betaFeedback

WorksMathematics

On the toric Bando–Kasue–Nakajima conjecture

Abstract

The conjecture of Bando, Kasue, and Nakajima asserts that every simply connected Ricci-flat 44-manifold that is asymptotically locally Euclidean (ALE), that is, asymptotic at infinity to a flat cone R4/Γ\mathbb{R}^4/Γ, is hyperkähler for one of its orientations. A toric ALE instanton is a Ricci-flat ALE 44-manifold with an effective isometric action of the 22-torus; such a manifold is simply connected. We prove the conjecture for the toric ALE instantons whose group at infinity ΓΓ is conjugate in O(4)\mathrm O(4) to a subgroup of SU(2)\mathrm{SU}(2): each of them is a multi-Eguchi–Hanson metric with collinear centres. In particular, the conjecture holds for toric ALE instantons with ∣Γ∣∈{1,2,3,4,6}|Γ|\in\{1,2,3,4,6\}. The proof uses a closed formula for toric gravitational instantons, which expresses the metric through Hankel determinants of a positive measure on the fixed points of the torus action. With it, the problem becomes a question about lattice points in a parallelogram determined by the end.

Introduction

A gravitational instanton is a complete, non-compact, oriented, Ricci-flat Riemannian 4-manifold whose curvature decays at least quadratically [26 Def. 1.1]. It is asymptotically locally Euclidean (ALE) if, outside a compact set, it is asymptotic to a flat cone R4/Γ\mathbb{R}^4/\Gamma, where Γ⊂O(4)\Gamma\subset\mathrm{O}(4) is a finite group acting freely on the unit sphere, the group at infinity. The hyperkähler ALE instantons are completely understood. Their group Γ\Gamma is a finite subgroup of SU(2)\mathrm{SU}(2), and for every such Γ\Gamma Kronheimer constructed and classified them [20, 21]. The simplest examples are the Eguchi–Hanson metric [9], with Γ=Z2\Gamma= \mathbb{Z}_2, and the multi-Eguchi–Hanson metrics of Gibbons and Hawking [14], with Γ=Zm\Gamma= \mathbb{Z}_m.

A conjecture of Bando, Kasue, and Nakajima [6 Rem. (1.8)(3)] asserts that there are no others: every simply connected Ricci-flat ALE 4-manifold should be hyperkähler for one of its two orientations. See also [26 Conj. 11], where it is stated without simple connectivity, as the assertion that the curvature is anti-self-dual for one orientation. Gibbons had conjectured earlier that every ALE instanton is half flat and that the group at infinity of its universal covering lies in SU(2)\mathrm{SU}(2) [13 Conjecture (1)]. A Ricci-flat metric on a simply connected manifold is half flat exactly when it is hyperkähler for one orientation, so for simply connected manifolds the two formulations agree.

Simple connectivity cannot be dropped: there are free quotients of Gibbons–Hawking metrics that are Ricci-flat Kähler ALE but hyperkähler for neither orientation [34, 32].

The conjecture has been proved by Nakajima [29 Cor. 3.4] and the second author [8 Thm. 6.2] when the manifold is spin and Γ⊂SU(2)\Gamma\subset\mathrm{SU}(2), with the trivial spin structure on the end. Li and Sun proved that a toric ALE instanton that is spin is hyperkähler for one orientation, with no condition on Γ\Gamma [25 Prop. A.21].

Further cases are known under additional geometric assumptions [23, 28, 12]; without such assumptions the conjecture is open, even for Γ⊂SU(2)\Gamma\subset\mathrm{SU}(2). In this paper we consider toric ALE

2020 Mathematics Subject Classification. Primary 53C25; Secondary 53C26, 83C15.

Key words and phrases. Gravitational instantons, Ricci-flat ALE metrics, toric metrics, Bando–Kasue–Nakajima conjecture, Hankel determinants. instantons, which carry an effective isometric action of the torus T2T^2 (Definition 2.1); for the literature on toric gravitational instantons see [26 §§5, 6, 7.1] and [22, 5].

The group at infinity of a toric ALE instanton is cyclic, of order mm say, and the cross-section S3/ΓS^3/\Gamma of its end is a lens space. The group Γ\Gamma is conjugate in O(4)\mathrm{O}(4) to a subgroup of SU(2)\mathrm{SU}(2) exactly when this lens space is L(m,1)L(m,1) (Lemma 2.6). Every toric ALE instanton is simply connected, whatever its group at infinity (Lemma 2.3), so the conjecture of Bando, Kasue, and Nakajima applies to all of them. Our main result is the following.

Theorem A. A toric ALE instanton (M,g)(M,g) (Definition 2.1) whose group at infinity Γ\Gamma is conjugate in O(4)\mathrm{O}(4) to a subgroup of SU(2)\mathrm{SU}(2) is a multi-Eguchi–Hanson metric with ∣Γ∣|\Gamma| collinear centres, hyperkähler for one of its orientations.

Theorem A assumes no spin structure, no Hermitian structure, and no topological condition on MM. See Theorem 5.4 for a more detailed statement. Li and Sun proved the case Γ≅Z2\Gamma\cong\mathbb{Z}_2 [25 Prop. A.24], as well as the spin case recalled above.

The proof rests on the solitonic form of Geroch’s conjecture for toric gravitational instantons proved in [3]. Its proof gives the metric in closed form, which allows algebraic and combinatorial methods to be applied: the rod vectors of the finite rods of a non-hyperkähler instanton lie in the interior of a parallelogram in the Lie algebra of the torus that is determined by the end [4]. What is new here is the arithmetic of the lattice points in this parallelogram. When Γ\Gamma is conjugate in O(4)\mathrm{O}(4) to a subgroup of SU(2)\mathrm{SU}(2), the lattice points in its interior lie on one line, too few to accommodate the rod vectors of a non-hyperkähler instanton, and the remaining hyperkähler case is the multi-Eguchi–Hanson metric.

For small groups at infinity the hypothesis of Theorem A holds automatically. For m∈{1,2,3,4,6}m \in\{1,2,3,4,6\} every integer prime to mm is congruent to 11 or −1-1 modulo mm, so every lens space L(m,q)L(m,q) is diffeomorphic to L(m,1)L(m,1), and Theorem A has the following consequence (Corollary 5.5).

Theorem B. A toric ALE instanton (M,g)(M,g) (Definition 2.1) whose group at infinity Γ\Gamma has order 11, 22, 33, 44, or 66 is a multi-Eguchi–Hanson metric with ∣Γ∣|\Gamma| collinear centres.

Toric ALE instantons and rod structures

A toric ALE instanton is described by two kinds of data. The continuous data are the positions of the fixed points of the torus action on the boundary of the orbit space. The discrete data are the stabilisers of the action along this boundary, a finite sequence of lattice vectors, which we encode in a chain of integers. The group at infinity and the topology of MM are determined by the chain. We recall these notions, following [4 §2.1] and [3 §2], and fix the conventions used throughout.

Toric ALE instantons

Let T2T^2 be the 2-torus, t\mathfrak{t} its Lie algebra, and Λ⊂t\Lambda\subset\mathfrak{t} the lattice of the elements XX with exp⁡(2πX)=1\exp(2\pi X)=1, so that T2=t/2πΛT^2=\mathfrak{t}/2\pi\Lambda. A Z\mathbb{Z}-basis of Λ\Lambda is a basis of Λ\Lambda as a free abelian group. The flat space R4=C2\mathbb{R}^4=\mathbb{C}^2 carries the standard action of the torus (S1)2(S^1)^2 by (z1,z2)↦(eiϕ1z1,eiϕ2z2)(z_1,z_2)\mapsto(e^{i\phi_1}z_1,e^{i\phi_2}z_2).

Definition 2.1. A toric ALE instanton is a complete, connected, orientable Ricci-flat Riemannian 4-manifold (M,g)(M,g) with an effective isometric action of T2T^2, such that (M,g)(M,g) is ALE in the following sense. There are a compact set K⊂MK\subset M, a finite group Γ⊂O(4)\Gamma\subset\mathrm{O}(4) that acts freely on S3S^3, and a diffeomorphism FF of M∖KM\setminus K onto the complement of a closed ball in R4/Γ\mathbb{R}^4/\Gamma such that ∣∇k(g−F∗gR4)∣=O(r−k−δ)|\nabla^k(g-F^*g_{\mathbb{R}^4})|=O(r^{-k-\delta}) for some δ>0\delta>0 and all k≥0k\geq0, with the norms and covariant derivatives of F∗gR4F^*g_{\mathbb{R}^4}, where rr is the Euclidean distance from the origin. The group Γ\Gamma is the group at infinity.

Remark 2.2 (invariant cross-sections). The end E=M∖KE=M\setminus K is diffeomorphic to (R,∞)×S3/Γ(\mathbb{R},\infty)\times S^3/\Gamma. The chart FF is not assumed to be compatible with the action, but it can be chosen so on the universal covering E~\widetilde{E} of the end, to which the action lifts through a finite covering torus T~2\widetilde{T}^2 of T2T^2. By the argument in the proof of [24 Prop. 4.38], which applies the centre-of-mass construction of Grove and Karcher [15] on the end, the lift of FF can be replaced by a chart F~\widetilde{F} that is equivariant for an isomorphism of T~2\widetilde{T}^2 onto (S1)2(S^1)^2 and, after composing it with (z1,z2)↦(z1,z‾2)(z_1,z_2)\mapsto(z_1,\overline{z}_2) if necessary, orientation-preserving for a given orientation of MM. So the end is ALE in the sense of [3 Def. 2.9] for either orientation of MM. The chart F~\widetilde{F} differs from the lift of FF, on which the deck group acts through Γ⊂O(4)\Gamma\subset O(4), by O(r1−ϵ)O(r^{1-\epsilon}) for some ϵ>0\epsilon>0, with the corresponding bounds on derivatives, which the centre-of-mass construction provides. Averaging the function r∘F~r\circ\widetilde{F} over the deck group gives a T2T^2-invariant function on EE that differs from r∘Fr\circ F by O(r1−ϵ)O(r^{1-\epsilon}), and its gradient by O(r−ϵ)O(r^{-\epsilon}). Its large level sets are invariant cross-sections S3/ΓS^3/\Gamma of the end, and beyond each of them the end is a collar [c,∞)×S3/Γ[c,\infty)\times S^3/\Gamma.

Lemma 2.3. Let (M,g)(M,g) be a toric ALE instanton. Then the action has finitely many fixed points, at least one; MM is simply connected; and all stabilisers of the action are connected. So (M,g)(M,g), with either orientation, is a simply connected toric gravitational instanton in the sense of [3 Defs. 2.1, 2.2] with an ALE end in the sense of [3 Def. 2.9], and the results of [3, 4] apply to it.

Proof. The fixed points are isolated and lie in a compact set, since the model end contains none, so there are finitely many of them [3 Rem. 2.10], nn say. The manifold MM deformation retracts onto a compact T2T^2-invariant domain Ω\Omega that contains them and is bounded by an invariant cross-section of the end, since M∖ΩM\setminus\Omega is a collar of ∂Ω\partial\Omega (see Remark 2.2). The Euler characteristic of a closed manifold with a torus action equals that of its fixed-point set, by [19 Ch. II, Thm. 5.5] applied to a Killing field of an invariant metric whose flow is dense in the torus. The double of Ω\Omega carries the induced action of T2T^2, which is smooth for an equivariant collar of ∂Ω\partial\Omega. It has 2n2n fixed points and the Euler characteristic 2χ(Ω)2\chi(\Omega), since ∂Ω\partial\Omega is a closed 3-manifold, so χ(M)=χ(Ω)=n\chi(M)=\chi(\Omega)=n. The fundamental group of MM is finite, since gg is Ricci-flat with Euclidean volume growth [27, 2], so b1(M)=0b_1(M)=0. Since MM is connected, orientable, and non-compact, Poincaré duality gives b4(M)=0b_4(M)=0 and b3(M)=dim⁡Hc1(M;R)b_3(M)=\dim H^1_c(M;\mathbb{R}), which vanishes since b1(M)=0b_1(M)=0 and the end of MM is connected. Hence n=χ(M)=1+b2(M)≥1n=\chi(M)=1+b_2(M)\geq1.

When the stabilisers are connected and the action has a fixed point, simple connectivity follows from [3 Rem. 3.9]. Here connected stabilisers are not assumed. Let M~→M\widetilde{M}\to M be the universal covering, which is finite, with the lifted metric g~\widetilde{g} and orientation. The Killing fields of the action lift to complete commuting Killing fields of g~\widetilde{g}. For X∈ΛX\in\Lambda the time-2π2\pi flow of the lift of XX covers the identity of MM, so it is a deck transformation. It fixes the points over a fixed point of MM, where the lifted fields vanish, so it is the identity. Hence the lifted fields generate an effective isometric action of T2T^2 on M~\widetilde{M}, whose fixed points are the finitely many points over the fixed points of MM. Since M~\widetilde{M} is simply connected and non-compact, its stabilisers are connected [4 Lemma 2.1]. By Definition 2.1, ∣Rm⁡g∣=O(r−2−δ)|\operatorname{Rm}_g|=O(r^{-2-\delta}), so gg has finite curvature energy ∫M∣Rm⁡g∣2\int_M|\operatorname{Rm}_g|^2, and that of g~\widetilde{g} is ∣π1(M)∣|\pi_1(M)| times as large. So (M~,g~)(\widetilde{M},\widetilde{g}) is a gravitational instanton in the sense of [3 Def. 2.2]; it is simply connected and toric, and its action has finitely many fixed points, at least one. Hence the Weyl coordinates (ρ~,z~)(\widetilde{\rho},\widetilde{z}) of a Z\mathbb{Z}-basis of Λ\Lambda map its orbit space homeomorphically onto the closed half-plane {ρ~≥0}\{\widetilde{\rho}\geq0\} [3 Prop. 3.8(ii)].

A deck transformation γ\gamma is an orientation-preserving isometry that commutes with the lifted Killing fields, hence with the action. So it preserves the Gram matrix of the Killing fields and ρ~\widetilde{\rho}, and it induces an orientation-preserving isometry of the orbit space. Hence z~∘γ\widetilde{z}\circ\gamma is again a harmonic conjugate of ρ~\widetilde{\rho}, so z~∘γ=z~+c\widetilde{z}\circ\gamma=\widetilde{z}+c for a constant cc. Since γ\gamma has finite order kk, z~=z~∘γk=z~+kc\widetilde{z}=\widetilde{z}\circ\gamma^{k}=\widetilde{z}+kc, so c=0c=0. Since (ρ~,z~)(\widetilde{\rho},\widetilde{z}) is injective on the orbit space, γ\gamma maps every orbit onto itself, and in particular it fixes the fixed points of the action on M~\widetilde{M}. A deck transformation with a fixed point is the identity, so the deck group is trivial, MM is simply connected, and its stabilisers are connected [4 Lemma 2.1]. □\square

Fix orientations of MM and t\mathfrak{t}. For a positively oriented basis of t\mathfrak{t}, let GG be the Gram matrix of the corresponding Killing fields, and let ϕ1,ϕ2\phi^{1},\phi^{2} be the dual angles. Since gg is Ricci-flat and the action has fixed points, the distribution orthogonal to the orbits is integrable, and the function ρ=(det⁡G)1/2\rho=(\det G)^{1/2} is harmonic on the interior of the orbit space M/T2M/T^{2} with its quotient metric [3 §2.1]. Where dρ≠0d\rho\ne0, a harmonic conjugate zz of ρ\rho gives isothermal coordinates [3 §2.1]. As in [3 §2.1] and [4 §2.1], we choose the harmonic conjugate for which

dρ∧dz∧dϕ1∧dϕ2>0.(1)d\rho\wedge dz\wedge d\phi^{1}\wedge d\phi^{2}>0. \tag*{(1)}

Then (ρ,z)(\rho,z) are the Weyl coordinates of the basis, and zz is determined up to an additive constant. If the basis is replaced by a basis of determinant 11 relative to it, then ρ\rho is unchanged and zz changes by an additive constant. We use Z\mathbb{Z}-bases of Λ\Lambda, except in Section 4. Since MM is simply connected (Lemma 2.3), the Weyl coordinates of a positively oriented Z\mathbb{Z}-basis of Λ\Lambda are global: they map the orbit space homeomorphically onto the closed half-plane {ρ≥0}\{\rho\ge0\}, and its interior conformally onto the open half-plane [3 Prop. 3.8(ii)].

The fixed points map to nn points w1<⋯<wnw_{1}<\cdots<w_{n} of the axis {ρ=0}\{\rho=0\}, the turning points. They divide the axis into the rods I0=(−∞,w1)I_{0}=(-\infty,w_{1}), Ij=(wj,wj+1)I_{j}=(w_{j},w_{j+1}) for 1≤j≤n−11\le j\le n-1, and In=(wn,∞)I_{n}=(w_{n},\infty). Over the open half-plane the orbits are tori, over each rod the stabiliser is a circle subgroup of T2T^{2}, and over a turning point it is T2T^{2} (Figure 1). The convention (2.1) fixes the direction of the axis, so the numbering of the rods depends on the orientations of MM and t\mathfrak{t}. Over the open half-plane the metric takes the Weyl–Papapetrou form

Orbit space of a toric ALE instanton

Figure 1. The orbit space of a toric ALE instanton with n=4n=4 turning points, in Weyl coordinates. Over the rod IjI_j the circle generated by the rod vector vjv_j collapses, and over a turning point the whole torus collapses to a fixed point. The preimage of the closure of a finite rod IjI_j is a 22-sphere SjS_j.

g=Gab(ρ,z)dϕadϕb+e2ν(ρ,z)(dρ2+dz2),det⁡G=ρ2,(2)g=G_{ab}(\rho,z)d\phi^{a}d\phi^{b}+e^{2\nu(\rho,z)}(d\rho^{2}+dz^{2}),\qquad\det G=\rho^{2}, \tag*{(2)}

and Ricci-flatness amounts to a harmonic map equation for ρ−1G\rho^{-1}G and a quadrature for the conformal factor e2νe^{2\nu} [3 (2.1)–(2.4)].

Rod vectors and chains

2.2. Rod vectors and chains. Let det⁡\det be the area form on t\mathfrak{t} with det⁡(e,f)=1\det(e,f)=1 for every positively oriented Z\mathbb{Z}-basis (e,f)(e,f) of Λ\Lambda. The stabiliser over the rod IjI_j is the circle subgroup generated by a primitive vector vj∈Λv_j\in\Lambda, that is, one that is not of the form kwkw with w∈Λw\in\Lambda and an integer k≥2k\ge2. It is the rod vector, and it is determined up to sign. At the fixed point over wjw_j the vectors vj−1v_{j-1} and vjv_j form a basis of Λ\Lambda [31]. So the signs can be chosen, uniquely up to one common sign, such that

det⁡(vj−1,vj)=−1,1≤j≤n.(3)\det(v_{j-1},v_j)=-1,\qquad1\le j\le n. \tag*{(3)}

Then (vj−1,vj)(v_{j-1},v_j) and (vj,vj+1)(v_j,v_{j+1}) are bases with the same determinant, so vj+1+vj−1v_{j+1}+v_{j-1} is an integer multiple of vjv_j:

vj+1=κjvj−vj−1,κj=−det⁡(vj−1,vj+1)∈Z,1≤j≤n−1.(4)v_{j+1}=\kappa_jv_j-v_{j-1},\qquad\kappa_j=-\det(v_{j-1},v_{j+1})\in\mathbb{Z},\qquad1\leq j\leq n-1. \tag*{(4)}

The turning points w1<⋯<wnw_1<\cdots<w_n together with the rod vectors v0,…,vnv_0,\ldots,v_n form the rod structure of (M,g)(M,g). Since the Weyl coordinates are oriented by (1), the rod vectors and the chain of MM are those of [3, 4], and the results of these papers apply to MM as they are stated. Finally, we identify Λ\Lambda with Z2\mathbb{Z}^2 by the basis (v1,v0)(v_1,v_0), which is positively oriented by (3). Then v0=(0,1)v_0=(0,1), v1=(1,0)v_1=(1,0), and det⁡(a,b)=a1b2−a2b1\det(a,b)=a_1b_2-a_2b_1. The rod vectors are therefore determined by the integers κj\kappa_j, and we consider such sequences on their own.

Definition 2.4. A chain is a vector κ=(κ1,…,κn−1)∈Zn−1\kappa=(\kappa_1,\ldots,\kappa_{n-1})\in\mathbb{Z}^{n-1}, n≥1n\geq1. Its rod vectors are

v0=(0,1),v1=(1,0),vj+1=κjvj−vj−1(1≤j≤n−1),(5)v_0=(0,1),\qquad v_1=(1,0),\qquad v_{j+1}=\kappa_jv_j-v_{j-1}\quad(1\leq j\leq n-1), \tag*{(5)}

and we write vj=(pj,qj)v_j=(p_j,q_j) and (p,q)=(pn,qn)(p,q)=(p_n,q_n), the last rod vector. The intersection form of κ\kappa is the symmetric tridiagonal (n−1)×(n−1)(n-1)\times(n-1) matrix

Qκ=(κ111κ2⋱⋱⋱11κn−1),(6)Q_\kappa= \begin{pmatrix} \kappa_1 & 1 & & & \\ 1 & \kappa_2 & \ddots& & \\ & \ddots& \ddots& 1 & \\ & & 1 & \kappa_{n-1} \end{pmatrix}, \tag*{(6)}

and the degrees of κ\kappa are the numbers rr and rˉ\bar r of negative and of positive eigenvalues of QκQ_\kappa. We call rr the degree of κ\kappa.1 The chain of a toric ALE instanton (M,g)(M,g), with orientations of MM and t\mathfrak{t}, is the chain defined by (4), and its degrees are the degrees of (M,g)(M,g).

By induction, the rod vectors of a chain satisfy (3), and pj=−det⁡(v0,vj)p_j=-\det(v_0,v_j). In particular p=−det⁡(v0,vn)p=-\det(v_0,v_n). Our pp is minus the pp of [4], and from [4] we use only statements about ∣p∣|p|. The chain determines MM with its torus action up to equivariant diffeomorphism [31], [30, proof of Thm. 1], and it has a direct topological meaning [4 §2.1]. The preimage of the closure of a finite rod IjI_j is an embedded 2-sphere $S_j\subset M (Figure 1). Consecutive spheres meet transversally in one fixed point, spheres that are not consecutive are disjoint, the classes of $S_1,\ldots,S_{n-1}$ form a basis of $H_2(M;\mathbb{Z})$, and by {citep}[4 (2.6)]`

Sj⋅Sj=κj,1≤j≤n−1.(7)S_j\cdot S_j=\kappa_j,\qquad1\leq j\leq n-1. \tag*{(7)}

So QκQ_\kappa is the intersection form of MM in the basis (Sj)(S_j), for suitable orientations of the spheres, and the degrees are r=b2−(M)r=b_2^-(M) and rˉ=b2+(M)\bar r=b_2^+(M). Reversing the orientation of MM replaces κ\kappa by −(κn−1,…,κ1)-(\kappa_{n-1},\ldots,\kappa_1) [4 §2.1] and exchanges the two degrees. Reversing the orientation of t\mathfrak{t} replaces κ\kappa by (κn−1,…,κ1)(\kappa_{n-1},\ldots,\kappa_1) and leaves the degrees unchanged.

Remark 2.5 (orientation conventions). With the opposite convention dρ∧dz∧dϕ1∧dϕ2<0\mathrm{d}\rho\wedge\mathrm{d}z\wedge\mathrm{d}\phi^1\wedge\mathrm{d}\phi^2<0, the chain assigned to MM is −(κn−1,…,κ1)-(\kappa_{n-1},\ldots,\kappa_1), so its entries are minus the self-intersection numbers of the spheres, in reverse order. In that convention the chain of the minimal resolution of a cyclic quotient singularity C2/Γm,a\mathbb{C}^2/\Gamma_{m,a} (Section 2.3), with its complex orientation, consists of the entries of the Hirzebruch–Jung continued fraction of m/am/a, all of which are ≥2\geq2. This is the usual convention in algebraic geometry [18], [10, §2.6], and the multi-Eguchi–Hanson metrics have the chain (2,…,2)(2,\ldots,2) in their hyperkähler orientation. We follow the convention of [3, 4], in which they have the chain (−2,…,−2)(-2,\ldots,-2) (Example 2.7).

The end

The two semi-infinite rods determine the end of MM. We recall the arithmetic needed to describe it. For an integer m≥1m \ge1, two integers aa and a′a' are congruent modulo mm, a≡a′a \equiv a' (mod mm), if mm divides a−a′a-a'. A unit modulo mm is an integer prime to mm. The units are exactly the integers aa for which aa′≡1aa' \equiv1 (mod mm) for some integer a′a', an inverse of aa modulo mm, and the number of units among 1,…,m1,\ldots,m is Euler’s function φ(m)\varphi(m). The integers 11 and −1-1 are always units, and they are the only ones up to congruence exactly when φ(m)≤2\varphi(m) \le2, that is, when m∈{1,2,3,4,6}m \in\{1,2,3,4,6\}. For m=5m=5, for instance, the unit 22 is congruent to neither 11 nor −1-1.

For a unit aa modulo mm let Γm,a⊂U(2)\Gamma_{m,a} \subset\mathrm{U}(2) be the cyclic group generated by diag⁡(ζ,ζa)\operatorname{diag}(\zeta,\zeta^a), ζ=e2πi/m\zeta=e^{2\pi i/m}. It acts freely on S3S^3, and the quotient is the lens space L(m,a)L(m,a). The lens spaces L(m,a)L(m,a) and L(m,a′)L(m,a') are homeomorphic if and only if a′≡±aa' \equiv\pm a or aa′≡±1aa' \equiv\pm1 (mod mm) [7], and then they are diffeomorphic. Since det⁡(vn−1,vn)=−1\det(v_{n-1},v_n)=-1, the coordinates pp and qq of the last rod vector are coprime, so that qq is a unit modulo ∣p∣|p| if p≠0p \ne0.

Lemma 2.6. Let (M,g)(M,g) be a toric ALE instanton with the chain κ\kappa, and let Γ⊂O(4)\Gamma\subset\mathrm{O}(4) be its group at infinity. Then p≠0p \ne0, Γ≅Z∣p∣\Gamma\cong\mathbb{Z}_{|p|}, the cross-section S3/ΓS^3/\Gamma of the end is diffeomorphic to L(∣p∣,q)L(|p|,q), and Γ\Gamma is conjugate in O(4)\mathrm{O}(4) to a subgroup of SU(2)\mathrm{SU}(2) if and only if q≡±1q \equiv\pm1 (mod ∣p∣|p|).

Proof. Let Σ≅S3/Γ\Sigma\cong S^3/\Gamma be an invariant cross-section of the end (see Remark 2.2). Its orbit space is a compact 1-manifold in the closed half-plane [31 §2]. It is not a circle, since then Σ\Sigma would be a torus bundle over the circle, whose fundamental group is infinite. So it is an arc with its end points on the axis. Since Σ\Sigma separates the fixed points from infinity, the arc separates the turning points from infinity in the half-plane, so its end points lie on the rods I0I_0 and InI_n, with the stabilisers generated by v0v_0 and vnv_n, and the stabiliser is trivial in between. If p=0p=0, then vn=±v0v_n=\pm v_0, and Σ\Sigma is diffeomorphic to S1×S2S^1 \times S^2, whose fundamental group is infinite. Since Γ\Gamma is finite, p≠0p \ne0. A 3-manifold with such an action is determined up to equivariant diffeomorphism by the lattice and the two stabilisers [31 §2], and comparison with a linear action on a lens space shows that Σ\Sigma is L(∣p∣,q)L(|p|,q) [30 Lemma 3]. Hence Γ=π1(Σ)≅Z∣p∣\Gamma=\pi_1(\Sigma)\cong\mathbb{Z}_{|p|}.

No non-trivial element of Γ\Gamma has the eigenvalue 11, since Γ\Gamma acts freely on S3S^3. Every element of O(4)∖SO(4)\mathrm{O}(4)\setminus\mathrm{SO}(4) has the eigenvalue 11, so Γ⊂SO(4)\Gamma\subset\mathrm{SO}(4), and Γ\Gamma is conjugate in O(4)\mathrm{O}(4) to Γ∣p∣,a\Gamma_{|p|,a} for a unit aa. By the classification of lens spaces, a≡±qa\equiv\pm q or aq≡±1aq\equiv\pm1. A generator of Γ∣p∣,a\Gamma_{|p|,a} rotates two orthogonal planes by the angles 2π/∣p∣2\pi/|p| and 2πa/∣p∣2\pi a/|p|, while an element of SU(2)\mathrm{SU}(2) rotates two orthogonal planes by opposite angles. So Γ∣p∣,a\Gamma_{|p|,a} is conjugate in O(4)\mathrm{O}(4) into SU(2)\mathrm{SU}(2) if and only if a≡±1a\equiv\pm1, that is, if and only if q≡±1q\equiv\pm1. □

We say that MM has the end L(m,a)L(m,a) if the cross-section S3/ΓS^3/\Gamma is diffeomorphic to L(m,a)L(m,a), without reference to an orientation. For a chain with p≠0p \ne0 we likewise call L(∣p∣,q)L(|p|,q) its end. By Lemma 2.6 and the classification of lens spaces, the following three conditions are therefore equivalent: Γ\Gamma is conjugate in O(4)\mathrm{O}(4) to a subgroup of SU(2)\mathrm{SU}(2), the end is L(m,1)L(m,1) with m=∣Γ∣m=|\Gamma|, and q≡±1q\equiv\pm1 (mod ∣p∣|p|).

Example 2.7. The multi-Eguchi–Hanson metrics with nn collinear centres are the hyperkähler ALE metrics obtained from the Gibbons–Hawking ansatz with nn centres on a line in R3\mathbb{R}^3 [14], [26 Ex. 4.3], [4 (2.10)]. For n=1n=1 the metric is flat, and for n=2n=2 it is the Eguchi–Hanson metric [9]. They are toric ALE instantons whose turning points are the centres [3 §11.2]. In their hyperkähler orientation the chain is (−2,…,−2)(-2,\ldots,-2), and QRQ_R is minus the Cartan matrix of An−1A_{n-1}, so that r=n−1r=n-1 and rˉ=0\bar r=0 [4 Rem. 3.6]. With the opposite orientation the chain is (2,…,2)(2,\ldots,2), and the two degrees are exchanged. The rod vectors and the end L(n,1)L(n,1) are given in Table 1.

chain κ\kappannrod vectors v0,…,vnv_0,\ldots,v_n(p,q)(p,q)end(r,rˉ)(r,\bar r)
(2,…,2)(2,\ldots,2)nnvj=(j,1−j)v_j=(j,1-j)(n,1−n)(n,1-n)L(n,1)L(n,1)(0,n−1)(0,n-1)
(−2,…,−2)(-2,\ldots,-2)nnvj=(−1)j+1(j,j−1)v_j=(-1)^{j+1}(j,j-1)(−1)n+1(n,n−1)(-1)^{n+1}(n,n-1)L(n,1)L(n,1)(n−1,0)(n-1,0)
(2,2,2)(2,2,2)44(0,1),(1,0),(2,−1),(3,−2),(4,−3)(0,1),(1,0),(2,-1),(3,-2),(4,-3)(4,−3)(4,-3)L(4,1)L(4,1)(0,3)(0,3)
(1,−1,1,−1)(1,-1,1,-1)55(0,1),(1,0),(1,−1),(−2,1),(−3,2),(5,−3)(0,1),(1,0),(1,-1),(-2,1),(-3,2),(5,-3)(5,−3)(5,-3)L(5,2)L(5,2)(2,2)(2,2)
(2,−2)(2,-2)33(0,1),(1,0),(2,−1),(−5,2)(0,1),(1,0),(2,-1),(-5,2)(−5,2)(-5,2)L(5,2)L(5,2)(1,1)(1,1)

Table 1. The chains used as examples, with nn the number of turning points, the rod vectors (5), the last rod vector (p,q)(p,q), the end (Lemma 2.6), and the degrees rr and rˉ\bar r (Definition 2.4). In the first two rows 0≤j≤n0\leq j\leq n.

Table 1 collects the data of the chains that serve as examples in this paper: the chains of Example 2.7, the chain (2,2,2)(2,2,2) of the multi-Eguchi–Hanson metrics with four centres, with the orientation opposite to the hyperkähler one, and the chains (1,−1,1,−1)(1,-1,1,-1) and (2,−2)(2,-2), whose end L(5,2)L(5,2) is not covered by our argument (Example 4.8). The rod vectors follow from (5), the ends from Lemma 2.6 and the classification of lens spaces, and the degrees from the eigenvalues of QκQ_\kappa, or more quickly from Lemma 3.1 below.

Chains and their degrees

The degree rr of a chain was defined in Definition 2.4 as the number of negative eigenvalues of the intersection form QκQ_\kappa of (6). It can also be read off from the rotation of the rod vectors, and this description identifies rˉ\bar r with the degree of Li and Sun. By (3), each vjv_j is obtained from vj−1v_{j-1} by a clockwise rotation through an angle θj∈(0,π)\theta_j\in(0,\pi), followed by a dilation. Here clockwise refers to the orientation of Λ=Z2\Lambda=\mathbb{Z}^2, and the angles are measured in the Euclidean structure of R2\mathbb{R}^2. The total turning of the rod vectors is

Θ=θ1+⋯+θn,(8)\Theta=\theta_1+\cdots+\theta_n, \tag*{(8)}

the clockwise angle from v0v_0 to vnv_n swept out by the rod vectors. We also put ε=(−1)r\varepsilon=(-1)^r.

Lemma 3.1. Let κ\kappa be a chain with p≠0p\ne0, with the intersection form QκQ_\kappa of (6), the degrees rr and rˉ\bar r of Definition 2.4, and the total turning Θ\Theta of (8). Then det⁡Qκ=p\det Q_\kappa=p, so that QκQ_\kappa is non-degenerate and r+rˉ=n−1r+\bar r=n-1. Moreover Θ∉πZ\Theta\notin\pi\mathbb{Z}, and

r=⌊Θπ⌋(9)r=\left\lfloor\frac{\Theta}{\pi}\right\rfloor \tag*{(9)}

is the number of multiples of π\pi in the interval (0,Θ)(0,\Theta). Equivalently, rr is the number of sign changes in the sequence (p1,…,pn)(p_1,\ldots,p_n) of first coordinates of the rod vectors, where a vanishing term counts as one sign change between its neighbours. In particular pp has the sign ε=(−1)r\varepsilon=(-1)^r.

Proof. Let DjD_j be the leading principal j×jj\times j minor of QκQ_\kappa, with D0=1D_0=1. Expanding along the last row gives D1=κ1D_1=\kappa_1 and Dj=κjDj−1−Dj−2D_j=\kappa_jD_{j-1}-D_{j-2}, while (5) gives p0=0p_0=0, p1=1p_1=1, and pj+1=κjpj−pj−1p_{j+1}=\kappa_jp_j-p_{j-1}. By induction,

Dj=pj+1,0≤j≤n−1.(10)D_j=p_{j+1},\qquad0\leq j\leq n-1. \tag*{(10)}

In particular det⁡Qκ=Dn−1=p≠0\det Q_\kappa=D_{n-1}=p\ne0.

Let Θj=θ1+⋯+θj\Theta_j=\theta_1+\cdots+\theta_j be the clockwise angle from v0v_0 to vjv_j, so that Θn=Θ\Theta_n=\Theta. Since v0=(0,1)v_0=(0,1), pj=∣vj∣sin⁡Θjp_j=|v_j|\sin\Theta_j. The angles Θj\Theta_j increase in steps less than π\pi. So a vanishing pjp_j with 0<j<n0<j<n has neighbours pj−1,pj+1p_{j-1},p_{j+1} of opposite signs, and two consecutive terms do not vanish. Since Θ1∈(0,π)\Theta_1\in(0,\pi) and pn=p≠0p_n=p\ne0, the number r′r' of sign changes in (p1,…,pn)(p_1,\ldots,p_n), counted as in the statement, is the number of multiples of π\pi in (0,Θ)(0,\Theta), and Θ≠πZ\Theta\ne\pi\mathbb{Z}. It remains to show that r=r′r=r'. We use Jacobi’s rule for the signature of a real symmetric matrix, with Gundelfinger’s supplement for a vanishing minor [11 Ch. X, §3, Thm. 2 and Note 1]. For a symmetric tridiagonal matrix with non-zero entries next to the diagonal and non-zero determinant, the leading principal minors (D0,…,Dn−1)(D_0,\ldots,D_{n-1}) have no two consecutive zeros, and a vanishing minor has neighbours of opposite signs, since Dj+1=−Dj−1D_{j+1}=-D_{j-1} when Dj=0D_j=0. The rules then say that the number of negative eigenvalues is the number of sign changes in this sequence, where a vanishing minor counts as one sign change. By (10) this sequence is (p1,…,pn)(p_1,\ldots,p_n), with the same convention for vanishing terms, so r=r′r=r'. Finally p=∣vn∣sin⁡Θp=|v_n|\sin\Theta and Θ∈(rπ,(r+1)π)\Theta\in(r\pi,(r+1)\pi), so pp has the sign (−1)r(-1)^r. □

For the chains of Table 1 the lemma can be checked directly. The rod vectors of the chain (1,−1,1,−1)(1,-1,1,-1) have the first coordinates 1,1,−2,−3,51,1,-2,-3,5, which change sign twice, so r=2r=2. The chains (2,…,2)(2,\ldots,2) have the positive first coordinates 1,2,…,n1,2,\ldots,n, so r=0r=0, while the chains (−2,…,−2)(-2,\ldots,-2) have the first coordinates (−1)j+1j(-1)^{j+1}j, so r=n−1r=n-1.

Remark 3.2 (chains as walks in the Farey graph). The vertices of the Farey graph are the points of Q∪{∞}\mathbb{Q}\cup\{\infty\}, and two fractions a2/a1a_2/a_1 and b2/b1b_2/b_1 in lowest terms are adjacent if ∣a1b2−a2b1∣=1|a_1b_2-a_2b_1|=1 [16 Ch. 1–3]. By (3), consecutive rod vectors form a basis of Z2\mathbb{Z}^2. So their slopes xj=qj/pjx_j=q_j/p_j, with 1/0=∞1/0=\infty, are fractions in lowest terms, and consecutive slopes are adjacent. Hence a chain is a walk x0=∞x_0=\infty, x1=0x_1=0, x2,…,xn=q/px_2,\ldots,x_n=q/p in the Farey graph, that is, a sequence of vertices in which consecutive ones are adjacent, from the slope of v0v_0 to that of vnv_n. The slope decreases along each step on the circle R∪{∞}\mathbb{R}\cup\{\infty\}, since vj+1v_{j+1} is obtained from vjv_j by a clockwise rotation through an angle in (0,π)(0,\pi). By Lemma 3.1, rr is the number of steps from xjx_j to xj+1x_{j+1}, 1≤j≤n−11\leq j\leq n-1, with pjpj+1<0p_jp_{j+1}<0 or pj+1=0p_{j+1}=0, that is, of the steps after the first that cross or reach ∞\infty. For the chain (1,−1,1,−1)(1,-1,1,-1) the walk is ∞,0,−1,−12,−23,−35\infty,0,-1,-\frac{1}{2},-\frac{2}{3},-\frac{3}{5}, and these are the steps from −1-1 to −12-\frac{1}{2} and from −23-\frac{2}{3} to −35-\frac{3}{5}.

Remark 3.3 (the degree of Li and Sun). For a toric ALE instanton, rˉ\bar r is the degree that Li and Sun assign to the rod structure of (M,g)(M,g) [24 (3.2), Def. 3.11], and rr is the degree of the rod structure of MM with the opposite orientation. We use the description of their degree in [4 §2.1], in the orientation convention (1). Each of the lines Rv0,…,Rvn\mathbb{R}v_0,\ldots,\mathbb{R}v_n spanned by the rod vectors is rotated counter-clockwise through an angle in (0,π)(0,\pi) onto the next, and finally Rvn\mathbb{R}v_n onto Rv0\mathbb{R}v_0. Since vjv_j is obtained from vj−1v_{j-1} by a clockwise rotation through θj\theta_j, the first nn angles are π−θ1,…,π−θn\pi-\theta_1,\ldots,\pi-\theta_n. The last angle, counter-clockwise from Rvn\mathbb{R}v_n to Rv0\mathbb{R}v_0, is the clockwise angle from Rv0\mathbb{R}v_0 to Rvn\mathbb{R}v_n, which is Θ−rπ∈(0,π)\Theta-r\pi\in(0,\pi) by Lemma 3.1, as p≠0p\ne0 by Lemma 2.6. The sum is (n−r)π(n-r)\pi, and for an ALE end the degree is this sum divided by π\pi, minus 11, that is, n−1−r=rˉn-1-r=\bar r. Reversing the orientation of MM exchanges rr and rˉ\bar r (§ 2.2), which gives the second statement.

The closed formula and the diamond inequality

The proof of the main theorem needs one consequence of the analytic theory of toric gravitational instantons: for a toric ALE instanton whose degree rr satisfies 1≤r≤n−21\leq r\leq n-2, the interior rod vectors v1,…,vn−1v_1,\ldots,v_{n-1} lie in the open parallelogram with the vertices ±v0\pm v_0 and ±vn\pm v_n, the diamond of the end. This inequality is due to [4]. It follows from the closed formula for the metric [3], which expresses the rod vectors through Hankel determinants of a positive measure on the turning points, and from Heine’s formula, which writes these determinants as sums of positive terms. We recall Hankel determinants, state the closed formula, and include the short proof of the inequality for completeness.

Hankel determinants and Heine’s formula

Let λ=∑k=1Nλkδxk\lambda= \sum_{k=1}^{N} \lambda_k\delta_{x_k} be a measure with complex weights λk\lambda_k on distinct real points x1,…,xNx_1,\ldots,x_N. Its moments are hi(λ)=∑kλkxkih_i(\lambda)=\sum_k\lambda_kx_k^i, i≥0i\geq0, and its Hankel determinants are

Hs(λ)=det⁡(hi+l(λ))i,l=0s−1(s≥1),H0(λ)=1.(11)H_s(\lambda)=\det\left(h_{i+l}(\lambda)\right)_{i,l=0}^{s-1}\quad(s\geq1),\qquad H_0(\lambda)=1. \tag*{(11)}

For a set S⊂{1,…,N}S\subset\{1,\ldots,N\} the Heine term is

PS(λ)=∏k∈Sλk∏k,l∈Sk<l(xk−xl)2,(12)P_S(\lambda)=\prod_{k\in S}\lambda_k\prod_{\substack{k,l\in S\\k<l}}(x_k-x_l)^2, \tag*{(12)}

with empty products equal to 1.

Lemma 4.1 (Heine’s formula). For every s≥0s\geq0,

Hs(λ)=∑S⊂{1,…,N}∣S∣=sPS(λ).(13)H_s(\lambda)=\sum_{\substack{S\subset\{1,\ldots,N\}\\|S|=s}}P_S(\lambda). \tag*{(13)}

Proof. For s=0s=0 both sides equal 1. For s≥1s\geq1 let WW be the N×sN\times s matrix with the entries Wkl=xklW_{kl}=x_k^l, 0≤l≤s−10\leq l\leq s-1. Then WTdiag⁡(λ1,…,λN)W=(hi+l(λ))i,lW^{T}\operatorname{diag}(\lambda_1,\ldots,\lambda_N)W=(h_{i+l}(\lambda))_{i,l}, and the Cauchy–Binet formula gives Hs(λ)=∑∣S∣=s∏k∈Sλkdet⁡(WS)2H_s(\lambda)=\sum_{|S|=s}\prod_{k\in S}\lambda_k\det(W_S)^2, where WSW_S is the s×ss\times s submatrix of the rows in SS. This is a Vandermonde matrix, with det⁡(WS)2=∏k<l∈S(xk−xl)2\det(W_S)^2=\prod_{k<l\in S}(x_k-x_l)^2. □

Formula (13) goes back to Heine [17]. See also [33 Ch. II] and [3 Lemma 5.1]. In particular, if all weights are positive, then Hs(λ)>0H_s(\lambda)>0 for 0≤s≤N0\leq s\leq N, and Hs(λ)=0H_s(\lambda)=0 for s>Ns>N.

Now let μ=∑k=1nωkδwk\mu=\sum_{k=1}^{n}\omega_k\delta_{w_k} be a positive measure on nn points w1<⋯<wnw_1<\cdots<w_n, so that ωk>0\omega_k>0. For 0≤j≤n0\leq j\leq n let

μj=−∑k≤jωkδwk+∑k>jωkδwk(14)\mu_j=-\sum_{k\leq j}\omega_k\delta_{w_k}+\sum_{k>j}\omega_k\delta_{w_k} \tag*{(14)}

be μ\mu with the signs of its first jj weights reversed. By Heine’s formula, Hs(μj)H_s(\mu_j) is a sum of the positive Heine terms PS(μ)P_S(\mu) with signs, and this bounds its size. For two points, for example, H1(μ1)=ω2−ω1H_1(\mu_1)=\omega_2-\omega_1 lies strictly between −H1(μ)-H_1(\mu) and H1(μ)H_1(\mu).

Lemma 4.2 ([4 Lemma 3.1]). Let μ\mu be a positive measure on n≥1n\geq1 points and 0≤s≤n0\leq s\leq n. Then Hs(μ)>0H_s(\mu)>0, and the following hold.

(i) Hs(μn)=(−1)sHs(μ)H_s(\mu_n)=(-1)^sH_s(\mu), and Hn(μj)=(−1)jHn(μ)H_n(\mu_j)=(-1)^jH_n(\mu) for 0≤j≤n0\leq j\leq n.

(ii) If 1≤s≤n−11\leq s\leq n-1, then ∣Hs(μj)∣<Hs(μ)\lvert H_s(\mu_j)\rvert<H_s(\mu) for 0<j<n0<j<n.

Proof. By (12), PS(μj)=(−1)∣S∩{1,…,j}∣PS(μ)P_S(\mu_j)=(-1)^{|S\cap\{1,\ldots,j\}|}P_S(\mu), and PS(μ)>0P_S(\mu)>0. So by (13), Hs(μj)H_s(\mu_j) is the sum of the positive terms PS(μ)P_S(\mu), ∣S∣=s|S|=s, with these signs. For j=nj=n every set of ss atoms has the sign (−1)s(-1)^s, and for s=ns=n the only set is {1,…,n}\{1,\ldots,n\}, with the sign (−1)j(-1)^j. This proves (i). The triangle inequality gives ∣Hs(μj)∣≤Hs(μ)\lvert H_s(\mu_j)\rvert\leq H_s(\mu), with equality only if all signs agree. Let 0<j<n0<j<n and 1≤s≤n−11\leq s\leq n-1, and choose a set T⊂{2,…,n−1}T\subset\{2,\ldots,n-1\} of s−1s-1 atoms. The sets T∪{1}T\cup\{1\} and T∪{n}T\cup\{n\} carry opposite signs, since the first atom is reversed in μj\mu_j and the last one is not. Hence the inequality is strict, which proves (ii). □

The closed formula

The closed formula is built from the following functions. Let w1<⋯<wnw_1<\cdots<w_n be points of the axis, μ=∑kωkδwk\mu=\sum_k\omega_k\delta_{w_k} a positive measure, and 0≤s≤n−10\leq s\leq n-1. For ρ≥0\rho\geq0 and (ρ,z)(\rho,z) not one of the points (0,wk)(0,w_k) put

Rk=(ρ2+(z−wk)2)1/2,σk=wk−z−iρRk,(15)R_k=(\rho^2+(z-w_k)^2)^{1/2},\qquad\sigma_k=\frac{w_k-z-i\rho}{R_k}, \tag*{(15)}

so that ∣ϖk∣=1\lvert\varpi_k\rvert= 1. Let ϖμ\varpi\mu, Rμ\mathbf{R}\mu, and μ/R\mu/\mathbf{R} be the measures on the points wkw_k with the weights ϖkωk\varpi_k\omega_k, RkωkR_k\omega_k, and ωk/Rk\omega_k/R_k, and let

X=Hs(ϖμ),Y=Hs+1(ϖμ),F1=Hs(Rμ),F2=Hs+1(μ/R).(16)\mathcal{X}=H_s(\varpi\mu),\qquad\mathcal{Y}=H_{s+1}(\varpi\mu),\qquad F_1=H_s(\mathbf{R}\mu),\qquad F_2=H_{s+1}(\mu/\mathbf{R}). \tag*{(16)}

By Heine’s formula, F1>0F_1>0 and F2>0F_2>0, since s+1≤ns+1\le n. On the rod IjI_j, 0≤j≤n0\le j\le n, we have ρ=0\rho=0, ϖk=−1\varpi_k=-1 for k≤jk\le j, and ϖk=1\varpi_k=1 for k>jk>j, so that ϖμ=μj\varpi\mu=\mu_j there. For 0≤j≤n0\le j\le n put

Vj=(Hs(μj),Hs+1(μj))∈R2.(17)V_j=\left(H_s(\mu_j),H_{s+1}(\mu_j)\right)\in\mathbb{R}^2. \tag*{(17)}

Theorem 4.3 ([3 Thm. 10.1, (10.1)], in the form of [4 Thm. 2.4]). Let (M,g)(M,g) be a toric ALE instanton with nn turning points and the rod vectors v0,…,vnv_0,\ldots,v_n of § 2.2. Then there are an integer ss with 0≤s≤n−10\le s\le n-1, a positively oriented basis (f0,f1)(f_0,f_1) of t\mathfrak{t}, a positive measure μ=∑kωkδwk\mu=\sum_k\omega_k\delta_{w_k} on the turning points w1<⋯<wnw_1<\cdots<w_n, in the Weyl coordinates (ρ,z)(\rho,z) of this basis with zz chosen as in (2.1), and a constant c>0c>0 with the following properties, where X\mathcal{X}, Y\mathcal{Y}, F1F_1, and F2F_2 are defined by (4.5) and (4.6), and VjV_j by (4.4) and (4.7), with the Hankel determinants HsH_s and Hs+1H_{s+1} of the orders ss and s+1s+1.

(i) With the angles ψ0,ψ1\psi^0,\psi^1 dual to (f0,f1)(f_0,f_1),

g=∣Y dψ0−X dψ1∣2F1F2+cF1F2(dρ2+dz2).(18)g=\frac{\lvert\mathcal{Y}\,\mathrm{d}\psi^0-\mathcal{X}\,\mathrm{d}\psi^1\rvert^2}{F_1F_2}+cF_1F_2(\mathrm{d}\rho^2+\mathrm{d}z^2). \tag*{(18)}

(ii) After a change of the common sign of the rod vectors, which does not change the chain, the components of vjv_j with respect to (f0,f1)(f_0,f_1) are

vj=c1/2Vj,0≤j≤n.(19)v_j=c^{1/2}V_j,\qquad0\le j\le n. \tag*{(19)}

Theorem 4.3 is the detailed form of the theorem that toric gravitational instantons are gravitational solitons [3 Thm. 1], for an ALE end (for a Pfaffian form of the Belinski–Zakharov solitons see also [1 §§2.3–2.4]). Its hypotheses hold by Lemma 2.3. The soliton data of [3, 4] contain a further number γ≥0\gamma\ge0, the weight of an atom at infinity, which vanishes exactly for ALE ends [4 §2.3], so that [4 (2.16)] reduces to (4.6) and (4.8). Items (i) and (ii) are items (ii) and (iii) of [4 Thm. 2.4], and ss, the order of the Hankel determinants, is the degree index of [3, 4]. Since (f0,f1)(f_0,f_1) is positively oriented, its Weyl coordinates differ from those of a positively oriented Z\mathbb{Z}-basis of Λ\Lambda by a positive factor and a translation, so the turning points, the rods, and their numbering are those of § 2.1. Item (ii) says that along each rod the generator of the stabiliser is given by Hankel determinants of the measure μ\mu with some signs reversed. Item (i) of [4 Thm. 2.4] relates the degree index to the degree of Li and Sun, and with Remark 3.3 it identifies the degree index with our degree rr. The following lemma gives a direct argument from item (ii).

Lemma 4.4. The integer ss of Theorem 4.3 equals the degree rr of (M,g)(M,g) of Definition 2.4.

Proof. Let

V~j=(Hs(μj)Hs(μ),Hs+1(μj)Hs+1(μ)),0≤j≤n,(20)\widetilde{V}_j=\left(\frac{H_s(\mu_j)}{H_s(\mu)},\frac{H_{s+1}(\mu_j)}{H_{s+1}(\mu)}\right),\qquad0\le j\le n, \tag*{(20)}

be the normalised Hankel family of [3 §5]. It is the image of (Vj)(V_j) under a diagonal linear map with positive entries. By (4.9), and since (f0,f1)(f_0,f_1) is positively oriented, det⁡(vi,vj)\det(v_i,v_j) is therefore a positive multiple of the determinant det⁡(V~i,V~j)\det(\widetilde{V}_i,\widetilde{V}_j) of the components, for 0≤i,j≤n0\le i,j\le n. In particular det⁡(V~j−1,V~j)<0\det(\widetilde{V}_{j-1},\widetilde{V}_j)<0 by (2.3), so V~j\widetilde{V}_j is obtained from V~j−1\widetilde{V}_{j-1} by a clockwise rotation through an angle in (0,π)(0,\pi), followed by a dilation, and these nn angles add up to (s+12)π(s+\frac{1}{2})\pi [3 Prop. 5.5]. By Lemma 4.2(i), V~0=(1,1)\widetilde{V}_0=(1,1) and V~n=(−1)s(1,−1)\widetilde{V}_n=(-1)^s(1,-1), so det⁡(V~0,V~n)≠0\det(\widetilde{V}_0,\widetilde{V}_n)\ne0. The argument of the proof of Lemma 3.1, with V~0\widetilde{V}_0 in place of v0v_0, shows that the number of sign changes in the sequence (det⁡(V~0,V~1),…,det⁡(V~0,V~n))(\det(\widetilde{V}_0,\widetilde{V}_1),\ldots,\det(\widetilde{V}_0,\widetilde{V}_n)), counted as there, is the number of multiples of π\pi in (0,(s+12)π)(0,(s+\frac{1}{2})\pi), that is, ss. This sequence has the signs of (−p1,…,−pn)(-p_1,\ldots,-p_n), so by Lemma 3.1 the number of its sign changes is rr. Hence s=rs=r. □

The diamond inequality

Recall that p=−det⁡(v0,vn)p=-\det(v_0,v_n) is the first coordinate of the last rod vector vn=(p,q)v_n=(p,q). Its absolute value is the area of the parallelogram spanned by v0v_0 and vnv_n, and by Lemma 2.6 it is the order ∣Γ∣|\Gamma| of the group at infinity. Since p≠0p\ne0, the vectors v0v_0 and vnv_n form a basis of R2\mathbb{R}^2, and we write

v=αv0+βvn,α=−det⁡(v,vn)p,β=−det⁡(v0,v)p,(21)v=\alpha v_0+\beta v_n,\qquad\alpha=-\frac{\det(v,v_n)}{p},\qquad\beta=-\frac{\det(v_0,v)}{p}, \tag*{(21)}

for v∈R2v\in\mathbb{R}^2, and vj=αjv0+βjvnv_j=\alpha_jv_0+\beta_jv_n for the rod vectors. Recall that ε=(−1)r\varepsilon=(-1)^r. By Lemma 4.4, the order ss of the Hankel determinants in Theorem 4.3 is rr. The following comparison of components is taken from the proof of [4 Cor. 3.2(d)].

Lemma 4.5. Let (M,g)(M,g) be a toric ALE instanton with the degree rr, and let μ\mu be as in Theorem 4.3. Then for 0≤j≤n0\le j\le n,

Hr(μj)Hr(μ)=αj+εβj,Hr+1(μj)Hr+1(μ)=αj−εβj.(22)\frac{H_r(\mu_j)}{H_r(\mu)}=\alpha_j+\varepsilon\beta_j,\qquad\frac{H_{r+1}(\mu_j)}{H_{r+1}(\mu)}=\alpha_j-\varepsilon\beta_j. \tag*{(22)}

Proof. Let A\mathcal{A} be the linear map that sends v∈tv\in\mathfrak{t}, with the components (x0,x1)(x_0,x_1) with respect to (f0,f1)(f_0,f_1), to c−1/2(x0/Hr(μ),x1/Hr+1(μ))c^{-1/2}(x_0/H_r(\mu),x_1/H_{r+1}(\mu)). By (4.9), Avj\mathcal{A}v_j is the vector of the two quotients in (4.12). By Lemma 4.2(i), Av0=(1,1)\mathcal{A}v_0=(1,1) and Avn=ε(1,−1)\mathcal{A}v_n=\varepsilon(1,-1). Since A\mathcal{A} is linear, Avj=αj(1,1)+βjε(1,−1)\mathcal{A}v_j=\alpha_j(1,1)+\beta_j\varepsilon(1,-1), which is (4.12). □

Definition 4.6. For a chain with p≠0p\ne0 the open diamond of its end is the open parallelogram with the vertices ±v0\pm v_0 and ±vn\pm v_n,

D(v0,vn)={αv0+βvn:∣α∣+∣β∣<1}={v∈R2:∣det⁡(v0,v)∣+∣det⁡(v,vn)∣<∣p∣}.(23)D(v_0,v_n)=\{\alpha v_0+\beta v_n:|\alpha|+|\beta|<1\}=\{v\in\mathbb{R}^2:|\det(v_0,v)|+|\det(v,v_n)|<|p|\}. \tag*{(23)}

The two descriptions agree by (4.11). The open diamond is convex and invariant under v↦−vv\mapsto-v. Since ∣a∣+∣b∣=max⁡(∣a+b∣,∣a−b∣)|a|+|b|=\max(|a+b|,|a-b|), it is the set where ∣α+β∣<1|\alpha+\beta|<1 and ∣α−β∣<1|\alpha-\beta|<1. The end rod vectors v0v_0 and vnv_n are vertices of the diamond, where the sum in (4.13) equals ∣p∣|p|. The diamond inequality below concerns only the interior rod vectors v1,…,vn−1v_1,\ldots,v_{n-1}.

Proposition 4.7 (the diamond inequality, [4 Cor. 3.2(d)]). Let (M,g)(M,g) be a toric ALE instanton with nn turning points and the degree rr, and suppose that 1≤r≤n−21\le r\le n-2. Then the interior rod vectors lie in the open diamond of the end:

∣det⁡(v0,vj)∣+∣det⁡(vj,vn)∣<∣p∣,0<j<n.(24)|\det(v_0,v_j)|+|\det(v_j,v_n)|<|p|,\qquad0<j<n. \tag*{(24)}

Proof. Both rr and r+1r+1 lie in {1,…,n−1}\{1,\ldots,n-1\}. So Lemma 4.2(ii), applied to the orders rr and r+1r+1, and Lemma 4.5 give ∣αj+εβj∣<1|\alpha_j+\varepsilon\beta_j|<1 and ∣αj−εβj∣<1|\alpha_j-\varepsilon\beta_j|<1 for 0<j<n0<j<n, that is, ∣αj∣+∣βj∣<1|\alpha_j|+|\beta_j|<1. □

In the degrees r=0r=0 and r=n−1r=n-1 excluded in Proposition 4.7 the metric is hyperkähler (Proposition 5.3). For r=0r=0 the first quotient in (4.12) is H0(μj)/H0(μ)=1H_0(\mu_j)/H_0(\mu)=1, so αj+βj=1\alpha_j+\beta_j=1, and the second one has absolute value at most 11 by Lemma 4.2, so ∣αj−βj∣≤1|\alpha_j-\beta_j|\le1. Hence αj,βj≥0\alpha_j,\beta_j\ge0, and all rod vectors lie on the segment from v0v_0 to vnv_n, on the boundary of the diamond. For the chains (2,…,2)(2,\ldots,2), for instance, vj=(j,1−j)=(1−jn)v0+jnvnv_j=(j,1-j)=(1-\frac{j}{n})v_0+\frac{j}{n}v_n.

Example 4.8. We compare the diamonds of two ends (Figure 2). For v=(x,y)v=(x,y) we have det⁡(v0,v)=−x\det(v_0,v)=-x.

The open diamond with rod vectors and lattice points

Figure 2. The open diamond D(v0,vn)D(v_0,v_n) (shaded) with the rod vectors (arrows) and its non-zero lattice points (red dots), for two chains of Table 1. The end rod vectors v0v_0 and vnv_n are vertices of the diamond. (a) For the chain (2,2,2)(2,2,2) the rod vectors lie on the boundary of the diamond, on the segment from v0v_0 to v4v_4, and the only non-zero lattice points inside are ±u=±(1,−1)\pm u=\pm(1,-1), joined by a dashed segment, and uu is not a rod vector. (b) For the chain (1,−1,1,−1)(1,-1,1,-1) the diamond contains eight non-zero lattice points on four lines, and all four interior rod vectors lie inside.

The chain (2,2,2)(2,2,2) of Table 1 has p=4>0p=4>0. Here det⁡(v,v4)=−(3x+4y)\det(v,v_4)=-(3x+4y), so the open diamond is the set where ∣x∣+∣3x+4y∣<4|x|+|3x+4y|<4. Its only lattice points are 00 and ±(1,−1)\pm(1,-1), all on one line. By Lemma 5.2 below, no chain with this end, or with any end L(m,1)L(m,1), and n≥3n \ge3 has all its interior rod vectors in the open diamond.

The chain (1,−1,1,−1)(1,-1,1,-1) of Table 1 has the end L(5,2)L(5,2) and the degree r=2r=2, so that 1≤r≤n−21 \le r \le n-2. It is invariant under the change κ↦−(κn−1,…,κ1)\kappa\mapsto-(\kappa_{n-1},\ldots,\kappa_1) that reverses the orientation of MM (§ 2.2). Here det⁡(v,v5)=−(3x+5y)\det(v,v_5)=-(3x+5y), so the open diamond is the set where ∣x∣+∣3x+5y∣<5|x|+|3x+5y|<5. Its non-zero lattice points are ±(1,0)\pm(1,0), ±(1,−1)\pm(1,-1), ±(2,−1)\pm(2,-1), and ±(3,−2)\pm(3,-2), on four different lines. The interior rod vectors v1,…,v4v_1,\ldots,v_4 give the sums ∣det⁡(v0,vj)∣+∣det⁡(vj,v5)∣=4,3,3,4<5|\det(v_0,v_j)|+|\det(v_j,v_5)|=4,3,3,4<5, so they lie in the open diamond, and the chain satisfies (4.14). So the diamond inequality does not exclude a toric ALE instanton with this chain, and the argument of § 5 says nothing about the end L(5,2)L(5,2). We show this chain because the chains of toric ALE instantons that are hyperkähler for neither orientation have entries in {−1,0,1}\{-1,0,1\} [25 Lemma A.20], [4 Thm. 2.3], and such a chain with ∣p∣=5|p|=5 has at least four entries. Shorter chains of the end L(5,2)L(5,2) also satisfy (4.14) with 1≤r≤n−21 \le r \le n-2, for instance the chain (2,−2)(2,-2) of Table 1.

Proof of the main theorem

The proof combines the diamond inequality with an elementary fact about lattice points. If q≡±1(mod∣p∣)q \equiv\pm1 \pmod{|p|}, which by Lemma 2.6 means that the group at infinity is conjugate in O(4)O(4) to a subgroup of SU(2)SU(2), then all lattice points of the open diamond of the end lie on one line. Two consecutive rod vectors form a lattice basis, so they cannot both lie on this line. By Proposition 4.7, the degree of a toric ALE instanton with such an end is therefore 00 or n−1n-1. In these two degrees the metric is a multi-Eguchi–Hanson metric (Proposition 5.3). For r=0r=0 this can also be read off from the closed formula (4.8) of Theorem 4.3, which is then the Gibbons–Hawking form of the metric [3 Prop. 11.3].

The lattice lemma

We first determine the lattice points of the open diamond of an end with q≡±1(mod∣p∣)q \equiv\pm1 \pmod{|p|}, and then show that such a diamond cannot contain two consecutive rod vectors.

Lemma 5.1. Let κ\kappa be a chain with p≠0p \ne0, and suppose that q≡σ(mod∣p∣)q \equiv\sigma\pmod{|p|} with σ∈{1,−1}\sigma\in\{1,-1\}. Then vn−σv0=puv_n-\sigma v_0=p\mathbf{u} for a primitive vector u∈Z2\mathbf{u}\in\mathbb{Z}^2, and the lattice points of the open diamond D(v0,vn)D(v_0,v_n) are the multiples kuk\mathbf{u} with 2∣k∣<∣p∣2|k|<|p|. In particular all of them lie on the line Ru\mathbb{R}\mathbf{u}.

Proof. Since v0=(0,1)v_0=(0,1) and vn=(p,q)v_n=(p,q), the vector vn−σv0=(p,q−σ)v_n-\sigma v_0=(p,q-\sigma) is pp times u=(1,(q−σ)/p)∈Z2\mathbf{u}=(1,(q-\sigma)/p)\in\mathbb{Z}^2, which is primitive. Let v=αv0+βvnv=\alpha v_0+\beta v_n be a lattice point of the open diamond, so that ∣α∣+∣β∣<1|\alpha|+|\beta|<1. By (21) the numbers l=pαl=p\alpha and k=pβk=p\beta are integers, and ∣l∣+∣k∣<∣p∣|l|+|k|<|p|. The second coordinate of v=(lv0+kvn)/pv=(lv_0+kv_n)/p is (l+kq)/p(l+kq)/p, which is an integer, so l≡−kq≡−kσ(mod∣p∣)l\equiv-kq\equiv-k\sigma\pmod{|p|}. Since ∣l+kσ∣≤∣l∣+∣k∣<∣p∣|l+k\sigma|\leq|l|+|k|<|p|, this forces l=−kσl=-k\sigma. Hence v=k(vn−σv0)/p=kuv=k(v_n-\sigma v_0)/p=k\mathbf{u}, with 2∣k∣=∣l∣+∣k∣<∣p∣2|k|=|l|+|k|<|p|. Conversely every such kuk\mathbf{u} lies in the open diamond. □

For the chains (2,…,2)(2,\ldots,2) the vector u=(1,−1)\mathbf{u}=(1,-1) satisfies det⁡(u,vj)=1\det(\mathbf{u},v_j)=1 for all their rod vectors (Table 1), so its slope −1-1 is adjacent in the Farey graph to the slopes of all rod vectors (Remark 3.2 and Figure 2(a)). In general, the condition q≡σ(mod∣p∣)q\equiv\sigma\pmod{|p|} says that the slope of u\mathbf{u} is adjacent to the slopes of v0v_0 and vnv_n, since det⁡(v0,u)=−1\det(v_0,\mathbf{u})=-1 and det⁡(u,vn)=σ\det(\mathbf{u},v_n)=\sigma. Lemma 5.1 says that in the open diamond only the line Ru\mathbb{R}\mathbf{u} is available, which by the opening of this section is too little for two consecutive rod vectors.

Lemma 5.2 (lattice lemma). Let κ\kappa be a chain with n≥3n\geq3 and p≠0p\ne0 whose interior rod vectors v1,…,vn−1v_1,\ldots,v_{n-1} lie in the open diamond D(v0,vn)D(v_0,v_n). Then q≢±1(mod∣p∣)q\not\equiv\pm1\pmod{|p|}. In particular ∣p∣≥5|p|\geq5 and ∣p∣≠6|p|\ne6.

Proof. Suppose that q≡±1(mod∣p∣)q\equiv\pm1\pmod{|p|}. By Lemma 5.1, v1v_1 and v2v_2 are multiples of the same vector u\mathbf{u}, so det⁡(v1,v2)=0\det(v_1,v_2)=0, which contradicts (3). The last statement holds since every unit modulo m∈{1,2,3,4,6}m\in\{1,2,3,4,6\} is congruent to ±1\pm1 (§ 2.3). □

The degrees 0 and n−1n-1

In the two extreme degrees the metric is hyperkähler for one of the orientations, by a theorem of Li and Sun, and the toric hyperkähler ALE instantons are known. The following proposition is [4 Cor. 3.2(c), Rem. 3.6].

Proposition 5.3. Let (M,g)(M,g) be a toric ALE instanton with nn turning points and the degree r=0r=0. Then its chain is (2,…,2)(2,\ldots,2), p=np=n, and (M,g)(M,g) is a multi-Eguchi–Hanson metric with nn collinear centres that is hyperkähler for the orientation opposite to that of MM. If instead r=n−1r=n-1, then the chain is (−2,…,−2)(-2,\ldots,-2), p=(−1)n+1np=(-1)^{n+1}n, and (M,g)(M,g) is a multi-Eguchi–Hanson metric with nn collinear centres that is hyperkähler for the orientation of MM.

Proof. Let r=n−1r=n-1. By Remark 3.3, the degree of Li and Sun of (M,g)(M,g) is rˉ\bar r, and rˉ=n−1−r=0\bar r=n-1-r=0 by Lemma 3.1. By the theorem of Li and Sun [24 Thm. 7.4], in the form of [4 Thm. 2.2], gg is therefore hyperkähler for the orientation of MM, and is a Gibbons–Hawking metric with collinear centres. The centres are the nn turning points [3 §11.2], and since the end is ALE, gg is a multi-Eguchi–Hanson metric [4 Rem. 3.6]. That remark is stated for n≥2n\geq2, and for n=1n=1 a Gibbons–Hawking metric with one centre and an ALE end is flat, the multi-Eguchi–Hanson metric with one centre. The chain is (−2,…,−2)(-2,\ldots,-2) by [4 Cor. 3.2(c)], since the degree index of [4] is s=r=n−1s=r=n-1 (Lemma 4.4), and p=(−1)n+1np=(-1)^{n+1}n by Table 1. If r=0r=0, then MM with the opposite orientation has degree n−1n-1, hence reversing the orientation exchanges rr and rˉ\bar r (§ 2.2), and the case just treated applies to it. So gg is a multi-Eguchi–Hanson metric that is hyperkähler for the orientation opposite to that of MM, the chain of MM is −(−2,…,−2)=(2,…,2)-(-2,\ldots,-2)=(2,\ldots,2), and p=np=n by Table 1. □`

The main theorem

We now combine the diamond inequality, the lattice lemma, and the description of the extreme degrees.

Theorem 5.4. Let (M,g)(M,g) be a toric ALE instanton whose group at infinity Γ\Gamma is conjugate in O(4)O(4) to a subgroup of SU(2)SU(2). Then the action has exactly ∣Γ∣|\Gamma| fixed points. For one of the orientations of MM, (M,g)(M,g) is a multi-Eguchi–Hanson metric with ∣Γ∣|\Gamma| collinear centres that is hyperkähler for this orientation, and its chain is (−2,…,−2)(-2,\ldots,-2). For ∣Γ∣=1|\Gamma|=1 it is the flat R4\mathbb{R}^{4}, and for ∣Γ∣=2|\Gamma|=2 it is the Eguchi–Hanson metric.

Proof. By Lemma 2.3, MM is simply connected, the stabilisers are connected, and the action has n≥1n\geq1 fixed points. Fix orientations of MM and t\mathfrak{t}, and let κ\kappa be the chain, (p,q)=vn(p,q)=v_n, and rr the degree. By Lemma 2.6, ∣p∣=∣Γ∣|p|=|\Gamma|, and the hypothesis says that q≡±1(mod∣p∣)q\equiv\pm1\pmod{|p|} (§ 2.3). Suppose that 1≤r≤n−21\leq r\leq n-2. Then n≥3n\geq3, and by Proposition 4.7 the interior rod vectors lie in the open diamond D(v0,vn)D(v_0,v_n). By Lemma 5.2, q≢±1(mod∣p∣)q\not\equiv\pm1\pmod{|p|}, which contradicts the hypothesis. Hence r=0r=0 or r=n−1r=n-1, and Proposition 5.3 shows that (M,g)(M,g) is a multi-Eguchi–Hanson metric with nn collinear centres, with the chain (−2,…,−2)(-2,\ldots,-2) in its hyperkähler orientation. In both cases n=∣p∣=∣Γ∣n=|p|=|\Gamma|. A multi-Eguchi–Hanson metric with one centre is flat, and with two centres it is the Eguchi–Hanson metric. □

The hypothesis on Γ\Gamma enters only through the lattice lemma.

Corollary 5.5. Let (M,g)(M,g) be a toric ALE instanton whose group at infinity has order ∣Γ∣∈{1,2,3,4,6}|\Gamma|\in\{1,2,3,4,6\}. Then (M,g)(M,g) is a multi-Eguchi–Hanson metric with ∣Γ∣|\Gamma| collinear centres.

Proof. By Lemma 2.6, ∣Γ∣=∣p∣|\Gamma|=|p|, and qq is a unit modulo ∣p∣|p|. For ∣p∣∈{1,2,3,4,6}|p|\in\{1,2,3,4,6\} every unit modulo ∣p∣|p| is congruent to ±1\pm1 (§ 2.3), so q≡±1q\equiv\pm1, and Theorem 5.4 applies. □

Our argument does not cover ∣Γ∣=5|\Gamma|=5. In that case, the end L(5,2)L(5,2) is possible a priori, and for this end the diamond inequality alone does not suffice (Example 4.8). Computer calculations by the authors, which use a reduction to finitely many chains and interval arithmetic and are not part of this paper, indicate that the conclusion of Corollary 5.5 holds for ∣Γ∣=5|\Gamma|=5 as well.

Computation and the use of an AI assistant

The project was carried out with the help of an AI assistant, Claude (Anthropic), used through the Claude Code environment. The assistant took part in developing the arguments and in writing the text. We also used it to review the literature and to read the drafts critically. Computer calculations were used only for the remark on the order ∣Γ∣=5|\Gamma|=5 at the end of § 5. The proofs in this paper do not use them. The authors take responsibility for the content.

(Lars Andersson) Department of Mathematics, KTH Royal Institute of Technology, 100 44 Stockholm, Sweden Email address: [email protected]

(Mattias Dahl) Department of Mathematics, KTH Royal Institute of Technology, 100 44 Stockholm, Sweden Email address: [email protected]

(Gustav Nilsson) Mathematisches Institut, Universität Münster, Einsteinstrasse 62, 48149 Münster, Germany Email address: [email protected]

References

  1. [1]Steffen Aksteiner, Pure spinor factorization of Belinski–Zakharov solitons and regular gravitational instantons, 2026, arXiv:2611.03750v1.arxiv.org/abs/2611.03750
  2. [2]Michael T. Anderson, On the topology of complete manifolds of nonnegative Ricci curvature, Topology 29 (1990), no. 1, 41–55.DOI
  3. [3]Lars Andersson, Mattias Dahl, and Gustav Nilsson, On the Geroch conjecture for toric gravitational instantons, 2026, arXiv:2610.41875v1.arxiv.org/abs/2610.41875
  4. [4]———, Toric gravitational instantons with four fixed points, 2026, arXiv:2611.05333v1.arxiv.org/abs/2611.05333
  5. [5]Bernardo Araneda and James Lucietti, All toric Hermitian ALE gravitational instantons, J. Geom. Anal. 36 (2026), no. 11, Paper No. 345, arXiv:2510.09291.DOI
  6. [6]Shigetoshi Bando, Atsushi Kasue, and Hiraku Nakajima, On a construction of coordinates at infinity on manifolds with fast curvature decay and maximal volume growth, Invent. Math. 97 (1989), no. 2, 313–349.DOI
  7. [7]E. J. Brody, The topological classification of the lens spaces, Ann. of Math. (2) 71 (1960), 163–184.DOI
  8. [8]Mattias Dahl, The positive mass theorem for ALE manifolds, Mathematics of Gravitation, Part I (Warsaw, 1996) (Piotr T. Chruściel, ed.), vol. 41, Banach Center Publ., no. 1, Polish Acad. Sci. Inst. Math., Warsaw, 1997, pp. 133–142.DOI
  9. [9]Tohru Eguchi and Andrew J. Hanson, Asymptotically flat self-dual solutions to Euclidean gravity, Phys. Lett. B 74 (1978), no. 3, 249–251.
  10. [11]F. R. Gantmacher, The theory of matrices. Vol. 1, Chelsea Publishing Co., New York, 1959.
  11. [12]Partha Ghosh, Mass and rigidity in almost Kähler geometry, 2026, arXiv:2603.08627v2.arxiv.org/abs/2603.08627
  12. [13]G. W. Gibbons, Gravitational instantons: a survey, Mathematical Problems in Theoretical Physics (Proc. Internat. Conf. Math. Phys., Lausanne, 1979) (K. Osterwalder, ed.), Lecture Notes in Phys., vol. 116, Springer, Berlin, 1980, pp. 282–287.DOI
  13. [14]G. W. Gibbons and S. W. Hawking, Gravitational multi-instantons, Phys. Lett. B 78 (1978), no. 4, 430–432.
  14. [15]Karsten Grove and Hermann Karcher, How to conjugate C¹-close group actions, Math. Z. 132 (1973), 11–20.
  15. [16]Allen Hatcher, Topology of numbers, American Mathematical Society, Providence, RI, 2022.
  16. [17]Eduard Heine, Handbuch der Kugelfunktionen. Theorie und Anwendungen. Erster Band, second ed., G. Reimer, Berlin, 1878.
  17. [19]Shoshichi Kobayashi, Transformation groups in differential geometry, Ergebnisse der Mathematik und ihrer Grenzgebiete, vol. 70, Springer-Verlag, New York–Heidelberg, 1972.
  18. [20]P. B. Kronheimer, The construction of ALE spaces as hyper-Kähler quotients, J. Differential Geom. 29 (1989), no. 3, 665–683.
  19. [21]⎯⎯⎯⎯, A Torelli-type theorem for gravitational instantons, J. Differential Geom. 29 (1989), no. 3, 685–697.DOI
  20. [22]Hari K. Kunduri and James Lucietti, On the existence of toric ALE and ALF gravitational instantons, Ann. Henri Poincaré, online first, 2026, arXiv:2604.15159.DOI
  21. [23]Mingyang Li, On 4-dimensional Ricci-flat ALE manifolds, 2024, arXiv:2304.01609v4.arxiv.org/abs/2304.01609
  22. [24]Mingyang Li and Song Sun, Gravitational instantons and harmonic maps, 2025, arXiv:2507.15284v1.arxiv.org/abs/2507.15284
  23. [25]⎯⎯⎯⎯, Compact toric Einstein four-manifolds, 2026, arXiv:2610.10425v2.arxiv.org/abs/2610.10425
  24. [26]⎯⎯⎯⎯, Geometry of gravitational instantons, 2026, arXiv:2606.10443v1.arxiv.org/abs/2606.10443
  25. [27]Peter Li, Large time behavior of the heat equation on complete manifolds with nonnegative Ricci curvature, Ann. of Math. (2) 124 (1986), no. 1, 1–21.DOI
  26. [28]Michael T. Lock and Jeff A. Viaclovsky, Quotient singularities, eta invariants, and self-dual metrics, Geom. Topol. 20 (2016), no. 3, 1773–1806, arXiv:1501.03234.
  27. [29]Hiraku Nakajima, Self-duality of ALE Ricci-flat 4-manifolds and positive mass theorem, Recent Topics in Differential and Analytic Geometry, Adv. Stud. Pure Math., vol. 18-I, Kinokuniya, Tokyo; Academic Press, Boston, MA, 1990, pp. 385–396.
  28. [30]Gustav Nilsson, Topology of toric gravitational instantons, Differential Geom. Appl. 96 (2024), Paper No. 102171, arXiv:2301.10212.DOI
  29. [31]Peter Orlik and Frank Raymond, Actions of the torus on 4-manifolds. I, Trans. Amer. Math. Soc. 152 (1970), no. 2, 531–559.DOI
  30. [32]Ioana Șuvaina, ALE Ricci-flat Kähler metrics and deformations of quotient surface singularities, Ann. Global Anal. Geom. 41 (2012), no. 1, 109–123, arXiv:1102.2672.arxiv.org/abs/1102.2672
  31. [33]Gábor Szegő, Orthogonal polynomials, 4th ed., Amer. Math. Soc. Colloquium Publications, vol. 23, American Mathematical Society, Providence, RI, 1975.
  32. [34]Evan P. Wright, Quotients of gravitational instantons, Ann. Global Anal. Geom. 41 (2012), no. 1, 91–108, arXiv:1102.2442.arxiv.org/abs/1102.2442

Paper details

vlnc:W_ajdk8s2r
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents