{"id":"7f9aed6e-6234-463e-8f91-85c897af3503","arxiv_id":"2508.17332","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The maximal abelian cover of a regular multi-graph has no flat-band eigenvalues, characterized completely by roots of generalized matching polynomials.","lead":"This paper gives a complete rule for when an infinite periodic graph built from a finite graph has special flat band eigenvalues, and proves that regular graphs never produce them. It settles a 2009 conjecture about the spectra of these periodic covers.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.2 is false as stated for regular graphs of degree 0 and 1; the proof only covers degree at least 2, so the theorem needs an explicit degree lower bound.","rationale":"I agree with the CONDITIONAL verdict, but the most load-bearing issue is different from the one named as the reader's weakest assumption. The reader's weakest_assumption is the unproved bridge-less check in Proposition 4.8; while that is a real omitted proof, a short argument (the deleted edge lies on a cycle, and the new gadget reconnects the endpoints) shows the check is true, so it does not threaten the main result for d≥3. The direct threat is the statement of Theorem 1.2 itself: K2 is 1-regular and its maximal abelian cover is K2, which has eigenvalues. Hence the headline theorem, as written, is false. This is a boundary-condition correction, not a refutation of the intended result for regular graphs of degree at least 2, so the appropriate verdict remains CONDITIONAL/UNCHANGED rather than REJECT: the authors must add 'degree at least 2' and confirm the proof covers exactly that range. The reader's rationale already noted the degree 0/1 problem, but because the formal weakest_assumption field points elsewhere, I mark agreement as partial.","tokens_in":23864,"tokens_out":23216,"duration_ms":222962,"concrete_test":"Apply the paper's own criterion, Theorem 1.1, to G=K2: the only degree-2 subgraph is the empty subgraph, so the criterion requires λ to be a root of m_G(λ)=(λ-V_a)(λ-V_b)-|w|^2. For the adjacency operator this gives λ=±1, and indeed Hab=K2 has eigenvalues ±1. This explicit computation falsifies Theorem 1.2 exactly as stated. A second verification is to insert the hypothesis deg(G)≥2 and check that the proof's cases then cover every regular graph: even d≥2 by the 2-factor argument and odd d≥3 by Lemma 4.11 plus the bridge-less case via Lemma 4.6.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim, Theorem 1.2, is stated without a lower bound on the degree: 'If G is regular, then Hab has no eigenvalues.' This is false for degree 1. Let G=K2. Its fundamental group is trivial, so the maximal abelian cover Gab is K2 itself; the pullback Hab is a 2x2 self-adjoint matrix with eigenvalues ±1 for the adjacency operator. Thus the theorem as printed has a genuine counterexample. The proof never covers this case: even degrees are handled by 2-factors (which requires d≥2) and odd degrees by the whole Section 4 machinery with d≥3 odd. In particular, Lemma 4.6 and Lemma 4.11 both assume d≥3, and the final proof of Theorem 1.2 starts with 'any odd degree regular graph G that is not bridge-less,' excluding d=1. The theorem should be restated with 'degree at least 2' in the abstract, introduction, and Theorem 1.2. The bridge-less check in Proposition 4.8 is also asserted without proof, but that is a repairable omitted justification rather than a counterexample to the claimed result.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the maximal abelian cover Gab of a finite multi-graph G and characterizes the eigenvalues (flat bands) of the pullback Hab of a Schrödinger operator in terms of the combinatorics of the base graph. Theorem 1.1 states that λ is an eigenvalue of Hab if and only if λ is a root of the generalized matching polynomial of G\\γ for every degree-2 subgraph γ of G. The proof uses a periodic graph Gper whose connected components are copies of Gab, Floquet theory, and a determinant expansion in terms of oriented degree-2 subgraphs and matchings. Theorem 1.2 then claims that if G is regular, Hab has no eigenvalues, resolving a conjecture of Higuchi and Nomura. The proof treats even-regular graphs by 2-factors and odd-regular graphs by an induction on bridge-blocks, using a new combinatorial result (Proposition 4.8) on degree-2 subgraphs in bridge-less blocks. An appendix shows that the Banks–Garza-Vargas–Mukherjee criterion for eigenvalues of the universal cover implies the new criterion, so that universal-cover eigenvalues are also eigenvalues of the maximal abelian cover.","tokens_in":24076,"tokens_out":16641,"duration_ms":159957,"significance":"If the degree qualification is fixed, this is a substantial contribution: it provides a complete combinatorial characterization of flat bands for maximal abelian covers and confirms a conjecture for regular graphs of degree at least 2. The determinant expansion in Proposition 3.8 and the generalized matching polynomial framework are clean and likely to be useful beyond this paper. The proof of the odd-regular case is a delicate and novel induction, and the appendix gives a useful comparison with the existing universal-cover criterion. The proofs are detailed and the main line is convincing, but the false statement for degree 0 and 1 and the unproved bridge-less preservation claim need to be addressed.","major_comments":[{"comment":"Theorem 1.2 as stated is false for regular graphs of degree 0 and 1. For G=K2, the fundamental group is trivial, so Gab is isomorphic to G and the adjacency operator on Gab has eigenvalues ±1. Similarly, a single vertex with no edges is 0-regular and its maximal abelian cover has eigenvalue 0. The proof of Theorem 1.2 only covers degrees where a 2-factor can be used for even d (requiring d≥2) and the Section 4 machinery for odd d≥3 (Lemmas 4.6, 4.7, and 4.11 all assume d≥3). The theorem, the abstract, and the introductory claims should be restated for regular multi-graphs of degree at least 2.","section":"Section 1, Theorem 1.2 (and Abstract)"},{"comment":"The assertion after Figure 4 that the degree-correction procedure yields a graph G′ that is 'also bridge-less' is not proved. This property is load-bearing because the induction hypothesis is applied to G′. In the even-degree subcase, after removing the edge {v1,u}, the new vertex v′1 is attached by multiple parallel edges, and the proof should verify that no bridge is created in the resulting graph. Please provide a rigorous proof of the preservation of the bridge-less property, or identify precisely which configurations of parallel edges and self-loops could make the claim false and how they are excluded.","section":"Section 4.2, proof of Proposition 4.8"},{"comment":"The proof of Theorem 1.2 begins with 'Given any odd degree regular graph G that is not bridge-less', but it does not explicitly treat odd-regular bridge-less graphs. A bridge-less d-regular graph has 0 ≤ d−1 bridges, so Lemma 4.6 supplies a 2-factor and Theorem 1.1 already gives the conclusion. This case should be stated explicitly in the proof; without it, the written proof does not cover all odd-regular graphs.","section":"Section 4.4, proof of Theorem 1.2"}],"minor_comments":[{"comment":"In the displayed formula for A(z), the second term should read z_e^{-1} w_e A_{e^{-1}} (or equivalently \\(\\bar z_e w_e A_{e^{-1}}\\)), not z_e w_e A_{e^{-1}}. As printed, the formula contradicts the later convention z_{e^{-1}} = z_e^{-1} used in the proof of Theorem 1.1.","section":"Section 2.3, definition of A(z)"},{"comment":"The statements 'the maximal abelian cover of any regular multi-graph has no eigenvalues' (Abstract) and 'If G is regular, then Hab has no eigenvalues' (Theorem 1.2) should be qualified to regular graphs of degree at least 2, both in the abstract and in every summarizing sentence in the introduction.","section":"Abstract and Introduction"},{"comment":"The step applying Proposition 4.8 to leaf-blocks says 'with I containing only the vertex v incident to the outgoing bridge'; it would be helpful to state explicitly why this block is Type II (maximum degree d and bridge-less by Lemma 4.3), so that the hypotheses of Proposition 4.8 are visibly satisfied.","section":"Section 4.4, final paragraph"},{"comment":"The proposition states 'maximum degree d' but the proof uses Lemma 4.7 with 'maximum degree at most d'. In the context of Type II blocks the maximum degree is exactly d, but for the induction in Proposition 4.8 the intermediate graphs may have maximum degree less than d; please clarify the exact degree hypothesis in both statements.","section":"Section 4.2, Proposition 4.8 statement"}],"recommendation":"major_revision","confidential_remarks":"The counterexample at degree 1 is decisive and the theorem statement must be corrected before acceptance. The bridge-less preservation claim in Proposition 4.8 should be written out in full, and the bridge-less odd-regular case should be inserted into the proof of Theorem 1.2. Once these fixes are made, the main results appear sound and the paper is likely to be a strong contribution to the spectral theory of periodic graphs."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is worth taking seriously. Theorem 1.1 gives a complete characterization of flat bands of the maximal abelian cover in terms of roots of generalized matching polynomials of G with all degree-2 subgraphs removed. The Floquet/determinant proof is sound: expanding det(lambda - H(z)) and using orthogonality of the z^gamma characters is a neat way to turn a periodic operator problem into a finite matching-polynomial condition. Theorem 1.2 confirms Higuchi-Nomura's Conjecture 6.12 for regular multi-graphs of degree at least 2. The Section 4 induction is intricate and original; the degree-correction construction in Proposition 4.8 is a genuinely nice tool, and the way leaf-blocks are collapsed to leaves in Lemma 4.11 is convincing. The appendix connecting the B-GV-M universal-cover criterion to the abelian-cover criterion is a useful extra. No circularity, no invented entities, and the citations look right.\n\nNow the soft spots, in order of importance. First, Theorem 1.2 as stated is false. \"If G is regular\" includes degree 1: take G=K2. The fundamental group is trivial, the maximal abelian cover is K2 itself, and the adjacency operator has eigenvalues +/-1. The proof simply does not treat d<=1: even d needs a 2-factor (so d>=2), and the odd case assumes d>=3 in Lemma 4.6/4.11. The fix is cosmetic---add 'degree at least 2' to the abstract, introduction, and theorem---but it must be made. Second, in Proposition 4.8 the degree-correction step asserts that G' is also bridge-less with the words \"it is straightforward to check.\" That claim is load-bearing for the induction and is not proved. I don't think it is false, but a referee should ask for the argument. Minor: the empty graph's matching polynomial convention could be stated explicitly, since the 2-factor case relies on it.\n\nThe central ideas hold up; the problems are boundary cases and an unproved but plausible claim. This is a solid paper that deserves a serious referee, not a desk reject. I would send it out and ask for the statement correction and the Proposition 4.8 justification.","headline":"Returns a clean criterion and proves a long-open conjecture for regular graphs of degree at least 2, but the main theorem as stated is false for degree 0 and 1—a cheap fix.","tokens_in":24614,"tokens_out":3290,"would_cite":true,"duration_ms":34473,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05C50","47B39","05C70"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper gives a complete combinatorial description of the eigenvalues of a maximal abelian cover, and proves that regular base graphs have none.","keywords":["maximal abelian cover","flat bands","matching polynomial","regular graph","periodic Schrödinger operator","2-factor","bridge-block tree","eigenvalue criterion"],"falsifier":"For a fixed small regular multi-graph G with bridges, the criterion of Theorem 1.1 makes the question finite: enumerate every degree-2 subgraph γ, compute the generalized matching polynomials m^H_{G\\γ}, and check whether their root sets share a real number. A shared root would be an eigenvalue of Hab by the paper's own criterion, refuting Theorem 1.2. A natural candidate class to search is odd-regular multi-graphs assembled from several bridge-blocks, where Proposition 4.8 is the inductive engine.","tokens_in":23640,"feed_emoji":"🕸️","tokens_out":5579,"duration_ms":55128,"temperature":0.7,"pith_summary":"This paper pinpoints exactly when the maximal abelian cover of a finite multi-graph has eigenvalues, the so-called flat bands of the associated periodic Schrödinger operator. The answer is combinatorial: a real number is an eigenvalue precisely when it is a root of the generalized matching polynomial of the graph with every degree-2 subgraph removed. Using this criterion, the paper proves a 2009 conjecture of Higuchi and Nomura: if the base graph is regular, allowing multi-edges and self-loops, then its maximal abelian cover has no eigenvalues at all. This matters because flat bands are the obstruction to purely continuous spectra in periodic quantum graphs, and regular covers are among the most common periodic structures.","feed_headline":"Maximal abelian covers of regular graphs have no eigenvalues","feed_subtitle":"A matching-polynomial criterion identifies every flat band, settling a 2009 conjecture.","key_machinery":"The load-bearing object is the generalized matching polynomial m^H_G, a weighted sum over matchings, together with the set of oriented degree-2 subgraphs. Proposition 3.8 expands det(λI − H(z)) as a trigonometric polynomial whose coefficients are exactly m^H_{G\\γ}, and orthogonality of the characters z^γ in $L^{2}$($T^{{|E|}}$) forces every coefficient to vanish, yielding the criterion of Theorem 1.1. The proof of Theorem 1.2 rests on Proposition 4.8, which uses a degree-correction construction and a 2-factor theorem to build degree-2 subgraphs covering all high-degree vertices in bridge-less blocks of a regular graph, and on two lemmas that eliminate matching-polynomial roots by deleting leaves.","core_discovery":"The central discovery is a complete characterization: for a finite multi-graph G and a Schrödinger operator H, a real number λ is an eigenvalue of the maximal abelian cover Hab if and only if for every degree-2 subgraph γ, λ is a root of the generalized matching polynomial m^H_{G\\γ} of the graph obtained by deleting γ. This is Theorem 1.1, and it solves the Higuchi–Nomura problem. The paper then derives its main theorem: if G is regular, then Hab has no eigenvalues, confirming the corresponding conjecture. An appendix shows that whenever the universal cover has an eigenvalue according to the Banks–Garza-Vargas–Mukherjee criterion, the same value is also an eigenvalue of Hab.","pith_inferences":["The criterion suggests a concrete finite algorithm: enumerate degree-2 subgraphs, compute the generalized matching polynomials, and intersect their root sets; the open bi-regular case could be probed this way.","If the authors' open question—whether every eigenvalue of Hab is also an eigenvalue of the universal cover—has an affirmative answer, then Theorem 1.1 and the Banks–Garza-Vargas–Mukherjee criterion would be equivalent, unifying flat bands of abelian covers with point spectrum on trees.","The degree-correction tool behind Proposition 4.8 may be useful for other factor-existence questions in multi-graphs, since it adjusts vertex degrees while preserving the bridge-less property.","The leaf-attachment and self-loop corollaries suggest that the absence of eigenvalues is governed by a combination of degree bounds and bridge structure, rather than by regularity itself."],"forward_implications":["If G is even regular, Petersen's theorem supplies a 2-factor, so Theorem 1.1 immediately rules out eigenvalues; the new content lies in the odd-regular, bridge-heavy case.","All eigenvalues of Hab, when they exist, lie inside the Ramanujan bound: |λ| ≤ ρ where ρ is the spectral radius of the universal-cover operator.","Any graph satisfying the Banks–Garza-Vargas–Mukherjee criterion for a universal-cover eigenvalue also has the same eigenvalue in its maximal abelian cover.","The criterion turns the question of flat bands into a finite check: one must examine the roots of generalized matching polynomials associated to all degree-2 subgraphs."],"supporting_citations":[{"why":"Higuchi and Nomura posed the problem and conjecture that this paper solves.","marker":"[10]"},{"why":"Supplies the Floquet–Bloch lemmas used to relate eigenvalues of the periodic graph to the matrices H(z).","marker":"[17]"},{"why":"Heilmann–Lieb matching-polynomial root bounds, extended here to give the Ramanujan bound for eigenvalues of Hab.","marker":"[9]"},{"why":"The Banks–Garza-Vargas–Mukherjee criterion for universal-cover eigenvalues, which Appendix A compares to Theorem 1.1.","marker":"[4]"},{"why":"Kostochka et al. theorem bounding the number of bridges in regular multi-graphs that still admit a 2-factor, used in Lemma 4.6.","marker":"[14]"},{"why":"Arizmendi et al. result that universal-cover eigenvalues lift to the maximal abelian cover, motivating the appendix's comparison.","marker":"[2]"}],"fun_headline_variants":["No eigenvalues in abelian covers of regular graphs","Flat bands fully characterized for maximal abelian covers","Matching polynomial criterion settles abelian cover eigenvalues","Higuchi-Nomura problem resolved for maximal abelian covers"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument's load-bearing premise is that the degree-correction operation in Proposition 4.8 never creates a bridge where none existed; the paper says this is straightforward to check, and if it fails for some arrangement of multi-edges or self-loops, the induction breaks.","fun_headline_variants_meta":{"raw":{"variants":["No eigenvalues in abelian covers of regular graphs","Flat bands fully characterized for maximal abelian covers","Matching polynomial criterion settles abelian cover eigenvalues","Higuchi-Nomura problem resolved for maximal abelian covers"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000559,"raw_usage":{"total_tokens":2571,"prompt_tokens":774,"completion_tokens":1797,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":390,"completion_tokens_details":{"reasoning_tokens":1735}},"tokens_in":390,"tokens_out":1797,"duration_ms":15892,"temperature":1.0,"reasoning_tokens":1735,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T17:08:00.992680+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a fixed small regular multi-graph G with bridges, the criterion of Theorem 1.1 makes the question finite: enumerate every degree-2 subgraph γ, compute the generalized matching polynomials m^H_{G\\γ}, and check whether their root sets share a real number. A shared root would be an eigenvalue of Hab by the paper's own criterion, refuting Theorem 1.2. A natural candidate class to search is odd-regular multi-graphs assembled from several bridge-blocks, where Proposition 4.8 is the inductive engine.","supporting_citations":[{"cited_title":"Higuchi and Y","cited_arxiv_id":null,"evidence_quote":"Higuchi and Nomura posed the problem and conjecture that this paper solves."},{"cited_title":"Sabri and P","cited_arxiv_id":null,"evidence_quote":"Supplies the Floquet–Bloch lemmas used to relate eigenvalues of the periodic graph to the matrices H(z)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Heilmann–Lieb matching-polynomial root bounds, extended here to give the Ramanujan bound for eigenvalues of Hab."},{"cited_title":"Banks, J","cited_arxiv_id":null,"evidence_quote":"The Banks–Garza-Vargas–Mukherjee criterion for universal-cover eigenvalues, which Appendix A compares to Theorem 1.1."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Kostochka et al. theorem bounding the number of bridges in regular multi-graphs that still admit a 2-factor, used in Lemma 4.6."},{"cited_title":"Arizmendi, G","cited_arxiv_id":null,"evidence_quote":"Arizmendi et al. result that universal-cover eigenvalues lift to the maximal abelian cover, motivating the appendix's comparison."}],"review_version":1}