1: \documentclass[aps,showpacs,twocolumn]{revtex4}

2: \usepackage{graphicx}

3: \def\phii{$\Phi(z)$ }

4: \def\s{{}}

5: \def\c{\cite{}}

6: \def\pt{$p_T$ }

7: 

8: \begin{document}

9: \title{Fractional Energy Loss and Centrality Scaling}

10:  \author{Rudolph C. Hwa$^1$ and  C.\ B.\ Yang$^{2}$}

11: \affiliation{$^1$Institute of Theoretical Science and Department of Physics\\

12: University of Oregon, Eugene, OR 97403-5203, USA\\

13: $^2$Institute of Particle Physics, Hua-Zhong Normal University,

14: Wuhan 430079, P.\ R.\ China}

15: \begin{abstract}

16: The phenomenon of centrality scaling in the high-\pt spectra of $\pi^0$

17: produced in Au-Au collisions at $\sqrt s=200$ GeV is examined in

18: the framework of  relating fractional energy loss to fractional

19: centrality increase. A new scaling behavior is found where the scaling

20: variable is given a power-law dependence on $N_{\rm part}$. The exponent

21: $\gamma$ specifies the fractional proportionality relationship between

22: energy loss and centrality, and is a phenomenologically determined

23: number that characterizes the nuclear suppression effect. The

24: implication on the parton energy loss in the context of recombination is

25: discussed.

26: 

27: \pacs{25.75.Dw}

28: \end{abstract}

29: 

30: \maketitle

31: 

32: The production of hadrons in Au-Au collisions at the

33: Relativistic Heavy-Ion Collider is found at high $p_T$

34: to depend sensitively on centrality \cite{de}.  At $\sqrt s=200$ GeV

35: the scaled inclusive cross section of pions at $p_T\approx$ 3-4 GeV/c

36: decreases by a factor of 4-5 when the centrality is varied from the most

37: peripheral to the most central \cite{ph,st}.  In a previous paper

38: \cite{hy} we reported the finding of a universal function \phii that

39: can describe all the inclusive cross sections at all

40: centralities. Such a scaling behavior is achieved by use of a

41: scaling variable $z$ that combines \pt with $N_{\rm part}$, the

42: number of participants.  In this paper we investigate the

43: origin of that scaling.  In particular, we consider the nature of

44: energy loss that can give rise to such a behavior.

45: 

46: Although the scaling behavior can be extended to include

47: energy dependence also \cite{hy2}, we restrict our

48: consideration here to centrality dependence only, and

49: emphasize the phenomenological implication of the data at

50: $\sqrt{s} = 200$ GeV, measured by PHENIX \cite{ph}.  We

51: shall not discuss energy loss at the parton level except near

52: the end, since perturbative QCD is not reliable for $p_T < 6$

53: GeV/c. Indeed, soft partons have been found to be important in

54: the hadronization process through recombination

55: \cite{hy3,fr,gr,li}.  We shall stay mainly at the hadronic

56: level that is phenomenological and uncontroversial, and

57: consider the nuclear suppression effect on the observed

58: hadrons.

59: 

60: Let us recall the scaling behavior found in \cite{hy}.  With

61: the definition of the variable

62: \begin{eqnarray}

63: z = p_T/K(N)

64: \label{1}

65: \end{eqnarray}

66: we find that the function $\Phi(z)$,

67: \begin{eqnarray}

68: \Phi(z)  = A(N) K^2(N) {1 \over 2\pi p_T} {dN_{\pi} \over

69: d\eta dp_T } \ ,

70: \label{2}

71: \end{eqnarray}

72: exhibits scaling behavior.  We have used the notation $N =

73: N_{\rm part}$, for brevity, and

74: \begin{eqnarray}

75: K(N) = 1.226 - 6.36 \times 10^{-4} N\ ,

76: \label{3}

77: \\

78: A(N) = 530 N_c(N)^{-0.9}, \qquad N_c(N)=0.44N^{1.33} \ ,

79: \label{4}

80: \end{eqnarray}

81: where both $K(N)$ and $A(N)$ are normalized to 1 at $N =

82: 350$.  At all $p_T$ and $N$, the data of $dN_{\pi} / p_T

83: d\eta dp_T$ at midrapidity collapse to one universal curve $\Phi(z)$,

84: which is parametrized in \cite{hy} by

85: \begin{equation}

86: \Phi(z)=1200\ (z^2+2)^{-4.8}\ (1+25 e^{-4.5 z}) \ .  \label{4.5}

87: \end{equation}

88: 

89: 

90: It is clear that what gives rise to the scaling behavior must be related

91: to an universal property in the medium effect on the production of

92: pions.  Since it is not possible to determine experimentally  the

93: degradation of parton momentum as the medium size is increased, and

94: since whether hadronization is by means of fragmentation or

95: recombination is still controversial, we choose to stay at

96: the hadronic level and examine energy loss.  Since the produced pions

97: do not themselves traverse the dense medium, energy loss here does not

98: refer to the evolutionary process of a pion, as one can for a parton.

99: Instead, it refers to the shift of the pion distribution, as the medium

100: size quantified by $N$ is increased.

101: 

102: Let us now consider the implications of a scaling function

103: $\Phi(z)$.  Let $z$ be defined as

104: \begin{eqnarray}

105: z = x J(N) \ ,            \label{5}

106: \end{eqnarray}

107: where $x$ is a dimensionless momentum variable identified as

108: $x = p_T/p_0$ with $p_0$ chosen at $p_0 = 1$  GeV/c so that $x$ is

109: numerically the same as $p_T$.     We now examine the consequences of

110: writing \phii in terms of $x$ and $N$ explicitly

111: \begin{eqnarray}

112: \Phi(z) = F(x, N) \ .

113: \label{7}

114: \end{eqnarray}

115: 

116: In pQCD, such as in  Ref.\ \cite{ba}, one compares a distribution in

117: medium with one in vacuum in order to emphasize the medium

118: effect.  We prefer, however, to stay away from the $pp$ collision case,

119: since we want to

120: consider incremental changes of the medium size.  From our

121: perspective of dealing only with the observables, it is very natural to

122: ask the $\epsilon$-$\delta$ type question.  That is, given a medium that

123: is not too small, in which pions are produced, if its size is increased

124: by an $\epsilon$ amount, what is the corresponding downward shift

125: $\delta$ in momentum in order to maintain the same probability of

126: producing the pions?  In terms of $x$ and $N$, the proposition can

127: essentially be stated as

128: \begin{eqnarray}

129: F(x,N) = F ( x - \delta, N + \epsilon) \ .

130: \label{8}

131: \end{eqnarray}

132: The $+\epsilon$ and $- \delta $ relationship is a consequence

133: of the suppression effect.

134: 

135: For infinitesimal $\epsilon=\delta N$ and $\delta=\delta x$ we can

136: expand the right-hand side of Eq.\ (\ref{8}) and keep only the first

137: order terms, getting

138: \begin{eqnarray}

139: {\delta x \over \delta N}  = {\partial F / \partial N \over \partial

140: F / \partial x} \ .

141: \label{9}

142: \end{eqnarray}

143: If $F(x,N)$  satisfies Eqs.\ (\ref{5}) and (\ref{7}), then we

144: have

145: \begin{eqnarray}

146: {\delta x\over \delta N}  = {x \over J} {dJ \over dN} \ .

147: \label{10}

148: \end{eqnarray}

149: Since $\delta x$ is proportional to $x$, it is more sensible to consider

150: the fractional energy loss

151: \begin{eqnarray}

152: {\delta x \over x}  =  {d \ln J \over d \ln N} {\delta N \over

153: N} \ .

154: \label{11}

155: \end{eqnarray}

156: If the fractional energy loss is proportional to the fractional change of

157: centrality, a notion that seems extremely reasonable, i.e.,

158: \begin{eqnarray}

159: {\delta x\over x}  =  \gamma\ {\delta N \over N} \ ,

160: \label{12}

161: \end{eqnarray}

162: where $\gamma$ characterizes the suppression effect, then it is

163: necessary that

164: \begin{eqnarray}

165: J(N) = \left({N \over N_0

166:      } \right)^{\gamma}

167: \label{13}

168: \end{eqnarray}

169: for some normalization $N_0$.  Clearly, Eq.\ (\ref{12}) does not make

170: sense for $N$ very small, such as $N = 2$, since $\epsilon$ cannot be

171: made infinitesimal compared to 2.

172: 

173: The power-law behavior in Eq.\ (\ref{13}) is a necessary consequence of

174: scaling and fractional proportionality, Eq.\ (\ref{12}).

175: Comparing $J(N)$ with $1/K(N)$ in Eq.\ (\ref{3}), one finds that Eq.\

176: (\ref{13}) differs enough from our first scaling parametrization to cast

177: some doubt on whether Eq.\ (\ref{13}) is sufficient for scaling.

178: However, the fractional proportionality relationship is so compelling

179: that we have been motivated to reexamine the data, especially since the

180: original preliminary data have by now been finalized.  It should  be

181: noted that Eq.\ (\ref{13}) is obtained without relying on any specific

182: form for

183: $\Phi(z)$; it depends only on the structure of the scaling variable

184: expressed in Eq.\ (\ref{5}).  Thus one expects Eqs.\ (\ref{12}) and

185: (\ref{13}) to be very general properties of centrality scaling.

186: 

187: \begin{figure}[tbph]

188: \includegraphics[width=0.45\textwidth]{pi0scal.eps}

189: \caption{ Scaling distribution $\Phi(z)$ showing the coalescence of

190:   all data points of 9 centrality bins for

191: $\pi^0$  and the most central collisions for $\pi^+$ production in Au+Au

192: collisions at $\sqrt s=200$ GeV as measured by PHENIX

193: \cite{ht,tc}. The solid line is a fit parametrized by Eq.\ (\ref{15}).}

194: \end{figure}

195: 

196: Using the data on $\pi^0$ from PHENIX tabulated on the web \cite{ht}

197: we have assembled the spectra for all measured centralities, and made

198: appropriate horizontal and vertical shifts in the log-log plot to

199: obtain a universal behavior.  The result is shown in Fig.\ 1. The

200: $\pi^+$ data for the most central collisions are used to supplement

201: $\pi^0$ in the low-$p_T$ region \cite{tc}.

202: The horizontal shift determines the scaling factor $J(N)$ shown in

203: Fig.\ 2(a). The vertical shift determines the normalization factor

204: $B(N)$ shown in Fig.\ 2(b).  The resultant scaling distribution

205: $\Phi(z)$ is now related to the measured inclusive distribution by

206: \begin{eqnarray}

207: \Phi(z) = {B(N) \over J^2(N)}{p_0^2 \over 2\pi p_T} {dN_{\pi} \over

208: d\eta dp_T} \ .

209: \label{14}

210: \end{eqnarray}

211: Evidently, the scaling behavior exhibited in Fig.\ 1 is very good.  The

212: solid line is a fit using the formula

213: \begin{equation}

214: \Phi(z) = 150\ (z^2 + 1.05)^{-4.18} \ (1 + 7e^{-4z} ) \ .

215: \label{15}

216: \end{equation}

217: There are many more points included in Fig.\ 1 than those of the

218: preliminary data used in Ref.\ \cite{hy}.

219: 

220: The behavior of $J(N)$ in the

221: log-log plot in Fig.\ 2 can be fitted by a straight line according

222: to Eq.\ (\ref{13}) with

223: \begin{equation}

224: \gamma = 0.077 \ .

225: \label{16}

226: \end{equation}

227:    The region $N<10$ is not considered. The normalization point is chosen

228: to be $N_0=325$, which corresponds to the most central 0-10\%

229: collisions \cite{ph}. The normalization factor $B(N)$ can be fitted by

230: two straight lines as shown by the solid lines. The

231: implication of this result will be discussed in the following.

232: 

233: \begin{figure}[tbph]

234: \includegraphics[width=0.45\textwidth]{jb.eps}

235: \caption{(a) The scaling factor $J(N)$. Note that the vertical scale is

236: logarithmic. The solid line is a straightline fit, yielding

237: $\gamma=0.077$. (b) The normalization factor $B(N)$.

238: The solid lines are fits showing two scaling regions.}

239: \end{figure}

240: 

241: Although Eqs.\ (\ref{13}) and (\ref{16}) do not differ too much from

242: the inverse of $K(N)$ given in Eq.\ (\ref{3}), we have investigated the

243: source of the difference. Our conclusion is that in \cite{hy} we read

244: the preliminary data from a figure given in a conference talk \cite{dde}

245: and estimated the central points, whereas here we use the finalized data

246: in tabulated form, which differ slightly from the original. Thus our

247: present result is more reliable. Moreover, since in our new description

248: the medium suppression effect is characterized by one and only one

249: parameter $\gamma$, which plays the crucial role of specifying the

250: fractional proportionality relation in Eq.\ (\ref{12}), the resultant

251: scaling behavior has the distinction of being physically

252: motivated. As we have shown, it is not so much what the scaling function

253: \phii is as what the scaling variable $z$ is. Since $z$ quantifies the

254: difficulty of producing transverse motion, it can be termed {\it

255: transversality} that gives a universal description of that difficulty at

256: any centrality. The $z$ dependence of \phii in

257: Eq.\ (\ref{15}) differs little from that of Eq.\ (\ref{4.5}), except at

258: very high $z$ where the new data extend beyond that of the

259: preliminary result.

260: 

261: With the value of $\gamma$ in Eq.\ (\ref{16}) now determined

262: phenomenologically, we can return to Eq.\ (\ref{12}) and claim that

263: the notion of fractional proportionality has direct support by the

264: data.  Note that the independence of the fractional

265: energy loss $\delta x/x$ on $x$ is a result that differs from that of

266: pQCD  on the momentum shift of hard partons where

267: $\Delta p_T$ is proportional to $p_T^{1/2}$ \cite{ba}, and is

268: consistent with the independence of $R_{AA}$ on $p_T$ \cite{ph}.

269: Equation (\ref{12}) also implies that

270: $\delta x/\delta N$ is proportional to $x/N$ which is very different from

271: the assumption that the energy loss per unit length traversed by a

272: parton, $dE/dL$, is a constant. Although our phenomenological result on

273: the produced pions has no direct implication on the evolutionary

274: properties of the partons propagating through a medium, one should keep

275: these differences in mind, when the measurable consequences of what can

276: be calculated in pQCD are inferred.

277: 

278: Since the method by which we obtained $J(N)$ is by means of data

279: fitting, it is desirable to find an alternative method that is more

280: direct. Given $\Phi(z)$, one can determine the average

281: \begin{equation}

282: \left< z\right> = \int dz\ z^2 \Phi(z) \left/ \int dz\ z\,\Phi(z)\right.

283: = 0.42 \ .           \label{17}

284: \end{equation}

285: Then from Eq.\ (\ref{5}) we have $J(N)=0.42/\left< x\right>(N)$, where

286: $\left< x\right>$ is related to the inclusive cross section as

287: \begin{equation}

288: \left< x\right>(N)=\int dp_T\ p_T {dN_\pi\over d\eta dp_T}\left/ p_0\int

289: dp_T{dN_{\pi}\over d\eta dp_T} \right. \ .   \label{18}

290: \end{equation}

291: Since $J(N)$ is normalized to 1 at $N=N_0$, we now have

292: \begin{equation}

293: J(N)=\left< x\right>(N_0) / \left< x\right>(N)\ ,    \label{19}

294: \end{equation}

295: thus eliminating the reference to the scaling variable $z$.

296:    The $N$ dependence of $\left< x\right>(N)$ can be determined directly

297: from the data when the average is calculated at various centralities.

298: The large \pt part of the integration is important, for it is that part

299: of the $dN_{\pi}/d\eta dp_T$ that led us to the scaling factor $J(N)$ in

300: the first place. The low-\pt part of the spectra should be accurately

301: parametrized. Now, it is a matter of evaluating Eq.\ (\ref{18}) instead

302: of shifting and rescaling in Fig.\ 1. We recommend that experimental

303: groups that have the data, not only of $\pi^0$, but also of charged

304: hadrons, can use this method to check whether $J(N)$ indeed has the form

305: given in Eq.\ (\ref{13}).

306: 

307: The scaling behavior that we have obtained is for the produced

308: $\pi^0$. We have found in \cite{hy3} that the anomalously high $p/\pi$

309: ratio at $p_T\sim$ 3-4 GeV/c can be understood in terms of

310: recombination without fragmentation, when hard partons are allowed to

311: recombine with the  soft ones. That is possible if the recombination

312: function for the pion does not restrict the recombining

313: $q$ and $\bar q$ to have roughly the same momentum \cite{hhy}. Indeed,

314: in our view jet fragmentation is included in recombination because a

315: large-$p_T$ hard parton initiates a parton shower that hadronizes by

316: recombination \cite{rh}. The use of fragmentation function is only a

317: phenomenological way of parametrizing that process, and does not stand

318: for an independent hadronization mechanism. In heavy-ion collisions the

319: parton shower on the surface has low-$p_T$ components that mingle with

320: the soft partons with hydrodynamical origin. Since separating them would

321: be artificial, we treat them on equal footing in the recombination process

322: that can involve hard partons as well. The purpose of this discussion is

323: to prepare our way to descend to the parton level without fragmentation.

324: The scaling behavior that we have found supports this view, since our

325: rescaling procedure is universal and does not separate the high-\pt from

326: the low-\pt regions using different transversality variables. We note that

327: this view differs from that taken in \cite{fr}, in which recombination

328: and fragmentation are important in different regions. We also note that

329: our application of the recombination mechanism to all partons encounters

330: no inconsistency with the two experimental facts: (a) $p/\pi$ ratio is

331: roughly 1 at $p_T \sim$ 3-4 GeV/c, and (b) the jet structure in Au-Au

332: collisions is similar to that in $pp$ collisions in the same \pt region

333: \cite{jac}.

334: 

335: With the above discussion we have laid the basis for the expectation

336: that the quark distribution $F_q$ for all \pt contributes to the

337: determination of the pion distribution at all $p_T$. The

338: recombination equation derived in Refs.\ \cite{hy2,hy3} has the form

339: \begin{equation}

340: \Phi(z)=\int dz_1dz_2\,{z_1z_2\over z}F_q(z_1)F_{\bar

341: q}(z_2)\delta(z_1+z_2-z) \ ,    \label{20}

342: \end{equation}

343: in which the \pt variables have all been transformed to the scaling

344: variables with

345: \begin{equation}

346: z_i={p_{i_T}\over p_0}J(N),  \qquad i=1,2\ .   \label{21}

347: \end{equation}

348:    The $\delta$ function comes from the recombination

349: function and guarantees the conservation of momentum. Note that $z_1$

350: and $z_2$ are integrated over all values; in particular, they are not

351: restricted to the region $z/2$. Indeed, since the quark distributions

352: fall rapidly with $z_i$ it is necessary for one $z_i$ to be large and

353: the other $z_i$ to be small in order to give the highest contribution to

354: \phii at large $z$. The essential remark we want to make now is that in

355: our formalism for hadronization the $q$ and $\bar q$ distributions

356: must have their $p_{i_T}$ scaled by the same factor as shown in Eq.\

357: (\ref{21}). Since the $J(N)$ in Eq.\ (\ref{21}) is the same as that for

358: the pion, we conclude that the fractional energy loss for the quarks

359: (and antiquarks) satisfies the same proportionality relationship as in

360: Eq.\ (\ref{12}). It should immediately be emphasized that these

361: partons are at the end of their evolution (hydrodynamical and/or

362: branching in showers) just before recombination. As mentioned earlier

363: the energy loss discussed here does not refer to the radiative energy

364: loss of a parton traversing a medium \cite{wang}. Our point is that  the

365: fractional energy loss of the partons at hadronization satisfies the

366: same property as for the pions. Since  in this formalism of

367: hadronization we have been able to obtain the correct

368: $p/\pi$ ratio \cite{hy3}, it follows that the produced protons should

369: have the same property  in fractional energy loss also. This prediction

370: should be checked experimentally by studying the proton spectra and

371: seeing whether centrality scaling can be achieved with the same $J(N)$.

372: We expect, however, the proton mass effect to break the scaling at low

373: $z$.

374: 

375: Our final remark is a speculative one. The universality of the single

376: exponent $\gamma$ for all centralities (except the very peripheral

377: collisions) raises the question whether a hot and dense medium is any

378: different from a less dense medium in its effect on pion production at

379: high $p_T$. If not, the high \pt spectra would not be a fruitful place

380: to find the signature of plasma formation. A possible escape from that

381: conclusion is the observation that the result on $B(N)$ in Fig.\ 2(b)

382: can be fitted by two straight lines, which can be parametrized by

383: \begin{eqnarray}

384: B(N)&=&(N/N_1)^{-\beta_1},  \qquad N<38,   \nonumber \\

385:     &=&(N/N_2)^{-\beta_2},  \qquad N>38 \ ,

386: \label{22}

387: \end{eqnarray}

388: where $\beta_1=0.744$ and $\beta_2=1.292$ with $N_1=1610$ and

389: $N_2=325$. It  suggests that there are two regimes of $N$,

390: requiring different exponents $\beta_1$ and $\beta_2$ to achieve

391: scaling. If this break can be ascertained by more detailed analysis at

392: different energies, then it offers a way out of the strict universality in

393: which there is no hint of any essential diffference between hot and cold

394: media, leaving no room for any signal for the formation of quark-gluon

395: plasma.

396: 

397: To summarize, we have found a relationship between the centrality

398: scaling behavior of the observed pion spectra and the fractional energy

399: loss of the pions that is proportional to the fractional centrality

400: change. That proportionality is specified by an exponent $\gamma$,

401: which characterizes the medium suppression effect. In the recombination

402: model the same value of $\gamma$ is valid for the fractional energy

403: loss of the light quarks just before hadronization.  The existence of

404: the  universal scaling function of transversality suggests that

405: there is no essential difference in how the low- and high-$p_T$

406: hadronization processes should be treated. The possibility of a

407: single exponent $\gamma$ that can be directly extracted from the data

408: to summarize the nuclear suppression effect offers a very succinct

409: description of a complicated dynamical process.  Universal behavior

410: in centrality may be broken by the existence of two scaling regions in

411: the normalization factor, thus providing the possibility of a

412: threshold for a new distinctive regime.

413: 

414: We are grateful to Dr.\ D.\ d'Enterria for informing us of the PHENIX

415: data and for very useful comments. We also thank Drs.\ P.\ Jacobs, B.\

416: M\"uller and X.-N.\ Wang for helpful communication. This work was

417: supported, in part,  by the U.\ S.\ Department of Energy under Grant No.

418: DE-FG03-96ER40972.

419: 

420: 

421: \begin{thebibliography}{99}

422: 

423: %1

424: \bibitem{de}

425: K.\ Adcox {\it et al} (PHENIX Collaboration), Phys.\ Rev.\ Lett.\ {\bf

426: 88}, 022301 (2002); C.\ Adler {\it et al.} (STAR     Collaboration),

427: Phys.\ Rev.\ Lett.\ {\bf 89}, 202301 (2002).

428: 

429: %2

430: \bibitem{ph}

431: S.\ S.\ Adler {\it et al.} (PHENIX Collaboration), nucl-ex/0304022.

432: 

433: %3

434: \bibitem{st}

435: J.\ Adams {\it et al.} (STAR Collaboration), nucl-ex/0305015.

436: 

437: %4

438: \bibitem{hy}

439: R.\ C.\ Hwa, and C.\ B.\ Yang,

440: Phys.\ Rev.\ Lett.\ {\bf 90}, 212301 (2003).

441: 

442: %5

443: \bibitem{hy2}

444: R.\ C.\ Hwa, and C.\ B.\ Yang, Phys.\ Rev.\ C {\bf 67}, 064902 (2003).

445: 

446: %6

447: \bibitem{hy3}

448:      R.\ C.\ Hwa, and C.\ B.\ Yang,

449: Phys.\ Rev.\ C {\bf 67}, 034902 (2003).

450: 

451: %7

452: \bibitem{fr}

453: R.\ J.\ Fries, B. M\"{u}ller, C.\ Nonaka and S.\ A.\ Bass, Phys.\

454: Rev.\ Lett.\ {\bf 90}, 202303 (2003); nucl-th/0306027.

455: 

456: %8

457: \bibitem{gr}

458: V.\ Greco, C.\ M.\ Ko and P.,\ L\'{e}vai, Phys.\ Rev.\

459: Lett.\ {\bf 90}, 202302 (2003); nucl-th/0305024.

460: 

461: %9

462: \bibitem{li}

463: D.\ Moln\'ar and S.\ A.\ Voloshin, nucl-th/302014; Z.-W.\ Lin and D.\

464: Moln\'ar, nucl-th/0305049.

465: 

466: %10

467: \bibitem{ba}

468: R.\ Baier, Yu.\ Dokshitzer, A.\ Mueller, and D.\ Schiff, J.\ High

469: Energy Physics 0109, 033 (2001).

470: 

471: %11

472: \bibitem{ht}

473: http://www.phenix.bnl.gov/phenix/WWW/publish/enterria\\ /AuAu pi0 200GeV

474: data/

475: 

476: %12

477: \bibitem{tc}

478: T.\ Chujo, Nucl.\ Phys.\ {\bf A715}, 151c (2003).

479: 

480: %13

481: \bibitem{dde}

482: D.\ d'Enterria (PHENIX Collaboration), talk given at Quark Matter 2002,

483: Nante, France. Figures used in \cite{hy} are not included in the

484: conference proceedings.

485: 

486: %14

487: \bibitem{hhy}

488:     R.\ C.\ Hwa, Phys.\ Rev.\ D{\bf 22}, 1593 (1980); R.\ C.\ Hwa, and

489: C.\ B.\ Yang, Phys.\ Rev.\ C {\bf 66}, 025205 (2002).

490: 

491: %15

492: \bibitem{rh}

493: R.\ C.\ Hwa, talk given at Heavy-Ion Collisions 03, McGill University,

494: Montreal, CA, June 2003.

495: 

496: %16

497: \bibitem{jac}

498: C.\ Adler {\it et al.} (STAR Collaboration), Phys.\ Rev.\ Lett.\ {\bf

499: 90}, 082302 (2003).

500: 

501: %17

502: \bibitem{wang}

503: X.-N.\ Wang, nucl-th/0307036.

504: 

505: \end{thebibliography}

506: 

507: 

508: \end{document}

509: