{"id":"97304889-22a4-4456-ade1-70868242edea","arxiv_id":"2611.16810","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Surfaces of revolution split into four types, with maximal loxodromic curves behaving as non-self-intersecting, closed, dense, or pole-reaching curves according to the type.","lead":"This mathematics paper classifies the global shapes of loxodromes, the constant-angle curves used in navigation, on surfaces like spheres, cylinders, tori, and paraboloids. It shows exactly which types of these curves close up, which ones wrap densely, and which ones can cross the poles, and it describes the maps that preserve them.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek/deepseek-v4-flash","headline":"Forward direction of Theorem 6.18 (paraboloid-like) is under-justified: the argument that an azimuth-consistent diffeomorphism must satisfy (6.11)–(6.12) relies on the claim that it preserves maximal meridians and parallels, but the proof of that claim (Corollary 6.17) itself depends on the…","rationale":"The reader identified the same gap: the forward direction of Theorem 6.18 is under-justified because the argument from Theorem 6.8 does not directly apply to paraboloid-like surfaces. The paper's current proof relies on Corollary 6.17, which itself is not rigorously justified for the forward direction. The reader's weakest_assumption was about pole regularity, which is a different concern; the reader's rationale explicitly mentions the under-justified forward direction of Theorem 6.18 as one of two concerns. I agree with that assessment. The concern is load-bearing because the characterization of azimuth-consistent diffeomorphisms on paraboloid-like surfaces (and the conformality result) depends on Theorem 6.18. The concrete test I propose would settle whether the gap can be filled. I recommend keeping the verdict CONDITIONAL, as the reader did, because the gap is fixable but currently unresolved.","tokens_in":40286,"tokens_out":1067,"duration_ms":11915,"concrete_test":"Independently derive the differential conditions (6.11)–(6.12) from the definition of azimuth-consistency for a paraboloid-like surface without relying on the claim that maximal meridians and parallels are preserved. Specifically: take an azimuth-consistent diffeomorphism f between paraboloid-like surfaces. For an arbitrary point p in S*, consider the vector fields e_u and e_v. Show that f_*(e_u) must be parallel to ˆe_u (i.e., no ˆe_v component) and f_*(e_v) must be parallel to ˆe_v (i.e., no ˆe_u component) using only the fact that f maps loxodromes to loxodromes and that the azimuth of the image depends only on the azimuth of the source. If this derivation succeeds, the theorem stands; if it fails or requires additional assumptions, the theorem is incomplete.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Theorem 6.18 states that a diffeomorphism f between paraboloid-like surfaces is azimuth-consistent iff its differential on S* satisfies (6.11)–(6.12). The forward direction (azimuth-consistent ⇒ (6.11)–(6.12)) is argued as follows: 'By Corollary 6.17, f maps maximal meridians to maximal meridians and maximal parallels to maximal parallels. Therefore, by the same argument as in Theorem 6.8, its differential on S* has the stated form.' However, Corollary 6.17 itself is proved by saying: 'The maximal meridians of S are the only maximal loxodromic curves that pass through q. Since f is a loxodromic equivalence, the image of each maximal meridian is a maximal loxodromic curve that passes through ˆq. It is therefore a maximal meridian on ˆS.' This reasoning assumes that the only maximal loxodromic curves passing through a pole are maximal meridians—which is true by Corollary 3.12—but it does not establish that f preserves the direction of the meridian (i.e., that the image curve is a meridian rather than some other curve that also passes through the pole). On a paraboloid-like surface, a maximal loxodromic curve that passes through the pole must have azimuth 0 mod π (Lemma 3.11), but that does not force it to be a fixed-longitude meridian: the image could be a loxodrome that passes through the pole with azimuth 0 mod π but is not a meridian (i.e., it could have a varying longitude coordinate that still yields a curve passing through the pole). The argument conflates 'passes through the pole' with 'is a meridian.' The same issue affects the parallel preservation claim. Thus the forward direction of Theorem 6.18 is not adequately justified; the differential form (6.11)–(6.12) may not follow from azimuth-consistency alone without additional reasoning.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek/deepseek-v4-flash","summary":"The paper studies loxodromes and loxodromic curves on surfaces of revolution in R^3. It classifies surfaces into four types (cylinder-like, torus-like, paraboloid-like, sphere-like) based on number of poles and closedness of profile curve. It determines global behavior of maximal loxodromic curves on each type: on cylinder-like, nonparallel curves are non-self-intersecting; on torus-like, closed/dense dichotomy via rotation number; on paraboloid-like and sphere-like, only maximal meridians pass through poles, others have poles as endpoints. It solves direct and inverse problems for loxodromic segments, characterizes when maximal loxodromic curves are geodesics, and studies diffeomorphisms preserving maximal loxodromic curves (loxodromic equivalences). It introduces azimuth-consistent diffeomorphisms and characterizes them on cylinder-like, paraboloid-like, sphere-like surfaces, showing that on surfaces with poles they are conformal. For torus-like surfaces, it constructs loxodromic equivalences from GL(2,Z) matrices that need not preserve meridians or parallels.","tokens_in":40641,"tokens_out":8887,"duration_ms":50824,"significance":"The paper provides a systematic classification of loxodromic curves on surfaces of revolution, extending known results (e.g., Blackwood et al.) to a global setting including poles. The closed-dense dichotomy on torus-like surfaces is a standard result but placed in a unified framework. The characterization of loxodromic equivalences and azimuth-consistent diffeomorphisms is new and interesting, especially the conformality result on surfaces with poles. The paper is thorough and includes many explicit formulas and examples. The results are likely to be of interest to differential geometers and those studying curves on surfaces.","major_comments":[{"comment":"The forward direction of Theorem 6.18 (azimuth-consistent ⇒ (6.11)–(6.12)) is argued by saying 'By Corollary 6.17, f maps maximal meridians to maximal meridians and maximal parallels to maximal parallels. Therefore, by the same argument as in Theorem 6.8, its differential on S* has the stated form.' However, the argument in Theorem 6.8 uses the fact that f is azimuth-consistent to derive the form of the differential. Here, we are assuming azimuth-consistency, so the argument should be: from azimuth-consistency, we can deduce that f preserves maximal meridians and parallels (Corollary 6.17), but then we need to derive the differential form. The 'same argument as in Theorem 6.8' would require that we know f* maps eu to a multiple of e_u_hat and ev to a combination. But Theorem 6.8's derivation used the azimuth-consistency condition directly, not the preservation of meridians/parallels. The forward direction of Theorem 6.18 should be justified by directly applying the definition of azimuth-consistency, not by invoking Corollary 6.17 and then Theorem 6.8. The current reasoning is circular or at least insufficiently explained. Please clarify the logical flow: either derive (6.11)–(6.12) directly from azimuth-consistency, or explain how preservation of meridians/parallels (which itself follows from azimuth-consistency) together with the definition yields the differential form. As written, the argument is too brief and risks a gap.","section":"§6.2, Theorem 6.18 (forward direction)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The paragraph following Theorem 6.18 claims that smoothness at the pole forces |c|=1, and therefore every azimuth-consistent diffeomorphism between paraboloid-like surfaces is conformal. The argument uses a limiting process along two perpendicular meridians. However, the definition of azimuth-consistency is given on S* (away from poles). The extension to the pole is not automatic; the paper assumes f is a diffeomorphism (so smooth at the pole). The argument that |c|=1 is plausible but should be made rigorous: the limits of the ratios |f*(ev)|/|f*(eu)| along two different meridians give |c| and 1/|c|, so |c|^2=1. But one must ensure that the limits exist and are independent of the chosen meridian. Since f is smooth, the differential at the pole is a linear map, and the ratio of norms of images of orthogonal unit vectors is well-defined. The argument can be made precise using the differential at the pole. I recommend adding a short justification using the linear map f_* at the pole, rather than a limiting argument along curves. As it stands, the reasoning is heuristic; it should be tightened.","section":"§6.2, after Theorem 6.18","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The converse direction of Theorem 6.8 (if differential has form (6.3)-(6.4) then f is azimuth-consistent) includes a maximality argument to show that f preserves maximal loxodromic curves. However, the definition of azimuth-consistency does not require preservation of maximality; it only requires that the image of any loxodrome (not necessarily maximal) is a loxodrome with constant azimuth. The paper then argues that f preserves maximal loxodromic curves and is a loxodromic equivalence. This is fine, but the proof that f preserves maximal loxodromic curves uses a contradiction argument that assumes the image of a maximal loxodromic curve is not maximal, then extends it, etc. This part is somewhat involved and may be simplified. More importantly, the proof that f maps loxodromes to loxodromes (which is needed for azimuth-consistency) is given by the computation of tan(theta_hat) showing constancy. That part is correct. The additional maximality argument is not necessary for the forward direction of the theorem statement (azimuth-consistency characterization). It might be better to separate the characterization of azimuth-consistency from the proof that such maps are loxodromic equivalences. The theorem as stated claims equivalence between azimuth-consistency and the differential form, but the proof of the converse includes extra steps to show f is a loxodromic equivalence. This is acceptable but could be streamlined. No major error, but the presentation could be clearer.","section":"§6.1, Theorem 6.8","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The proof of Lemma 3.11(2) uses an argument that if theta is not 0 mod pi, then u(s) diverges to ±infinity as s approaches s_p, leading to a contradiction with smoothness at the pole. The argument constructs sequences s_n and t_n such that u(s_n)=2nπ and u(t_n)=2nπ+π, and then shows that lambda'(s_n) and lambda'(t_n) converge to different limits, contradicting continuity of the derivative at s_p. This is a valid argument, but it relies on the fact that u(s) is strictly monotonic and unbounded. The unboundedness follows from the divergence of the integral of 1/r(v) as v approaches the pole. The paper states that 'there exists delta>0 such that 0<r(v)<2(v-a) for every v in (a, a+delta)' and then argues that the integral diverges. This is correct given |r'(a)|=1. However, the argument that u(s) diverges to ±infinity uses tan(theta) ≠ 0. But the sign of tan(theta) determines whether u goes to +∞ or -∞. The construction of sequences s_n, t_n works in either case. The proof is sound. I have no major concern here.","section":"§3.3, Lemma 3.11","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The proof of denseness uses a mapping to the torus T^2 and a standard result about irrational flows. The argument is: set alpha = 2π/Omega, which is irrational, then the set D_alpha = {(x,y): x = (1/alpha) y + (n - (1/alpha)m)} is dense in R^2. Then define f(x,y) = (2π x, Omega_w y). Then f(D_alpha) = Psi^{-1}(lambda(R)). Since D_alpha is dense and f is continuous and surjective, the image is dense. However, the definition of D_alpha seems to be the preimage under F of the irrational line on the torus. The paper references Example 6.15 in Boothby. The reasoning is standard but the notation is a bit confusing. The map F: R^2 -> T^2 is (x,y) -> (exp(2π i x), exp(2π i y)). Then G_alpha(t) = (t, alpha t). Then H = F ∘ G_alpha. The set D_alpha = F^{-1}(H(R)) is the set of all (x,y) such that (x mod 1, y mod 1) lies on the line y = alpha x mod 1. The paper writes D_alpha = {(x,y): x = (1/alpha) y + (n - (1/alpha)m) for some n,m in Z}. This is equivalent. The argument that D_alpha is dense in R^2 is correct for irrational alpha. Then f(D_alpha) = Psi^{-1}(lambda(R)). But note that Psi is not globally injective; it factors through the torus. The denseness of Psi^{-1}(lambda(R)) in R^2 implies that lambda(R) is dense in S because Psi is continuous and surjective. This is correct. However, the paper uses the linearizing parametrization Psi, which is defined on R^2 and is a covering map. The argument is sound. No major issue.","section":"§3.2, Proposition 3.8","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The proof that f defined by (6.18) preserves maximal loxodromic curves is brief: 'The curve T∘c is also defined on R and has a straight line as its image. Hence (f∘λ)(t) = Ψ(T(c(t))) is a loxodrome on S. Since S is compact and f is a diffeomorphism, the speed of f∘λ is bounded below by a positive constant. Therefore, an arc-length parametrization of f∘λ is defined on R and is maximal.' The claim that the speed is bounded below by a positive constant needs justification: because f is a diffeomorphism on a compact manifold, its differential is bounded and has bounded inverse, so the speed of f∘λ is comparable to the speed of λ. Since λ is a straight line in the (u,w) coordinates, its speed in the covering space is constant. The pushforward by Ψ may vary, but the composition with T and Ψ yields a curve whose speed in S is bounded away from zero because the covering map is a local isometry in the induced metric? Actually, the metric in (u,w) coordinates is not Euclidean; the speed of λ in S is not constant in general. However, the paper uses arc-length parametrization for λ on S. The curve λ is a maximal loxodromic curve, so it is parametrized by arc length on S. Its image under f is then reparametrized by arc length. The argument that the arc-length parametrization is defined on R and is maximal relies on the fact that f∘λ is a loxodrome (since T∘c is a straight line, and Ψ maps it to a loxodrome). The maximality follows because the domain of the straight line is R, and the loxodrome cannot be extended because any extension would correspond to extending the straight line, which is already maximal. The speed argument is not needed; the key is that the straight line in the covering space is defined for all t∈R, and the loxodrome is defined for all s∈R after arc-length reparametrization. The paper's mention of speed bounded below is perhaps to ensure that the arc-length reparametrization has domain R. But if the loxodrome is defined for all t∈R and its speed is always positive (since it's regular), then the arc-length parameter will also be defined on R (possibly with finite total length if the speed decays? But on a compact surface, a curve defined for all t∈R with positive speed can have infinite length; arc-length parametrization will have domain R if the length is infinite, or a bounded interval if total length is finite.","section":"§6.4, Proposition 6.22","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The example claims that the map f defined with M = [[1,1],[1,2]] and b=0 preserves neither parallels nor meridians. It shows that the image of a parallel (t,0) is Ψ(t, (Ω_w/(2π)) t). Since the w-coordinate is not constant, it is not a parallel. That is correct. Similarly for meridian. However, the example does not verify that f is a loxodromic equivalence (it follows from Proposition 6.22). The example is clear. No issue.","section":"§6.4, Example 6.23","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Theorem 5.4 states: 'Let S be a surface of revolution with r>0, and assume that r is not constant on any open interval. Then a maximal loxodromic curve is a geodesic if and only if it is either a maximal meridian or a maximal parallel at a stationary point of r.' The proof uses Clairaut's theorem. The converse direction (if it is a maximal meridian or maximal parallel at stationary point, then it is a geodesic) is given by Proposition 5.1. The forward direction: assume λ is a geodesic and a maximal loxodromic curve. The proof considers cases: if sinθ=0 then λ is a maximal meridian; if cosθ=0 then λ is a maximal parallel, and since it is a geodesic, r'(v0)=0 by Proposition 5.1. If both sinθ≠0 and cosθ≠0, then Clairaut gives r(v(s)) sinθ = constant, so r(v(s)) is constant, implying r'(v(s))=0 for all s, contradicting that r is not constant on any open interval. This is correct. However, the assumption that r is not constant on any open interval is used to rule out the possibility that r is constant on the image of v(s), which is an open interval because v is strictly monotonic. This is valid. The theorem is correctly stated and proved.","section":"§5, Theorem 5.4","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Theorem 5.3: 'Every maximal loxodromic curve on a cylinder is a geodesic.' The proof: if λ is a maximal parallel, then r is constant, so r'(v)=0, hence by Proposition 5.1 it is a geodesic. If λ is nonparallel, then r and sinθ are constant, so r sinθ is constant, and no part is a parallel, so by Clairaut it is a geodesic. This is correct. However, note that Clairaut's theorem gives a sufficient condition for a curve to be a geodesic if it is not part of a parallel. The paper cites Proposition 5.2 (Clairaut's Theorem) which states: 'If γ is a unit-speed curve on a surface of revolution S, let ρ:S→R be the distance of a point of S from the axis of rotation, and let θ be the angle between γ' and the meridians of S. If γ is a geodesic, then ρ sin(θ) is constant along γ. Conversely, if ρ sin(θ) is constant along γ, and if no part of γ is part of some parallel of S, then γ is a geodesic.' The converse is used. The condition 'no part of γ is part of some parallel' is satisfied because cosθ≠0 implies v is not constant, so it is not a parallel. So the proof is valid.","section":"§5, Theorem 5.3","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Proposition 6.7: 'Let S and Ŝ be cylinder-like surfaces. Suppose that f:S→Ŝ is a loxodromic equivalence. If λ is a maximal parallel at v=v0 on S, then f∘λ is a maximal parallel on Ŝ.' The proof: λ is closed, so f∘λ is closed. Since f is a loxodromic equivalence, f∘λ is a maximal loxodromic curve. By Proposition 3.3, on a cylinder-like surface, the only closed maximal loxodromic curves are maximal parallels. Therefore f∘λ is a maximal parallel. This reasoning assumes that f∘λ is closed. But is it necessarily closed? λ is a parallel, which is a closed curve (circle). Since f is a diffeomorphism, the image is also a closed curve (topologically a circle). However, the parametrization of f∘λ may not be periodic in the arc-length parameter, but the set is closed. The paper says 'Since λ is closed, f∘λ is also closed.' This is true as sets. But maximal loxodromic curves are defined as arc-length parametrized curves; the image set being closed does not guarantee that the arc-length parametrization is periodic. However, on a cylinder-like surface, a maximal parallel is a closed curve (circle) and its arc-length parametrization is periodic. The image of a parallel under a diffeomorphism is a simple closed curve. Could it be that the image is not a parallel but some other closed curve that is a maximal loxodromic curve? Proposition 3.3 says that on a cylinder-like surface, the only kinds of maximal loxodromic curves are maximal parallels and non-self-intersecting (non-closed) curves. So any closed maximal loxodromic curve must be a maximal parallel. Therefore, f∘λ, being closed, must be a maximal parallel. This is valid. The only subtlety: is it guaranteed that f∘λ, when reparametrized by arc length, is a maximal loxodromic curve? Since f is a loxodromic equivalence, by definition, an arc-length parametrization of f∘λ is a maximal loxodromic curve. So yes. The argument is sound.","section":"§6.1, Proposition 6.7","authors_acknowledge":null,"authors_acknowledge_p":null}],"minor_comments":[{"comment":"The paper is generally well-written and organized. However, there are several typographical and formatting issues. For example, in the abstract and introduction, the word 'parametrized' is used consistently, but in some places 'parametrization' is used; both are acceptable but consistency is preferred. Also, there are occasional missing spaces after punctuation (e.g., 'surface.Since' in the proof of Lemma 3.11). These should be corrected.","section":"General","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The vector field W_θ is defined as sinθ e_u + cosθ e_v. But e_u = (1/r) ∂_u, e_v = ∂_v. So W_θ = (sinθ/r) ∂_u + cosθ ∂_v. This is consistent with the differential equations (2.6)-(2.7). However, in the text after (2.5), it says 'Note that W_θ is defined only on S\\(P), since e_u is not well-defined at the poles.' This is correct. But in the linearizing parametrization, they use a different scaling. It might be helpful to note that the vector field is not globally defined on S, but its integral curves are loxodromes.","section":"§2, equation (2.5)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The proof of Lemma 3.2 uses the existence and uniqueness theorem for ODEs to argue that the maximal solution (u,v) has domain containing intJ. This is standard. However, the proof could be more precise about the fact that the ODE system is defined on an open set where r(v)>0. Since S has no poles, r(v)>0 for all v in I, so the system is smooth on I. The argument is fine.","section":"§3.1, Lemma 3.2","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The definition of Ω = tanθ ∫_a^b (1/r(τ)) dτ. This is the longitude increment after one full circuit of the profile curve. It might be helpful to note that Ω depends on θ. In Proposition 3.7, the condition for closedness is Ω/(2π) ∈ Q. This is standard.","section":"§3.2, equation (3.6)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The example uses the upper hemisphere. The formula for u(v) involves ln|sec(v)+tan(v)|, which is the inverse Gudermannian function. This is correct. The figure is helpful.","section":"§3.3, Example 3.13","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Tables 2 and 3 summarize the direct and inverse problems. They are useful but could be more self-contained. For instance, the equations referenced (4.1)-(4.19) are in the text, but the table does not indicate the conditions on d. The text provides those conditions. It might be helpful to add a note in the caption or in the table itself. However, as is, it is acceptable.","section":"§4, Tables 2 and 3","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Remark 6.10 gives the relation tan(θ_hat) = c1/c2 + (1/c2) tanθ. This is derived from (6.6). It might be useful to note that this is an affine transformation of the tangent of the azimuth. This is a nice observation.","section":"§6.1, Remark 6.10","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Corollary 6.17 states that a loxodromic equivalence between paraboloid-like surfaces maps maximal meridians to maximal meridians and maximal parallels to maximal parallels. The proof is brief: 'The maximal meridians of S are the only maximal loxodromic curves that pass through q. Since f is a loxodromic equivalence, the image of each maximal meridian is a maximal loxodromic curve that passes through q_hat. It is therefore a maximal meridian on Ŝ.' This reasoning assumes that the only maximal loxodromic curves passing through a pole are maximal meridians. This is true by Corollary 3.12. However, the skeptic's note points out that the image could be a loxodrome that passes through the pole with azimuth 0 mod π but is not a meridian (i.e., could have varying longitude). But Corollary 3.12 says that on a paraboloid-like surface, the only maximal loxodromic curves that pass through the pole are maximal meridians. So any maximal loxodromic curve passing through the pole must be a maximal meridian. Therefore, the image, being a maximal loxodromic curve passing through q_hat, must be a maximal meridian. This is logically sound. The skeptic's concern is that the image could be a loxodrome that passes through the pole but is not a meridian; but Corollary 3.12 explicitly states that such a curve does not exist on a paraboloid-like surface. So the argument is valid. The skeptic's objection is based on a misunderstanding: they say 'the image could be a loxodrome that passes through the pole with azimuth 0 mod π but is not a meridian (i.e., it could have a varying longitude coordinate that still yields a curve passing through the pole).' However, Corollary 3.12 says that on a paraboloid-like surface, the only maximal loxodromic curves that pass through the pole are maximal meridians. So such a loxodrome with varying longitude cannot pass through the pole (except as an endpoint). Therefore, the image must be a maximal meridian. The argument is correct. I will not raise this as a major comment because the paper's reasoning is sound. However, the skeptic's note might indicate that the proof could be expanded to explicitly cite Corollary 3.12. The paper does cite it implicitly. I think it's fine.","section":"§6.2, Corollary 6.17","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"As noted in major comments, the forward direction of Theorem 6.18 is under-justified. The paper says 'By Corollary 6.17, f maps maximal meridians to maximal meridians and maximal parallels to maximal parallels. Therefore, by the same argument as in Theorem 6.8, its differential on S* has the stated form.' But Theorem 6.8's argument used azimuth-consistency to derive the differential form. Here, we are assuming azimuth-consistency, so we can directly apply the definition of azimuth-consistency to derive the differential form, without invoking Corollary 6.17. The reference to Theorem 6.8 is misleading because Theorem 6.8 also assumes azimuth-consistency and derives the differential form. So the logic is: assume azimuth-consistency; then by definition, f maps loxodromes to loxodromes with constant azimuth. In particular, it maps meridians (θ=0) to loxodromes with some constant azimuth θ_hat(0). But we also know from Corollary 6.17 that f maps maximal meridians to maximal meridians, so θ_hat(0) must be 0 mod π. Similarly for parallels. Then we can use the same computation as in Theorem 6.8 to get the form. However, the paper's phrasing 'by the same argument as in Theorem 6.8' is ambiguous because Theorem 6.8's argument started from the assumption that f is azimuth-consistent and derived the differential form using the fact that f* maps eu to a multiple of e_u_hat and ev to a combination. That derivation did not use preservation of meridians/parallels; it used the azimuth-consistency condition directly. So the same argument can be applied here directly, without invoking Corollary 6.17. The mention of Corollary 6.17 might be intended to justify that the constants c1 and c2 are such that c1=0? Actually, in Theorem 6.18, the differential form has c1=0 (only diagonal terms). In Theorem 6.8, the form includes c1 and c2. So the 'same argument' would give a form with possibly nonzero c1. But then the paper claims that the differential has the form (6.11)-(6.12) with c1=0. How do we get c1=0? It must come from the fact that f maps meridians to meridians (θ=0 maps to θ_hat=0 mod π). In Theorem 6.8, the relation (6.10) gives tan(θ_hat) = c1/c2 + (1/c2) tanθ. For θ=0, tan(θ_hat)=c1/c2. For the image to be a meridian, we need θ_hat ≡ 0 mod π, so tan(θ_hat)=0, hence c1=0.","section":"§6.2, Theorem 6.18 forward direction","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Similar issue as Theorem 6.18: the forward direction should explicitly use the preservation of meridians and parallels to conclude c1=0. The paper says 'by the same argument as in Theorem 6.8'. Again, the same argument would give a form with possibly nonzero c1, but preservation of meridians forces c1=0. The paper should clarify this.","section":"§6.3, Theorem 6.20","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The map T is defined as T = D M D^{-1} (u,w)^T + (b1,b2)^T, where D = diag(2π, Ω_w). This is an affine transformation. The paper claims that for any (u1,w1) and (u2,w2) with Ψ(u1,w1)=Ψ(u2,w2), we have (u2,w2) = (u1+2πm, w1+Ω_w n). This is true because Ψ is periodic with those periods. Then T(u2,w2) = T(u1,w1) + (2π(m11 m + m12 n), Ω_w(m21 m + m22 n)). Since M is integer matrix, the vector is in the period lattice. Therefore Ψ(T(u2,w2)) = Ψ(T(u1,w1)). So f is well-defined. This is correct.","section":"§6.4, equation (6.15)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The proof that f is a diffeomorphism uses local diffeomorphism and bijectivity. The bijectivity is argued by constructing an inverse using T^{-1}. Since M ∈ GL(2,Z), T is invertible with integer matrix inverse. The local diffeomorphism argument is standard. This is correct.","section":"§6.4, Proposition 6.21","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"As noted in major comments, the proof that f preserves maximal loxodromic curves is acceptable but could be more rigorous. The statement 'the speed of f∘λ is bounded below by a positive constant' is not necessary; the key is that the straight line in the covering space is defined for all t∈R, and its projection is a loxodrome defined for all t∈R. The arc-length parametrization will have domain R if the length is infinite, or can be extended periodically if finite. Since the loxodrome is maximal, its domain is maximal. The argument that the arc-length parametrization is maximal because the straight line cannot be extended is valid. However, the paper should ensure that the loxodrome f∘λ is indeed maximal: if it could be extended, then the straight line in the covering space could be extended, contradiction. This is correct because any extension of the loxodrome would lift to an extension of the straight line (since the covering map is a local diffeomorphism). So the proof is sound.","section":"§6.4, Proposition 6.22","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The proof of Proposition 3.7 uses the fact that the least period of the periodic extension of the profile curve is L. This is true because the profile curve is closed and traversed exactly once on [a,b]. However, if the profile curve is not simple (e.g., it could have self-intersections), the least period might be a fraction of L. But the paper assumes the profile curve is an embedding? In the definition of torus-like surface, it says 'generated by a profile curve α: [a,b] → R^3 parametrized by arc length. By the definition of a torus-like surface, we have α(a)=α(b) and α'(a)=α'(b).' It does not explicitly say that the profile curve is simple. However, for the surface to be a regular surface, the profile curve should be an embedding (or at least injective on [a,b) ). The paper mentions in Remark 3.5 that if the profile curve is an embedding, then the surface is homeomorphic to I×S^1. For torus-like surfaces, the profile curve is closed, so it is a closed curve. If it is not simple, the surface may have self-intersections and may not be a regular surface. The paper likely assumes the profile curve is simple (embedding) for torus-like surfaces. This is a reasonable assumption. The proof of Proposition 3.7 uses that the least period is L, which holds if the profile curve is simple and closed. If the profile curve is multiply covered, the period could be L/k. But the paper's classification likely excludes such degenerate cases. I think it's fine.","section":"§3.2, Proposition 3.7","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The example uses a profile curve with r(v)=cos(v)+c, z(v)=sin(v) on I=[-π,π]. This is a circle of radius 1 centered at (c,0) in the rz-plane. Revolving gives a torus. The formula for I_c(v) involves arctan and the floor function. This is correct. The longitude increment Ω = 2π tanθ / sqrt(c^2-1). The example is illustrative.","section":"§3.2, Example 3.9","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The proof of part (1) uses a contradiction argument to show that the pole is unique in the domain. The argument constructs a sequence t_n and uses continuity and derivative to get a contradiction. This is rigorous. However, there is a small typo: 'convergencet n →c' should be 'convergence t_n → c'. Also, the sentence 'The definition of the derivative therefore gives λ'(c) = lim (λ(t_n)-λ(c))/(t_n-c) = 0' assumes that λ is differentiable at c. Since c is an interior point of J, λ is smooth, so this is valid. The argument is correct.","section":"§3.3, Lemma 3.11","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Corollary 3.14 states that on a sphere-like surface, there are three kinds of maximal loxodromic curves: closed maximal parallels, maximal meridians passing smoothly through both poles, and non-self-intersecting maximal loxodromic curves having the two poles as endpoints. This is consistent with the classification. The proof is given in the preceding paragraphs. It is correct.","section":"§3.4, Corollary 3.14","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The inverse problem for paraboloid-like surfaces includes cases where points are on the same maximal meridian. The formulas for d_q and d_0 are given. The paper distinguishes between same branch and opposite branches. This is clear.","section":"§4.2, inverse problem for paraboloid-like","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Proposition 5.5 proves that every maximal meridian is a geodesic, including at poles. The proof uses the fact that r''(0)=0 at the pole, which follows from the regularity condition |r'(0)|=1 and z'(0)=0. The computation of left and right limits of the difference quotient shows that λ''(0) is vertical, which is normal to the tangent plane (horizontal). This is correct. The note at the end that a different parametrization may not satisfy the geodesic definition is a good caution.","section":"§5, Proposition 5.5","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Example 6.13 shows a diffeomorphism that preserves meridians and parallels but is not azimuth-consistent and does not preserve maximal loxodromic curves. The example uses f(x,y,z) = (x,y,z^3) on the upper unit cylinder. This is a diffeomorphism? The map z -> z^3 is a diffeomorphism from (0,∞) to (0,∞), but at z=0 it is not a diffeomorphism because derivative is 0. The cylinder is defined for v>0, so it's fine. The example is valid. It illustrates that preservation of meridians and parallels does not imply azimuth-consistency.","section":"§6.1, Example 6.13","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The argument that |c|=1 uses limits along two perpendicular meridians. As noted in major comments, this should be made rigorous using the differential at the pole. The paper states: 'Indeed, let X_q be an arbitrary unit tangent vector at q. Approach q along a meridian whose tangent at q is X_q. Up to signs, the vectors e_v and e_u approach X_q and R X_q, respectively, where R denotes rotation by π/2 in the tangent plane. Taking the limit in |f_*(e_v)|/|f_*(e_u)| = |c| gives |f_{*q}(X_q)|/|f_{*q}(R X_q)| = |c|. Now choose another meridian whose tangent at q is R X_q. Along this meridian, up to signs, e_v approaches R X_q, while e_u approaches R(R X_q) = -X_q. Applying the same argument yields |f_{*q}(R X_q)|/|f_{*q}(X_q)| = |c|. Multiplying gives |c|^2=1, hence c=±1.' This argument assumes that the limits of the ratios exist and are equal to the ratio of the norms of the images of the limit vectors. Since f is smooth, the differential at q is a linear map L = f_{*q}. The vector fields e_u and e_v are not defined at q, but they have limits along curves. The argument can be made precise by considering a smooth curve γ(t) approaching q with γ'(t) approaching X_q, and using that f_*(e_v) along γ tends to L(X_q) (up to scaling). The paper's heuristic is acceptable, but a more rigorous treatment would strengthen the paper. I will note this as a minor comment because the conclusion is correct and the argument is plausible.","section":"§6.2, after Theorem 6.18, conformality argument","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The paper claims that every azimuth-consistent diffeomorphism between sphere-like surfaces is conformal, with c=±1. The same limiting argument applies at either pole. This is consistent.","section":"§6.3, sphere-like surfaces","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The references are appropriate and cover the relevant literature. However, reference [12] (Tu) is used for the definition of maximal integral curve. The definition is standard. Reference [13] (Pressley) is used for basic differential geometry. Reference [14] (Boothby) is used for the dense curve on torus. All are standard texts.","section":"References","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"major_revision"},"author_rebuttal":null,"desk_editor":{"model":"deepseek/deepseek-v4-flash","letter":"The paper does what it says on the tin: it classifies maximal loxodromic curves on surfaces of revolution into four types (cylinder-like, torus-like, paraboloid-like, sphere-like) and works out their global behavior. The genuinely new content is the treatment of curves that pass through poles—introducing loxodromic curves (as opposed to loxodromes) and proving that only maximal meridians can pass smoothly through a pole. The closed/dense dichotomy on tori is not new (Blackwood–Dukehart–Javaheri already had it), but the paper places it in a clean framework and adds the pole-passing analysis, which is nontrivial. The linearizing coordinate w is a nice expository device. The direct/inverse problem tables are encyclopedic and will be useful to practitioners. The geodesic characterization (Theorems 5.3–5.4, Proposition 5.5) is straightforward but solid. The diffeomorphism analysis in Section 6 is ambitious; the torus-like case using GL(2,Z) is clever. The paper is well written, the proofs are generally clear, and the assumptions (regularity at poles, arc-length parametrization) are stated upfront. The authors do not overclaim. Now the soft spots. The stress-test note on Theorem 6.18 is correct: the forward direction (azimuth-consistent ⇒ differential form (6.11)–(6.12)) is under-justified. The argument that f preserves maximal meridians and parallels (Corollary 6.17) is fine, but the leap from that to the differential form on S* is not fully spelled out. The proof says 'by the same argument as in Theorem 6.8,' but Theorem 6.8 uses the fact that on a cylinder-like surface the only closed maximal loxodromic curves are parallels, which is not true on a paraboloid-like surface (meridians are also closed). The authors need a separate argument that the preservation of both families forces the differential to be diagonal with constant ratio. I think the conclusion is still correct, but the proof as written is incomplete. This is a fixable gap, not a fatal error. A minor concern: the pole-regularity assumption (|r'(v)|=1, z'(v)=0 at poles) excludes cone-like singularities. The authors acknowledge this, but it means the results on pole behavior do not apply to surfaces with a conical tip. That is fine—it is a stated assumption—but readers should be aware. The direct/inverse problem tables are dense and some branch cases are only summarized; verifying every entry is tedious but I did not find errors. The paper is for differential geometers working on curves on surfaces, and for geodesists who want explicit formulas. It deserves a serious referee. I would accept peer review. The main gap in Theorem 6.18 needs to be addressed, but the rest is solid. Recommend revision with a request to expand the proof of the forward direction of Theorem 6.18.","headline":"A thorough, well-organized classification of loxodromic curves on surfaces of revolution, with real new results on pole behavior and a clean four-type framework. The forward direction of Theorem 6.18 has a genuine gap, but it is fixable and does not undermine the core contributions.","tokens_in":41347,"tokens_out":745,"duration_ms":9618},"pith_extraction":null,"created_at":"2026-11-13T12:47:22.774337+00:00","model_set":{"reader":"deepseek/deepseek-v4-flash"},"pipeline_version":"pith-pipeline@v1.4.1-daily-deepseek-or-medium","input_kind":null,"input_chars":null,"input_total_chars":null,"passes_missing":null,"published_standing":null,"published_confidence":null,"objection_audit":null,"run_record":null,"basis":{"kind":"not_recorded","source":null,"chars_given":null,"chars_total":null},"receipt":null,"falsifier":null,"supporting_citations":[],"lane_accuracy":{"status":"not yet measured","lane":{"model":"deepseek/deepseek-v4-flash","input_kind":null},"url":"https://pith.science/accuracy"},"review_version":1}