{"id":"dc5e9a01-785f-4369-ba32-7f357625b4fb","arxiv_id":"1908.11577","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For small area-constrained Willmore surfaces, an adapted geometric center of mass satisfies |∇Sc(p0)| ≤ C area, improving the previous C sqrt(area) bound and implying the enclosed region contains a single critical point of scalar curvature.","lead":"This paper proves a sharper location estimate for small area-constrained Willmore surfaces in three-dimensional Riemannian manifolds: the scalar-curvature gradient at a suitably chosen center of mass is bounded by the surface area, one power better than earlier bounds. The result implies that these surfaces enclose a single critical point of scalar curvature under a Morse assumption.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 7.1 hinges on an unshown parity cancellation in §7.2–7.3: after frame contraction every term in δU and δV must contain an odd number of factors ν or y/R, but the mixed term in (32) is only asserted, not verified.","rationale":"The reader's weakest assumption correctly identifies the load-bearing point: the estimates (33) and (36) are what convert the variation computation into |∇ Sc(p0)| ≤ C R^2, and these estimates depend entirely on the claim that every surviving top-order integrand has odd parity with respect to the factors ν and y/R. I checked the structure of the principal terms and the parity argument is plausible: the explicit resolutions of |ω|² and |T°|² in §7.2 do produce odd factors once multiplied by g(b,ν), and the frame-completeness relation converts products of tangent-frame vectors into δ^{αβ} - ν^α ν^β, which preserves oddness. However, the paper does not carry out this resolution for all mixed terms, especially ω_i g(b,e_j) T°_{ij} and the error terms from replacing Ric by Ric(0) and g by δ. Because the theorem's quantitative gain from R to R² depends on these cancellations being exact, the proof is incomplete as printed. The other issues noted by the reader, the sign discrepancy near (17) and the integration-by-parts display after (23), appear to be repairable: Proposition 5.3 and 5.4 can be obtained from the L² error bounds with constants that still supply the needed R-powers, and the displayed inequality in Proposition 5.4 can be replaced by a Young-inequality argument using ‖E‖_{L²} ≤ C R. Those are expositional defects rather than threats to the architecture. The center-of-mass construction and Corollary 3.3 are sound, and the final volume comparison Vol(Ω) ≥ c R^3 is standard. Thus no change to the reader's conditional verdict is warranted; the paper should be accepted only after the omitted parity computation is supplied explicitly.","tokens_in":18755,"tokens_out":21595,"duration_ms":195975,"concrete_test":"Expand (32) and the analogous expression in (34) to order R^4, using the normal-coordinate expansions g = δ + O(|y|^2), Ric = Ric(0) + O(|y|), ∇b from (24), and the frame completeness relation ∑_i e_i^α e_i^β = g^{αβ} - ν^α ν^β. Enumerate every monomial in the variables {ν^α, y^β/R} that survives after contraction, and verify that each contains an odd number of these factors and that the total R^4 coefficient vanishes. This check can be done by hand or with a computer-algebra parity routine; if any even monomial with nonzero coefficient appears, Theorem 7.1(iii) fails as stated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of Theorem 7.1(iii) reduces to the estimates |δ_f U| ≤ C R^5 in (33), |δ_f V| ≤ C R^5 after (36), and |λ∫ fH| ≤ C R^5 in (29). The δU and δV estimates are not actually demonstrated. After substituting the top-order relations H^{-1} ≈ R/2, A° ≈ -(4/3)H^{-1}T°, ∇H ≈ (2/3)ω, and div A° ≈ (4/3)ω, the text in §7.2 asserts that all tangential contractions can be resolved and that every remaining term is a product of an odd number of factors ν or y/R, so Corollary 3.3 applies. No explicit resolution is given for the mixed terms, for example -(1/3)R^2 ω_i g(b,e_j) T°_{ij} in (32), nor for the error terms generated by replacing Ric and g by Ric(0) and δ. If any such term contains an even number of normal or position factors after full frame contraction, it would contribute at order R^4 rather than R^5, and the conclusion would be only |∇ Sc(p0)| ≤ C R, not C R^2. Thus the central claim rests on an unverified parity cancellation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies compact two-sided surfaces in a three-dimensional Riemannian manifold with CB-bounded geometry that are critical points of the Willmore functional under an area constraint. For surfaces with small area and Willmore energy below 4π+a0, the author derives refined curvature corrections to the intrinsic geometry (Propositions 5.1, 5.3, 5.4) and, using a geometrically defined center of mass introduced in Section 3, proves Theorem 7.1: there is a point p0 with |dist(p0,x)-R(Σ)| ≤ C R(Σ)^3 for all x∈Σ, with ∫ y dμg = 0 in normal coordinates, and with |∇Sc(p0)| ≤ C R(Σ)^2. This improves the earlier bound |∇Sc| ≤ C R(Σ). The paper closes with corollaries asserting that small solutions enclose a unique nondegenerate critical point of the scalar curvature.","tokens_in":19013,"tokens_out":30010,"duration_ms":272019,"significance":"If correct, the estimate |∇Sc(p0)| ≤ C area is a genuine improvement over the square-root bound and leads to the attractive geometric conclusion that area-constrained Willmore minimizers enclose points where the scalar curvature is maximal. The paper's architecture is coherent: the curvature corrections are parameter-free, the geometric center-of-mass construction is natural, and the main theorem gives explicit, falsifiable quantitative claims. The proof relies on previously published results as black boxes, but the new theorem does not reduce to those results by construction, and no fitting or self-referential normalization is introduced. The main risk is the completeness of the parity-cancellation argument in Sections 7.2–7.3, which is load-bearing for Theorem 7.1(iii).","major_comments":[{"comment":"The estimates |δ_f U| ≤ C R^5 and |δ_f V| ≤ C R^5 rest on the asserted parity cancellation: after resolving all tangential contractions, every surviving integrand must contain an odd number of factors chosen from {ν^α, y^α/R}. The text demonstrates this only for the first term in (32) and for the |ω|^2 and |T°|^2 terms, and states that the mixed term -(1/3)R^2 ω_i g(b,e_j) T°_{ij} and the term R g(∇_i b,e_j) T°_{ij} \"have a similar structure.\" These are exactly the terms where an even number of normal or position factors could survive, and the bracket in (34) is dismissed with \"by inspection.\" I checked that the contraction identity Σ_i e_i^α e_i^β = δ^{αβ} - ν^α ν^β together with the expansion (24) does indeed give an odd total number of factors for these terms, so the claim is true; nevertheless the proof as written does not provide this verification. Because Theorem 7.1(iii) would fail at order R^4 if any even-parity term contributed, this is a load-bearing step. The authors should include the explicit contraction computation for the mixed term in (32) and for the analogous terms in (34), at least in an appendix.","section":"§7.2, eqs. (32)–(33); §7.3 after eq. (34)"}],"minor_comments":[{"comment":"The displayed identity is algebraically incorrect: with Δ Ric(ν,ν) = -(3/2)H^2 Ric(ν,ν) + (1/2)H^2 Sc + O_{L2}(R^{-1}), the expression simplifies to -H^2(λ + (1/3)Sc) + O_{L2}(R^{-1}), not -H^2(λ + Sc). The final conclusion |Δw| ≤ C R^{-1} still follows from Proposition 2.5, which gives |λ + Sc/3| ≤ C R, so this is a typographical/algebraic slip rather than a fatal flaw, but it should be corrected.","section":"§5.1, eq. (17)"},{"comment":"The integration by parts step omits constants and a sign explanation. From ΔS = (1/2)H^2 S + E with ‖E‖_{L2} = O(1), one obtains ∫|∇S|^2 + (1/2)∫H^2|S|^2 = -∫⟨E,S⟩ ≤ C∫|S|, not directly ≤ ∫|S|. The missing constant is harmless for the subsequent absorption argument, but the displayed chain should be corrected.","section":"§5.3, after eq. (23)"},{"comment":"The replacement div A° ≈ (4/3)ω is used without derivation. It follows from the Codazzi equation ∇^i A_ij - ∇_j H = ω_j together with Proposition 5.3 (∇H ≈ (2/3)ω), but this should be stated explicitly because it is not quoted from earlier sections.","section":"§7.2, around eq. (32)"},{"comment":"The statements refer to \"the region enclosed by Σ\" without explicitly assuming or proving that Σ is embedded. The hypotheses allow immersed surfaces, and the proof uses Stokes' theorem on the enclosed region Ω. The author should either restrict the theorems to embedded surfaces or cite a result that the small-area, small-Willmore critical points under consideration are embedded.","section":"Theorems 1.4 and 7.1"},{"comment":"The notation g is used both for the ambient metric and for the quantity |∇Sc(p0)|. Renaming the latter would remove a source of confusion in a proof that otherwise depends on metric computations.","section":"§8, proof of Corollary 1.5"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take on Metzger's arXiv:1908.11577. The main theorem is a genuine step beyond [6]: with the adapted center of mass, the position estimate improves from CR to CR², and the corollary that a small constrained Willmore sphere encloses exactly one nondegenerate critical point of Sc is clean and worth having. The curvature corrections in Propositions 5.1–5.4 are new, parameter-free, and built from prior results without circularity; the black boxes from [6,7] are cited appropriately. The center-of-mass lemma is a good tool.\n\nThat said, I would not take the proof as printed. Equation (17) does not simplify the way the text claims: combining (1) with (11) gives -H²(λ + Sc/3), not -H²(λ + Sc), and the terms -H²|A°|² and |∇H|² that are swept into O(R^{-1}) are order one under the paper's own L∞ estimates (|A°| ≤ CR and |∇H| bounded). So the proof of Proposition 5.1 has a real gap. There may be a hidden cancellation, but it is not in the text, and the displayed calculation is not trustworthy as written.\n\nThe second soft spot is the O(R^5) control of δU and δV in §7.2–7.3. The whole argument hangs on the assertion that after frame contraction every surviving term has an odd number of ν or y/R factors. I checked the representative mixed term with ω_i T°_{ij} and, using γ = g - ν⊗ν to resolve tangent contractions, the parity does come out odd. So the concern is not obviously fatal. But the paper leaves this load-bearing cancellation as a matter of inspection, and the error terms from replacing Ric and g by their values at p0 are not written down. That should be a real lemma, not a remark.\n\nOne point flagged by another reader does not bother me: the integration by parts after (23) works. The boundary term moves to the left with the right sign, and the displayed estimate is fine after rearrangement.\n\nBottom line: this is a promising paper for geometric analysts working on Willmore-type surfaces and foliations. It deserves peer review, but a serious referee should ask for a repaired expansion in Proposition 5.1 and a written parity/cancellation argument. Without those, Theorem 7.1 is not established as printed.","headline":"A genuine R² improvement over the earlier R-position estimate, but the printed proof needs real repair: equation (17) is algebraically off and the key parity cancellation in §7.2 is asserted rather than shown.","tokens_in":19530,"tokens_out":14454,"would_cite":false,"duration_ms":147100,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53C42","53C21","58E12"],"pacs":[],"model":"deepseek-v4-flash","headline":"Every small area-constrained Willmore surface has an adapted center at which the scalar-curvature gradient is bounded by a constant times the surface area, improving the earlier square-root bound and, when scalar curvature is Morse…","keywords":["Willmore functional","area constraint","position estimates","scalar curvature","geometric center of mass","nearly umbilical surfaces","Riemannian 3-manifolds","Euler-Lagrange equation"],"falsifier":"Take a concrete nearly-spherical family, such as geodesic spheres of radius $R$ centered at a point with nonzero $\\nabla\\mathrm{Sc}$ in a fixed 3-manifold, and explicitly evaluate the variation identity $\\delta_f W(\\Sigma) = \\lambda \\int_\\Sigma fH\\, d\\mu$ for $f = H^{-1}g(b,\\nu)$ in normal coordinates centered at the adapted center of mass. If the terms dropped as $O(R^5)$ in equations (32) and (34) contain an even-factor integral that survives at order $C R^4$, the claimed cancellation fails and $|\\nabla\\mathrm{Sc}(p_0)| \\le C R^2$ would be violated.","tokens_in":91,"feed_emoji":"🎯","tokens_out":8668,"duration_ms":199669,"temperature":0.7,"pith_summary":"The paper studies closed surfaces that are critical points of the Willmore energy subject to fixed area, in a three-dimensional Riemannian manifold with uniformly bounded geometry. When the surface is small and its Willmore energy is close to $4\\pi$, the surface is nearly spherical, and the paper computes how ambient curvature shifts its mean curvature, trace-free second fundamental form, and gradient. The main result is that an adapted geometric center of mass can be chosen so that the surface lies within radius $R$ plus $O(R^3)$ of that center, and the gradient of the ambient scalar curvature at the center is $O(R^2)$, equivalently $O(|\\Sigma|)$. This improves the previous $O(R)$ square-root-area bound and, when scalar curvature is a Morse function, forces the region enclosed by the surface to contain exactly one critical point of scalar curvature. A reader should care because it sharpens the known concentration of minimizers near maxima of scalar curvature and gives a direct relation between surface area and the ambient curvature gradient.","feed_headline":"Gradient of scalar curvature at Willmore center is O(area)","feed_subtitle":"For area-constrained Willmore surfaces, the center of mass sits within radius cubed, improving the previous square-root bound.","key_machinery":"The load-bearing construction is an adapted geometric center of mass: $p_0$ minimizes $w(p) = \\int_\\Sigma \\mathrm{dist}(p,x)^2\\, d\\mu$, so in normal coordinates centered at $p_0$ the first moment $\\int_{\\psi(\\Sigma)} y\\, d\\mu_g$ vanishes and the Euclidean first moment is only $O(d^3 |\\Sigma|)$. From this, Corollary 3.3 controls every integral of an odd product of the functions $\\{\\nu_\\alpha, y_\\alpha/R\\}$ by $O(R^3 + R\\|\\mathring{A}\\|_{L^2})$, which is exactly what lets the paper discard the many curvature-correction terms as $O(R^5)$. The second ingredient is the expansion package $H \\approx 2/R$, $\\mathring{A} \\approx -\\tfrac43 H^{-1} \\mathring{T}$, $\\nabla H \\approx \\tfrac23 \\omega$, and $\\mathrm{div}\\mathring{A} \\approx \\tfrac43 \\omega$, derived from the Euler-Lagrange equation and the Simons identity, which converts the first variation into integrals to which Corollary 3.3 applies.","core_discovery":"For every solution $\\Sigma$ of the constrained Willmore Euler-Lagrange equation $\\Delta H + H|\\mathring{A}|^2 + H\\,\\mathrm{Ric}(\\nu,\\nu) + \\lambda H = 0$ with $|\\Sigma| \\le a_0$ and $W(\\Sigma) \\le 4\\pi + a_0$ in a $C_B$-bounded 3-manifold, there exists a point $p_0$ in the enclosed region with $|\\mathrm{dist}(p_0,x) - R(\\Sigma)| \\le C R(\\Sigma)^3$ for all $x \\in \\Sigma$, a vanishing first moment $\\int_{\\psi(\\Sigma)} y\\, d\\mu_g = 0$ in normal coordinates, and $|\\nabla \\mathrm{Sc}(p_0)| \\le C R(\\Sigma)^2$. Since $R(\\Sigma)^2 = |\\Sigma|/(4\\pi)$, the scalar-curvature gradient at the geometric center is controlled linearly by area, improving the square-root bound of earlier work. The proof varies the Willmore functional with the test function $f = H^{-1} g(b,\\nu)$ and shows, after substituting the top-order relations $H \\approx 2/R$, $\\mathring{A} \\approx -\\tfrac{4}{3} H^{-1} \\mathring{T}$, and $\\nabla H \\approx \\tfrac{2}{3}\\omega$, that both parts of the first variation are $O(R^5)$, leaving only the term $-\\tfrac12 \\mathrm{Vol}(\\Omega)\\, g(b,\\nabla\\mathrm{Sc}(p_0))$ at leading order; dividing by the enclosed volume, which is at least a constant times $R^3$, gives the estimate.","pith_inferences":["The $O(R^2)$ order is likely the best the method can give in general: the terms dropped as $O(R^5)$ in equations (32) and (34) contain even-factor integrals of $\\nu$ and $y/R$, and only the specific cancellation pattern asserted there prevents an $O(R^4)$ contribution to the variation.","A direct check on an explicit family such as geodesic spheres in a metric with nonzero $\\nabla\\mathrm{Sc}$ could isolate the size of those even-factor error terms and test whether the cancellation is exact or merely an artifact of the expansion.","The same center-of-mass and odd-product strategy should apply to other variational problems whose small-area critical surfaces are $W^{2,2}$-close to round spheres, such as small constant-mean-curvature spheres, whenever the ambient curvature enters only through the same combination of Ricci terms."],"forward_implications":["For every admissible surface there is a point $p_0$ inside the enclosed region with all surface points at distance $R(\\Sigma) + O(R(\\Sigma)^3)$, and $|\\nabla\\mathrm{Sc}(p_0)| \\le C R(\\Sigma)^2$; since $R^2$ is proportional to area, the gradient bound is linear in area.","If the scalar curvature is a Morse function with non-degenerate critical points, then for sufficiently small area the region enclosed by any solution of the constrained Willmore equation contains exactly one critical point of $\\mathrm{Sc}$.","The area-constrained minimizers $\\Sigma_{\\min}^a$ of the Willmore functional enclose a single point of maximal scalar curvature when the Hessian there is non-degenerate.","With the adapted center, the first variation $\\delta_f W(\\Sigma)$ for $f = H^{-1}g(b,\\nu)$ is $O(R^5)$, one power of $R$ better than the estimate used in earlier position estimates.","Near a non-degenerate critical point of scalar curvature, the bound $|\\nabla\\mathrm{Sc}(p_0)| \\le C R^2$ together with the Morse lower bound puts the geometric center within $O(R^2)$ of the critical point, so the surface concentrates in a ball of radius $R + O(R^2)$ around it."],"supporting_citations":[{"why":"Supplies the base position estimate, the formulas for $\\delta_f U$ and $\\delta_f V$, the Willmore expansion, and the volume and diameter bounds used throughout.","marker":"[6]"},{"why":"Provides the existence of minimizers and the a priori $L^2$ estimates for $\\nabla^2 H$, $H\\nabla A$, and $H^2\\mathring{A}$, as well as the bounds $H \\approx 2/R$ and $\\|\\mathring{A}\\|_{L^2} = O(R^2)$.","marker":"[7]"},{"why":"The De Lellis-Muller optimal rigidity estimate for nearly umbilical surfaces, quoted as Theorem 2.11, is the Euclidean approximation behind Lemma 3.2 and Corollary 3.3.","marker":"[2]"},{"why":"Supplies the Poincare and eigenvalue estimates for nearly umbilical surfaces used in Lemma 2.8 and in the proof of Lemma 3.2.","marker":"[3]"},{"why":"The Simons identity in the form of equation (8) there and the proof of its Lemma 15 are used to derive the $L^\\infty$ bound for $\\mathring{A}$ in Lemma 2.10.","marker":"[8]"},{"why":"Provides the original Simons identity on which the Laplacian computations for $\\mathring{A}$ in Lemma 2.10 and Section 5.3 rest.","marker":"[13]"}],"fun_headline_variants":["Linear bound beats square root for Willmore center","Willmore surfaces: gradient of scalar curvature now O(area)","Willmore surfaces: scalar curvature gradient drops linearly with area","For Willmore surfaces, curvature gradient now O(area) not sqrt"],"cache_read_input_tokens":21632,"weakest_assumption_plain":"The conclusion rests on the assumption that after expanding the test-function variations in powers of the surface radius, every remaining term contains an odd number of normal or position factors, so that the center-of-mass estimates force those integrals to be tiny; if any even-factor term actually survives at the critical order, the gradient bound would be one power larger.","fun_headline_variants_meta":{"raw":{"variants":["Linear bound beats square root for Willmore center","Willmore surfaces: gradient of scalar curvature now O(area)","Willmore surfaces: scalar curvature gradient drops linearly with area","For Willmore surfaces, curvature gradient now O(area) not sqrt"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000736,"raw_usage":{"total_tokens":3544,"prompt_tokens":954,"completion_tokens":2590,"prompt_tokens_details":{"cached_tokens":896},"prompt_cache_hit_tokens":896,"prompt_cache_miss_tokens":58,"completion_tokens_details":{"reasoning_tokens":2521}},"tokens_in":58,"tokens_out":2590,"duration_ms":55990,"temperature":1.0,"reasoning_tokens":2521,"cache_read_input_tokens":896,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:13:12.992584+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a concrete nearly-spherical family, such as geodesic spheres of radius $R$ centered at a point with nonzero $\\nabla\\mathrm{Sc}$ in a fixed 3-manifold, and explicitly evaluate the variation identity $\\delta_f W(\\Sigma) = \\lambda \\int_\\Sigma fH\\, d\\mu$ for $f = H^{-1}g(b,\\nu)$ in normal coordinates centered at the adapted center of mass. If the terms dropped as $O(R^5)$ in equations (32) and (34) contain an even-factor integral that survives at order $C R^4$, the claimed cancellation fails and $|\\nabla\\mathrm{Sc}(p_0)| \\le C R^2$ would be violated.","supporting_citations":[{"cited_title":"Lamm and J","cited_arxiv_id":null,"evidence_quote":"Supplies the base position estimate, the formulas for $\\delta_f U$ and $\\delta_f V$, the Willmore expansion, and the volume and diameter bounds used throughout."},{"cited_title":"Lamm and J","cited_arxiv_id":null,"evidence_quote":"Provides the existence of minimizers and the a priori $L^2$ estimates for $\\nabla^2 H$, $H\\nabla A$, and $H^2\\mathring{A}$, as well as the bounds $H \\approx 2/R$ and $\\|\\mathring{A}\\|_{L^2} = O(R^2)$."},{"cited_title":"De Lellis and S","cited_arxiv_id":null,"evidence_quote":"The De Lellis-Muller optimal rigidity estimate for nearly umbilical surfaces, quoted as Theorem 2.11, is the Euclidean approximation behind Lemma 3.2 and Corollary 3.3."},{"cited_title":"De Lellis and S","cited_arxiv_id":null,"evidence_quote":"Supplies the Poincare and eigenvalue estimates for nearly umbilical surfaces used in Lemma 2.8 and in the proof of Lemma 3.2."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The Simons identity in the form of equation (8) there and the proof of its Lemma 15 are used to derive the $L^\\infty$ bound for $\\mathring{A}$ in Lemma 2.10."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the original Simons identity on which the Laplacian computations for $\\mathring{A}$ in Lemma 2.10 and Section 5.3 rest."}],"review_version":1}