% ****** Start of file template.aps ******
%   This file is part of the APS files in the REVTeX 3.0 distribution.
%   Version 3.0 of REVTeX,! November 10, 1992.
%
%   Copyright (c) 1992 The American Physical Society.
%
%   See the REVTeX 3.0 README file for restrictions and more information.
%
%
% This is a template for producing files for use with REVTEX 3.0.
% Copy this file to another name and then work on that file.
% That way, you always have this original template file to use.
%

%
% THIS FILE: dndeta200-draft33.tex
%
%            October 13-2001. 23.00 GMT+1
%            Based on first text by SS and JJG
% REVISED:
%            Oct 14-2001 JJG. Update text
%            Oct 16-2001 JJG. Update text
%            Oct 17-2001 JJG. Include hard/soft fit results + update text
%            Oct 17-2001 JJG. Fill out Table + update text, author list, 
figure 4.
%            Oct 18-2001 JJG. update references and include eta = 1.5 
for fits.
%            Oct 19-2001 JJG. smaller updates and corrections after 
input from   CEJ
%                             and TST
%            Nov 12-2001 JJG. text adjusted to HIRO's new figures
%            Nov 15-2001 JJG. text shortened. Prepare for 5 figures.
%            Nov 16-2001 JJG. various text revisions
%            Nov 19-2001 JJG text revision with various input from others. New  fig 5 not described yet.
%            Nov 20-2001 MM + SS changed to Revtex 4. SS update numbers + some text.
%                                  JJG update text. Go back to Npart figure 5. Keep figure 2.
% **************************************************************************

%documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
\documentclass[preprint,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
% Some other (several out of many) possibilities
%\documentclass[preprint,aps]{revtex4}
%\documentclass[preprint,aps,draft]{revtex4}
%\documentclass[prb]{revtex4}% Physical Review B

\usepackage{graphicx}% Include figure files
\usepackage{dcolumn}% Align table columns on decimal point
\usepackage{bm}% bold math
\usepackage{epsf,epsfig,latexsym}

%Revtex3   \documentstyle[prl,aps,epsfig,preprint]{revtex}

%\documentstyle[prl,aps,epsfig,multicol]{revtex}

  \newcommand{\Version}       {Version: dndeta200-draft33.tex, Nov. 19, 2001}
  \newcommand{\bnl}           {$\rm^{1}$}
  \newcommand{\ires}          {$\rm^{2}$}
  \newcommand{\kraknuc}   {$\rm^{3}$}
  \newcommand{\krakow}   {$\rm^{4}$}
  \newcommand{\baltimore}  {$\rm^{5}$}
  \newcommand{\newyork}  {$\rm^{6}$}
  \newcommand{\nbi}           {$\rm^{7}$}
  \newcommand{\texas}         {$\rm^{8}$}
  \newcommand{\bergen}        {$\rm^{9}$}
  \newcommand{\bucharest}     {$\rm^{10}$}
  \newcommand{\kansas}        {$\rm^{11}$}
  \newcommand{\oslo}          {$\rm^{12}$}

  \begin{document}
% \draft command makes pacs numbers print
%reftex3   \draft
% repeat the \author\address pair as needed


%  \title{Centrality and pseudorapidity dependence of
%charged particle densities for
%$\sqrt{s_{NN}} = 200$ Gev Au+Au collisions at RHIC}


 \title{Role of partonic collisions in the production of charged particles 
         in 100 AGeV + 100 AGeV Au+Au collisions at RHIC}

  \author{
    I.~G.~Bearden\nbi, % \and
    D.~Beavis\bnl, % \and
    C.~Besliu\bucharest, % \and
    Y.~Blyakhman\newyork, % \and
    J.~Brzychczyk\krakow, % \and
    B.~Budick\newyork, % \and
    H.~B{\o}ggild\nbi, % \and
    C.~Chasman\bnl, % \and
    C.~H.~Christensen\nbi, % \and
    P.~Christiansen\nbi, % \and
    J.~Cibor\kraknuc, % \and
    R.~Debbe\bnl, % \and
    E. Enger\oslo, %\and
    J.~J.~Gaardh{\o}je\nbi, % \and
    K.~Grotowski\krakow, % \and
    K.~Hagel\texas, % \and
    O.~Hansen\nbi, % \and
    A.~Holm\nbi, % \and
    A.~K.~Holme\oslo, % \and
    H.~Ito\kansas, % \and
    E.~Jakobsen\nbi, % \and
    A.~Jipa\bucharest, % \and
    J.~I.~J{\o}rdre\bergen, % \and
    F.~Jundt\ires, % \and
    C.~E.~J{\o}rgensen\nbi, % \and
    T.~Keutgen\texas, % \and
    E.~J.~Kim\baltimore, % \and
    T.~Kozik\krakow, % \and
    T.~M.~Larsen\oslo, % \and
    J.~H.~Lee\bnl, % \and
    Y.~K.~Lee\baltimore, % \and
    G.~L{\o}vh{\o}iden\oslo, % \and
    Z.~Majka\krakow, % \and
    A.~Makeev\texas, % \and
    B.~McBreen\bnl, % \and
    M.~Mikelsen\oslo, % \and
    M.~Murray\texas, % \and
    J.~Natowitz\texas, % \and
    B.~S.~Nielsen\nbi, % \and
    J.~Norris\kansas, % \and
    K.~Olchanski\bnl, % \and
    J.~Olness\bnl, % \and
    D.~Ouerdane\nbi, % \and
    R.~P\l aneta\krakow, % \and
    F.~Rami\ires, % \and
    D.~R{\"o}hrich\bergen, % \and
    B.~H.~Samset\oslo, % \and
    D.~Sandberg\nbi, % \and
    S.~J.~Sanders\kansas, % \and
    R.~A.~Sheetz\bnl, % \and
    Z.~Sosin\krakow, % \and
    P.~Staszel\nbi, % \and
    T.~F.~Thorsteinsen\bergen$^+$,% \and
    T.~S.~Tveter\oslo, % \and
    F.~Videb{\ae}k\bnl, % \and
    R.~Wada\texas, % \and
    A.~Wieloch\krakow, and
    I.~S.~Zgura\bucharest\\% \and
    (BRAHMS Collaboration )\\[1ex]
    \bnl~Brookhaven National Laboratory, Upton,New York 11973,
    \ires~Institut de Recherches Subatomiques and Universit{\'e} Louis 
Pasteur,
  Strasbourg, France,
    \kraknuc~Institute of Nuclear Physics, Krakow, Poland,
    \krakow~Jagiellonian University, Krakow, Poland,
    \baltimore~Johns Hopkins University, Baltimore, Maryland 21218,
    \newyork~New York University, New York, New York 10003,
    \nbi~Niels Bohr Institute, University of Copenhagen, Denmark,
    \texas~Texas A$\&$M University, College Station,Texas 77843,
    \bergen~University of Bergen, Department of Physics, Bergen,Norway,
    \bucharest~University of Bucharest,Romania,
    \kansas~University of Kansas, Lawrence, Kansas 66049,
    \oslo~University of Oslo, Department of Physics, Oslo,Norway,
    $^+ Deceased$}


  \date{\Version}

  \begin{abstract}
% insert abstract here
  {
  We present charged particle densities as a function of pseudorapidity and
  collision centrality for the $^{197}$Au+$^{197}$Au reaction
  at $\sqrt{s_{NN}}$=200 GeV, the maximum energy for RHIC.
The charged particle
  multiplicity for the 5\% most central events is  632 $\pm$ 1 (stat)  
$\pm$55
  (syst), i.e. a 14\% increase
  relative to $\sqrt{s_{NN}}$=130 GeV collisions. The total multiplicity of 
charged
  particles for
  $-4.7\le \eta \le 4.7$ is 4630  $\pm$370, an increase
  by 20\%  over the lower energy. The data show an increase from 2.9 to
  3.7 in the
  production of charged particles per pair of participant nucleons from
  peripheral (40-50\%) to central (0-5\%) collisions  around midrapidity.
  These results constrain current models based on high density QCD gluon 
  saturation and on the superposition of particle production from soft hadronic 
  and hard partonic collisions.
  }

  \end{abstract}
% insert suggested PACS numbers in braces on next line
\pacs{25.75.Dw} % Revtex 4
\maketitle

  A central question in the study of collisions between heavy nuclei at the
  maximum energy afforded by the RHIC collider, $\sqrt{s_{NN}}$=200 GeV,
  is the  role of hard scatterings between partons and the interactions of these
  partons in a high density environment. Indeed, it has been conjectured
  that new phenomena related to non-perturbative QCD may come into play at
  this energy. Among these,  a saturation of the number of parton (mainly gluon, 
  since $N_g \propto 1/ \alpha_s$)
  collisions in central nucleus-nucleus collisions
  has been predicted to limit the production of charged
  particles~\cite{partonsat83,Eskola00,Kharzeev_and_Levin}.
  Recently, indications for a reduction in the
  number of hadrons at high transverse momentum
  for $\sqrt{s_{NN}}$=130 GeV collisions have been presented that may
  hint at suppression of hadronic jets at high matter densities
  ~\cite{Phenix-jets,Star-jets}.

  The present Letter addresses these issues via the first
  comprehensive investigation of multiplicity distributions of emitted
  charged particles in relativistic heavy ion collisions between $^{197}$Au
  nuclei at the maximum RHIC energy. In particular,
  we have measured pseudorapidity distributions of charged particles in 
the range   $-4.7 \le \eta \le 4.7$ for such collisions at 
$\sqrt{s_{\small{NN}}}$=200 GeV   and as a function of collision centrality. 
The production of charged particles  in these highly energetic nuclear collisions can be due to hadronic 
and partonic collision processes and thus depends on the presence of
  gluon shadowing effects and, in general, on the relative importance of 
soft and hard scattering processes. We find in this work that the production
  of charged particles at midrapidity increases by about 14\% for the most
  central collisions relative to $\sqrt{s_{NN}}$=130 GeV collisions. 
This result  is in agreement with first results of the PHOBOS experiment at
  midrapidity~\cite{Phobos-mult200-1}. At larger rapidities our results
suggest a saturation of the fragment excitations. 
  
  The BRAHMS experiment consists of two magnetic spectrometers for measuring
  exclusive charged particle spectra over a wide range of pseudorapidity and
  transverse momentum and a number of global detectors  for determining the
  location
  of the collision vertex, the time of the collision, and for characterizing
  general reaction properties such as the collision centrality and
  the inclusive charged particle densities.
  The present data were obtained using three of the global
  detector systems at BRAHMS: the Multiplicity Array (MA),
  the Beam-Beam Counter arrays (BBC), and the Zero-degree Calorimeters 
(ZDC).
  A detailed discussion of the BRAHMS experimental
  arrangement can be found in ref.~\cite{bearden01b}. In addition, an 
analysis of
  charged particle densities for Au+Au reactions
  at $\sqrt{s_{NN}}$=130 GeV that is very similar in method
  to that presented here is described fully in ref.~\cite{bearden01a}.

  The MA determines charged particle densities around mid-rapidity and
  consist of a hexagonal-sided double barrel arrangment
  with a modestly segmented
  Si strip detector array (SiMA)  surrounded by an outer plastic
  scintillator tile array (TMA). Four sides of the SiMA
  were fully populated with six Si detectors each, one side had one detector,
  and one side was left unpopulated.  Each Si wafer is dimensioned
  4~cm x 6~cm x 300~$\mu$m,  is subdivided into seven active strips, and
  is located 5.3~cm
  from the beam axis.  The plastic scintillator tiles, dimensioned
  12~cm x 12~cm x 0.5~cm and located 13.9 cm from the beam axis, were
  arranged with four sides of the hexagonal barrel populated with
  eight detectors, each, and the remaining two sides populated with
  two detectors and one detector, respectively.
  Both the SiMA and TMA cover a pseudorapidity range of 
$-2.2\le\eta\le2.2$ for
  collisions occurring at the nominal interaction vertex. Using an 
extended range
  of collision vertices, the effective coverage of the array
  is $-3.0\le\eta\le3.0$ . Particle densities are deduced
  from the observed energy loss in the SiMA and TMA elements using
  GEANT simulations ~\cite{Geant} to relate energy loss to the number of particles
  hitting a given detector element~\cite{bearden01a}.

  The BBC Arrays consist of two sets of Cherenkov UV transmitting
  plastic radiators coupled to photomultiplier tubes. They are positioned
  around the beam pipe on either side of the nominal interaction point at a
  distance of 2.20~m.  The time resolution of the BBC elements is
  sufficient to locate the interaction
  point with an accuracy of $\approx$ 0.9~cm.
  Charged particle multiplicities in the pseudorapidity range
  $2.1\le |\eta| \le 4.7$ are deduced from the number of particles
  hitting each tube, as found by dividing the measured ADC signal by
  that corresponding to a single particle hitting the detector.

  The ZDC detectors~\cite{adler00}are located  $\approx\pm$18m from the
  nominal interaction vertex and measure neutrons that are
  emitted at small angles with respect to the beam direction.
  The ZDCs provide the minimum bias trigger for the experiment (estimated to
  include 97\% of the nuclear reaction cross section) and can also be 
  used to  locate the interaction vertex with an accuracy of $\approx$~3.6~cm.

  The SiMA and TMA total multiplicities are averaged after
  accounting for the different geometric acceptances of the two arrays
  and used together with the BBC total multiplicity information to determine
  the reaction centrality by assuming that a cut on total
  multiplicity translates to a cut on collision centrality. In analyzing
  particle densities in dN/d$\eta$, the centrality dependence of the
  MA and BBC distributions are based on the centrality measurements of the
  corresponding array. In the pseudorapidity range of 3.0$\le\eta\le$4.2,
  where it was possible to analyze the BBC data using both
  centrality selections, the two centrality analyses give identical results
  to within 1-2\%. In general, statistical error on the measurements are less than
  1\%,   while we estimate that the systematic errors are 8\% and 10\% for the SiMA
  and BBC arrays, respectively.

  In figure~\ref{dndeta} we show the measured pseudorapidity
  distributions for charged particles for centrality cuts 0-5\%, 5-10\%,
  10-20\%, 20-30\%, 30-40\% and 40-50\% of the minimum bias distribution.
  The $dN/d\eta$ values for these cuts at $\eta$=0, 1.5, 3.0, 4.5
  are listed in Table 1, together with the
  number of participating baryons estimated from the HIJING model.
  For the most central collisions (0-5\%) the multiplicities reach
  $dN/d\eta$=632 $\pm 55$ at midrapidity. This corresponds, by division 
  with the number of participating baryon pairs, to $3.7 \pm 0.3$ charged particles per 
  pair.
  This value indicates an increase in the multiplicities of about
  14\% relative to $^{197}$Au+$^{197}$Au reactions at 
$\sqrt{s_{NN}}$=130 GeV
  ~\cite{bearden01a,back00,Phobos-mult200-1,adcox01,Star-mult130}.
  By integrating the most central distribution we deduce that $N=4630 
\pm 370$
  charged particles are emitted in the considered rapidity range, the 
largest
  number of particles observed so far in energetic nuclear collisions. 
This value
  is  20\% higher than for $\sqrt{s_{NN}}$=130 GeV reactions. More detailed
  comparison shows
  that the distributions at the two energies are quite similar in shape. 
Indeed,
  the FWHM of the most central distributions is $\Delta \eta = 7.5 \pm 0.5$
  for $\sqrt{s_{NN}}$=200 GeV, as compared to $\Delta \eta = 7.2 \pm 0.8$
  for $\sqrt{s_{NN}}$=130 GeV collisions. For the most  peripheral 
collisions
  analyzed here (40-50\%) the multiplicities at $\eta=0$
  reach $dN/d\eta=110\pm 10$ while the corresponding value
  scaled to the number of participating pairs is 2.9$\pm 0.3$. For 
  comparison,
  the similar number for proton-proton collisions at this energy is 2.5, also a
  14\% increase
  as compared to the lower energy.

  Figure~\ref{dndeta_fragment} shows, on the other hand, that the charged
  particle multiplicities in an interval of approximately 0.5-1.5 units below the
  beam rapidity are independent of the collision centrality and energy, from
  CERN-SPS energy ($\sqrt{s_{NN}}$=17 GeV) \cite{deines00}
  to the present RHIC energy.
  This is consistent with a limiting fragmentation picture in which the
  excitations
  of the fragment baryons saturate already at moderate collision energies
  independently of the system size ~\cite{bearden01a}. In contrast, the
  increased projectile kinetic energy is utilized for particle production
  in the region around midrapidity, as evidenced by the observed increase
  of the multiplicities per participant pair around the center of mass 
rapidity.  The observed multiplicity excess of $\approx50\%$ above the 
corresponding value for p+p 
collisions clearly demonstrates significant medium effects.

  Figure~\ref{dndeta_models} presents the $dN/d\eta$ distributions
  obtained by mirroring and averaging the negative and
  positive halves of the measured distributions to further decrease
  errors. We also compare the distributions with model calculations. The full
  drawn lines are calculations using the model of Kharzeev and
  Levin~\cite{Kharzeev_and_Levin}
  which is based on a classical QCD calculation using parameters fixed to
  the $\sqrt{s_{NN}}$=130 GeV data. This approach is
  able to reproduce the magnitude and shape of the observed multiplicity
  distributions quite well. Also shown in Figure~\ref{dndeta_models}
  (dotted lines) are the results of calculations with the AMPT
  model~\cite{zhang01,lin01a,lin01b}, which is a cascade model based on
  HIJING~\cite{wang91} but including final state rescattering of produced
  particles.
  The AMPT model is also able to account for the general trend of the measured
  distributions, particularly for the most central collisions,
  but tends to overpredict the data for the more peripheral collisions.

  The ratio of the pseudorapidity densities measured at
  $\sqrt{s_{NN}}$=130 GeV and $\sqrt{s_{NN}}$=200 GeV for different 
centralities
  are
  shown in Figure~\ref{dndeta_ratios}. The figure shows a constant 
increase of
  about 14\% in
  multiplicity as a function of energy
  for a central plateau  in the range $\eta= 0-2.5 $. The upturn of the 
ratios at
  the forward rapidities is due to
  the widening of the multiplicity distribution at the higher energy 
consistent
  with the increase in beam rapidity
  ($\Delta y = 0.45$). The overlaid curves show the corresponding ratios 
resulting
  from the two
  model calculations.

  Finally in Figure~\ref{Npart} we plot the dependence of the multiplicity
  of charged particles per pair of participant baryons as a function of the
  number of participants, $N_{part}$, for 3 narrow pseudrapidity regions
  ($\Delta \eta \approx 0.2$ around $\eta$ =0, 3.0 and 4.5. While the 
  figure shows  that particle production per participant pair is remarkably constant 
at the  forward
  rapidities characteristic of the fragmentation region and close to 
unity, this
  is not the
  case for the central rapidities. Indeed, we find a significant increase of
  particle production
  per pair of participant nucleons for the more central collisions at  
$\eta=0$.
  Plotted using the $N_{part}$ values listed in table 1, the curves for 
these
  rapidities rise as a function of collision centrality. This has 
previously been
  attributed
  to the onset of hard scatterings dependent, not on the number of 
participants,
  but
  on the number of binary parton collisions, $N_{coll}$. Using for 
$N_{coll}$ the
  values estimated in ~\cite{Kharzeev_and_Nardi} we fit the observed
  dependencies to a functional $dN/d\eta/(N_part/2)=\alpha\cdot
  N_{part}+\beta \cdot N_{coll}$. For rapidities $\eta=$ 0 and 3.0 we 
obtain:
  $\alpha=0.98 \pm 0.10 and , 1.04 \pm 0.08$ and
  $\beta=0.21 \pm 0.4,  0.08 \pm 0.03$ respectively.
  For comparison we find $\alpha=0.99 \pm 0.09, 0.99 \pm 0.07, $
  and $\beta=0.18 \pm 0.03, 0.02 \pm 0.02 $ at
  $\sqrt{s_{NN}}$=130 GeV. At $\eta=0.0$ we find that the hard scattering
  component to the charged particle production increases from 36\% 
 $\pm$7\% at
  the lower energy to about 43\% $\pm$ 7\% at $\sqrt{s_{NN}}$=200 GeV.
  It should be stressed, however, that this interpretation is highly model
  dependent. Using the $N_{part}$ values from ~\cite{Kharzeev_and_Nardi} 
  which are smaller
  than the corresponding HIJING numbers for the more peripheral 
 collisions the
  curves
  become practically flat ($\beta \approx 0$) and thus inconstent with a 
mixing
  of soft and hard scatterings.

  In conclusion, we find that the charged particle production scales
  smoothly from $\sqrt{s_{NN}}$=130 GeV to $\sqrt{s_{NN}}$=200 GeV 
  in a wide region around midrapidity.  The data are well reproduced by calculations 
  based on high density QCD and  by the AMPT/HIJING 
  microscopic parton model. A phenomenological  two component analysis in terms of a 
  superposition of particle production due to soft/hard scatterings also accounts well for the 
  data but does not show significant  differences between the two energies. We find
  good consistency with the gluon saturation model of Kharzeev and Levin,
  but stress that within errors of models and data alike, the data can 
  be equally well reproduced by other models not requiring saturation effects in the 
  description  of  parton collisions. While the current work establishes the baseline for 
  particle  production at the maximum energy available for nucleus-nucleus 
  collisions for several years to come, the full understanding of these energetic collisions must await
  more  detailed analyses of hadronic and leptonic observables over a wide region
  of phase space and rapidity.

  \vskip 1.2cm

%\section{Acknowledgments}

  We thank the RHIC collider team for their support to the experiment.
  This work was supported by the Division of Nuclear Physics of the Office
  of Science of the U.S. Department of Energy under
  contracts DE-AC02-98-CH10886, DE-FG03-93-ER40773, DE-FG03-96-ER40981, and
  DE-FG02-99-ER41121, the Danish Natural Science Research Council, the
  Research Council of Norway, the Jagiellonian University Grants,
  the Korea Research Foundation,  and the Romanian Ministry of
  Research (5003/1999,6077/2000). We are grateful to Drs. D Kharzeev,  BNL, and E.
  Levin, Tel Aviv, and Zi-Wei Lin, Texas A\&M , for stimulating 
 discussions and for supplying
  us with the
  model calculations shown and discussed in this article.

  \vskip 0.8cm


%Revtex3   \begin{references}
\begin{thebibliography}{99}
%Parton saturation.
\bibitem{partonsat83} L. V. Gribov, E. M. Levin and M. G. Ryskin, 
Phys.Rep.100 (1983) 1.
% ``Centrality dependence of multiplicities in ultrarelativistic nuclear 
collisions''
\bibitem{Eskola00} K. J. Eskola, K. Kajantie and K. Tuominen, Phys. 
Lett.    %
  {\bf B497}, 39 (2001). %; hep-ph/0009246.
\bibitem{Kharzeev_and_Levin} D. Kharzeev and E. Levin. nucl- th/0108006 
and private
  communication. %
%jet quenching
\bibitem{Phenix-jets} K. Adcox {\it et al.}, subm. to Phys. Rev. Lett. 
(2001),
nucl-ex/0109003.
\bibitem{Star-jets}
J. C. Dunlop, STAR Collaboration, Nucl. Phys. {\bf A698} 515c-518c 
(2002) and
B. Lasiuk, BNL  Workshop  on High $p_T$ phenomenen at RHIC,  Nov. 2001
http://skipper.physics.sunysb.edu/highpt/lasiuk.ppt
%
%Phobos 200 Gev mult at midrapidity 2001. subm to PRL
\bibitem{Phobos-mult200-1} B. B. Back {\it et al.}, submitted. to Phys. 
Rev. Lett.
  (2001), nucl.exp/0108009.
%
%BRAHMS NIM reference
\bibitem{bearden01b} I. G. Bearden {\it et al.}, submitted to Nucl. 
Instrum. Meth A
%
%GEANT reference
\bibitem{Geant} GEANT 3.2.1, CERN program library.
%
%BRAHMS mult 130 reference PLB
\bibitem{bearden01a} I. G. Bearden {\it et al.}, accepted for publ. in
  Phys. Lett. B.; nucl-ex/0108016.
%ZDC reference
\bibitem{adler00} C. Adler, A. Denisov, E. Garcia, M. Murray, H. 
Stroebele and
  S. White, Nucl. Inst. Meth., {\bf A470}  488 (2001). % nucl-ex/0008005.
%
%Phobos dN/deta paper 2000 (56 and 130)
\bibitem{back00}  B. B. Back {\it et al.}, Phys. Rev. Lett.  {\bf 85}, 
3100 (2000).
%STAR paper on mult at 130GeV
\bibitem{Star-mult130} C. Adler {\it et al.}, Phys. Rev Lett. {\bf 87}, 
112303 (2001)
%
%PHENIX dN/deta paper 130GeV
\bibitem{adcox01} K. Adcox {\it et al.}, Phys. Rev. Lett. {\bf 86}, 3500 
(2001).
%
%SPS multiplicity
\bibitem{deines00} P. Deines-Jones {\it et al.} Phys. Rev. C {\bf 62},
  014903(2000). %hep-ex/9912008.
%
\bibitem{zhang01} Bin Zhang, C. M. Ko, Bao-An Li and Zi-wei Lin,
  Phys. Rev. C  {\bf 61} 067901 (2001).
\bibitem{lin01a} Zi-wei Lin, Subrata Pal, C. M. Ko, Bao-An Li and Bin 
Zhang,  %
Phys. Rev. C   {\bf 64} 011902R (2001).
\bibitem{lin01b} Zi-wei Lin, Subrata Pal, C.M. Ko, Bao-An Li and Bin Zhang,
  Nucl. Phys. {\bf A698} 375c-378c (2002),  nucl-th/0105044; and Zi-wei 
Lin, private communication.
%Hijing Paper
\bibitem{wang91} X. N. Wang and M. Gyulasy, Phys. Rev. D {\bf 44}, 3501 
(1991).
%
%AMPT model
%
\bibitem{Kharzeev_and_Nardi} D. Kharzeev and M. Nardi.
Phys. Lett. B507,  121 (2001). %, nucl-th/0012025.
%SPS Pb+Pb   referenced in Fig2
%limiting fragmentation
%not used \bibitem{benecke69} J. Benecke, T. T. Chou, C. N. Yang, and E. 
Yen,  %
% Phys. Rev. {\bf 188}, 2159 (1969).
%
\end{thebibliography}
%SPS Pb+Pb
%\bibitem{aggarwal01} M.M.~Aggaral {\it et al.} Eur. Phys. J. {\bf C18},
%651 (2001); nucl-ex/0008004.
%STAR reference on scintillator tiles
%\bibitem{aota95} S. Aota {\it et al.}, NIM {\bf A352}, 557(1995).
%
%Wounded nucleon
%\bibitem{bialas76} A.~Bialas, M.~Bleszynski, and W.~Czyz, Nucl. Phys.
%{\bf B111}, 461(1976).
%
%Fritiof reference:  Check if to correct version...
%\bibitem{pi92} H. Pi, Comput. Phys. Commun. {\bf 71}, 173(1992).
%
%Glauber paper
%\bibitem{glauber} R.J. Glauber and G. Matthiae, Nucl. Phys. {\bf B21}, 
135(1970).
%
%
%\bibitem{urqmd} S.A. Bass {\it et al.}, Prog. Part. Nucl. Phys.
%{\bf 41}, 225(1998); nucl-th/9803035.
% ``Energy and Centrality Dependence of Rapidity Densities at RHIC''
%\bibitem{wang00} X. N. Wang and M. Gyulassy, Phys. Rev. Lett. {\bf 86}, 
3496 (2001).
  %
%UA5 p-pbar paper
%\bibitem{alner86} G. Alner {\it et al.}, Z. Phys. C, {\bf 33}, 1 (1986).

%revtex3   \end{references}

%   \begin{table} [h!]
  \begin{table} [ht]
%   \caption{\label{TABLE}\textit{\sl Charged particle densities in
  \caption{\label{TABLE} Charged particle densities in
  $dN_{ch}/d\eta$ as a function of
  centrality and pseudorapidity. Total uncertainties, dominated by
  the systematics,  are indicated. The average
  number of participants $<N_{part}>$ is given for each centrality
  class based on HIJING model calculations. The last column gives the
  integral charged particle multiplicity within the pseudorapidity
  range $-4.7 \le \eta \le 4.7$.}
  \begin{tabular}{|c|c|c|c|c|c|c|c|}
  \hline
  Centrality &
  $<N_{coll}>$&
  $<N_{part}>$&
  $\eta = 0$ &
  $\eta =1.5$ &
  $\eta = 3.0$ &
  $\eta = 4.5$ &
  $N_{ch}$
  \\
  0-5\%   & 1086& 345 & 632$\pm$55 & 628$\pm$57 & 453$\pm$41
  & 181 $\pm$21 & 4630$\pm$370 \\
  5-10\%  & 866 & 293 & 498$\pm$44 & 509$\pm$46 & 379$\pm$37
  &  156$\pm$17 & 3810$\pm$300 \\
  10-20\% & 5613 & 228 & 373$\pm$33 & 385$\pm$35 & 296$\pm$29
  &  124$\pm$13 &  2920$\pm$230 \\
  20-30\% & 389 & 164 & 257$\pm$23 & 265$\pm$24 & 207$\pm$21
  &  89 $\pm$10 &  2020$\pm$160 \\
  30-40\% & 232 & 114 & 170$\pm$15 & 178$\pm$16 & 140$\pm$15
  &  62 $\pm$7  & 1380$\pm$110  \\
  40-50\% & 1137 & 75  & 110$\pm$10 & 115$\pm$10 &  90$\pm$9
  &  42 $\pm$5  &  890$\pm$70 \\
  \hline
  \end{tabular}
  \end{table}

  \begin{figure}
  \epsfig{file=fig1.eps,width=8.5cm}
  \caption{
  Top panel: Distributions of  $dN_{ch}/d\eta$ for centrality ranges of,
  top to bottom, 0-5\%, 5-10\%, 10-20\%, 20-30\%, 30-40\%, and
  40-50\%.  The SiMA and BBC results are indicated by the circles and
  triangles, respectively. Statistical error are shown for all points where
  smaller than the symbol size. Isolated large error bars illustrate the systematic
  errors.  }
  \label{dndeta}
  \end{figure}

  \begin{figure}
  \epsfig{file=fig2.eps,width=8.5cm}
  \caption{
  Charged particle densities normalized to the number of
  participant pairs (see table) for the present 0-5\% central (open 
circles) and
  30-40\% central (open squares) Au+Au results at  $\sqrt{s_{NN}}$=200 
GeV, the
  BRAHMS Au+Au results at  $\sqrt{s_{NN}}$=130 GeV (closed circles) and the
  9.4\% central Pb+Pb data at  $\sqrt{s_{NN}}$=17 GeV(closed triangles)
  of ref~\cite{deines00}. Data at different beam energies are 
plotted as a
  function
  of the pseudorapidity shifted by the relevant beam rapidity. Total 
uncertainties
  (statistical and systematic) are shown. }
  \label{dndeta_fragment}
  \end{figure}


  \begin{figure}
  \epsfig{file=fig3.eps,width=8.5cm}
  \caption{
  (a-d) Distribution of the measured $dN_{ch}/d\eta$ for  centrality
  ranges 0-5\%, 10-20\%, 20-30\% and 40-50\%.  Total uncertainties
  (statistical and systematic) are indicated. Theoretical
  predictions by Kharzeev and Levin (full drawn line)
  and by the AMPT model (dashed line) are also shown.
  }
  \label{dndeta_models}
  \end{figure}


  \begin{figure}
  \epsfig{file=fig4.eps,width=8.5cm}
  \caption{
  Ratio of particle densities at $\sqrt{s_{NN}}$=200 GeV and 130 GeV
  compared to the models. Points are only shown with statistical errors
  as systematic errors cancel out
  }
  \label{dndeta_ratios}
  \end{figure}

  \begin{figure}
  \epsfig{file=fig5.eps,width=8.5cm}
  \caption{
  Distributions of $dN_{ch}/d\eta$
  per participant pair as a function of the  number of participants (see 
table)
  for  $\eta$= 0,1.5, 3.0 and 4.5. The curves show predictions by the
  Kharzeev and Levin model (full drawn line) and the AMPT model (dashed 
line).
  }
  \label{Npart}
  \end{figure}

  \end{document}