1: \documentclass[debug,overfull,info]{epl}

2: \usepackage{amsmath}

3: \usepackage{epsf}

4: \usepackage{epsf}

5: \title{

6: Universal crossing probability in anisotropic systems

7: }

8: \author{L. Turban}

9: \institute{ 

10:   Laboratoire de Physique des Mat\'eriaux, UMR CNRS 7556, 

11:   Universit\'e Henri Poincar\'e (Nancy 1) 

12:   - BP 239, 54506 Vand\oe uvre l\`es Nancy Cedex, France

13: }

14: 

15: \pacs{64.60.Ak}{Renormalization-group, fractal, 

16: and percolation studies of phase transitions}

17: \pacs{05.50.+q}{Lattice theory and statistics (Ising, Potts, etc.)}

18: \pacs{02.50.-r}{Probability theory, stochastic processes, and 

19: statistics}

20: 

21: \begin{document}

22: 

23: \maketitle

24: 

25: \begin{abstract}

26: Scale-invariant universal crossing probabilities are studied for critical 

27: anisotropic systems in two dimensions. For weakly anisotropic standard 

28: percolation in a rectangular-shaped system, Cardy's exact formula is 

29: generalized using a length-rescaling procedure. For strongly anisotropic 

30: systems in $1+1$ dimensions, exact results are obtained for the random walk 

31: with absorbing boundary conditions, which can be considered as a linearized 

32: mean-field approximation for directed percolation. The bond and site 

33: directed percolation problem is itself studied numerically via Monte Carlo 

34: simulations on the diagonal square lattice with either free or periodic 

35: boundary conditions. A scale-invariant critical crossing probability is 

36: still obtained, which is a universal function of the effective aspect ratio 

37: $r_\ab{eff}=c\, r$ where $r=L/t^z$, $z$ is the dynamical exponent and $c$ is 

38: a non-universal amplitude. 

39: \end{abstract}

40: 

41: \section{Introduction}

42: The study of {\it universal crossing probabilities} in the standard 

43: random percolation problem at criticality has been the subject of 

44: much 

45: interest over the last decade~\cite{langlands92,cardy92,cross-perco,%

46: rigorous,span-clus-mc,span-clus-ex}

47: (see ref.~\cite{cardy01} for a review).

48: At the percolation threshold, $p_\ab{c}$, on a rectangular lattice 

49: with length $L_\parallel$ and width $L_\perp$, the probability to have 

50: at least one cluster crossing the system in one of the two directions 

51: (or in both) is a {\it scale-invariant} quantity in the {\it continuum 

52: limit} where the lattice spacing $a\to 0$. Actually, the crossing 

53: probability is a function $\pi(r)$ of the aspect ratio, 

54: $r=L_\perp/L_\parallel$, and depends also on the boundary conditions. It 

55: remains universal in the sense that it does not depend on the lattice 

56: type or whether one is considering site, bond or site-bond percolation. 

57: 

58: Following the pioneering numerical work of Langlands 

59: \etal~\cite{langlands92}, Cardy~\cite{cardy92} obtained an analytical 

60: expression for the crossing probability between opposite sides of a 

61: rectangle. He combined the correspondence between percolation and the 

62: limit $q\to1$ of the $q$-state Potts model~\cite{kasteleyn69} with 

63: boundary conformal field theory techniques to calculate the crossing 

64: probability between two segments on the boundary of a half-plane at 

65: criticality. A conformal mapping could then be used to transform the 

66: half-plane result into the rectangular geometry. Cardy's and another 

67: crossing probability formula have been recently proven 

68: rigorously~\cite{rigorous}.

69: Monte Carlo~\cite{span-clus-mc} and exact results~\cite{span-clus-ex} 

70: have been obtained for the probability to find $n$ incipient spanning 

71: clusters connecting two disjoint segments at the boundary of a finite 

72: system. More recently, crossing probabilities on same-spin Ising 

73: clusters in two dimensions have been also 

74: considered~\cite{cross-ising}. 

75: 

76: Let us notice that the existence of a {\it scale-invariant} crossing probability 

77: is not a trivial result. In the case of ordinary percolation, for example, it is 

78: linked to the vanishing, in the limit $q\to1$, of the scaling dimension $x(q)$ 

79: of a boundary condition changing operator of the Potts model~\cite{cardy92}.

80: 

81: In this letter, we investigate the behaviour of crossing 

82: probabilities in anisotropic critical systems. We first give an exact treatment 

83: of weakly anisotropic standard percolation. We show how a length rescaling 

84: technique, which has been used in other contexts~\cite{barber84}, allows to 

85: generalize Cardy's exact results in this situation.

86: Then we turn to crossing probabilities in strongly anisotropic systems. To our 

87: knowledge, such problems have not been considered before, at least with the 

88: appropriate anisotropic scaling analysis. We start with the random walk with 

89: absorbing boundary conditions in the continuum limit which can be considered as 

90: a linearized mean-field approximation~\cite{kaiser95} for directed 

91: percolation~\cite{broadbent57,kinzel83,hinrichsen00} and illustrates on 

92: an exactly solvable example the behaviour of the crossing probability in 

93: strongly anisotropic systems. Finally we present the results of Monte Carlo 

94: simulations for the crossing probability in directed percolation, a 

95: strongly anisotropic problem which has so far defied all attempts

96: at an exact solution, even in two dimensions. 

97: 

98: \section{Weakly anisotropic percolation}

99: We consider the random bond percolation on a square lattice with 

100: anisotropic bond occupation probabilities. The system is rectangular and 

101: the bond probability is $p_\parallel$ ($p_\perp$) in the direction 

102: parallel to the side with length $L_\parallel$ ($L_\perp$). 

103: The correlation lengths, $\xi_\parallel=\hat{\xi}_\parallel t^{-\nu}$, 

104: $\xi_\perp=\hat{\xi}_\perp t^{-\nu}$, diverge along the critical 

105: line where $t=0$ with the same exponent $\nu$ for both directions but 

106: different amplitudes $\hat{\xi}_\parallel$ and $\hat{\xi}_\perp$. The 

107: isotropy can be restored~\cite{barber84} by rescaling the lattice 

108: parameter anisotropically so that 

109: $\xi_\parallel a_\parallel=\xi_\perp a_\perp$. The aspect ratio $r$ 

110: is then changed into an effective one 

111: \begin{equation}

112: r_\ab{eff}=\frac{L_\perp a_\perp}{L_\parallel a_\parallel}

113: =\frac{L_\perp/\xi_\perp}{L_\parallel/\xi_\parallel}

114: =\frac{\hat{\xi}_\parallel}{\hat{\xi}_\perp}r\,.

115: \label{e-reff}

116: \end{equation}

117: In the second expression the effective aspect ratio involves the 

118: lengths measured in units of the corresponding correlation lengths.

119: 

120: The critical amplitude ratio 

121: $c=\hat{\xi}_\parallel/\hat{\xi}_\perp$, which is non-universal, 

122: is known for the $q$-state Potts model on the square 

123: lattice~\cite{kim87}. In the percolation limit, $q\to1$, one obtains:

124: \begin{equation}

125: c=\tan\frac{3 u}{2}\,,\qquad \frac{\sin u}{\sin(\pi/3-u)}

126: =\frac{p_{\ab{c}\parallel}}{1-p_{\ab{c}\parallel}}

127: =\frac{1-p_{\ab{c}\perp}}{p_{\ab{c}\perp}}\,,

128: \label{e-rescal}

129: \end{equation}

130: which solves the problem for a weak anisotropy. When the anisotropy 

131: axis is not parallel to one edge of the rectangle, the rescaling 

132: leads to an isotropic system with the shape of a parallelogram.

133: 

134: \section{Random walk}

135: Standard conformal methods do not apply in strongly anisotropic systems.

136: The simplest example of such a system is the random 

137: walk in $1+1$~dimensions, which may be considered as a directed walk 

138: in space-time. In the continuum limit, the probability density 

139: $P(x,t)$ 

140: satisfies the diffusion equation

141: \begin{equation}

142: \frac{\partial P}{\partial t}=D\frac{\partial^2 P}{\partial x^2}\,.  

143: \label{e-dif}

144: \end{equation}

145: We assume a uniform initial probability density $P(x,0)=1/L$ and 

146: absorbing boundary conditions at~$x=0$ and~$L$. The solution is then 

147: obtained as the following eigenvalue expansion:

148: \begin{equation}

149: P(x,t)=\sum_{p=0}^\infty\frac{4}{(2p+1)\pi L}

150: \sin\left[(2p+1)\frac{\pi x}{L}\right]

151: \exp\left[-\frac{(2p+1)^2\pi^2Dt}{L^2}\right]\,.

152: \label{e-P}

153: \end{equation}

154: On the rectangle $L\times t$, the crossing probability in the time 

155: direction is 

156: given by:

157: \begin{equation}

158: \pi^\ab{rw}_t=\int_0^LP(x,t)\,\upd x

159: =\sum_{p=0}^\infty\frac{8}{(2p+1)^2\pi^2}

160: \exp\left[-\frac{(2p+1)^2\pi^2Dt}{L^2}\right]\,.

161: \label{e-cross-rw}

162: \end{equation}

163: Thus the crossing probability shown in fig.~\ref{f.1} is a universal 

164: function of the scaled variable $cL^z/t$, product of the aspect ratio 

165: $r=L^z/t$, where $z=2$ is the dynamical 

166: exponent for the random walk, by the non-universal inverse diffusion 

167: constant $c=D^{-1}$. 

168: 

169: In strongly anisotropic systems, such a behaviour generally follows 

170: from anisotropic scaling~\cite{binder89}. Under the anisotropic scale 

171: transformation, $L'=L/b$, $t'=t/b^z$, the crossing probability 

172: behaves as:

173: \begin{equation}

174: \pi_t(L,t)=b^{-x_{\pi_t}}\,\pi_t\left(\frac{L}{b},\frac{t}{b^z}\right

175: )

176: \,.

177: \label{e-scaling}

178: \end{equation}

179: For a dimensionless crossing probability, $x_{\pi_t}=0$, one obtains  

180: the scale-invariant behaviour mentioned above with $b=L$.

181: 

182: \begin{figure}

183: \onefigure[width=7cm]{fig1-dp.eps}

184: \caption{Crossing probability in the time direction for a random walk 

185: with  

186: absorbing boundary conditions at $l=0$ and $L$ and a uniform initial 

187: probability density at $t=0$. $r=L^2/t$ is the scale-invariant aspect 

188: ratio 

189: of the system and $D=1$.}

190: \label{f.1}

191: \end{figure}

192: 

193: 

194: The probability density for a walker starting at $t=0$ with uniform 

195: probability on the segment $[0,L]$ to die at time $t$ is given by

196: \begin{equation}

197: -\frac{\upd \pi^\ab{rw}_t}{\upd t}

198: =\frac{8D}{L^2}\sum_{p=0}^\infty 

199: \exp\left[-\frac{(2p+1)^2\pi^2Dt}{L^2}\right]

200: =\frac{4D}{L^2}\,\vartheta_2\left[

201: \exp\left(-\frac{4\pi^2Dt}{L^2}\right)\right]\,,

202: \label{e-surv-rw}

203: \end{equation}

204: where $\vartheta_2(u)=\sum_{p=-\infty}^{+\infty}u^{(p+1/2)^2}$ is a 

205: Jacobi theta function.

206: 

207: Since the linearized mean-field approximation for directed 

208: percolation~\cite{kinzel83,kaiser95} is also

209: governed by the diffusion equation~(\ref{e-dif}),

210: the mean-field crossing probability for directed percolation is given 

211: by~(\ref{e-cross-rw}) when the aspect ratio is sufficiently small.

212: 

213: \section{Directed percolation}

214: The critical crossing probability in the time direction, 

215: $\pi^\ab{dp}_t$, 

216: has been studied via Monte Carlo simulations for bond and site 

217: directed percolation on the square lattice, with the time axis along the 

218: diagonal direction. Space and time coordinates take integer and 

219: half-integer values alternatively on successive spatial rows and 

220: $t\geq0$. There are two directed bonds leaving a site at $(x,t)$ and 

221: terminating on the sites at $(x\pm1/2,t+1/2)$. 

222: 

223: Sites may be wet or dry and a directed percolation cluster is a 

224: collection of connected wet sites starting from some source at 

225: $t=0$~\cite{broadbent57,kinzel83,hinrichsen00}. In the bond problem, bonds 

226: are open with probability $p$. A site is wet at time $t+1/2$ when it is 

227: connected through an open bond to a site which was wet at time $t$. In the 

228: site problem, sites are occupied with probability $p$ and all the bonds 

229: are open. A site is wet at time $t+1/2$ when it is occupied and connected 

230: to a site which was wet at time $t$. A {\it scale-invariant crossing 

231: probability} is obtained when all the sites are wet in the initial 

232: state at $t=0$.

233: 

234: \begin{figure}

235: \twoimages[width=7cm]{fig2a-dp.eps}{fig2b-dp.eps}

236: \caption{(a) Crossing probability $\pi^\ab{dp}_t$ in the time 

237: direction 

238: as a function of the logarithm of the aspect ratio $r=L^z/t$ for site 

239: and 

240: bond directed percolation on the diagonal square lattice with either 

241: periodic or free boundary conditions in the space direction. The $L$ 

242: sites are wet at $t=0$. (b) Shifting the curves for the site problems 

243: by 

244: the same amount, the data collapse on a single scale-invariant 

245: universal 

246: curve for each type of boundary conditions.}

247: \label{f.2ab}

248: \end{figure}

249: 

250: The critical crossing probability was calculated for $10^6$ samples 

251: with 

252: sizes $L=2^3$ to $2^8$ and different values of the aspect ratio, 

253: $r=L^z/t$, ranging from $r\simeq 2^{-5}$ to $2^5$. The percolation 

254: thresholds $p_\ab{c}^\ab{bond}=0.644700185(5)$, 

255: $p_\ab{c}^\ab{site}=0.70548522(4)$ and the dynamical exponent 

256: $z=1.580745(10)$ were taken from ref.~\cite{jansen99}. Either 

257: periodic 

258: boundary conditions (PBC) or free boundary conditions (FBC) were used 

259: at 

260: $l=1$ and $L$. The samples were grouped into 20 packets in order to 

261: evaluate the errors on $\pi^\ab{dp}_t$. The raw data $\pi^\ab{dp}_t$ 

262: versus $\ln r$ are shown in fig.~2a for the largest sizes. The 

263: statistical 

264: errors are actually smaller than the size of the symbols. The scale 

265: invariance is manifest and finite size corrections to the leading 

266: behaviour are quite weak. There is a shift in $\ln r$ between the 

267: curves 

268: for site and bond percolation which is the same for PBC and FBC. As 

269: shown 

270: in fig.~2b, an universal curve is obtained for each type of boundary 

271: condition after the curves corresponding to the site problem have 

272: been 

273: shifted by the same amount.

274: 

275: In this strongly anisotropic problem, the correlation lengths, 

276: $\xi_\perp$ in the space direction and $\xi_\parallel$ in the time 

277: direction, diverge as 

278: \begin{equation}

279: \xi_\perp=\hat{\xi}_\perp \vert p-p_\ab{c}\vert^{-\nu}\;,

280: \qquad \xi_\parallel=\hat{\xi}_\parallel \vert 

281: p-p_\ab{c}\vert^{-z\nu}\;.

282: \label{e-xi}

283: \end{equation}

284: Thus, generalizing the behaviour for the weakly anisotropic system in 

285: eq.~(\ref{e-reff}), for each type of boundary conditions, $i=FBC$ or 

286: $PBC$, one expects the crossing probability $\pi^\ab{dp}_t$ to be an 

287: universal function $f_i(r_\ab{eff})$ of the scale-invariant ratio

288: \begin{equation}

289: r_\ab{eff}=\frac{(L/\xi_\perp)^z}{t/\xi_\parallel}=c\, r\;,

290: \qquad c=\frac{\hat{\xi}_\parallel}{(\hat{\xi}_\perp)^z}\;,

291: \label{e-rlt}

292: \end{equation}

293: in agreement with anisotropic scaling in eq.~(\ref{e-scaling}).

294: As above for weakly anisotropic percolation and for the random-walk,   

295: the constant $c$ is non-universal and takes different values, 

296: $c_\ab{s}$ 

297: and $c_\ab{b}$, for the site and bond directed percolation on the 

298: diagonal square lattice. 

299: 

300: For a given type of boundary conditions, the crossing probability 

301: $\pi^\ab{dp}_t$ is the same for couples of values of the aspect ratio 

302: $r_\ab{s}$ and $r_\ab{b}$ such that the arguments of the universal 

303: function $f_i(r_\ab{eff})$ are the same for the site and bond 

304: problems. 

305: This happens when $c_\ab{s}r_\ab{s}=c_\ab{b}r_\ab{b}$. Thus, with a 

306: logarithmic scale, the two curves have a relative shift  

307: \begin{equation}

308: \ln r_\ab{s}-\ln 

309: r_\ab{b}=\ln\left(\frac{c_\ab{b}}{c_\ab{s}}\right)\;.

310: \label{e-shift}

311: \end{equation}

312: 

313: \begin{figure}

314: \onefigure[width=7cm]{fig3-dp.eps}

315: \caption{Non-universal amplitude ratio $c_\ab{b}/c_\ab{s}$ as defined 

316: in 

317: eq.~(\protect\ref{e-rlt}) as a function of the inverse of the system 

318: size 

319: $L$. The dotted line corresponds to a cubic (quadratic) fit of the 

320: data 

321: for periodic (free) boundary conditions.}

322: \label{f.3}

323: \end{figure}

324: 

325: This shift was estimated for the different sizes for 6 points in the 

326: ascending part of the the curves. Since the points for the site and 

327: bond 

328: problems correspond to slightly different ordinates, a direct 

329: estimation 

330: of the shift was not possible. Instead we used a polynomial 

331: interpolation 

332: for the curve corresponding to the bond problem and the value of the 

333: abscissa associated with the common value of the ordinate was 

334: determined through the tangent method. The results for the ratio of 

335: non-universal amplitudes as a function of the inverse size are shown 

336: in 

337: fig.~\ref{f.3} for PBC and FBC. They converge from above and from 

338: below 

339: to the common asymptotic value

340: \begin{equation}

341: \frac{c_\ab{b}}{c_\ab{s}}=1.281(1)\;,

342: \label{e-amprat}

343: \end{equation}

344: corresponding to the shift $\ln r_\ab{s}-\ln r_\ab{b}=0.2476(8)$ 

345: which was 

346: used to obtain the data collapse in fig.~2b.

347: \section{Final remarks}

348: The same universal crossing probability is obtained for any finite 

349: homogeneous density of wet sites in the initial state at $t=0$ 

350: although the 

351: finite-size corrections to the leading scale-invariant behaviour is 

352: stronger at large values of $r$ (small $t$) when the initial density 

353: decreases.

354: With a single wet site at $l=L/2$ in the initial state, i.e., for an 

355: asymptotically vanishing initial density, the crossing probability 

356: remains a function of $r_\ab{eff}$ but the scale-invariance is lost, 

357: $\pi^\ab{dp}_t$ decaying as $L^{-\beta/\nu}$. The probability 

358: $\pi^\ab{dp}_t(n)$ to find $n$ clusters crossing the system in the time 

359: direction at $p_\ab{c}$ has been also studied. Details will be given elsewhere.

360: 

361: This work can be extended by considering the crossing probability on same-spin 

362: clusters in strongly anisotropic spin systems, like the three-dimensional 

363: uniaxial ANNNI model at the Lifshitz point, for which presumably exact results 

364: for the form of the correlation functions, following from a generalization of 

365: local scale invariance~\cite{henkel02}, have been obtained 

366: recently~\cite{pleimling01}.

367:  

368: \acknowledgments

369: The author wishes to thank {\scshape Lev Shchur} for a stimulating 

370: introductory talk about spanning clusters.

371: 

372: \begin{thebibliography}{0}

373: \bibitem{langlands92}

374:   \Name{Langlands R. P., Pichet C., Pouliot P. \and Saint-Aubin Y.}

375:   \REVIEW{J. Stat. Phys.}{67}{1992}{553}.

376: 

377: \bibitem{cardy92}

378:   \Name{Cardy J. L.}

379:   \REVIEW{J. Phys. A}{25}{1992}{L201}.

380: 

381: \bibitem{cross-perco}

382:   \Name{Langlands R. P., Pouliot P. \and Saint-Aubin Y.}

383:   \REVIEW{Bull. AMS}{30}{1994}{1};

384:   \Name{Pinson T. H.}

385:   \REVIEW{J. Stat. Phys.}{75}{1994}{1167};

386:   \Name{Saint-Aubin Y.}

387:   \REVIEW{Physica A}{221}{1995}{41};

388:   \Name{Ziff R. M.}

389:   \REVIEW{J. Phys. A}{28}{1995}{1249};

390:   \Name{Ziff R. M.}

391:   \REVIEW{J. Phys. A}{28}{1995}{6479};

392:   \Name{Watts G. M. T.}

393:   \REVIEW{J. Phys. A}{29}{1996}{L363};

394:   \Name{Hu C.-K. \and Lin C.-Y.}

395:   \REVIEW{Phys. Rev. Lett.}{77}{1996}{8};

396:   \Name{Hovi J.-P. \and Aharony A.}

397:   \REVIEW{Phys. Rev. E}{53}{1996}{235};

398:   \Name{Ziff R. M., Lorenz C. D. \and Kleban P.}

399:   \REVIEW{Physica A}{266}{1999}{17};

400:   \Name{Kleban P.}

401:   \REVIEW{Physica A}{281}{2000}{242};

402: 

403: \bibitem{rigorous}

404:   \Name{Smirnov S.}

405:   \REVIEW{C. R. Acad. Sci. Paris, S\'erie I}{333}{2001}{239};

406:   \Name{Schramm O.}

407:   \REVIEW{Elect. Comm. Probab.}{6}{2001}{115}.

408:   

409: \bibitem{span-clus-mc}

410:   \Name{Sen P.}

411:   \REVIEW{Int. J. Mod. Phys. C}{7}{1996}{603};

412:   \REVIEW{ibid}{8}{1997}{229};

413:   \Name{Shchur L. N. \and Kosyakov S. S.}

414:   \REVIEW{Int. J. Mod. Phys. C}{8}{1997}{473};

415:   \REVIEW{Nucl. Phys. B (Proc. Suppl.)}{63A-C}{1998}{664};

416:   \Name{Grassberger P. \and Nadler W.} cond-mat/0010265;

417:   \Name{Grassberger P.} cond-mat/0201313.  

418: 

419: \bibitem{span-clus-ex}

420:   \Name{Aizenman M.}

421:   \REVIEW{Nucl. Phys. B}{485}{1997}{551};

422:   \Name{Aizenman M., Duplantier B. \and Aharony A.}

423:   \REVIEW{Phys. Rev. Lett.}{83}{1999}{1359};

424:   \Name{Cardy J. L.}

425:   \REVIEW{J. Phys. A}{31}{1998}{L105}.

426: 

427: \bibitem{cardy01}

428:   \Name{Cardy J. L.}

429:   {\it Lectures on Conformal Invariance and Percolation}, 

430:   Bunkyo-ku, Tokyo (2001), math-ph/0103018.

431: 

432: \bibitem{kasteleyn69}

433:   \Name{Kasteleyn P. W. \and Fortuin C. M.}

434:   \REVIEW{J. Phys. Soc. Japan (Suppl.)}{26}{1969}{11}.

435: 

436: \bibitem{cross-ising}

437:   \Name{Langlands R. P., Lewis M.-A. \and Saint-Aubin Y.}

438:   \REVIEW{J. Stat. Phys.}{98}{2000}{131};

439:   \Name{Lapalme E. \and Saint-Aubin Y.}

440:   \REVIEW{J. Phys. A}{34}{2001}{1825};

441:   \Name{Arguin L.-P. \and Saint-Aubin Y.} hep-th/0109138.

442: 

443: \bibitem{barber84}

444:   \Name{Barber M. N., Peschel I. \and Pearce P. A.}

445:   \REVIEW{J. Stat. Phys.}{37}{1984}{497}.

446: 

447: \bibitem{kaiser95}

448:   \Name{Kaiser C. \and Turban L.}

449:   \REVIEW{J. Phys. A}{28}{1995}{351}.

450: 

451: \bibitem{broadbent57}

452:   \Name{Broadbent S. R. \and Hammersley J. M.}

453:   \REVIEW{Proc. Camb. Philos. Soc.}{53}{1957}{629}.

454: 

455: \bibitem{kinzel83}

456:   \Name{Kinzel W.}

457:   \Book{Percolation Structures and Processes, Annals of the Israel

458:   Physical Society}

459:   \Editor{G. Deutscher, R. Zallen \and J. Adler}

460:   \Vol{5}

461:   \Publ{Israel Physical Society, Haifa}

462:   \Year{1983}

463:   \Pages{425}{445}.

464: 

465: \bibitem{hinrichsen00}

466:   \Name{Hinrichsen H.}

467:   \REVIEW{Adv. Phys.}{49}{2000}{815}.

468: 

469: \bibitem{kim87}

470:   \Name{Kim D. \and Pearce P. A.}

471:   \REVIEW{J. Phys. A}{20}{1987}{L451}.

472: 

473: \bibitem{binder89}

474:   \Name{Binder K. \and Wang J.-S.}

475:   \REVIEW{J. Stat. Phys.}{55}{1989}{87}.

476:   See also \Name{Hucht A.} cond-mat/0202142.

477: 

478: \bibitem{jansen99}

479:   \Name{Jensen I.}

480:   \REVIEW{J. Phys. A}{32}{1999}{5233}.

481: 

482: \bibitem{henkel02}

483: \Name{Henkel M.} hep-th/0205256.

484: 

485: \bibitem{pleimling01}

486:   \Name{Pleimling M. \and Henkel M.}

487:   \REVIEW{Phys. Rev. Lett.}{87}{2001}{125702}.

488: 

489: \end{thebibliography}

490: 

491: \end{document}

492: