Brauer-Manin obstruction for Erd\H{o}s-Straus surfaces

Daniel Loughran; Martin Bright

REVIEW 1 major objections 3 minor 24 references

Brauer-Manin obstruction for Erd\H{o}s-Straus surfaces

T0 review · 1 major / 3 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read The paper proves that the Brauer–Manin obstruction leaves the Erdős–Straus conjecture open but blocks strong approximation.

desk verdict Solid Brauer-Manin analysis of Erdos-Straus surfaces: new necessary condition, full Brauer group computation, and a minor n=1 edge case in Theorem 1.8. read the letter →

arxiv 1908.02526 v2 pith:5PEZ24FG submitted 2019-08-07 math.NT math.AG

Martin Bright , Daniel Loughran This is my paper · ORCID

classification math.NTmath.AG MSC 14G0511D6811D2514F22

keywords Erdős–StrausconjectureBrauer–ManinobstructionstrongapproximationHilbertsymbolunitfractionslogK3surfacesrationalsingularitiesquaternionalgebras

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

The Erdős–Straus conjecture asks whether every rational number $4/n$ can be written as $1/u_1+1/u_2+1/u_3$ with positive integers $u_i$. The paper approaches this through the affine cubic surface $U_n$ cut out by $4u_1u_2u_3=n(u_1u_2+u_1u_3+u_2u_3)$ and proves that its Brauer group is generated over $\mathbb{Q}$ by a single quaternion algebra. Using that algebra, it shows that no Brauer–Manin obstruction prevents natural-number solutions from existing, yet a Brauer–Manin obstruction does prevent strong approximation: for odd $n$, every natural solution satisfies $\prod_{p\mid n}(-u_1/u_3,-u_2/u_3)_p=-1$, while signed solutions give the opposite sign. The explicit product recovers several previously known necessary conditions as special cases. If the paper is right, the Erdős–Straus conjecture can only be proved by a mechanism genuinely different from a Brauer–Manin obstruction.

What carries the argument

The load-bearing object is the quaternion algebra $\alpha=(-u_1/u_3,-u_2/u_3)$, whose local evaluation is the Hilbert symbol $(-u_1/u_3,-u_2/u_3)_v$. The identity (2.5), $-\frac{u_i}{u_j}=1/(1+u_j/u_k-4u_j/n)$, is what lets $\alpha$ descend to $\mathbb{Q}$ by killing the residues along the boundary lines. The identification $\operatorname{Br}U_n\cong\operatorname{Br}\widetilde{U}_n$ comes from a general theorem for surfaces with rational singularities, so the entire obstruction reduces to a single reciprocity product over the primes dividing $n$.

What would settle it

Enumerate all natural-number solutions for a fixed odd composite $n$, such as $n=45$ or $n=225$, and compute $\prod_{p\mid n}(-u_1/u_3,-u_2/u_3)_p$ using the Hilbert-symbol laws in the paper; a single tuple with product $+1$ would refute Theorem 1.2, while a computation showing the Brauer–Manin set on the positive real component is empty would refute Theorem 1.8.

Watch

Extended reading notes

Core claim

On the singular affine surface $U_n$, the paper computes $\operatorname{Br} U_n/\operatorname{Br} \mathbb{Q} \cong \mathbb{Z}/2\mathbb{Z}$, generated by the quaternion algebra $\alpha=(-u_1/u_3,-u_2/u_3)$, and shows that this class is transcendental rather than algebraic. The local invariant of $\alpha$ at a place $v$ is the Hilbert symbol $(-u_1/u_3,-u_2/u_3)_v$, and global Hilbert reciprocity turns a global point into a single product over primes. Evaluating the invariant at infinity separates $U_n(\mathbb{R})_+$, where all coordinates are positive, from the other real component. The result is an exact dichotomy: natural solutions have product $-1$ over the primes dividing $n$, non-natural integer solutions have product $1$, and the Brauer–Manin set for natural solutions is always non-empty but strictly smaller than the full relevant adelic set. In particular, the paper establishes new cases of a transcendental Brauer–Manin obstruction on a log K3 surface.

Load-bearing premise

The load-bearing premise is that the Brauer group of the singular surface $U_n$ is faithfully captured by the Brauer group of its desingularisation; if that identification failed, the Hilbert-product identity would still make sense on the smooth model but would no longer be an obstruction on $U_n$ itself.

Editorial extensions

If this is right

The Erdős–Straus conjecture cannot be settled by a Brauer–Manin obstruction; any successful proof must find an obstruction or descent mechanism outside this framework.
For odd $n$, every natural solution must satisfy the explicit Hilbert-symbol product condition $\prod_{p\mid n}(-u_1/u_3,-u_2/u_3)_p=-1$, a new local necessary condition that can be checked prime by prime.
Known necessary conditions for solutions, including those for odd primes and odd squares, follow uniformly from the Brauer–Manin formalism rather than from ad hoc reciprocity computations.
The failure of strong approximation is genuine but incomplete: rational points are not dense in the Brauer–Manin set, so further obstructions beyond Brauer–Manin must control strong approximation on these surfaces.

Reading between the lines

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

The explicit $-1$ product could be used as a fast sieve: search algorithms for unit-fraction decompositions can discard any tuple failing the product identity before more expensive arithmetic checks.
Because all $U_n$ are $\mathbb{Q}$-isomorphic to $U_1$, the sharp difference between models suggests that strong approximation for this class of affine surfaces depends delicately on the chosen integral model; analogous rescalings of other surfaces might be worth testing.
The general theorem on Brauer groups of surfaces with rational singularities has potential use elsewhere: an affine surface whose desingularisation has a computable Brauer group can inherit transcendental classes without a separate purity calculation at the singular locus.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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

First thing to know: this is a solid paper, not a breakthrough. It shows that the Brauer-Manin obstruction does not rule out natural solutions to Erdos-Straus, but does obstruct strong approximation. The genuinely new pieces are Theorem 1.2, a universal Hilbert-symbol condition for all odd n, and Theorem 1.7, a surjectivity result for Brauer groups of rational singularities. The paper also computes the full Brauer group of the singular surface, including the transcendental part; that is rare and gives the paper independent value. The derivation is honest: the quaternion algebra is constructed from the equation, and the local computations are explicit. The authors are careful to note where they rely on existing rational-singularity machinery (Theorem 1.7) and supply a fallback via the desingularisation in Remark 2.11, so the central claim does not depend on the most delicate identification. I checked the reciprocity step and the valuation arguments; they hold.

The only concrete error I find is small and in the statement of Theorem 1.8: for n=1, equation (1.4) is false. The stress-test note is right that for n=1 the Brauer kernel is empty because the invariant at infinity is -1 and all finite invariants are 1. The Erdos-Straus conjecture starts at n>=2, and Theorem 1.8 for n>=2 is correct. This should be a one-line correction in a final version (exclude n=1 or adjust the statement). Also, some local invariant tables in Lemma 3.8 are asserted after a case check rather than shown in full; this is a minor exposition issue, not a gap. The dependence on external results for Theorem 1.7 is real but not load-bearing, given Remark 2.11. The paper cites prior work appropriately, including the Yamamoto and Elsholtz-Tao results it recovers.

Who is this for? Arithmetic geometers working on Brauer-Manin and integral points, and number theorists interested in Erdos-Straus. The paper deserves a serious referee. I would accept it for peer review with a request to fix the n=1 statement. The proof of Theorem 1.7 is a bit condensed but the argument is standard. Overall, the paper is exactly what a good arithmetic geometry paper should look like: it does not overclaim, the main theorem is explicit and checkable, and the unified derivation of known conditions is a real service.

Referee Report

1 major / 3 minor

Summary. The paper studies the affine cubic surface U_n defined by the Erdős–Straus equation 4u1u2u3 = n(u1u2+u1u3+u2u3), computes its Brauer group via a desingularisation, and determines the resulting Brauer–Manin set for integral points. The main theorems assert: (i) there is no Brauer–Manin obstruction to the existence of natural-number solutions to 4/n = 1/u1 + 1/u2 + 1/u3 (Theorem 1.1, made precise as non-emptiness of the Brauer set in Theorem 1.8(1.4)); (ii) for odd n, every natural solution satisfies an explicit product formula ∏_{p|n}(-u1/u3, -u2/u3)_p = -1 (Theorem 1.2), while non-natural integer solutions satisfy the opposite product = 1 (Theorem 1.5); and (iii) strong approximation for integral points fails, with a Brauer–Manin obstruction to strong approximation (Theorem 1.8(1.5)) but not every failure explained by Brauer–Manin (Theorem 1.9). The proofs go through a desingularisation of U_n, a computation of Br U_n via a new general result on Brauer groups of surfaces with rational singularities (Theorem 1.7), and explicit local Hilbert-symbol computations including a 2-adic case analysis (Lemma 3.8).

Significance. If correct, the paper is a significant contribution to both the arithmetic of log K3 surfaces and the Erdős–Straus problem. It provides one of the few computations of the Brauer group of a singular surface, exhibits a transcendental Brauer–Manin obstruction to strong approximation on a surface of arithmetic interest, and yields an explicit, falsifiable necessary condition (Theorem 1.2) that recovers and unifies earlier results of Yamamoto and Elsholtz–Tao. The main derivation is self-contained and checkable: the Brauer element is constructed from the defining equation, the local invariants are computed by explicit Hensel lifting and Hilbert-symbol formulae, and no parameters are fitted to data. Theorem 1.7 on rational surface singularities is of independent interest. The only concrete error I found is the n=1 edge case in Theorem 1.8(1.4), which lies outside the Erdős–Straus range n≥2 and is easily corrected.

major comments (1)

[Theorem 1.8, Eq. (1.4)] The assertion (U_n(R)_+ × ∏ U_n(Z_p))^Br ≠ ∅ is false for n=1. For n=1, Lemma 3.5 applies to every odd prime (since p ∤ 2n) and Lemma 3.8 applies to p=2, so inv_p α = 1 for all p and all u ∈ U_1(Z_p); Lemma 3.1 gives inv_∞ α = −1 on U_1(R)_+. Hence for every point of U_1(R)_+ × ∏ U_1(Z_p) the product of local invariants is −1, so no point lies in the right kernel of the Brauer pairing and the Brauer set is empty. This contradicts (1.4). The error is an edge case outside the Erdős–Straus range n ≥ 2; the statement should be restricted to n ≥ 2, and the proof of Theorem 1.8 should be adjusted accordingly.

minor comments (3)

[Abstract] The abstract contains 'stro ng approximation' with an erroneous space; it should read 'strong approximation'.
[Lemma 3.8] The table of cases in Lemma 3.8 is asserted without explanation; the rows for (r1,r2) ≡ (5,1) and (7,3) mod 8 are omitted by symmetry, which should be stated, and a brief derivation of how the entries follow from equation (3.4) would aid the reader.
[Proof of Theorem 1.8] The first sentence of the proof, 'as U_n(Z)≠∅ and n>0, we have (3.5)≠∅', is terse: for n≥2 one should explicitly use the known integer solution for the p-adic components and the positive real solution u_i = 3n/4 for the real component.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity found; the central claim is derived from Hilbert reciprocity and explicit local computations, not fitted or assumed.

full rationale

The paper's central result, Theorem 1.2, is derived rather than assumed: it follows from Hilbert reciprocity for the quaternion algebra (-u1/u3, -u2/u3), the real invariant computed in Lemma 3.1, and the local computations in Lemmas 3.5 and 3.8 showing trivial invariants at places not dividing n. No parameter is fitted to solutions, and the product condition is not introduced as an ansatz. The Brauer group computation in Theorem 1.6 is obtained from a geometric analysis of the desingularisation, the Picard group calculation, and standard purity facts; it does not presuppose the Erdős–Straus conjecture or the product formula. Theorem 1.7, while novel, rests on external results of Artin and Lipman on rational singularities, and Remark 2.11 explicitly explains how one could work on the desingularisation instead, so the identification Br Un = Br of the desingularisation is not load-bearing for the local invariant argument. The recovery of Yamamoto's conditions and Elsholtz–Tao results in Corollaries 1.3 and 1.4 is presented as a consequence of Theorem 1.2, not as input, so it constitutes independent confirmation rather than circularity. The few self-citations present (e.g., [17]) are contextual or methodological and do not carry the argument. The possible n=1 edge-case failure of Theorem 1.8 identified in review is a correctness point outside the Erdős–Straus range n ≥ 2 and does not indicate circularity. Overall, the derivation chain is self-contained relative to standard Hilbert symbol and Brauer group facts.

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

No free parameters are fitted and no new entities are invented. All inputs are standard theorems from algebraic geometry, Galois cohomology, and local number theory, plus the definitions of the Erdos-Straus surface and the quaternion algebra constructed from its equation. The Brauer group computations rely on external results about rational singularities and purity, which are cited and used with their hypotheses checked.

assumptions (5)

standard math Hilbert reciprocity law for the Hilbert symbol over Q
Used in Section 3.4 to combine local invariants into the global product condition of Theorem 1.2.
standard math Grothendieck purity and the Leray spectral sequence for the sheaf G_m
Used in Section 2.5.1 to compute Br of the desingularisation from the torus V_n, and in Theorem 1.7 to pass from Pic of the exceptional fibre to Brauer groups.
standard math Rational singularity structure of the exceptional fibre as a tree of P^1s, after Artin and Lipman
Used in the proof of Theorem 1.7 in Section 2.5.2 to compute R^1 f_* G_m and show that Br U maps surjectively to Br of the desingularisation.
standard math Hensel's lemma in the stated nondegenerate situations
Used in Lemmas 3.6 and 3.7 to lift local solutions modulo p or modulo 8 to Z_p points with prescribed Hilbert symbols.
domain assumption Lang-Weil estimates and the lemma [6, Lem. 6.5] on dense images imply density of integral points
Used in Section 3.9 to prove Theorem 1.9 that the Brauer-Manin set is not enough for strong approximation. These are external results whose hypotheses are checked for the desingularisation.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Brauer-Manin obstruction for Erd\H{o}s-Straus surfaces." pith.science (2026). https://pith.science/paper/5PEZ24FG

@misc{pith2026190802526,
  author       = {Pith},
  title        = {Pith review of: Brauer-Manin obstruction for Erd\Hos-Straus surfaces},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5PEZ24FG}},
  note         = {Machine review of arXiv:1908.02526}
}

read the original abstract

We study the failure of the integral Hasse principle and strong approximation for the Erd\H{o}s-Straus conjecture using the Brauer-Manin obstruction.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 24 canonical work pages

[1]

Artin, On isolated rational singularities of surface s, Amer

M. Artin, On isolated rational singularities of surface s, Amer. J. Math. 88 (1966), no. 1, 129–136

work page 1966
[2]

Bright, Brauer groups of singular del Pezzo surfaces, Michigan Math

M. Bright, Brauer groups of singular del Pezzo surfaces, Michigan Math. J. , 62(3) (2013), 657–664

work page 2013
[3]

Y. Cao, F. Xu, Strong approximation with Brauer-Manin ob struction for toric varieties, Ann. Inst. Fourier , 68(5) (2018), 1879–1908

work page 2018
[4]

Colliot-Thélène, A

J.-L. Colliot-Thélène, A. Skorobogatov, The Brauer–Grothendieck group , http://wwwf.imperial.ac.uk/~anskor/brauer.pdf

work page
[5]

Colliot-Thélène, O

J.-L. Colliot-Thélène, O. Wittenberg, Groupe de Brauer et points entiers de deux familles de surfaces cubiques aﬃnes, Amer. J. Math. 134 (2012), no. 5, 1303–1327

work page 2012
[6]

Colliot-Thélène, D

J.-L. Colliot-Thélène, D. Wei, F. Xu, Brauer–Manin obst ruction for Markoﬀ surfaces, Annali della Scuola Normale di Pisa , to appear

work page
[7]

Colliot-Thélène, F

J.-L. Colliot-Thélène, F. Xu, Strong approximation for the total space of certain quadric ﬁbrations, Acta Arith. 157 (2013), 169–199

work page 2013
[8]

Elsholtz, T

C. Elsholtz, T. Tao, Counting the number of solutions to t he Erdős–Straus equation on unit fractions, J. Aust. Math. Soc. 94 (2013), no. 1, 50–105

work page 2013

Show all 24 references

[9]

Grothendieck, Le groupe de Brauer, III

A. Grothendieck, Le groupe de Brauer, III. In Dix Exposés sur la Cohomologie des Sché- mas, North-Holland, 1968

1968
[10]

Gille, T

P. Gille, T. Szamuely, Central Simple Algebras and Galois Cohomology , Cambridge studies in advanced mathematics vol. 101, Cambridge University Pre ss, Cambridge, 2006

2006
[11]

Harpaz, Geometry and arithmetic of certain log K3 sur faces

Y. Harpaz, Geometry and arithmetic of certain log K3 sur faces. Ann. Inst. Fourier 67 (2017), no. 5, 2167–2200

2017
[12]

Harpaz, Integral points on conic log K3 surfaces

Y. Harpaz, Integral points on conic log K3 surfaces. J. Eur. Math. Soc. 21 (2019), no. 3, 627–664

2019
[13]

Jahnel, D

J. Jahnel, D. Schindler, On integral points on degree fo ur del Pezzo surfaces, Israel J. Math. 222 (2017), no. 1, 21–62

2017
[14]

J. H. Jaroma, On expanding 4/n into three Egyptian fractions, Crux Mathematicorum, 30(1) (2014), 36–37

2014
[15]

S. Lang, A. Weil, Number of points of varieties in ﬁnite ﬁ elds. Amer. J. Math. 76 (1954), 819–827

1954
[16]

Lipman, Rational singularities, with applications to algebraic surfaces and unique fac- torization

J. Lipman, Rational singularities, with applications to algebraic surfaces and unique fac- torization. Inst. Hautes Études Sci. Publ. Math. 36, (1969), 195–279

1969
[17]

Loughran, V

D. Loughran, V. Mitankin, Integral Hasse principle and strong approximation for Markoﬀ surfaces, Int. Math. Res. Not. , to appear
[18]

J. S. Milne, Étale cohomology. Princeton University Press, Princeton, New Jersey, 1980

1980
[19]

L. J. Mordell, Diophantine equations. Pure and Applied Mathematics, Vol. 30 Academic Press, London-New York 1969

1969
[20]

Poonen, Rational points on varieties, Graduate Studies in Mathematics, vol

B. Poonen, Rational points on varieties, Graduate Studies in Mathematics, vol. 186, Amer- ican Mathematical Society, Providence, RI, 2017

2017
[21]

Serre, A course in arithmetic , Graduate Texts in Mathematics, No

J.-P. Serre, A course in arithmetic , Graduate Texts in Mathematics, No. 7. Springer- Verlag, 1973. BRAUER–MANIN OBSTRUCTION FOR ERDŐS–STRAUS SURF ACES 17

1973
[22]

Sierpiński, Sur les décompositions de nombres ratio nnels en fractions primaires, Math- esis 65 (1956), 16–32,

W. Sierpiński, Sur les décompositions de nombres ratio nnels en fractions primaires, Math- esis 65 (1956), 16–32,

1956
[23]

Wei, Strong approximation for a toric variety, arXiv:1403.1035

D. Wei, Strong approximation for a toric variety, arXiv:1403.1035

arXiv
[24]

Yamamoto, On the Diophantine equation 4 n = 1 x + 1 y + 1 z , Mem

K. Yamamoto, On the Diophantine equation 4 n = 1 x + 1 y + 1 z , Mem. Fac. Sci. Kyushu Univ. Ser. A 19 (1965), 37–47. Mathematisch Instituut, Niels Bohr weg 1, 2333 CA Leiden, Neth erlands E-mail address : m.j.bright@math.leidenuniv.nl Department of Mathematical Sciences, Univ...

1965

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Artin, On isolated rational singularities of surface s, Amer

M. Artin, On isolated rational singularities of surface s, Amer. J. Math. 88 (1966), no. 1, 129–136

work page 1966

[2] [2]

Bright, Brauer groups of singular del Pezzo surfaces, Michigan Math

M. Bright, Brauer groups of singular del Pezzo surfaces, Michigan Math. J. , 62(3) (2013), 657–664

work page 2013

[3] [3]

Y. Cao, F. Xu, Strong approximation with Brauer-Manin ob struction for toric varieties, Ann. Inst. Fourier , 68(5) (2018), 1879–1908

work page 2018

[4] [4]

Colliot-Thélène, A

J.-L. Colliot-Thélène, A. Skorobogatov, The Brauer–Grothendieck group , http://wwwf.imperial.ac.uk/~anskor/brauer.pdf

work page

[5] [5]

Colliot-Thélène, O

J.-L. Colliot-Thélène, O. Wittenberg, Groupe de Brauer et points entiers de deux familles de surfaces cubiques aﬃnes, Amer. J. Math. 134 (2012), no. 5, 1303–1327

work page 2012

[6] [6]

Colliot-Thélène, D

J.-L. Colliot-Thélène, D. Wei, F. Xu, Brauer–Manin obst ruction for Markoﬀ surfaces, Annali della Scuola Normale di Pisa , to appear

work page

[7] [7]

Colliot-Thélène, F

J.-L. Colliot-Thélène, F. Xu, Strong approximation for the total space of certain quadric ﬁbrations, Acta Arith. 157 (2013), 169–199

work page 2013

[8] [8]

Elsholtz, T

C. Elsholtz, T. Tao, Counting the number of solutions to t he Erdős–Straus equation on unit fractions, J. Aust. Math. Soc. 94 (2013), no. 1, 50–105

work page 2013