1: \documentclass[12pt]{article}

2: \usepackage{amsmath}

3: \usepackage{amssymb}

4: \usepackage{amscd}

5: % \usepackage{euler}

6: \usepackage{latexsym}

7: \usepackage{theorem}

8: \usepackage{doublespace}

9: \usepackage{epsfig}

10: \usepackage{psfrag}

11: \usepackage{amsfonts}

12: \usepackage{enumerate}

13: 

14: \newtheorem{theorem}{Theorem}[section]

15: \newtheorem{lemma}[theorem]{Lemma}

16: \newtheorem{proposition}[theorem]{Proposition}

17: \newtheorem{corollary}[theorem]{Corollary}

18: \newtheorem{claim}[theorem]{Claim}

19: \newtheorem{conjecture}[theorem]{Conjecture}

20: 

21: {\theorembodyfont{\rmfamily}

22: \theoremstyle{plain}

23: \newtheorem{definition}[theorem]{Definition}

24: \newtheorem{example}[theorem]{Example}

25: \newtheorem{problem}[theorem]{Problem}

26: \newtheorem{remark}[theorem]{Remark}

27: }

28: 

29: % Margin stuff from jason

30: \oddsidemargin=0pt

31: \evensidemargin=0pt

32: \topmargin=0in

33: \headheight=0pt

34: \headsep=0pt

35: \setlength{\textheight}{9in}

36: \setlength{\textwidth}{6.5in}

37: 

38: % \renewcommand{\familydefault}{ppl}

39: 

40: \newcommand{\bd}{\bf \boldmath}

41: \newcommand{\HH}{{\mathbb H}}

42: \newcommand{\T}{{\mathbb T}}

43: \newcommand{\gl}{{\mathfrak gl}}

44: \newcommand{\symp}{{\mathfrak sp}}

45: \newcommand{\liet}{{\mathfrak t}}

46: \newcommand{\lieu}{{\mathfrak u}}

47: \newcommand{\V}{{\mathbb V}}

48: \newcommand{\F}{{\mathbb F}}

49: \newcommand{\e}{\varepsilon}

50: \newcommand{\Id}{\bf 1}

51: \newcommand{\PP}{{\mathbb P}}

52: 

53: \renewcommand{\k}{\kappa}

54: \newcommand{\ktd}{\kappa_{\Td}}

55: \newcommand{\ktdso}{\kappa_{\Td\times S^1}}

56: \newcommand{\kso}{\kappa_{S^1}}

57: 

58: \newcommand{\<}{\left<}

59: \renewcommand{\>}{\right>}

60: 

61: \renewcommand{\Re}{\operatorname{Re}}

62: \renewcommand{\Im}{\operatorname{Im}}

63: \newcommand{\im}{\operatorname{im}}

64: \newcommand{\id}{\operatorname{id}}

65: \newcommand{\Coker}{\operatorname{Coker}}

66: \newcommand{\Ker}{\operatorname{Ker}}

67: \renewcommand{\dim}{\operatorname{dim}}

68: 

69: \newcommand{\Ot}{\displaystyle\bigotimes}

70: \newcommand{\Od}{\displaystyle\bigoplus}

71: \newcommand{\Pd}{\displaystyle\prod}

72: \newcommand{\Sd}{\displaystyle\sum}

73: \newcommand{\Cd}{\displaystyle\bigcap}

74: 

75: \renewcommand{\H}{{\mathbb{H}}}

76: \newcommand{\C}{{\mathbb{C}}}

77: \newcommand{\Zt}{{\mathbb{Z}_2}}

78: \newcommand{\Z}{{\mathbb{Z}}}

79: \newcommand{\Q}{{\mathbb{Q}}}

80: \newcommand{\R}{{\mathbb{R}}}

81: 

82: \newcommand{\la}{\lambda}

83: \renewcommand{\a}{\alpha}

84: \renewcommand{\b}{\beta}

85: \renewcommand{\t}{\theta}

86: \newcommand{\s}{\sigma}

87: 

88: \newcommand{\otd}{\{1,\ldots,d\}}

89: \newcommand{\Hn}{{\H^n}}

90: \newcommand{\Cn}{{\C^n}}

91: \newcommand{\Cns}{(\C^n)^{\text{ss}}}

92: \newcommand{\Cno}{\C^{n+1}}

93: \renewcommand{\cot}{T^*\Cn}

94: \newcommand{\Tk}{T^k}

95: \newcommand{\Tn}{T^n}

96: \newcommand{\Tdrs}{T_{\R}^d\times\Zt}

97: \newcommand{\Tdr}{T_{\R}^d}

98: \newcommand{\Td}{T^d}

99: \newcommand{\Tkd}{(\tk_{\Z})^*}

100: \newcommand{\Tnd}{(\tn_{\Z})^*}

101: \newcommand{\Tdd}{(\td_{\Z})^*}

102: \newcommand{\tk}{\mathfrak{t}^k}

103: \newcommand{\tn}{\mathfrak{t}^n}

104: \newcommand{\td}{\mathfrak{t}^d}

105: \newcommand{\tkd}{(\tk)^*}

106: \newcommand{\tnd}{(\tn)^*}

107: \newcommand{\tdd}{(\td)^*}

108: \newcommand{\Tdso}{\Td\times S^1}

109: \newcommand{\Tnso}{\Tn\times S^1}

110: 

111: \newcommand{\subs}{\subseteq}

112: \newcommand{\sups}{\supseteq}

113: \newcommand{\hookto}{{\hookrightarrow}}

114: \newcommand{\onto}{{\twoheadrightarrow}}

115: \renewcommand{\iff}{\Leftrightarrow}

116: \newcommand{\impl}{\Rightarrow}

117: 

118: \newcommand{\hm}{H^*(M)}

119: \newcommand{\htm}{H^*_{\Td}(M)}

120: \newcommand{\htsm}{H^*_{\Td\times S^1}(M)}

121: \newcommand{\hsm}{H^*_{S^1}(M)}

122: \newcommand{\hr}{H^*_{\Tdr}(M_{\R};\Zt)}

123: \newcommand{\hrs}{H^*_{\Tdr\times\Zt}(M_{\R};\Zt)}

124: \newcommand{\hscomp}{H^*_{\Zt}(\comp;\Zt)}

125: \newcommand{\hcomp}{H^*(\comp;\Zt)}

126: 

127: \newcommand{\becircled}{\mathaccent "7017}

128: \newcommand{\hs}{\hspace{3pt}}

129: \newcommand{\eps}{\varepsilon}

130: \renewcommand{\S}{\sum}

131: \renewcommand{\ss}{\substack}

132: \renewcommand{\bd}{\partial}

133: \newcommand{\gd}{\g^*}

134: \newcommand{\g}{\mathfrak{g}}

135: \newcommand{\half}{\frac{1}{2}}

136: \renewcommand{\mod}{{/\!\!/}}

137: \newcommand{\mmod}{{/\!\!/\!\!/\!\!/}}

138: \newcommand{\mc}{{\mathcal{C}}}

139: \newcommand{\arr}{{\mathcal{H}}}

140: \newcommand{\comp}{\mathcal{M}(\arr)}

141: \newcommand{\bomp}{\Cn\setminus\cup_{i=1}^n H_i^{\C}}

142: \newcommand{\os}{\mathcal{OS}}

143: \newcommand{\ost}{\mathcal{OS}\otimes\Zt}

144: \newcommand{\mhk}{\mu_{\text{HK}}}

145: \renewcommand{\mc}{\mu_{\C}}

146: \newcommand{\mr}{\mu_{\R}}

147: 

148: 

149: \newcommand{\qed}{\hfill \mbox{$\Box$}\medskip\newline}

150: % \newcommand{\qed}{\hfill $\Box$\newline}

151: \newenvironment{proofc}{\noindent {\bf Proof of \ref{coorientation}:}}{\qed\par}

152: \newenvironment{proofp}{\text{ }\newline\noindent 

153: {\bf Proof of \ref{placement}:}}{\qed \par}

154: \newenvironment{proofhtsm}{\text{ }\newline\noindent 

155: {\bf Proof of \ref{htsm}:}}{\qed \par}

156: \newenvironment{proofhsm}{\text{ }\newline\noindent 

157: {\bf Proof of \ref{hsm}:}}{\qed \par}

158: \newenvironment{proof}{\noindent {\bf Proof:}}{\qed \par}

159: \newenvironment{noproof}{\noindent {\bf Proof:}}{\newline}

160: \newenvironment{sketch}{\noindent {\bf Sketch of Proof:}}{\qed \par}

161: 

162: \begin{document}

163: \begin{spacing}{1.1}

164: 

165: \noindent

166: {\LARGE \bf Properties of 

167: the residual circle action \smallskip \\

168: on a hypertoric variety}\bigskip\\

169: {\bf Megumi Harada} \\

170: Department of Mathematics, University of California,

171: Berkeley, CA 94720\smallskip \\

172: {\bf Nicholas Proudfoot } \\

173: Department of Mathematics, University of California,

174: Berkeley, CA 94720

175: \bigskip

176: {\small

177: \begin{quote}

178: \noindent {\em Abstract.}

179: We 

180: consider an orbifold $X$ obtained by a 

181: K\"ahler reduction of $\Cn$,

182: and we define its ``hyperk\"ahler analogue'' $M$

183: as a hyperk\"ahler reduction of $\cot\cong\Hn$ by the same group.

184: In the case where the group is abelian and $X$ is a toric variety,

185: $M$ is a toric hyperk\"ahler orbifold, as defined in

186: \cite{BD}, and further studied in \cite{K1,K2} and \cite{HS}.  

187: The variety $M$ carries a natural action of $S^1$, induced

188: by the scalar action of $S^1$ on the fibers of $\cot$.

189: In this paper we study this action, computing its fixed points

190: and its equivariant cohomology.

191: As an application, we use the associated $\Zt$ action

192: on the real locus of $M$ to compute a deformation of the

193: Orlik-Solomon algebra of a smooth, real hyperplane arrangement $\arr$,

194: depending nontrivially on the affine structure of the arrangement.

195: This deformation is given

196: by the $\Zt$-equivariant cohomology of the complement of the complexification

197: of $\arr$, where $\Zt$ acts by complex conjugation.

198: \end{quote}

199: }

200: \bigskip

201: 

202: In order to construct a toric variety as a K\"ahler quotient of $\Cn$ by a 

203: torus,

204: one begins with the combinatorial data of an arrangement $\arr$ of $n$ cooriented,

205: rational, affine hyperplanes in $\R^d$.

206: The normal vectors to these hyperplanes determine

207: a subtorus $\Tk\subs\Tn$ ($k=n-d$), and the affine structure determines a 

208: value $\a\in\tkd$ at which to reduce, so that we may define $X=\Cn\mod_{\!\!\a}\Tk$.

209: Using the same combinatorial data, one can also construct a {\em hypertoric 

210: variety},\footnote{In 

211: \cite{BD,K1,K2,HS} $M$ is called a ``toric hyperk\"ahler'' variety,

212: but as it is a complex variety that is not toric in

213: the standard sense, we prefer the term ``hypertoric.''}

214: which is defined as the hyperk\"ahler quotient $M = \Hn\mmod_{\!\!(\a,0)}\Tk$ of 

215: $\Hn\cong\cot$ by the

216: induced action of the same subtorus $\Tk\subs\Tn$ \cite{BD}.  It is well known 

217: that the toric variety $X$ does not retain all of the information of $\arr$; indeed, it 

218: depends only on the polyhedron

219: $\Delta$ obtained by intersecting the half-spaces associated to each of the 

220: cooriented hyperplanes. Thus it is always possible to add an extra hyperplane to $\arr$ 

221: without changing $X$.  In contrast, the hypertoric variety $M$ remembers the 

222: number of hyperplanes in $\arr$, but its equivariant diffeomorphism type 

223: depends neither on the coorientations nor on the affine structure of $\arr$ (see Theorem \ref{htm}, 

224: and Lemmas \ref{placement} and \ref{coorientation}).

225: 

226: The purpose of this paper is to study the hamiltonian $S^1$ action on $M$ descending

227: from the scalar action of $S^1$ on the fibers of $\cot$.  This action is sensitive to

228: both the coorientations and the affine structure of $\arr$,

229: even on the level of equivariant cohomology (Section \ref{Konno}).

230: One can recover the toric variety $X$ as the minimum of the $S^1$ moment map,

231: hence the geometric structure of $M$ along with its circle action carries strictly more

232: information than either $X$ or $M$ alone.

233: In Section \ref{coreflow} we give an explicit description of this action

234: when restricted to the core $C$, a deformation retract of

235: $M$ which is a union of projective subvarieties.

236: In Section \ref{Konno} we compute the $S^1$ and $\Tdso$-equivariant cohomologies

237: of $M$, using the full combinatorial data of $\arr$. 

238: 

239: In Section \ref{defos}, we 

240: examine the {\em real locus} $M_{\R}\subs M$, i.e. the fixed point set 

241: of an involution

242: of $M$ that is anti-holomorphic with respect to the first complex structure.

243: By studying the topology of $M_{\R}$, we interpret the results of Section \ref{Konno}

244: in terms of the Orlik-Solomon algebra 

245: $\mathcal{OS}=H^*(\comp)$,

246: where $\comp$

247: is the complement of the complexification of $\arr$.

248: % In particular, we show that $\mathcal{OS}\otimes\Zt$ is isomorphic to a quotient

249: % of $H^*_{\Td}(M;\Zt)$, and that the equivariant cohomology ring $H^*_{\Zt}(\comp;\Zt)$

250: % is isomorphic to the analogous quotient of $H^*_{\Tdso}(M;\Zt)$,

251: % where $\Zt$ acts on $\comp$ by complex conjugation.

252: We show how to interpret Theorem \ref{htsm} as a computation

253: of $H^*_{\Zt}(\comp;\Zt)$,

254: a deformation of the Orlik-Solomon algebra of a smooth, real arrangement

255: that depends nontrivially on the affine structure.\footnote{A more

256: general computation of $H^*_{\Zt}(\comp;\Zt)$, in which $\arr$

257: is not assumed to be simple, rational, or smooth, will appear in \cite{Pr}.}

258: \newline

259: 

260: \noindent {\em Acknowledgments.}

261: We are very grateful to Tam\'as Hausel for introducing us to this problem

262: and answering many questions.

263: We would also like to thank Allen Knutson for guiding our research.

264: 

265: \begin{section}{Hyperk\"ahler reductions}\label{reduction}

266: A hyperk\"ahler manifold is a smooth manifold, 

267: necessarily of real dimension $4n$,

268: which admits

269: three complex structures $J_1,J_2,J_3$

270: %$i,j,k$

271: satisfying the usual quaternionic relations, in a manner compatible

272: with a metric. 

273: Just as in the K\"ahler case, we can define three different

274: symplectic forms on $N$ as follows:

275: % by using the compatibility of the

276: % metric and the complex structures, namely

277: \[

278: \omega_1(v,w) = g(J_{1}v, w), \hs\omega_2(v,w) = g(J_{2}v,w), \hs\omega_3(v,w) =

279: g(J_{3}v,w).

280: \]

281: % As happens in K\"ahler geometry, the integrability of the complex

282: % structures $J_i$ implies the closedness of the two-forms $\omega_i$.

283: Note that the complex-valued two-form \(\omega_2 + i

284: \omega_3\) is non-degenerate and covariant constant, hence closed and

285: {\em holomorphic} with respect to the complex structure $J_1$. 

286: Any hyperk\"aher manifold can therefore be considered as a holomorphic

287: symplectic manifold with complex structure $J_1$, real symplectic

288: form $\omega_{\R} := \omega_1$, and holomorphic symplectic form

289: $\omega_{\C} := \omega_2 + i\omega_3$.

290: This is the point of view that we will adopt in this paper.

291: 

292: % Just as K\"ahler 

293: % reduction is a powerful tool for producing 

294: % examples of K\"ahler manifolds, hyperk\"ahler reduction produces

295: % a plethora of 

296: % many

297: % interesting hyperk\"ahler manifolds. The method

298: % was first introduced 

299: % in a classic paper of 

300: % by Hitchin, Karlhede,

301: % Lindstrom, and Ro\v{c}ek \cite{HKLR},

302: % and is a natural ``hyperk\"ahler'' version of K\"ahler reduction.

303: We will refer to an action of $G$ on a hyperk\"ahler manifold

304: $N$ as {\em hyperhamiltonian}

305: if it is hamiltonian with respect to $\omega_{\R}$ and holomorphic

306: hamiltonian with respect to $\omega_{\C}$, with $G$-equivariant

307: moment map

308: $$\mhk:=\mr\oplus\mc:N\to\gd\oplus\g_{\C}^*.$$ 

309: % where the action of $G$ on $\gd\oplus\g_{\C}^*$ is given by 

310: % the diagonal embedding \(G \to G\times G.\) 

311: 

312: \begin{theorem}{\em\cite{HKLR}}

313: Let \((N^{4n},g)\) be a hyperk\"ahler manifold with real symplectic form

314: $\omega_{\R}$ and holomorphic symplectic form $\omega_{\C}$.

315: Suppose that $N$ is equipped with a hyperhamiltonian action of a 

316: compact Lie group $G$, with moment map

317: $\mhk=\mr\oplus\mc$. 

318: Suppose $\xi = \xi_{\R}\oplus \xi_{\C}$ is a

319: central regular value of $\mhk$. Then there is a unique 

320: hyperk\"ahler structure

321: on the hyperk\"ahler quotient 

322: \(M = N \mmod_{\!\!\xi}G := \mhk^{-1}(\xi)/G\), 

323: with associated symplectic and holomorphic symplectic forms

324: $\omega^{\xi}_{\R}$ and $\omega^{\xi}_{\C}$, such that

325: $\omega^{\xi}_{\R}$ and $\omega^{\xi}_{\C}$ pull back to the restrictions

326: of $\omega_{\R}$ and $\omega_{\C}$ to $\mhk^{-1}(\xi)$.

327: \end{theorem}

328: 

329: % If we fix a complex structure $J_1$, then as mentioned above, the

330: % latter two symplectic forms $\omega_2, \omega_3$ fit together to make

331: % a holomorphic symplectic form $\omega_{\C}$. The latter two moment

332: % maps $\mu_2, \mu_3$ also fit together to give a holomorphic moment map

333: % $\mu_{\C}$. When using this perspective, We can also distinguish

334: % between the moment maps by calling $\mu_1$ the ``real'' moment map,

335: % $\mu_{\R}$. Similarly, we can split the target $\g^{*} \oplus \g^{*}

336: % \oplus \g^{*}$ into components \(\g^{*}_{\R} \oplus \g_{\C}^{*},\) and 

337: % we reduce at values \(\xi = (\xi_{\R}, \xi_{\C}).\) 

338: 

339: % Similarly, it

340: % follows from the Kirwan-Ness theorem that a hyperK\"ahler reduction is

341: % equivalent to a holomorphic symplectic reduction of $M$ by $G_{\C}$,

342: % where the complex group also acts since $G$ preserves the complex

343: % structure $J_1$. Namely, a hyperk\"ahler reduction is by definition

344: % \[

345: % M_{red} := (\mu_{\R}^{-1}(\xi_{\R}) \cap \mu_{\C}^{-1}(\xi_{\C}))/G,

346: % \]

347: % where $G$ is the compact group. 

348: 

349: %BEGIN NEW

350: % Since the hyperk\"ahler moment map $\mhk$ is $G$-equivariant, 

351: If $\xi\in \gd\oplus\g_{\C}^*$ is fixed by the coadjoint action of $G$, the inverse

352: image $\mc^{-1}(\xi_{\C})$ is preserved by $G$,

353: and is a (singular) K\"ahler subvariety with respect to $\omega_{\R}$.

354: Then by 

355: \cite{HL} 

356: (see also \cite[3.2]{Na} and \cite[2.5]{Sj}),

357: we have 

358: $$N\mmod_{\!\!\xi}G = 

359: \mc^{-1}(\xi_{\C})\mod_{\!\!\xi_{\R}}G

360: =\mu_{\C}^{-1}(\xi_{\C})^{\text{ss}}/G_{\C},$$

361: where 

362: $$\mc^{-1}(\xi_{\C})^{\text{ss}} = \{x\in\mc^{-1}(\xi_{\C})\mid Gx\cap\mu_{\R}^{-1}(\xi_{\R})

363: \neq\emptyset\}.$$

364: % is the dense open stratum

365: % of $\mc^{-1}(\xi_{\C})$ in the stratification defined by the gradient

366: % flow of the real-valued Morse-Bott-Kirwan function $|\mr-\xi_{\R}|^2:N\to\R$. 

367: % END NEW

368: 

369: We now specialize to the case where $G$ is a compact Lie group

370: acting linearly on $\Cn$ with moment map $\mu:\Cn \to \gd$, taking $0\in\C^n$

371: to $0\in\gd$.  This action 

372: induces an action of $G$ on the holomorphic cotangent bundle 

373: $\cot\cong\Cn\times(\Cn)^*$.

374: If we choose a bilinear inner product on $\Cn$, 

375: we can coordinatize this representation

376: as $\{(z,w)\mid z,w\in\Cn\}$ with $g (z,w)=(g z,g^{-1} w)$.

377: Choose an identification of $\Hn$ with $\cot$

378: such that the complex structure $J_1$ on $\Hn$ 

379: given by right multiplication by $i$

380: corresponds to the natural complex structure on $\cot$.

381: Then $\cot$ inherits a hyperk\"ahler, and 

382: therefore also a holomorphic symplectic, structure, 

383: with

384: $\omega_{\R}$ given by 

385: adding the standard symplectic structures on $\Cn$ and $(\Cn)^{*}

386: \cong \Cn$, and 

387: $\omega_{\C} = d\eta$, where $\eta$ is the canonical 

388: holomorphic 1-form on $\cot$.  

389: 

390: %Next we note that $\cot$ naturally inherits the structure of a

391: %holomorphic symplectic, and therefore hyperk\"ahler manifold,

392: %with $\omega_{\R}$ given by adding the symplectic forms on $\Cn$ and $(\Cn)^*$,

393: %and $\omega_{\C} = d\eta$, where $\eta$ is the complex valued canonical $1$-form

394: %on $\cot$.

395: % $\omega_{\R} = \frac{i}{2} dz\wedge d\bar{z} - \frac{i}{2} dw\wedge d\bar{w}$

396: %We thus identify $\cot$ with $\Hn$. 

397: 

398: Note that $G$ acts $\H$-linearly on $\cot\cong\Hn$

399: (where $n\times n$ matrices

400: act on the left on $\Hn$, and scalar multiplication by $\H$

401: is on the right), 

402: and does so hyperhamiltonianly

403: with moment map $\mhk=\mc\oplus\mr$,

404: where

405: $$\mr(z,w) = \mu(z)-\mu(w)\hspace{10pt}\text{and}

406: \hspace{10pt}\mc(z,w)(v) = w(\hat{v}_z)$$

407: for $w\in T^*_z\Cn,\hs v\in\g_{\C}$, and $\hat{v}_z$ the element

408: of $T_z\Cn$ induced by $v$.

409: Consider a central regular value $\a\in\gd$ for $\mu$,

410: and suppose that $(\a,0)\in\gd\oplus\gd_{\C}$ is a central regular value for $\mu_{\text{HK}}$.

411: We refer to the hyperk\"ahler reduction $M = \Hn \mmod_{\!\!(\a,0)} G$

412: as the {\em hyperk\"ahler analogue} of the corresponding K\"ahler reduction

413: $X=\Cn\mod_{\!\!\a} G$. 

414: The following proposition is proven for the case 

415: where $G$ is a torus in \cite[7.1]{BD}. 

416: 

417: \begin{proposition}\label{compactification}

418: %Let $M = \Hn \mmod_{\!\!(\a,0)} G$ 

419: %be the hyperk\"ahler reduction of $\Hn$ by $G$

420: %at a central regular value $\left(\a,0\right)\in\gd\oplus\gd_{\C}$.

421: The cotangent bundle $T^*X$ is isomorphic to an open subset of $M$.

422: \end{proposition}

423: 

424: \begin{proof}

425: Let $Y = \{(z,w)\in\mc^{-1}(0)^{\text{ss}}\mid

426: z \in (\Cn)^{\text{ss}}\}$,

427: where we ask $z$ to be semistable with respect to $\a$

428: for the action of

429: $G_{\C}$ on $\Cn$, so that $X\cong (\Cn)^{\text{ss}}/G_{\C}$.

430: Let $[z]$ denote the element of $X$ represented by $z$.

431: The tangent space $T_{[z]}X$ is equal to the quotient of $T_{z}\Cn$

432: by the tangent space to the $G_{\C}$ orbit through $z$,

433: hence

434: $$T^*_{[z]}X\cong\{w\in T^*_{[z]}\Cn\mid w(\hat{v}_z)=0\text{ for all }

435: v\in\g_{\C}\} = \{w\in(\Cn)^*\mid\mc(z,w)=0\}.$$

436: Then $$T^*X \cong \{(z,w)\mid z\in(\Cn)^{\text{ss}}\text{ and }

437: \mc(z,w)=0\}/G_{\C} = Y/G_{\C}$$

438: is an open subset of $M$.

439: \end{proof}

440: 

441: Consider the action

442: of $S^1$ on $\Hn\cong T^*\Cn$ given by 

443: ``rotating the fibers'' of the cotangent

444: bundle,  

445: given explicitly by $\tau(z,w) = (z,\tau w)$. 

446: This action is hamiltonian with respect to the real symplectic

447: structure $\omega_{\R}$ with moment map $\Phi(z,w) = \half |w|^2$.

448: Because it commutes with the action of $G$,

449: the action descends to a hamiltonian action on $M$, where we will

450: still denote the moment map by $\Phi$.

451: Since $S^1$ acts trivially on $z$, and by scalars on $w$,

452: it does {\em not} preserve the complex symplectic form

453: $\omega_{\C}(z,w) = dw\wedge dz$,

454: and does not act $\H$-linearly.

455: % Further, this $S^1$ does {\em not} 

456: % act on $\cot \cong \H^n$ $\H$-linearly, i.e. as elements of 

457: % $M(n,\H)$. 

458: 

459: \begin{proposition}\label{proper}

460: If the original moment map $\mu:\Cn\to\gd$ is proper,

461: then so is $\Phi:M\to\R$.

462: \end{proposition}

463: 

464: \begin{proof}

465: We would like to show that $\Phi^{-1}[0,R]$ is compact

466: for any $R$.  Since 

467: $$\Phi^{-1}[0,R] = \{(z,w)\mid \mr(z,w)=\a,\hs

468: \mc(z,w)=0,\hs\Phi(z,w)\leq R\}\big/G$$ and $G$ is compact,

469: it is sufficient to show that 

470: the set $\{(z,w)\mid \mr(z,w)=\a,\hs\Phi(z,w)\leq R\}$

471: is compact.  Since $\mr(z,w)=\mu(z)-\mu(w)$,

472: this set is a closed subset

473: of $$\mu^{-1}\left\{\a+\mu(w) \,\big|\, \half |w|^2\leq R\right\}

474: \times\left\{w \,\big|\,

475: \half |w|^2 \leq R\right\},$$

476: which is compact by the properness of $\mu$.

477: \end{proof}

478: 

479: % It is important to note that, while the set of non-regular values of $\mu$

480: % has codimension $1$, the set of non-regular values of $\mhk$

481: % has codimension $3$, hence the set of regular values is simply-connected.

482: % Connectivity tells us that the diffeomorphism type of $M$ is independent of $\a$,

483: % and simple-connectivity tells us that the cohomology rings of $M$ and $M'$

484: % are naturally isomorphic for any regular values $\a, \a'\in\gd$.

485: % It is not true, however, that the regular values $(\a,0)$ and $(\a',0)$ of $\mhk$

486: % can be connected by a path of regular values of the form $(\b,0)$.

487: % This means that while the cohomology of $M$ is independent of $\a$,

488: % the $S^1$ equivariant cohomology is not.  In Section \ref{s1}, 

489: % we will compute this equivariant cohomology ring in the case where $G$

490: % is abelian, and $X = \Cn\mod G$ is a toric variety.  

491: 

492: In the case where $G$ is abelian and $X$ is a nonempty toric 

493: variety, properness of $\mu$ (and therefore of $\Phi$) is equivalent 

494: to compactness of $X$. 

495: \end{section}

496: 

497: \begin{section}{Hypertoric varieties}\label{ht}

498: In this section we restrict our attention to 

499: {\em hypertoric}

500: varieties,

501: which are the hyperk\"ahler analogues

502: of toric varieties in the sense of

503: Section \ref{reduction}.  

504: % We

505: % will follow the exposition of \cite{BD,K1,K2}, but some of our notation

506: % will differ.

507: We begin with the full $n$-dimensional torus $\Tn$ acting on $\Cn$,

508: and the induced action on $\Hn\cong\cot$

509: given by $t(z,w) = (tz, t^{-1} w)$.

510: Let \(\{a_i\}_{1 \leq i \leq n}\) be nonzero primitive integer vectors in

511: \(\liet^d \cong \R^d\) defining a map \(\beta: \liet^n \longrightarrow

512: \liet^d\) by \(\e_i \mapsto a_i,\) where

513: $\{\e_i\}$ is the standard basis for \(\liet^n \cong \R^n\), dual to $\{u_i\}$.

514: This map fits into an exact sequence

515: \[

516: 0 \longrightarrow \liet^k \stackrel{\iota}{\longrightarrow} \liet^n

517: \stackrel{\beta}{\longrightarrow} \liet^d\longrightarrow 0,

518: \]

519: where \(\liet^k := {\mathrm ker}(\beta).\) Exponentiating, we

520: get the exact sequence

521: \[

522: 0 \longrightarrow T^k \stackrel{\iota}{\longrightarrow} T^n

523: \stackrel{\beta}{\longrightarrow} T^d \longrightarrow 0,

524: \]

525: whereas by dualizing, we get

526: \[

527: 0 \longrightarrow (\liet^d)^{*} \stackrel{\beta^{*}}{\longrightarrow}

528: (\liet^n)^{*} \stackrel{\iota^{*}}{\longrightarrow} (\liet^k)^{*} \longrightarrow 0,

529: \]

530: where we abuse notation by using $\iota$ and $\beta$

531: to denote maps on the level of groups as well as on the level of algebras.

532: Note that $T^k$ is

533: connected if and only if the vectors $\{a_1,\ldots,a_n\}$ span

534: $\mathfrak{t}^d$ over the integers.

535: 

536: Consider the restriction of the action of $T^n$ on $\HH^n$ to the

537: subgroup $T^k$. This action is hyperhamiltonian with hyperk\"ahler moment map

538: $$\bar{\mu}_{\R}(z,w) = \iota^* \left(\frac{1}{2} \sum_{i=1}^n (|z_i|^2 - |w_i|^2) u_i

539: \right) \hspace{15pt}

540: \text{and}\hspace{15pt}\bar{\mu}_{\C}(z,w) = \iota^* \left(\sum_{i=1}^n (z_i w_i)u_i\right),$$

541: where $\{u_i\}$ is the standard basis in \(\tnd \cong \R^n.\)

542: In contrast with the K\"ahler situation, the hyperk\"ahler moment

543: map is surjective onto $\tnd\oplus(\tn_{\C})^*$.

544: 

545: We denote by 

546: % \(M(\bar{a}, \alpha)\)

547: $M$ the hyperk\"ahler

548: reduction of $\HH^n$ by the subtorus $T^k$ 

549: % (determined by the \(\bar{a} = (a_i)\)) 

550: at \((\alpha, 0)

551: \in (\liet^k)^{*}\oplus (\tk_{\C})^*,\) which is the 

552: hyperk\"ahler analogue of the K\"ahler toric variety $X = \Cn\mod_{\!\!\a}T^k$. 

553: Choose a lift $\tilde\a\in\tnd$ of $\a$ along $\iota^*$.

554: Then $M$ has a natural residual

555: action of $T^d$ with hyperk\"ahler moment map \(\mhk = \mr\oplus\mc.\)

556: Note that the choice of subtorus $\Tk\subs\Tn$ is equivalent

557: to choosing a central arrangement of cooriented hyperplanes in $\tdd$,

558: where the $i^{\text{th}}$ hyperplane is the annihilator

559: of $a_i\in\td$.  (The coorientation comes from the fact that we know

560: for which $x$ we have $\< x,a_i\> > 0$.)

561: The choice of $\tilde\a$ corresponds to an 

562: affinization $\arr$ of this arrangement,

563: where the $i^{\text{th}}$ hyperplane is 

564: $$H_i = \{x\in\tdd\mid\< x,a_i\>=\< -\tilde\a,\e_i\>\}.$$

565: Changing $\tilde\a$ by an element $c\in\tdd$ has the effect of translating $\arr$ by $c$,

566: and adding $c$ to the residual moment map $\mr$.

567: In order to record the information about coorientations,

568: we define the half-spaces 

569: \begin{equation}\label{FG}

570: F_i = \{x\in\tdd\mid\< x,a_i\>\geq\< -\tilde\a,\e_i\>\}

571: \hspace{15pt}\text{and}\hspace{15pt} 

572: G_i = \{x\in\tdd\mid\< x,a_i\>\leq\< -\tilde\a,\e_i\>\},

573: \end{equation}

574: which intersect in the hyperplane $H_i$.

575: Our convention will be to draw pictures, as in Figure 1,

576: in which we specify the coorientations

577: of the hyperplanes by shading the polyhedron $\Delta=\cap_{i=1}^n F_i$

578: (which works as long as $\Delta\neq\emptyset$).

579: Note that the K\"ahler variety $X$ is precisely the 

580: K\"ahler toric variety determined by $\Delta$. 

581: 

582: \begin{figure}[h]

583: \centerline{\epsfig{figure=htv2.eps}}

584: \caption{A hypertoric variety of real dimension 8 obtained

585: by reducing $\H^4$ by $T^2$.}

586: \end{figure}

587: 

588: % Bielawski and Dancer \cite{BD} show that 

589: % $M(\bar{a}, \alpha)$ 

590: The variety $M$ is an orbifold if and only if 

591: $\arr$ is simple, i.e. if and only if

592: every subset of $m$ hyperplanes intersect in codimension $m$ \cite[3.2]{BD}.

593: Furthermore, $M$ is smooth if and only if whenever

594: some subset of $d$ hyperplanes $\{H_i\}$ has non-empty intersection, 

595: the corresponding

596: vectors $\{a_i\}$ form a $\Z$-basis for \(\Z^d \subseteq \liet^d.\)

597: In this case we will refer to the arrangement itself as {\em smooth}.

598: We will always assume that $\arr$ is simple, and at times we will also assume that it is smooth.

599: % We will always assume that $\arr$ 

600: % is smooth and simple, and when we discuss

601: % cohomology in Sections \ref{Konno} and \ref{s1},

602: % we will also assume that

603: 

604: The hyperplanes $\{H_i\}$

605: divide \((\liet^d)^{*} \cong \R^d\) into a finite family of closed,

606: convex polyhedra $$\Delta_A = (\cap_{i\in A}F_i)\cap(\cap_{i\notin A}G_i),$$

607: indexed by subsets $A\subs\{1,\ldots,n\}$.

608: % , i.e. $\Delta_A$

609: % is

610: % obtained from $\Delta$ by flipping the coorientations of the hyperplanes corresponding

611: % to the elements of $A$. 

612: % For example, $\Delta_{\emptyset} = \Delta$ is the shaded region

613: % in Figure 1, and $\Delta_{\{2\}}$ is the triangle above $\Delta$.

614: Consider the subset $I = 

615: \{A\subs\{1,\ldots,n\}\mid \Delta_A\text{  bounded}\}$

616: of the power set of $\{1,\ldots,n\}$.  

617: For each $A\subs\{1,\ldots,n\}$, let $$M_A = \mr^{-1}(\Delta_A)\cap\mc^{-1}(0).$$

618: The K\"ahler submanifold

619: $\left(M_A,\omega_{\R}|_{M_A}\right)$ of $\left(M,\omega_{\R}\right)$

620: is $d$-dimensional and invariant under the action of $\Td$,

621: and is therefore $\Td$-equivariantly isomorphic

622: to the K\"ahler toric variety determined by $\Delta_A$ \cite[6.5]{BD}.

623: We define the {\em core} $C$ and {\em extended core} $D$

624: of a hypertoric variety by setting $$C = \cup_{A\in I}M_A\hs\hs\text{ and }\hs\hs

625: D = \cup_{A}{M_A} = \mc^{-1}(0) =

626: \{[z,w]\mid z_iw_i=0\text{  for all  }i\},$$

627: where $[z,w]$

628: denotes the $T^k$-equivalence class in $M$ of the element

629: $(z,w)\in \bar{\mu}_{\text{HK}}^{-1}(\a,0)$.

630: Bielawski and Dancer \cite{BD} show that

631: $C$ and $D$ are each $\Td$-equivariant

632: deformation retracts of $M$.  See Corollary \ref{unstable} for a Morse theoretic proof.

633: 

634: We take a minute to discuss the differences between

635: the combinatorial data

636: determining a toric variety $X = \Cn\mod_{\!\!\a}\Tk$ and its hypertoric

637: analogue $M = \Hn\mmod_{\!\!(\a,0)}\Tk$.

638: Each is determined by $\arr$,

639: a simple, cooriented, affine arrangement of $n$ hyperplanes in $\tdd$,

640: defined up to simultaneous translation.

641: The toric variety $X$ is in fact determined by less information than

642: this; it depends only on the polyhedron $\Delta = \cap_{i=1}^n F_i$.

643: Thus if the last hyperplane $H_n$ has the property that

644: $\cap_{i=1}^{n-1}F_i\subs F_n$, then this hyperplane is superfluous to $X$.

645: This is not the case for $M$, which means that it is slightly

646: misleading to call $M$ the hyperk\"ahler analogue of $X$;

647: more precisely, it is the hyperk\"ahler analogue of

648: {\em a given presentation} of $X$ as a K\"ahler reduction of $\Cn$.

649: On the other hand, the $\Td$-equivariant diffeomorphism type of $M$ also 

650: does not depend on all of the information of $\arr$, as evidenced by the two following results.

651: 

652: \begin{lemma}\label{placement}

653: The hypertoric varieties \(M_{\a} = \Hn \mmod_{\!\!(\a,0)}\Tk\) 

654: and $M_{\a'}= \Hn \mmod_{\!\!(\a',0)}\Tk$ are $\Td$-equivariantly

655: diffeomorphic, and their

656: cohomology rings can be naturally identified.

657: \end{lemma}

658: 

659: \begin{lemma}\label{coorientation}

660: The hypertoric variety $M$ does not depend on the coorientations

661: of the hyperplanes $\{H_i\}$.

662: \end{lemma}

663: 

664: This means that, unlike that of $X$, 

665: the $\Td$-equivariant diffeomorphism type of $M$ depends only on the unoriented

666: central arrangement underlying $\arr$.

667: A weaker version of Lemma \ref{placement}, involving

668: the (nonequivariant) homeomorphism type of $M$, appears in \cite{BD}.

669: 

670: \begin{proofp}

671: The set of nonregular values for \(\bar{\mu}_{\text{HK}}\) 

672: has codimension 3 inside of \(\tdd

673: \oplus (\td_{\C})^*.\) This tells us that the set of regular values 

674: is simply connected, and we can

675: choose a path connecting any two regular values \((\alpha,0)\)

676: and \((\alpha',0),\) 

677: unique up to homotopy.

678: 

679: Since the moment map \(\bar{\mu}_{\text{HK}}\)

680:  is not proper, we must take some care in showing that

681: two fibers are diffeomorphic. To this end, we note that the

682: norm-square function \(\psi(z,w) = \|z\|^2 + \|w\|^2\) is $T^n$-invariant

683: and proper on $\Hn$. 

684: Let $\H_{reg}^n$ denote the open submanifold of $\Hn$ consisting of the

685: preimages of the regular values of \(\bar{\mu}_{\text{HK}}.\) 

686: By a direct computation, it is easy

687: to see that the kernels of $d\psi$ and \(d(\bar{\mu}_{\text{HK}})\) intersect

688: transversely at any point \(p \in \H_{reg}^n.\) Using the standard

689: $T^n$-invariant metric on $\Hn$, we define an Ehresmann connection

690: on $\H_{reg}^n$ with respect to \(\bar{\mu}_{\text{HK}}\) 

691: such that the horizontal subspaces

692: are contained in the kernel of $d\psi$.

693: %Note that this connection can be extended to an Ehresmann

694: %connection with respect to \(\iota^{*} \bar{\mu}_{\R} \oplus

695: %\iota_{\C}^{*} \bar{\mu}_{\C} \oplus \psi.\) 

696: 

697: This connection allows us to lift a path connecting the two

698: regular values to a horizontal vector field on its preimage in

699: $\H_{reg}^n$. Since the horizontal subspaces are tangent to

700: the kernel of $d\psi$, the flow preserves level sets of $\psi$. Note that the

701: function \[\bar{\mu}_{\text{HK}} \oplus \psi: \Hn \to \tdd \oplus (\td_{\C})^{*} \oplus

702: \R\] {\it is} proper. By a theorem of Ehresmann \cite[8.12]{BJ}, the

703: properness of this map

704: implies that the flow of this vector field exists for all time, and

705: identifies the inverse image of $(\alpha,0)$ with that of

706: $(\alpha',0)$. Since the metric, $\psi$, and \(\bar{\mu}_{\text{HK}}\) are

707: all $T^n$-invariant, the Ehresmann connection is also $T^n$-invariant,

708: therefore the diffeomorphism identifying the fibers

709: is $T^n$-equivariant, making 

710: the reduced spaces are $T^d$-equivariantly diffeomorphic. 

711: \end{proofp}

712: 

713: \vspace{-\baselineskip}

714: \begin{proofc}

715: It suffices to consider the case when we change the orientation of a

716: single hyperplane within the arrangement.  Changing the coorientation

717: of a hyperplane $H_l$ is equivalent to defining a new map

718: $\beta':\tn\to\td$, with $\beta'(\e_i)=a_i$ for $i\neq l$, and $-a_i$ for $i=l$.

719: This map exponentiates to a map $\beta':\Tn\to\Td$,

720: and we want to show that the hyperk\"ahler variety obtained

721: by reducing $\Hn$ by the torus $\ker(\beta')$ is isomorphic to $M$,

722: which is obtained by reducing $\Hn$ by the torus $\Tk=\ker(\beta)$.

723: To see this, note that $\ker(\beta')$ and $\ker(\beta)$

724: are conjugate inside of $\operatorname M(n,\H)$ by the element

725: $(1,\ldots,1,j,1,\ldots,1)\in \operatorname 

726: M(1,\H)^n\subs \operatorname M(n,\H)$,

727: where the $j$ appears in the $l^{\text{th}}$ slot. 

728: \end{proofc}

729: 

730: \vspace{-\baselineskip}

731: \begin{example}\label{same}

732: The three cooriented arrangements of Figure 2 all specify

733: the same hyperk\"ahler variety $M$ up to equivariant

734: diffeomorphism.  The first has $X\cong \mathbb{F}_1$ (the

735: first Hirzebruch surface) and the second and the third have $X\cong\C P^2$.

736: Note that if we flipped the coorientation of $H_3$ in Figure 2(a)

737: or 2(c), then

738: we would get a non-compact

739: $X\cong\tilde{\C}^2$, the blow-up of $\C^2$

740: at a point.  If we flipped the coorientation of $H_3$ in

741: Figure 2(b), then $X$ would be empty.

742: We make no assumptions about $X$ in this section.

743: 

744: \begin{figure}[h]

745: \centerline{\epsfig{figure=htv1.eps}}

746: \caption{Three arrangements

747: related by flipping coorientations and translating hyperplanes.}

748: \end{figure}

749: \end{example}

750: 

751: The purpose of this paper is to study not just the topology

752: of $M$, but the topology of $M$ along with the natural hamiltonian

753: $S^1$ action defined in Section \ref{reduction}.

754: In order to define this $S^1$ action, it is necessary that

755: we reduce at a regular value of the form $(\a,0)\in\tdd\oplus(\liet_{\C}^d)^*$,

756: and although the set of regular values of $\bar{\mu}_{\text{HK}}$ is simply connected,

757: the set of regular values of the form $(\a,0)$ is not even connected.

758: Furthermore, 

759: left multiplication by the diagonal matrix $(1,\ldots,1,j,1,\ldots,1) \in 

760: \operatorname U(n,\H)$ is not

761: an $S^1$-equivariant automorphism of $\Hn$, therefore the geometric

762: structure of $M$ along with a circle action may depend nontrivially

763: both on the affine structure

764: and the coorientations of the arrangement $\arr$.

765: Indeed it must, because we can recover $X$ from $M$ by taking the minimum

766: $\Phi^{-1}(0)$ of the $S^1$ moment map $\Phi:M\to\R$.

767: In this sense, the structure of a hypertoric variety $M$

768: along with a circle action is the universal geometric object from which

769: both $M$ and $X$ can be recovered.

770: % In Section \ref{s1}, we will see that

771: % the $S^1$-equivariant cohomology ring of $M$ is sensitive

772: % to the coorientations as well as the affine structure of $\arr$.

773: \end{section}

774: 

775: \begin{section}{Gradient flow on the core}\label{coreflow}

776: Although $S^1$ does not act on $M$ as a subtorus of $\Td$, we show

777: below that when restricted to any single 

778: component $M_A$ of the extended core, $S^1$

779: {\em does} act as a subtorus of $\Td$, with the subtorus depending

780: combinatorially on $A$.  This will allow us to give a combinatorial

781: analysis of the gradient flow of $\Phi$ on the extended core.

782: 

783: \begin{lemma}\label{sides}

784: Let $x$ be an element of $\tdd$, and consider a point $[z,w]\in\mr^{-1}(x)\cap D$.

785: We have $x \in F_i$ if and only if $w_i=0$, and $x\in G_i$

786: if and only if $z_i=0$.

787: \end{lemma}

788: 

789: \begin{proof}

790: The fact that $[z,w]\in D$ tells us that $z_iw_i=0$.

791: Then 

792: \begin{eqnarray*}

793: x\in F_i & \iff & \< -\tilde\a,\e_i\>\leq\<\mr[z,w],

794: 	           a_i\>  =\<\bar\mr(z,w)-\tilde\a,\e_i\>  \\ 

795:         & \iff  &  \half|z_i|^2-\half|w_i|^2=\<\bar\mr(z,w),

796: 		    \e_i\>\geq 0.                                  \\

797: \end{eqnarray*}

798: Since $z_iw_i=0$, this is equivalent to the condition $w_i=0$.

799: The second half of the lemma follows similarly.

800: \end{proof}

801: 

802: On the suborbifold $M_A\subs D\subs M$ we have $z_i = 0$ for all $i\in A$

803: and $w_i = 0$ for all $i\notin A$, therefore for $\tau\in S^1$

804: and $[z,w]\in M_A$,

805: $$\tau[z,w]=[z,\tau w] = 

806: [\tau_1 z_1,\ldots \tau_n z_n, \tau_1^{-1}w_1,\ldots \tau_n^{-1}w_n], \text{ where }

807: \tau_i =

808: \begin{cases}

809: \tau^{-1} & \text{if }i\in A,\\

810: 1 & \text{if }i\notin A.

811: \end{cases}

812: $$

813: In other words, the $S^1$ action on $M_A$

814: is given by the one dimensional subtorus 

815: $(\tau_1, \ldots, \tau_n)$

816: of the original torus $T^n$,

817: hence the moment map $\Phi|_{M_A}$ is given (up to

818: an additive constant) by 

819: $$\Phi[z,w] = \left<\mr[z,w],\sum_{i\in A}a_i\right>.$$

820: This formula allows us to compute the fixed point sets of

821: the $S^1$ action.

822: Since $S^1$ acts freely on \((\liet_{\C}^d)^{*} \backslash \{0\}\)

823: and \(\mc: M \to (\liet_{\C}^d)^{*}\) is $S^1$-equivariant,

824: we must have $M^{S^1} \subseteq \mc^{-1}(0) = D$.

825: For any subset $B\subs\{1,\ldots,n\}$, let $M_A^B$ be the toric subvariety

826: of $M_A$ defined by the conditions $z_i=w_i=0$ for all $i\in B$.

827: Geometrically, $M_A^B$ is defined by the (possibly empty) face

828: $\cap_{i\in B}H_i \cap \Delta_A$ of the polyhedron $\Delta_A$.

829: 

830: \begin{proposition}\label{fixed}

831: The fixed point set of the action of $S^1$ on $M_A$

832: is the union of those toric subvarieties $M_A^B$ such that

833: $\sum_{i\in A}a_i\in \liet^d_B:=\operatorname{Span}_{j\in B}a_j$.

834: \end{proposition}

835: 

836: \begin{proof}

837: The moment map $\Phi|_{M_A^B}$ will be constant if and only if

838: $\sum_{i\in A}a_i$ is perpendicular to 

839: $\ker\left(\tdd\onto(\liet^d_B)^*\right)$,

840: i.e. if $\sum_{i\in A}a_i$ lies in the kernel of the projection

841: $\td\onto\td/\liet^d_B$.

842: \end{proof}

843: 

844: \vspace{-\baselineskip}

845: \begin{corollary}

846: Every vertex $v\in\tdd$ of the polyhedral complex defined by $\arr$

847: is the image of an $S^1$-fixed point in $M$.

848: Every component of $M^{S^1}$ has dimension less than or equal to $d$,

849: and the only component of dimension $d$ is $M_{\emptyset}=X=\Phi^{-1}(0)$. 

850: \end{corollary}

851: 

852: For any point $p\in M^{S^1}$, the {\em stable orbifold} 

853: $S(p)$ at $p$ is defined to be the set of $x\in M$ such

854: that $x$ approaches $p$ when flowing along the vector field 

855: $-\operatorname{grad}(\Phi)$,

856: and the {\em unstable orbifold} $U(p)$ at $p$ is defined to be the stable orbifold with respect

857: to the function $-\Phi$.

858: For any suborbifold $Y\subs M^{S^1}$, the unstable orbifold $U(Y)$ at $Y$ is defined

859: to be the union of $U(y)$ for all $y\in Y$.  In general, for $y\in Y$, we have the identity

860: $\dim_{\R} U(Y) + \dim_{\R} S(y) = 4d$.

861: 

862: Let $Y\subs M^{S^1}$ be a component of the fixed point set of $M$.

863: Let $v\in\tdd$ be a vertex in the polehedron $\mr(Y)$, and let $y$

864: be the unique preimage of $v$ in $Y$.

865: 

866: \begin{proposition}

867: The unstable orbifold $U(Y)$ is a complex suborbifold of complex 

868: dimension at most $d$, contained in the core $C\subs M$.

869: If $\arr$ is smooth at $y$, then $\dim_{\C}U(Y)=d$, and the closure of $U(Y)$

870: is an irreducible component of $C$.

871: \end{proposition}

872: 

873: \begin{proof}

874: For simplicity, we will assume that $v=\cap_{j=1}^dH_j$.

875: For all $l\in\otd$, let $b_l\in\td_{\Z}$ be the smallest integer vector

876: such that $\left<a_j,b_l\>=0$ for $j\neq l$ and $\left<a_l,b_l\> >0$.

877: Geometrically, $b_l$ is the primitive integer vector on the line

878: $\cap_{j\neq l}H_j$ pointing in the direction of $\Delta$.

879: Note that $M$ is smooth at the $T^d$-fixed point above $v$

880: if and only if $\left<a_l,b_l\>=1$ for all $l\in\otd$.

881: Let $R_l\subs\tdd$ be the ray eminating from $v$ in the direction of $b_l$,

882: and ending before it hits another vertex.  Let $Q_l$ be the analogous ray in the opposite

883: direction.

884: 

885: Let $\Delta_A$ be a region (not necessarily bounded) of the polyhedral 

886: complex defined by $\arr$ adjacent to $R_l$.

887: The preimage $\mr^{-1}(R_l)\cap D$ of $R_l$ in $D$ is a complex line,

888: and it is contained in the unstable

889: orbifold at $U(Y)$ if and only if $\left<b_l,\sum_{i\in A}a_i\right>\geq 0$.

890: If $\left<b_l,\sum_{i\in A}a_i\right> <0$, it is contained in the stable orbifold $S(y)$.

891: The preimage $\mr^{-1}(Q_l)\cap D$ of $Q_l$ in $D$ is also a complex line,

892: contained in the unstable

893: orbifold $U(Y)$ if and only if $\<-b_l,a_l+\sum_{i\in A}a_i\>\geq 0$,

894: and otherwise in $S(y)$.

895: Since $\<b_l,\sum_{i\in A}a_i\>+\<-b_l,a_l+\sum_{i\in A}a_i\>

896: =-\left<a_l,b_l\> <0$, at most one of these two directions can be unstable.

897: In the smooth case, $\left<a_l,b_l\>=1$ for all $l$, and exactly one of

898: the two directions is unstable.

899: 

900: Consider the polytope $\Delta_v$ incident to $v$ and characterized by the property that its

901: edges at the vertex $v$ are exactly the unstable directions. 

902: The toric variety $X_{\Delta_v}\subs D$ is contained

903: in the closure of $U(Y)$, and a dimension count tells us that this containment is an equality.

904: In the smooth case, $\Delta_v$ is $d$-dimensional, and $X_{\Delta_v}$ is a component of the core.

905: \end{proof}

906: 

907: Note that, even in the smooth case, it is not necessarily the case that the

908: $R_l$ direction is stable and the $Q_l$ direction is unstable.  See, for example,

909: the vertex $v = H_1\cap H_2$ in Figure 2(c).

910: 

911: \begin{corollary}

912: There is a natural injection from the set of bounded regions 

913: $\{\Delta_A\mid A\in I\}$

914: to the set of connected components of $M^{S^1}$.

915: If $\arr$ is smooth, this map is a bijection.

916: \end{corollary}

917: 

918: \begin{proof}

919: To each $A\in I$, we associate the fixed subvariety $M_A^B$ corresponding

920: to the face of $\Delta_A$ on which the linear functional $\sum_{i\in A}a_i$

921: is minimized, so that $M_A$ is the closure of

922: $U(M_A^B)$.  If $\arr$ is smooth, then every connected component of the fixed point

923: set will have a component of the core as its closed unstable orbifold.

924: \end{proof}

925: 

926: \vspace{-\baselineskip}

927: \begin{corollary}\label{unstable}

928: The core of $M$ is equal to the union of the unstable orbifolds of

929: the connected components of $M^{S^1}$, hence $C$ is a $\Tdso$-equivariant

930: deformation retract of $M$.

931: \end{corollary}

932: 

933: \begin{example}

934: In Figure 3, representing a reduction of $\H^5$ by $T^3$,

935: we choose a metric on $(\liet^2)^*$ in order

936: to draw the linear functional $\sum_{i\in A}a_i$ as a vector

937: in each region $\Delta_A$.  We see that $M^{S^1}$ has three components,

938: one of them $X\cong\mathbb F_1$, one of them a projective line, with another $\mathbb F_1$

939: as its unstable manifold, and one of them a point, with a $\C P^2$ as its unstable manifold.

940: 

941: \begin{figure}[h]

942: \centerline{\epsfig{figure=flow.eps}}

943: \caption{The gradient flow of $\Phi:M\to\R$.}

944: \end{figure}

945: \end{example}

946: 

947: \begin{example}

948: The hypertoric variety represented by Figure 4 has a fixed point set with four

949: connected components (three points and a $\C P^2$), but only three components in its core.

950: This phenomenon can be blamed on the orbifold point represented by the

951: intersection of $H_3$ and $H_4$, which has a one-dimensional unstable orbifold.

952: 

953: \begin{figure}[h]

954: \centerline{\epsfig{figure=orbi.eps}}

955: \caption{A singular example.}

956: \end{figure}

957: \end{example}

958: \end{section}

959: 

960: \begin{section}{Equivariant cohomology}\label{Konno}

961: In this section we extend Konno's computations

962: of the ordinary and $\Td$-equivariant cohomologies of $M$

963: to the $S^1$-equivariant setting.

964: We follow Konno's approach of restricting to the smooth case

965: to simplify arguments involving line bundles on $M$.

966: Hausel and Sturmfels, however, prove

967: theorems analogous to \ref{htm} and \ref{hm} with rational coefficients 

968: in the orbifold case, and Theorems \ref{htsm} and \ref{hsm}

969: extend to this setting as well (see Remark \ref{lawrence}).

970: 

971: \begin{theorem}\label{htm}{\em\cite{K2}}

972: The $\Td$-equivariant cohomology ring of a smooth 

973: hypertoric variety $M$ is given by

974: $$\htm = \Z[u_1,\ldots,u_n]\bigg/

975: \left<\prod_{i\in S}u_i\hs\bigg{\vert}\hs\bigcap_{i\in S} H_i = \emptyset \right>.$$

976: \end{theorem}

977: 

978: \begin{remark}

979: This is precisely

980: the Stanley-Reisner ring of

981: the unoriented matroid determined by the arrangement $\arr$ \cite{HS}.

982: \end{remark}

983: 

984: Just as the cohomology of a toric variety is obtained

985: from the equivariant cohomology by introducing linear relations

986: that generate $\ker\iota^*=\left(\ker\beta\right)^{\bot}$, the same

987: is true for hypertoric varieties:

988: 

989: \begin{theorem}\label{hm}{\em\cite{K1}}

990: The ordinary cohomology ring of a smooth hypertoric variety $M$ is given by

991: $$\hm = \htm \big/ \<\Sigma a_iu_i\in \ker\iota^*\>.$$

992: \end{theorem}

993: 

994: The rest of this section will be devoted to the proof of the following two theorems.

995: % Given any set $S\subs\{1,\ldots,n\}$ such that $\cap_{i\in S} H_i = \emptyset$,

996: % there exists a splitting (unique if $S$ is minimal) $S = S_1\sqcup S_2$

997: % such that

998: % $$\big(\cap_{i\in S_1}G_i\big)\cap

999: % \big(\cap_{j\in S_2}F_j\big) = \emptyset.$$

1000: 

1001: \begin{theorem}\label{htsm}

1002: Let $M$ be the hypertoric variety corresponding to a

1003: smooth, cooriented arrangement $\arr$.

1004: Given any minimal set $S\subs \{1,\ldots,n\}$ such that 

1005: $\cap_{i\in S} H_i = \emptyset$, let $S = S_1\sqcup S_2$

1006: be the unique splitting of $S$ such that

1007: $\big(\cap_{i\in S_1}G_i\big)\cap

1008: \big(\cap_{j\in S_2}F_j\big) = \emptyset$ (see \eqref{FG}).

1009: Then the $\Tdso$-equivariant cohomology of $M$ is given by

1010: $$\htsm\cong\Z[u_1,\ldots,u_n,x]\bigg/

1011: \< \prod_{i\in S_1}u_i\times\prod_{j\in S_2}(x-u_j)

1012: \hs\bigg{\vert}\hs\bigcap_{i\in S} H_i = \emptyset\>.$$

1013: \end{theorem}

1014: 

1015: \begin{theorem}\label{hsm}

1016: In the notation of Theorem \ref{htsm},

1017: the $S^1$-equivariant cohomology ring

1018: of $M$ is given by

1019: $$\hsm \cong \htsm \big/ \<\Sigma a_iu_i\in \ker\iota^*\>.$$

1020: \end{theorem}

1021: 

1022: \begin{remark}

1023: Konno observes that the quotient map from the abstract

1024: polynomial ring $\Z[u_1,\ldots,u_n]\to\htm$ is precisely the $\Td$-equivariant

1025: Kirwan map $$\ktd:H^*_{\Tn}(\cot)\to\htm$$

1026: induced by the inclusion

1027: $\mu^{-1}(\a,0)\hookto\cot$.  

1028: Likewise, the map from $\Z[u_1,\ldots,u_n]/\ker\iota^*$ to $H^*(M)$

1029: is the ordinary Kirwan map

1030: $$\k:H^*_{\Tk}(\cot)\to H^*(M).$$

1031: The analogous maps for K\"ahler reductions

1032: are known to always be surjective \cite[5.4]{Ki}, but the hyperk\"ahler case

1033: remains open.  Thus Theorems \ref{htm} and \ref{hm} can be interpreted as saying

1034: that the Kirwan maps for hypertoric varieties are surjective, and computing the kernel.

1035: Likewise, Theorems \ref{htsm} and \ref{hsm} assert that the $S^1$-equivariant Kirwan maps

1036: $$\ktdso:H^*_{\Tn\times S^1}(\cot)\to\htsm$$ and

1037: $$\kso:H^*_{\Tk\times S^1}(\cot)\to\hsm$$ are surjective,

1038: and provide computations of their kernels.

1039: \end{remark}

1040: 

1041: In order to apply Konno's results, we will make use of

1042: the principle of {\em equivariant formality},

1043: proven for compact manifolds in \cite{Ki}, which we adapt to our situation

1044: in Proposition \ref{formality}.

1045: For the sake of simplicity, we will restrict our attention

1046: to the case where $X$ is compact and nonempty.

1047: This condition will be necessary for the application of

1048: Proposition \ref{formality} and the proof of Theorem \ref{tara},

1049: both of which require a proper Morse function, which we get from

1050: Proposition \ref{proper}.

1051: We note, however, that both Proposition \ref{formality} and Theorem

1052: \ref{tara} can be extended to the case of a general hypertoric

1053: variety by a Mayer-Vietoris argument, using the fact that the core

1054: $C\subs M$ is a compact $\Td\times S^1$-equivariant deformation retract.

1055: We present the slightly less general Morse theoretic proofs

1056: only because we find them more pleasant.

1057: 

1058: \begin{proposition}\label{formality}

1059: Let $M$ be a symplectic

1060: orbifold, possibly noncompact but of finite topological type.

1061: Suppose that $M$ admits a hamiltonian action of a torus $T\times S^1$,

1062: and that the $S^1$-component $\Phi:M\to\R$ of the moment map

1063: is proper and bounded below.  Then 

1064: $H^*_{T\times S^1}(M)$ 

1065: is a free module over $H^*_{S^1}(pt)$.

1066: \end{proposition} 

1067: 

1068: \begin{proof}

1069: Because $\Phi$ is a moment map, it is a Morse-Bott function

1070: such that all of the critical suborbifold and their normal bundles

1071: carry almost complex structures.  Thus we get a Morse-Bott stratification

1072: of $M$ into even-dimensional $T$-invariant suborbifolds.

1073: This tells us, as in [Ki, 5.8],

1074: that the spectral sequence associated to the fibration

1075: $M\hookto EG\times_G M\to BG$ collapses, and we get the desired result.

1076: \end{proof}

1077: 

1078: Consider the following commuting square of maps, where

1079: $\phi$ and $\psi$ are each given by setting $x$ to zero.

1080: 

1081: $$\begin{CD}

1082: H^*_{\Tn\times S^1}(\cot) @>\ktdso >> \htsm\\

1083: @V\phi VV @ VV\psi V\\

1084: H^*_{\Tn}(\cot) @>\ktd >> \htm\\

1085: \end{CD}$$

1086: 

1087: Proposition \ref{formality} has the following consequence.

1088: 

1089: \begin{corollary}\label{enough}

1090: Let $\mathcal{I}\subs\ker\ktdso$ be an ideal with $\phi(\mathcal{I}) = \ker\ktd$.

1091: Then $\mathcal{I}=\ker\ktdso$.

1092: \end{corollary}

1093: 

1094: \begin{proof}

1095: Suppose that $a\in\ker\ktdso\smallsetminus\mathcal{I}$ is a homogeneous class of minimal degree,

1096: and choose $b\in\mathcal{I}$ such that $\phi(a-b)=0$.  Then $a-b = cx$ for some

1097: $c\in H^*_{\Tn\times S^1}(\cot)$.  By Proposition \ref{formality},

1098: $cx\in\ker\ktdso\impl c\in\ker\ktdso$, hence $c\in\ker\ktdso\smallsetminus\mathcal{I}$

1099: is a class of lower degree than $a$.

1100: \end{proof}

1101: 

1102: \vspace{-\baselineskip}

1103: \begin{lemma}\label{surjective}

1104: The equivariant Kirwan map $\ktdso$ is surjective.

1105: \end{lemma}

1106: 

1107: \begin{proof}

1108: Suppose that $\gamma\in\htsm$ is a homogeneous class of minimal degree

1109: that is {\em not}

1110: in the image of $\ktdso$.  By Theorem \ref{htm} $\ktd$ is surjective, hence we may choose a class

1111: $\eta\in\phi^{-1}\ktd^{-1}\psi(\gamma)$.

1112: Then $\ktdso(\eta)-\gamma = x\delta$ for some $\delta\in\htsm$,

1113: and therefore $\delta$ is a class of lower degree that is not in the image of $\ktdso$.

1114: \end{proof}

1115: 

1116: \vspace{-\baselineskip}

1117: \begin{proofhtsm}

1118: For any element $h\in H^2_{T^n\times S^1}(\Hn;\Z)$,

1119: let $\tilde{L}_h = \Hn\times\C_{h}$ 

1120: be the $\Tn\times S^1$-equivariant line bundle on $\Hn$

1121: with equivariant Euler class $h$. 

1122: This gives 

1123: $\tilde{L}_h$, as well as its dual $\tilde{L}_h^*$, 

1124: the structure of a \(T^n \times S^1\)-equivariant bundle. 

1125: Let $$L_h = \tilde{L}_h|_{\bar{\mu}_{\C}^{-1}(0)^{\text{ss}}}/T_{\C}^k$$ 

1126: be the quotient $\Td\times S^1$-equivariant line bundle on $M$.

1127: Let $\{u_i\}$ be the standard basis of $\Tnd$.

1128: Identifying $H^2_{T^n\times S^1}(\Hn;\Z)$ with $\Tnd\oplus\Z x$,

1129: We will use $\tilde L_i$ to denote the bundle $\tilde L_{u_i\oplus 0}$, 

1130: and $\tilde K$ to denote the bundle $\tilde L_{0\oplus x}$,

1131: with quotients $L_i$ and $K$. 

1132: Since the $\Tdso$-equivariant Euler class $e(L_i)$

1133: is the image of $u_i\oplus 0$ under the hyperk\"ahler Kirwan map

1134: $H^*_{T^n\times S^1}(\Hn)\to \htsm$, we will abuse notation and 

1135: denote it by $u_i$.  Similarly, we will denote $e(K)$ by $x$.

1136: Corollary \ref{surjective} tells us that

1137: $\htsm$ is generated by $u_1,\ldots,u_n,x$.

1138: 

1139: Consider the $\Tn\times S^1$-equivariant section $\tilde s_i$ of $\tilde{L}_i$

1140: given by the function $\tilde s_i(z,w)=z_i$.

1141: This descends to a $\Tdso$-equivariant section $s_i$ of $L_i$

1142: with zero-set $Z_i

1143: := \{[z,w]\in M\mid z_i=0\}$. 

1144: Similarly, the function $\tilde t_i(z,w)=w_i$ defines a $\Tdso$-equivariant

1145: section of $L_i^*\otimes K$ with zero set $W_i := \{[z,w]\in M\mid w_i=0\}$.

1146: Thus the divisor $Z_i$ represents the cohomology class $u_i$,

1147: and $W_i$ represents $x-u_i$.

1148: Note, by the proof of Lemma \ref{sides},

1149: that $\mr(Z_i) = G_i$ and $\mr(W_i) = F_i$ for all $1\leq i \leq n$.

1150: 

1151: Let $S = S_1\sqcup S_2$ be a subset of $\{1,\ldots n\}$ such that

1152: $\big(\cap_{i\in S_1}G_i\big)\cap

1153: \big(\cap_{j\in S_2}F_j\big) = \emptyset,$ and hence

1154: $$\big(\cap_{i\in S_1}Z_i\big)\cap

1155: \big(\cap_{j\in S_2}W_j\big)\subs\mr^{-1}\bigg(

1156: \big(\cap_{i\in S_1}G_i\big)\cap

1157: \big(\cap_{j\in S_2}F_j\big)\bigg) = \emptyset.$$

1158: Now consider the vector bundle

1159: $E_S=\left(\oplus_{i\in S_1}L_i\right)\oplus\left(\oplus_{j\in S_2}L_j^*\otimes K\right)$

1160: with equivariant Euler class $$e(E_S)=\prod_{i\in S_1}u_i\times\prod_{j\in S_2}(x-u_j).$$ 

1161: The section $\left(\oplus_{i\in S_1}s_i\right)\oplus\left(\oplus_{i\in S_2}t_i\right)$ 

1162: is a nonvanishing equivariant

1163: global section of $E_S$,

1164: hence for any such $S$, $e(E_S)$ is trivial in $\htsm$.

1165: 

1166: The fact that $u_1,\ldots,u_n,x$ generate $\htsm$ is proven in Lemma \ref{surjective},

1167: and the fact that

1168: we have found all of the relations

1169: follows from Theorem \ref{htm} and Corollary \ref{enough}.

1170: \end{proofhtsm}

1171: 

1172: \vspace{-\baselineskip}

1173: \begin{proofhsm}

1174: The proof of this theorem is identical to the proof of Theorem \ref{htsm},

1175: making use of Theorem \ref{hm} rather than Theorem \ref{htm}.

1176: \end{proofhsm}

1177: 

1178: How sensitive are the invariants $\htsm$ and $\hsm$?

1179: % By Proposition \ref{formality}, 

1180: We can recover $\htm$ and $\hm$

1181: by setting $x$ to zero, hence they are at least as fine

1182: as the ordinary or $\Td$-equivariant cohomology rings.

1183: The ring $\htsm$ does {\em not} depend on coorientations,

1184: for if $M'$ is related to $M$ by flipping the coorientation

1185: of the $l^{\text{th}}$ hyperplane $H_k$, then the map taking $u_i$ to

1186: $u_i$ for $i\neq l$ and $u_l$ to $x-u_l$ is an isomorphism

1187: between $\htsm$ and $H^*_{\Td\times S^1}(M')$.

1188: It is, however, dependent on the affine structure of 

1189: the arrangement $\arr$.

1190: 

1191: \begin{example}\label{ts}

1192: We compute the equivariant cohomology ring $\htsm$

1193: for the hypertoric varieties $M_a$, $M_b$, and $M_c$

1194: defined by the arrangements in Figure 2(a), (b), and (c), respectively.

1195: $$H^*_{\Td\times S^1}(M_a)=\Z[u_1,\ldots,u_4,x]\big/

1196: \< u_2u_3, u_1(x-u_2)u_4, u_1u_3u_4 \>,$$

1197: $$H^*_{\Td\times S^1}(M_b)=\Z[u_1,\ldots,u_4,x]\big/

1198: \< (x-u_2)u_3, u_1u_2u_4, u_1u_3u_4 \>,$$

1199: $$H^*_{\Td\times S^1}(M_c)=\Z[u_1,\ldots,u_4,x]\big/

1200: \< u_2u_3, (x-u_1)u_2(x-u_4), u_1u_3u_4 \>.$$

1201: As we have already observed, $H^*_{\Td\times S^1}(M_a)$ and $H^*_{\Td\times S^1}(M_b)$

1202: are isomorphic by interchanging $u_2$ with $x-u_2$.

1203: % We now show that $H^*_{\Td\times S^1}(M_a)$ is not isomorphic

1204: % to $H^*_{\Td\times S^1}(M_c)$.  

1205: One can check that the annihilator of $u_2$ in $H^*_{\Td\times S^1}(M_a)$

1206: is the principal ideal generated by $u_3$, while the ring $H^*_{\Td\times S^1}(M_c)$

1207: has no degree $2$ element whose annihilator is generated by a single element of degree $2$.

1208: Hence $H^*_{\Td\times S^1}(M_c)$ is not isomorphic to the other two rings.

1209: % Since $(x-u_2)(x-u_3)$ is the unique

1210: % degree $4$ relation in both rings, any isomorphism 

1211: % between the two rings

1212: % would have to have to take $x-u_2$ and $x-u_3$ 

1213: % to scalar multiples of $x-u_2$ and $x-u_3$, possibly in the reverse order.

1214: % One can check that the only degree $4$ classes

1215: % in $H^*_{\Td\times S^1}(M_a)$ that annihilate $x-u_2$ are multiples of $x-u_3$,

1216: % whereas in $H^*_{\Td\times S^1}(M_c)$, $x-u_2$ is killed by $u_1u_4$

1217: % and $x-u_3$ is killed by $(x-u_1)(x-u_4)$.

1218: % Hence $x-u_2$ can be taken neither to a multiple of $x-u_2$ nor of $x-u_3$,

1219: % therefore the rings cannot be isomorphic.

1220: \end{example}

1221: 

1222: The ring $\hsm$, on the other hand, is sensitive to coorientations

1223: as well as the affine structure of $\arr$.

1224: 

1225: \begin{example}\label{s}

1226: We now compute the ring $\hsm$ for 

1227: $M_a$, $M_b$, and $M_c$ of Figure 2.

1228: Theorem \ref{hsm} tells us that we need only to quotient the ring

1229: $\htsm$ by $\ker(\iota^*)$.

1230: For $M_a$, the kernel of $\iota_a^*$ is generated by

1231: $u_1+u_2-u_3$ and $u_1-u_4$, hence we have

1232: \begin{eqnarray*}

1233: H^*_{S^1}(M_a)&=&\Z[u_2,u_3,x]\big/

1234: \< u_2u_3, (u_3-u_2)^2(x-u_2), (u_3-u_2)^2u_3 \>\\

1235: &\cong & \Z[u_2,u_3,x]\big/\< u_2u_3, (u_3-u_2)^2(x-u_2), u_3^3 \>.\\

1236: \end{eqnarray*}

1237: Since the hyperplanes of 2(c) have the same coorientations as those

1238: of 2(a), we have $\ker\iota_b^*=\ker\iota_a^*$, hence

1239: \begin{eqnarray*}

1240: H^*_{S^1}(M_c) &=& \Z[u_2,u_3,x]\big/

1241: \< u_2u_3, (x-u_3+u_2)^2u_2, (u_3-u_2)^2u_3 \>\\

1242: &\cong & \Z[u_2,u_3,x]\big/\< u_2u_3, (x-u_3+u_2)^2u_2, u_3^3 \>.\\

1243: \end{eqnarray*}

1244: Finally, since Figure 2(b) is obtained from 2(a) by flipping the coorientation

1245: of $H_2$, we find that $\ker(\iota_b^*)$ is generated by $u_1-u_2-u_3$

1246: and $u_1-u_4$, therefore

1247: $$H^*_{S^1}(M_b)=\Z[u_2,u_3,x]\big/

1248: \< (x-u_2)u_3, (u_2+u_3)^2u_2, (u_2+u_3)^2u_3 \>.$$

1249: As in Example \ref{ts}, $H^*_{S^1}(M_a)$ and $H^*_{S^1}(M_c)$

1250: can be distinguished by the fact that the annihilator

1251: of $u_2\in H^*_{S^1}(M_a)$ is generated by a single element of degree $2$,

1252: and no element of $H^*_{S^1}(M_c)$ has this property.

1253: On the other hand, $H^*_{S^1}(M_b)$ is distinguished from 

1254: $H^*_{S^1}(M_a)$ and $H^*_{S^1}(M_c)$ by the fact that neither $x-u_2$ nor $u_3$ cubes to zero.

1255: % Each of three rings have two linear elements, unique up to scale, that multiply to zero.

1256: % In $H^*_{S^1}(M_c)$, both of them cube to zero, while

1257: % in $H^*_{S^1}(M_a)$ and $H^*_{S^1}(M_b)$, only $(x-u_3)$ cubes to zero.

1258: % By changing coordinates, we can write

1259: % $$H^*_{S^1}(M_a) = \Z[x,y,z]\big/\<yz, (z-y+x)^2(x-y), z^3\>$$ and

1260: % $$H^*_{S^1}(M_b) = \Z[x,y,z]\big/\<yz, (z-y)^2(x-y), z^3\>.$$

1261: % Any isomorphism from $H^*_{S^1}(M_a)$ to $H^*_{S^1}(M_b)$ would have to take

1262: % $y$ and $z$ to multiples of themselves, and $x$ to an expression of the form

1263: % $\a x+\b y+ \gamma z$, with $\a\neq 0$.  Then a contradiction arises from the

1264: % fact that the first degree six relation in $H^*_{S^1}(M_a)$ is cubic in $x$,

1265: % while $H^*_{S^1}(M_b)$ has no such nontrivial relations.

1266: % Hence all three rings are nonisomorphic.

1267: \end{example}

1268: 

1269: \begin{remark}\label{lawrence}

1270: Theorems \ref{htsm} and \ref{hsm} can be interpreted

1271: in light of the recent work of Hausel and Sturmfels \cite{HS} 

1272: on Lawrence toric varieties.

1273: The Lawrence toric variety $N$ associated to the arrangement $\arr$

1274: is the K\"ahler reduction $\cot\mod\Tk$,

1275: so that $M$ sits inside of $N$ as the complete

1276: intersection cut out by the equation $\bar\mc(z,w)=0$.

1277: The residual torus acting on $N$ has dimension $d+n$,

1278: and includes the $(d+1)$-dimensional torus $\Td\times S^1$

1279: acting on $M$, and

1280: the inclusion of $M$ into

1281: $N$ induces an isomorphism on $\Td\times S^1$-equivariant cohomology.

1282: One can use geometric arguments similar to those that were applied to prove

1283: Theorem \ref{htsm},

1284: or the purely combinatorial approach of \cite{HS},

1285: to show that

1286: $$H_{T^{d+n}}^*(N)=\Q[u_1,\ldots,u_n,v_1,\ldots,v_n]\bigg/

1287: \< \prod_{i\in S_1}u_i\times\prod_{j\in S_2}v_j\hs\bigg{\vert}\hs\bigcap_{i\in S} H_i = \emptyset\>.$$

1288: From here we can recover $\htsm = H_{\Td\times S^1}^*(N)$

1289: by setting $u_i+v_i = u_j+v_j$ for all $i,j\leq n$.

1290: Note that Hausel and Sturmfels' work applies to the general orbifold case.

1291: \end{remark}

1292: \end{section}

1293: 

1294: \begin{section}{A deformation of the Orlik-Solomon algebra of \boldmath$\arr$}\label{defos}

1295: Let $M_{\R} \subs M$ be the {\em real locus}

1296: $\{[z,w]\in M\mid z,w \text{ real}\}$ of $M$ with respect to the complex structure $J_1$.

1297: The full group $\Td\times S^1$ does not act on $M_{\R} $,

1298: but the subgroup $\Tdr\times\Zt$ does act,

1299: where $\Tdr :=\Z_2^d\subs \Td$ is the fixed point set

1300: of the involution of $\Td$ given by complex

1301: conjugation.\footnote{It is interesting to note that

1302: the real locus with respect to the complex structure $J_1$

1303: is in fact a complex submanifold with respect to the

1304: one of the other complex structures on $M$.  

1305: The action of $\Tdr$ is holomorphic because $\Tdr$ is a subgroup of $\Td$,

1306: which preserves all of the complex structures on $M$.

1307: The action of $\Zt$, on the other hand, is anti-holomorphic,

1308: i.e. it can be thought of as complex conjugation.}

1309: In this section we will study the geometry of the real locus,

1310: focusing in particular on the properties of the residual $\Zt$ action.

1311: 

1312: A proof of a more general statement of the following theorem is 

1313: forthcoming in \cite{HH}. 

1314: 

1315: % BEGIN NEW

1316: % We begin by 

1317: % extending a theorem of

1318: % \cite{BGH, Sc}.

1319: % stating a theorem about the equivariant cohomology of $M_{\R} $,

1320: % the proof of which is forthcoming in \cite{HH}.

1321: 

1322: \begin{theorem}\label{tara}

1323: Let $G = \Tdso$ or $\Td$,

1324: and $G_{\R} = \Tdr\times\Zt$ or $\Tdr$.

1325: Then we have $H^{2*}_{G}(M;\Zt)\cong

1326: H^{*}_{G_{\R}}(M_{\R} ;\Zt),$

1327: i.e. the rings are isomorphic

1328: by an isomorphism that halves the grading.

1329: Furthermore, this isomorphism identifies the class 

1330: $u_i\in H^{*}_{G}(M;\Zt)$, 

1331: represented by the divisor $Z_i$, with the class in $H^*_{G_{\R}}(M_{\R} ;\Zt)$

1332: represented by the divisor $Z_i\cap M_{\R} $, and likewise takes

1333: $x-u_i$ (if $G=\Tdso$) or

1334: $-u_i$ (if $G=\Td$) to the class represented by $W_i\cap M_{\R} $. 

1335: \end{theorem}

1336: 

1337: \begin{sketch}

1338: Consider the injection $H^*_G(M;\Zt)\hookto H^*_G(M^G;\Zt)$

1339: given by the inclusion of the fixed point set into $M$.

1340: % Let \(M^{(1)}\) be the {\em one-skeleton} of $M$, i.e. the set of 

1341: % points in $M$ whose $G$-orbits are 0- or 1-dimensional. 

1342: % Let \(M^{(0)}\) denote the fixed points. Let 

1343: % \(j: M^{(1)} \hookrightarrow M\) and \(i: M^{(0)} \hookrightarrow M\) 

1344: % denote the natural inclusions. 

1345: The essential idea is to show that 

1346: % the image \(j^{*}(H^*_{G}(M^{(1)})

1347: % \subset H^{*}_{G}(M^{(0)})\) is the same as the image 

1348: a class in $H^*_G(M^G;\Zt)$ extends over $M$ if and only

1349: if it extends to the set of points on which $G$ acts with

1350: a stabilizer of codimension at most $1$,

1351: and then to show that a similar

1352: statement in $G_{\R}$-equivariant cohomology 

1353: also holds for the real locus $M_{\R} $ with its $G_{\R}$

1354: action. One then uses a canonical isomorphism \(H^2_G(pt,\Z_2) \cong 

1355: H^1_{G_{\R}}(pt,\Z_2)\) to give the result. 

1356: 

1357: The key to the proof is a noncompact $G_{\R}$ version

1358: of the proposition in 

1359: \cite{TW} stating that the $G_{\R}$-equivariant Euler class of 

1360: the negative normal bundle of a critical point $p$ 

1361: is not a zero divisor, which can be shown explicitly

1362: using a local normal form for the actions of $G$ and $G_{\R}$. 

1363: % Using a local normal form theorem for 

1364: % Hamiltonian $G$-manifolds equipped with a compatible 

1365: % antisymplectic involution $\sigma$ and properties of 

1366: % the $G$ weights and $G_{\R}$ weights at the fixed points, 

1367: % one can compute

1368: % explicitly the $G_{\R}$ equivariant Euler class and verify 

1369: %that it is not a zero divisor.  

1370: The proposition then follows 

1371: from standard $G_{\R}$ versions of the Thom isomorphism theorem

1372: with coefficients in $\Z_2$. 

1373: Since a component of the 

1374: moment map is proper, bounded below, and has finitely many

1375: fixed points, one can then check that the inductive argument, given 

1376: in Section 3 of \cite{TW}

1377: to complete the proof of \cite[Thm 1]{TW} also holds in this case.

1378: \end{sketch}

1379: 

1380: % What kind of smoothness assumptions do we need to make about 

1381: % $\arr$ and $M$?

1382: % In this section we work with cohomology with coefficients in $\Zt$,

1383: % therefore it is okay for $M$ to have orbifold points

1384: % as long as they are of odd order.  In terms of $\arr$, this means

1385: % that the primitive normal vectors to the hyperplanes at each vertex must

1386: % span a sublattice of odd index inside of $\liet^d_{\Z}$.

1387: % Such an arrangement will be referred to as an {\em odd} arrangement.

1388: % The oddness condition may seem only marginally more general than

1389: % the smoothness condition, but it is significant because of the following

1390: % important lemma:

1391: % 

1392: % \begin{lemma}\label{odd}

1393: % Every simple arrangement is combinatorially equivalent 

1394: % as an affine arrangement to

1395: % an odd arrangement.

1396: % \end{lemma}

1397: % 

1398: % \begin{proof}

1399: % \end{proof}

1400: % 

1401: % \noindent For the remainder of this section, we will assume that $\arr$ is odd.

1402: 

1403: Let us consider the restriction of the hyperk\"ahler moment map

1404: $\mhk=\mr\oplus\mc$ to $M_{\R} $.  Since $z$ and $w$ are real

1405: for every $[z,w]\in M_{\R} $, the map $\mc$

1406: takes values in $\td_{\R}\subs\td_{\C}$, which we will identify

1407: with $i\R^n$, so that $f = \mr|_{M_{\R}} \oplus\mc|_{M_{\R}}$ takes values

1408: in $\R^n\oplus i\R^n\cong\C^n$.

1409: Note that $f$ is $\Zt$-equivariant, with $\Zt$ acting on

1410: $\Cn$ by complex conjugation.

1411: 

1412: \begin{lemma}

1413: The map $f:{M_{\R}}\to\Cn$ is surjective, and the fibers are the orbits of $\Tdr$.

1414: The stabilizer of a point $x\in {M_{\R}}$ has order 

1415: $2^r$, where $r$ is the number of hyperplanes in the complexified

1416: arrangement $\arr_{\C}$ containing the point $f(x)$.

1417: \end{lemma}

1418: 

1419: \begin{proof}

1420: For any point $p=a+bi\in\Cn$, choose a point $[z,w]\in M$ such that

1421: $\mr[z,w]=a$ and $\mc[z,w]=b$.  We can hit $[z,w]$

1422: with an element of $\Td=\Tn/\Tk$ to make $z$ real, and the fact that

1423: $\mc[z,w]\in\R^n$ forces $w$ to be real as well, hence we may assume

1424: that $[z,w]\in {M_{\R}}$.

1425: Then $f^{-1}(p)=\mu_{\text{HK}}^{-1}\left(a,b\right)\cap {M_{\R}}

1426: = \Td[z,w]\cap {M_{\R}} = \Tdr[z,w]$.

1427: The second statement follows easily from \cite[3.1]{BD}.

1428: \end{proof}

1429: 

1430: Let $Y\subs {M_{\R}}$ be the locus of points on which $\Tdr$

1431: acts freely, i.e. the preimage under $f$ of the space 

1432: $\comp:=\bomp$.

1433: The inclusion map $Y\hookto {M_{\R}}$

1434: induces maps backward on cohomology, which we will denote

1435: $$\phi:\hr\to H^*_{\Tdr}(Y;\Zt)\cong H^*(\comp;\Zt)$$

1436: and

1437: $$\phi_2:\hrs\to H^*_{\Tdr\times\Zt}(Y;\Zt)\cong H^*_{\Zt}(\comp;\Zt).$$

1438: The ring $H^*(\comp;\Z)$ has been studied extensively, and is called

1439: the {\em Orlik-Solomon algebra} \cite{OT}, which we will denote

1440: by $\os$.

1441: A remarkable fact about the Orlik-Solomon algebra is that it depends

1442: only on the combinatorial structure of $\arr$; the following is a presentation

1443: in terms of {\it anticommuting} generators $e_1,\ldots,e_n$ \cite{OS}:

1444: $$\os\cong H^*(\comp;\Z)\cong\Z[e_1,\ldots,e_n]\big/

1445: \< \Pi_{i\in S}e_i\mid\cap_{i\in S} H_i = \emptyset \>.$$

1446: Rather than working with anticommuting generators, we can work over the ground field $\Zt$,

1447: in which case commutativity and anticommutativity are the same.

1448: Because $\os$ is torsion-free \cite[3.74]{OT}, we have

1449: $$\ost\cong H^*(\comp;\Zt)\cong\Zt[e_1,\ldots,e_n]\big/

1450: \< \Pi_{i\in S}e_i\mid\cap_{i\in S} H_i = \emptyset \>

1451: +\< e_i^2\mid i\leq n\>,$$

1452: where $\operatorname{deg}(e_i)=1$.

1453: 

1454: \begin{claim}

1455: The map $\phi:H^*_{\Tdr}(M_{\R};\Zt)\to\ost$ 

1456: takes $u_i$ to $e_i$, hence $\ker\phi$

1457: is generated by

1458: the set $\{u_i^2\mid i\leq n\}$.

1459: \end{claim}

1460: 

1461: \begin{proof}

1462: Recall from Section \ref{ht}

1463: that the hyperplane $H_i\subs \tdd$

1464: is defined by the equation $\< x,a_i\> = \< -\tilde\a,\e_i\>$.

1465: Let $\eta_i:\Cn\to\C$ be the affine map

1466: taking $x$ to $\< x,a_i\> + \< \tilde\a,\e_i\>$,

1467: so that $H_i^{\C}$ is cut out of $\Cn$ by $\eta_i$.

1468: Then $\eta_i$ restricts to a map $\comp\to\C^*$, and

1469: Orlik and Terao identify $e_i$ with the cohomology class

1470: represented by the pull-back

1471: of the submanifold $\R_-$ (the negative reals) along $\eta_i$ [OT, 5.90].

1472: Theorem \ref{tara} tells us that the cohomology

1473: class $u_i$ is represented by the divisor $Z_i\cap {M_{\R}}$.

1474: By Lemma \ref{sides}, $f(Z_i\cap {M_{\R}}) = G_i\cap \R^n$,

1475: hence $\phi(u_i)$ is the class represented by the submanifold

1476: $G_i\cap \R^n\cap \comp = 

1477: \eta_i^{-1}(\R_-)$.

1478: \end{proof}

1479: 

1480: \vspace{-\baselineskip}

1481: \begin{remark}\label{kernel}

1482: The fact that the classes $u_i^2$ lie in the kernel of $\phi$ can be seen

1483: by noting that $u_i$ is represented

1484: both by the divisor $Z_i = \{[z,w]\in {M_{\R}}\mid z_i=0\}$, 

1485: and, since $u_i = -u_i$ over $\Zt$, 

1486: also by the divisor $W_i = \{[z,w]\in {M_{\R}}\mid w_i=0\}$.

1487: The condition $x\in Z_i\cap W_i$ says exactly that

1488: $\mr(x)\in H_i$, therefore

1489: the intersection $Z_i\cap W_i \cap Y$ is empty.

1490: \end{remark}

1491: 

1492: In some sense we have cheated here; we have concluded that

1493: we can recover a presentation of $\ost$ from a presentation

1494: of $\htm$, but we used the fact that we already have a presentation

1495: of $\ost$.  In the $\Zt$-equivariant picture, however,

1496: our trivial observation turns magically into new information,

1497: giving us a presentation of the equivariant cohomology

1498: ring $\hscomp$.

1499: 

1500: \begin{theorem}\label{surj}

1501: The map $\phi_2$ is surjective, with kernel generated by 

1502: $\{u_i (x-u_i)\mid i\leq n\}$.

1503: Hence 

1504: $$\hscomp \cong H^*_{\Tdr\times\Zt}(M_{\R};\Zt)\big/\ker\phi_2$$

1505: $$\cong\Zt[u_1,\ldots,u_n,x]\big/

1506: \<\Pi_{i\in S_1}u_i\times\Pi_{j\in S_2}(x-u_j)

1507: \mid\cap_{i\in S} H_i = \emptyset\>

1508: +\<u_i(x-u_i)\mid i\leq n\>.$$

1509: \end{theorem}

1510: 

1511: \begin{proof}

1512: % Theorem \ref{tara} and Proposition \ref{formality}

1513: % imply that $$\hrs\cong\hr\otimes H_{\Zt}^*(pt)$$

1514: % as a module over $H_{\Zt}^*(pt)$.

1515: % On the other hand, 

1516: The fact that $\arr_{\C}$ is the 

1517: complexification of a real arrangement

1518: tells us that the linear forms $\eta_i$ have real coefficients,

1519: therefore the generators $e_i$ of $\hcomp$ are $\Zt$-invariant.

1520: This implies, by \cite[3.5]{Bo},

1521: that $\hscomp$

1522: is a free module over $H_{\Zt}^*(pt)$.

1523: Then surjectivity of $\phi_2$ follows from surjectivity of $\phi$

1524: using a formal argument identical to that of Corollary \ref{surjective}.

1525: By Theorem \ref{tara} and Proposition \ref{formality},

1526: $\hrs$ is a free module over $H_{\Zt}^*(pt)$,

1527: therefore $\ker\phi_2$ is a free 

1528: $H_{\Zt}^*(pt)$-module of rank $n$.

1529: The fact that $u_i(x-u_i)\in \ker\phi_2$ follows from the

1530: argument of Remark \ref{kernel}, hence we are done.

1531: \end{proof}

1532: 

1533: The ring $\hscomp$ is therefore a deformation

1534: of $\ost$ (over the base $\operatorname{Spec}\Zt[x]$)

1535: that depends nontrivially on the affine structure of $\arr$,

1536: rather than simply on the underlying matroid.

1537: 

1538: \begin{example}\label{first}

1539: Consider the arrangements $\arr_a$ and $\arr_c$ in Figure 2(a) and 2(c).

1540: By Theorem \ref{surj} and Example \ref{ts} we have

1541: $$H^*_{\Zt}(\mathcal{M}(\arr_a);\Zt)\cong

1542: \Zt [u_1,\ldots,u_4,x]\bigg/

1543: \< \begin{array}{c}

1544: u_1(x-u_1), u_2(x-u_2), u_3(x-u_3), u_4(x-u_4),\\

1545: u_2u_3, u_1(x-u_2)u_4, u_1u_3u_4 

1546: \end{array}\>$$

1547: and

1548: $$H^*_{\Zt}(\mathcal{M}(\arr_c);\Zt)\cong

1549: \Zt [u_1,\ldots,u_4,x]\bigg/

1550: \< \begin{array}{c}

1551: u_1(x-u_1), u_2(x-u_2), u_3(x-u_3), u_4(x-u_4),\\

1552: u_2u_3, (x-u_1)u_2(x-u_4), u_1u_3u_4

1553: \end{array}\>.$$

1554: The map $f:H^*_{\Zt}(\mathcal{M}(\arr_a);\Zt)\to H^*_{\Zt}(\mathcal{M}(\arr_b);\Zt)$

1555: given by

1556: $$f(u_1) = u_1+u_2, f(u_2) = u_2+u_3+x, f(u_3) = u_3,

1557: f(u_4) = u_2+u_4,\text{ and }f(x)=x$$

1558: is an isomorphism of graded $\Zt[x]$-algebras, hence the

1559: ring $\hscomp$ is not a complete invariant of smooth, rational, 

1560: affine arrangements up to combinatorial equivalence.\footnote{We thank

1561: Graham Denham for finding this isomorphism.}

1562: \end{example}

1563: 

1564: \begin{example}\label{second}

1565: Now consider the arrangements $\arr_a'$ and $\arr_c'$ obtained from $\arr_a$ and $\arr_c$

1566: by adding a vertical line on the far left, as shown below.

1567: \begin{figure}[h]

1568: \begin{center}

1569: \psfrag{Ha'}{$\arr_a'$}

1570: \psfrag{Hc'}{$\arr_c'$}

1571: \includegraphics{htv3.eps}

1572: \end{center}

1573: \end{figure}

1574: Again by Theorem \ref{surj}, we have

1575: $$H^*_{\Zt}(\mathcal{M}(\arr_a');\Zt)\cong

1576: \Zt [\vec{u},x]\bigg/

1577: \< \begin{array}{c}

1578: u_1(x-u_1), u_2(x-u_2), u_3(x-u_3), u_4(x-u_4),\\ u_5(x-u_5), 

1579: u_2u_3, (x-u_1)u_5, u_1(x-u_2)u_4,\\ u_1u_3u_4, (x-u_2)u_4u_5, u_3u_4u_5 

1580: \end{array}\>$$

1581: and

1582: $$H^*_{\Zt}(\mathcal{M}(\arr_c');\Zt)\cong

1583: \Zt [\vec{u},x]\bigg/

1584: \< \begin{array}{c}

1585: u_1(x-u_1), u_2(x-u_2), u_3(x-u_3), u_4(x-u_4),\\ u_5(x-u_5),

1586: u_2u_3, (x-u_1)u_5, (x-u_1)u_2(x-u_4),\\ u_1u_3u_4, (x-u_2)u_4u_5, u_3u_4u_5 

1587: \end{array}\>.$$

1588: We have used Macaulay 2 to check that the annihilator of the element 

1589: $u_2\in H^*_{\Zt}(\mathcal{M}(\arr_a');\Zt)$ is generated by two linear elements

1590: (namely $u_3$ and $x-u_2$) and nothing else, while there is no element

1591: of $H^*_{\Zt}(\mathcal{M}(\arr_c');\Zt)$ with this property.

1592: Hence the two rings are not isomorphic.

1593: \end{example}

1594: \end{section}

1595: 

1596: \footnotesize{

1597: \begin{thebibliography}{10}

1598: 

1599: \bibitem[BD]{BD}

1600: R.~Bielawski and A.~Dancer.

1601: \newblock The geometry and topology of toric hyperk\"ahler manifolds.

1602: \newblock {\em Comm. Anal. Geom.} 8 (2000), 727--760.

1603: 

1604: \bibitem[BGH]{BGH}

1605: D.~Biss, T.~Holm, and V.~Guillemin.

1606: \newblock The mod 2 cohomology of fixed point sets

1607: of anti-symplectic involutions.

1608: \newblock math.SG/0107151.

1609: 

1610: \bibitem[Bo]{Bo}

1611: A.~Borel.

1612: \newblock {\em Seminar on Transformation Groups}, chapter XII.

1613: \newblock Annals of Mathematical Studies 46,

1614: \newblock Princeton, 1960.

1615: 

1616: \bibitem[BJ]{BJ}

1617: T.~Brocker and K.~Janich.

1618: \newblock {\em Introduction to Differential Topology}.

1619: \newblock Cambridge University Press, 1982.

1620: 

1621: % \bibitem[Br]{Br}

1622: % R.~Bryant.

1623: % \newblock An introduction to Lie groups and symplectic geometry.

1624: % \newblock {\em Geometry and quantum field theory (Park City, UT, 1991)},

1625: % 5--181.

1626: % 

1627: \bibitem[HH]{HH}

1628: M.~Harada and T.~Holm.

1629: \newblock The equivariant cohomology of hypertoric varieties and their real loci.

1630: \newblock In preparation.

1631: 

1632: \bibitem[HS]{HS}

1633: T.~Hausel and B.~Sturmfels.

1634: \newblock Toric hyperk\"ahler varieties.

1635: \newblock math.AG/0203096.

1636: 

1637: \bibitem[HL]{HL}

1638: P.~Heinzner and F.~Loose.

1639: \newblock Reduction of complex Hamiltonian $G$-spaces. 

1640: \newblock {\em Geom. Funct. Anal.} 4 (1994), no. 3, 288--297.

1641: 

1642: % \bibitem[Hi]{Hi}

1643: % N.~Hitchin.

1644: % \newblock Hyper-K\"ahler manifolds. 

1645: % \newblock S\'eminaire Bourbaki, Vol. 1991/92. 

1646: % \newblock {\em Astérisque} no. 206 (1992), Exp. no. 748, 3, 137--166.

1647: % 

1648: \bibitem[HKLR]{HKLR}

1649: N.~Hitchin, A.~Karlhede, U.~Lindstr\"om, and M.~Ro\v{c}ek.

1650: \newblock Hyper-K\"ahler metrics and supersymmetry.

1651: \newblock {\em Comm. Math. Phys.} 108 (1987)

1652: no. 4, 535--589.

1653: 

1654: \bibitem[Ki]{Ki}

1655: F.~Kirwan.

1656: \newblock {\em Cohomology of quotients in symplectic and algebraic geometry}.

1657: \newblock Mathematical Notes 31,

1658: \newblock Princeton University Press, 1984.

1659: 

1660: \bibitem[K1]{K1}

1661: H.~Konno.

1662: \newblock Cohomology rings of toric hyperk\"ahler manifolds.

1663: \newblock {\em Int. J. of Math.} 11 (2000), 

1664: no. 8, 1001--1026.

1665: 

1666: \bibitem[K2]{K2}

1667: H.~Konno.

1668: \newblock Equivariant cohomology rings of toric hyperk\"ahler manifolds.

1669: \newblock {\em Quaternionic structures in mathematics and physics (Rome, 1999)},

1670: 231--240 (electronic).

1671: 

1672: % \bibitem[MFK]{MFK}

1673: % D.~Mumford, J.~Fogarty, and F.~Kirwan.

1674: % \newblock {\em Geometric Invariant Theory}.

1675: % \newblock Third edition. Ergebnisse der Mathematik und ihrer 

1676: % Grenzgebiete (2) [Results in Mathematics and Related Areas (2)], 34.

1677: % \newblock Springer-Verlag, 1994.

1678: % 

1679: \bibitem[Na]{Na}

1680: H.~Nakajima.

1681: \newblock Instantons on ALE spaces, quiver varieties, and Kac-Moody algebras.

1682: \newblock {\em Duke Math. J.} 76 (1994) no. 2, 365--416.

1683: 

1684: \bibitem[OS]{OS}

1685: P.~Orlik and L.~Solomon.

1686: \newblock Combinatorics and topology of complements of hyperplanes.

1687: \newblock {\em Invent. Math.} 56 (1980), 167--189.

1688: 

1689: \bibitem[OT]{OT}

1690: P.~Orlik and H.~Terao.

1691: \newblock {\em Arrangements of Hyperplanes}.

1692: \newblock Grundlehren der mathematischen Wissenschaften 300,

1693: \newblock Springer-Verlag, 1992.

1694: 

1695: \bibitem[Pr]{Pr}

1696: N.~Proudfoot.

1697: \newblock {\em The equivariant Orlik-Solomon algebra}.

1698: \newblock http://math.berkeley.edu/$\sim$proudf/os.ps

1699: 

1700: \bibitem[Sc]{Sc}

1701: C.~Schmid.

1702: \newblock {\em Cohomologie \'equivariante des certaines vari\'et\'es

1703: hamiltoniennes et de leur partie r\'eelle.}

1704: \newblock Th\`ese \`a Universit\'e de Gen\`eve.

1705: \newblock Available at http://www.unige.ch/math/biblio/these/theses.html

1706: 

1707: \bibitem[Sj]{Sj}

1708: R.~Sjamaar.

1709: \newblock Holomorphic slices, symplecitic reduction, and multiplicities of representations.

1710: \newblock {\em Ann. Math.} 141 (1995), 87--129.

1711: 

1712: \bibitem[TW]{TW}

1713: S.~Tolman and J.~Weitsman.

1714: \newblock On the cohomology rings of Hamiltonian T-spaces.

1715: \newblock {\em Proc. of the Northern California Symplectic Geometry Seminar},

1716: Y. Eliashberg {\em et al.} eds. AMS Translations Series 2, vol. 196: 

1717: Advances in Mathematical Sciences 45 (1999), 251--258. 

1718: \end{thebibliography}

1719: }

1720: 

1721: \end{spacing}

1722: \end{document}

1723: