any other statement in the paper. If Proposition 2 is intended as a theorem, these cases need either full proofs or a clear statement that they are computational observations. As written, the assertion is not established, and it is load-bearing for Conjecture 1.

minor comments (3)

[Theorem 4 statement] The second congruence is written as W_{9n+3} = (-1)^{n-1} without a modulus symbol; since the proof only establishes congruence modulo 27, it should be written as W_{9n+3} congruent to (-1)^{n-1} (mod 3^3).
[Section 3, proof of Theorem 4, W_{9n+6} part] The line 'W_{9n+6} congruent to -1 - 24 - 3 congruent to -1 (mod 3^3)' appears immediately after the computation for W_{9n+6} and seems to be a stray or erroneous remark; it should be removed or replaced with a correct intermediate congruence.
[Section 3.1 heading] The heading 'Lehmer-Euler numbers modulo powers of three' contains the typo 'Lemer-Euler numbers'; it should be 'Lehmer-Euler numbers'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the congruences are derived from the defining generating function of W_n, standard Lucas's theorem, and explicit root-of-unity binomial sums; nothing is fitted, renamed, or imported self-referentially as the target result.

full rationale

The derivation chain starts with the definition of the Lehmer-Euler numbers as coefficients of the reciprocal of sum t^{3l}/(3l)!, so the recurrences and explicit formulas quoted in Section 2 are direct algebraic consequences of that definition rather than fitted inputs. Theorem 3 and Theorem 4 use only the recurrence, Lucas's theorem (Lemma 1, an external classical result), and exact evaluations of sums like sum_k binom(3n,3k+1)(-1)^{k-1} in terms of powers of sqrt(-3); the binomial congruences are asserted rather than derived, and the claim that they hold modulo 3^j for all j is false as stated, but that is a correctness gap in the proof, not circularity, since the needed modulo-27 version is independent of the theorem being proved. Proposition 2 is an observed pattern, not used to prove Theorem 4, and Conjecture 1 is explicitly labeled a prediction. The polynomial Delta(x,k) is defined before its identities are proved, and Theorems 5 and 6 are proved by induction using the recurrence defining Delta and standard recurrences for central factorial numbers. The paper's self-citations (e.g., [1] and [10] for basic properties of W_n) are not load-bearing in a circular way: the cited recurrences are elementary and parameter-free, and the final congruences do not reduce to those citations by construction. No fitted constant, no calibration to a subset of data, and no renaming of a known pattern as a derivation was found.

Assumptions & free parameters 0 free parameters · 6 assumptions · 2 invented entities

No fitted parameters appear. The paper relies on standard binomial and root-of-unity identities, Lucas's theorem, Trudi's formula, inversion formulas, and central factorial recurrences. It also relies on one unproved binomial congruence modulo powers of 3 that is load-bearing in Theorem 4. The new objects are the incomplete Lehmer-Euler numbers and the polynomial sequence Delta(x,k), both defined within the paper without external falsifiable handles.

assumptions (6)

standard math Lucas's theorem for binomial coefficients modulo p
Lemma 1, used in proof of Theorem 4 to relate binom(9n,9k+3l) to binom(3n,3k+l) modulo 3.
standard math Trudi's formula for determinants
Used to derive explicit formulas in Theorem 2 and Proposition 4; cited as references [2,21,26].
standard math Inversion formula for the determinant relation
Proposition 1 is derived from an inversion formula cited as reference [11].
standard math Recurrences (10) and (13) for central factorial numbers
Used in proofs of Theorems 5 and 6; they follow from definitions (8), (9), (11), (12) cited to reference [22].
standard math Root-of-unity summation identities for binomial coefficients
Used extensively in Section 3 to evaluate sums such as sum_k binom(3n,3k+1)(-1)^{k-1}; standard roots-of-unity filter.
ad hoc to paper Unproved congruence assertion: binom(9n,9k+3l) congruent to binom(3n,3k+l) modulo powers of 3 for sufficiently large n
Stated in proof of Theorem 4 without proof; load-bearing for the modulo-27 congruences.

invented entities (2)

Incomplete Lehmer-Euler numbers W_{n,<=m} and W_{n,>=m}
purpose: Define partial-sum truncations of the generating function to generalize W_n
New mathematical objects introduced in Section 4; they reduce to W_n at m = infinity or m = 1, but are defined within the paper and have no external falsifiable handle.
Polynomial sequence Delta(x,k)
purpose: Encoding Euler-number identities via central factorial numbers, with recurrence (7)
Defined in Section 6 and used in Theorems 5 and 6; it is a new construct of the paper.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Congruence properties of Lehmer-Euler numbers." pith.science (2026). https://pith.science/paper/6ESPLULQ

@misc{pith2026250101178,
  author       = {Pith},
  title        = {Pith review of: Congruence properties of Lehmer-Euler numbers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6ESPLULQ}},
  note         = {Machine review of arXiv:2501.01178}
}

read the original abstract

Certain generalization of Euler numbers was defined in 1935 by Lehmer using cubic roots of unity, as a natural generalization of Bernoulli and Euler numbers. In this paper, Lehmer's generalized Euler numbers are studied to give certain congruence properties together with recurrence and explicit formulas of the numbers. We also show a new polynomial sequence and its properties. Some identities including Euler and central factorial numbers are obtained.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Ordered set partition posets
math.CO 2025-06 accept novelty 6.0 of 10

Ordered partition posets with block sizes divisible by d are Cohen-Macaulay with explicitly computed symmetric-group homology, and the modulo-1 variant has Catalan Möbius functions.
Higher-Order Congruence for Reciprocal Power Sums and Generalized Lehmer-Type Products
math.NT 2026-07 accept novelty 4.5 of 10

Odd-order reciprocal power sums satisfy a uniform Bernoulli-polynomial congruence modulo n, and Lehmer-type products admit truncated Bell-polynomial expansions modulo n^{K+1}.

Reference graph

Works this paper leans on

27 extracted references · 27 canonical work pages · cited by 2 Pith papers

[1]

Barman and T

R. Barman and T. Komatsu, Lehmer’s generalized Euler numbers in hypergeometric functions, J. Korean Math. Soc. 56 (2019), 485–505

work page 2019
[2]

Brioschi, Sulle funzioni Bernoulliane ed Euleriane , Annali di Mat., i

F. Brioschi, Sulle funzioni Bernoulliane ed Euleriane , Annali di Mat., i. (1858), 260–263; Opere Mat., i. pp. 343–347

work page
[3]

P. J. Cameron, Some sequences of integers , Discrete Math. 75 (1989), 89–102

work page 1989
[4]

Ernvall, Generalized Bernoulli numbers, generalized irregular primes, and class number , Ann

R. Ernvall, Generalized Bernoulli numbers, generalized irregular primes, and class number , Ann. Univ. Turku Ser.A I 178 (1979), 1– 72

work page 1979
[5]

F. G. Frobenius, ¨Uber die Bernoullischen Zahlen und die Eulerschen Polynome, in: Sitzungsberichte der K¨ oniglich Preussischen Akadem ie der Wissenschaften zu Berlin, 1910, pp. 809–847; also in: Ge sammelte Abhandlungen III, Springer-Verlag, 1968, pp. 440–478

work page 1910
[6]

J. W. L. Glaisher, Expressions for Laplace’s coeﬃcients, Bernoullian and Eulerian numbers etc. as determinants , Messenger (2) 6 (1875), 49-63

work page
[7]

Komatsu, Complementary Euler numbers , Period

T. Komatsu, Complementary Euler numbers , Period. Math. Hungar. 75 (2017), 302–314

work page 2017
[8]

Komatsu, Incomplete poly-Cauchy numbers , Monatsh

T. Komatsu, Incomplete poly-Cauchy numbers , Monatsh. Math. 180 (2016), No. 2, 271–288

work page 2016

Show all 27 references

[9]

Komatsu, K

T. Komatsu, K. Liptai and I. Mez˝ o, Incomplete poly-Bernoulli num- bers associated with incomplete Stirling numbers , Publ. Math. Debr. 88 (2016), No. 3–4, 357–368

2016
[10]

Komatsu and R

T. Komatsu and R. K. Pandey, On hypergeometric Cauchy numbers of higher grade, AIMS Mathematics 6 (2021), Issue 7, 6630–6646

2021
[11]

Komatsu and J

T. Komatsu and J. L. Ramirez, Some determinants involving incom- plete Fubini numbers , An. S ¸tiint ¸. Univ. “Ovidius” Constant ¸a Ser. Mat. 26 (2018), no.3, 143–170. DOI: 10.2478/auom-2018-0038 19

2018 doi
[12]

Komatsu and H

T. Komatsu and H. Zhu, Hypergeometric Euler numbers , AIMS Math. 5 (2020), 1284–1303

2020
[13]

D. H. Lehmer, Lacunary recurrence formulas for the numbers of Bernoulli and Euler , Ann. of Math. (2), 36 (1935), 637–649

1935
[14]

Liu, On congruences of Euler numbers modulo powers of two , J

G. Liu, On congruences of Euler numbers modulo powers of two , J. Number Theory 128 (2008), 3063–3071

2008
[15]

A. D. Loveless, A congruence for products of binomial coeﬃcients mod- ulo a composite , Integers 7 (2007), #A44

2007
[16]

Lucas, Th´ eorie des Fonctions Num´ eriques Simplement P´ eriodiques, Amer

E. Lucas, Th´ eorie des Fonctions Num´ eriques Simplement P´ eriodiques, Amer. J. Math. 1 (1878), 184–196, 197–240, 289–321
[17]

Luo, An explicit formula for the Euler numbers of higher order , Tamkang J

Q.-M. Luo, An explicit formula for the Euler numbers of higher order , Tamkang J. Math. 36 (2005), No. 4, 315–317

2005
[18]

Komatsu, J

T. Komatsu, J. L. Ram ´ ırez and D. Villamizar, A combinatorial ap- proach to the Stirling numbers of the ﬁrst kind with higher le vel, Studia Sci. Math. Hungar. 58 (2021), no.3, 293–?307

2021
[19]

Komatsu, J

T. Komatsu, J. L. Ram ´ ırez and D. Villamizar, A combinatorial ap- proach to the generalized central factorial numbers , Mediterr. J. Math. 18 (2021), no.5, Paper No. 192, 14 pp

2021
[20]

Liu and W.-P

G.-D. Liu and W.-P. Zhang, Applications of an explicit formula for the generalized Euler numbers, Acta Math. Sin. (Eng. Ser.) 24 (2008), no.2, 343–352

2008
[21]

Muir, The theory of determinants in the historical order of develop - ment, Four volumes, Dover Publications, New York, 1960

T. Muir, The theory of determinants in the historical order of develop - ment, Four volumes, Dover Publications, New York, 1960

1960
[22]

Riordan, Combinatorial identities , John Wiley & Sons, Inc., New York-London-Sydney, 1968

J. Riordan, Combinatorial identities , John Wiley & Sons, Inc., New York-London-Sydney, 1968

1968
[23]

N. J. A. Sloane, The on-line encyclopedia of integer sequences , available at oeis.org
[24]

M. A. Stern, Zur Theorie der Eulerschen Zahlen , J. Reine Angew. Math. 79 (1875), 67–98
[25]

Sun, On Euler numbers modulo powers of two , J

Z.-W. Sun, On Euler numbers modulo powers of two , J. Number Theory 115(2) (2005), 371–380. 20

2005
[26]

Trudi, Intorno ad alcune formole di sviluppo , Rendic

N. Trudi, Intorno ad alcune formole di sviluppo , Rendic. dell’ Accad. Napoli (1862), 135–143
[27]

S. S. Wagstaﬀ Jr., Prime divisors of the Bernoulli and Euler numbers , in: Number Theory for the Millennium, III, Urbana, IL, 2000, A K Peters, Natick, MA, 2002, pp. 357–374. 21

2000

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [1]

Barman and T

R. Barman and T. Komatsu, Lehmer’s generalized Euler numbers in hypergeometric functions, J. Korean Math. Soc. 56 (2019), 485–505

work page 2019

[2] [2]

Brioschi, Sulle funzioni Bernoulliane ed Euleriane , Annali di Mat., i

F. Brioschi, Sulle funzioni Bernoulliane ed Euleriane , Annali di Mat., i. (1858), 260–263; Opere Mat., i. pp. 343–347

work page

[3] [3]

P. J. Cameron, Some sequences of integers , Discrete Math. 75 (1989), 89–102

work page 1989

[4] [4]

Ernvall, Generalized Bernoulli numbers, generalized irregular primes, and class number , Ann

R. Ernvall, Generalized Bernoulli numbers, generalized irregular primes, and class number , Ann. Univ. Turku Ser.A I 178 (1979), 1– 72

work page 1979

[5] [5]

F. G. Frobenius, ¨Uber die Bernoullischen Zahlen und die Eulerschen Polynome, in: Sitzungsberichte der K¨ oniglich Preussischen Akadem ie der Wissenschaften zu Berlin, 1910, pp. 809–847; also in: Ge sammelte Abhandlungen III, Springer-Verlag, 1968, pp. 440–478

work page 1910

[6] [6]

J. W. L. Glaisher, Expressions for Laplace’s coeﬃcients, Bernoullian and Eulerian numbers etc. as determinants , Messenger (2) 6 (1875), 49-63

work page

[7] [7]

Komatsu, Complementary Euler numbers , Period

T. Komatsu, Complementary Euler numbers , Period. Math. Hungar. 75 (2017), 302–314

work page 2017

[8] [8]

Komatsu, Incomplete poly-Cauchy numbers , Monatsh

T. Komatsu, Incomplete poly-Cauchy numbers , Monatsh. Math. 180 (2016), No. 2, 271–288

work page 2016