Building up the globular cluster system of the Milky Way. The contribution of the Sagittarius galaxy
Extracted source text ยท arXiv astro-ph/0210596
%% use the following for submission to AJ
%\documentclass{aastex}
%% use the following for preprints
%
%% preprint produces a one-column, single-spaced document:
%
\documentclass[preprint]{aastex}
\usepackage{emulateapj5, psfig, epsfig}
%
%% preprint2 produces a double-column, single-spaced document:
%
%\documentclass[preprint2]{aastex}
\slugcomment{Accepted by The Astronomical Journal}
\shorttitle{Globular Clusters in the Sgr Stream}
\shortauthors{Bellazzini, Ferraro \& Ibata}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% LATEX DEFINITIONS
\def\ltsima{$\; \buildrel < \over \sim \;$}
\def\simlt{\lower.5ex\hbox{\ltsima}}
\def\gtsima{$\; \buildrel > \over \sim \;$}
\def\simgt{\lower.5ex\hbox{\gtsima}}
%
% MATH FUNCTIONS:
\def\erf{\mathop{\rm erf}\nolimits} %error function
\def\sech{ \mathop{\rm sech}\nolimits} %hyperbolic sec
\def\csch{ \mathop{\rm csch}\nolimits} %hyperbolic csc
\def\arcsinh{\mathop{\rm arcsinh}\nolimits} %arc hyperbolic sin
\def\arccosh{\mathop{\rm arccosh}\nolimits} %arc hyperbolic cos
\def\arctanh{\mathop{\rm arctanh}\nolimits} %arc hyperbolic tan
\def\arccoth{\mathop{\rm arccoth}\nolimits} %arc hyperbolic cot
\def\arcsech{\mathop{\rm arcsech}\nolimits} %arc hyperbolic sec
\def\arccsch{\mathop{\rm arccsch}\nolimits} %arc hyperbolic csc
\def\arccot{\mathop{\rm arccot}\nolimits} %arc cot
\def\arcsec{\mathop{\rm arcsec}\nolimits} %arc sec
\def\arccsc{\mathop{\rm arccsc}\nolimits} %arc csc
\def\ylm{\mathop{\rm Y}_l^m\nolimits} %spherical harmonic
\def\ylmp{\mathop{\rm Y}_{l'}^{m'}\nolimits} %spherical harmonic primed
\def\real{\Re e} %real part
\def\imag{\Im m} %imaginary part
% UNITS:
\def\km{{\rm\,km}}
\def\kms{{\rm\,km\,s^{-1}}}
\def\mas{{\rm\,mas}}
\def\masyr{{\rm\,mas/yr}}
\def\kpc{{\rm\,kpc}}
\def\mpc{{\rm\,Mpc}}
\def\msun{{\rm\,M_\odot}}
\def\lsun{{\rm\,L_\odot}}
\def\rsun{{\rm\,R_\odot}}
\def\pc{{\rm\,pc}}
\def\cm{{\rm\,cm}}
\def\yr{{\rm\,yr}}
\def\au{{\rm\,AU}}
\def\g{{\rm\,g}}
\def\om{\Omega_0}
\def \ca {{\it ca.\/}}
\def \r {r$^{1/4}$ }
\def \magnitude {$^{\rm m}$}
\def\kr{${\cal K}_r$}
\def\kz{${\cal K}_z$}
\def\kzz{${\cal K}_z(z)$}
\def\mss{{\rm M}_\odot \rm pc^{-2}}
\def\msss{{\rm M}_\odot \rm pc^{-3}}
\newcommand{\fmmm}[1]{\mbox{$#1$}}
\newcommand{\scnd}{\mbox{\fmmm{''}\hskip-0.3em .}}
\newcommand{\scnp}{\mbox{\fmmm{''}}}
\newcommand{\mcnd}{\mbox{\fmmm{'}\hskip-0.3em .}}
\def\Aa{\; \buildrel \circ \over {\rm A}}
\def\AA{$\; \buildrel \circ \over {\rm A}$}
\def\yr{{\rm yr}}
\def\CompactFigs{0}
\def\UseFigs{1}
% MISCELLANEOUS:
% angles in degrees
\def\deg{^\circ}
%\degg produces degree symbol so that 3\sec5 produces 3.`5 with the degree
%symbol and the period aligned.
\def\degg{\hbox{$\null^\circ$\hskip-3pt .}}
%\sec produces arcsec symbol so that 3\sec5 produces 3."5 with the second
%symbol and the period aligned.
\def\sec{\hbox{"\hskip-3pt .}}
\def\half{{\scriptstyle{1\over2}}}
%\s produces tilde in mathmode or horizontal mode.
\def\s{\ifmmode \widetilde \else \~\fi}
\def\={\overline}
\def\scre{{\cal E}}
\def\spose#1{\hbox to 0pt{#1\hss}}
\def\larrow{\leftarrow}
\def\rarrow{\rightarrow}
\def\llangle{\langle\langle}
\def\rrangle{\rangle\rangle}
\def\etal{{\it et al.\ }}
\def\cf{{\it cf.\ }}
\def\eg{{ e.g.,\ }}
\def\ie{{ i.e.,\ }}
%\lta and \gta produce > and < signs with twiddle underneath
\def\lta{\mathrel{\spose{\lower 3pt\hbox{$\mathchar"218$}}
\raise 2.0pt\hbox{$\mathchar"13C$}}}
\def\gta{\mathrel{\spose{\lower 3pt\hbox{$\mathchar"218$}}
\raise 2.0pt\hbox{$\mathchar"13E$}}}
%\Dt and \dt put Newton's notation dots above upper and lower case chars
\def\Dt{\spose{\raise 1.5ex\hbox{\hskip3pt$\mathchar"201$}}} % upper case
\def\dt{\spose{\raise 1.0ex\hbox{\hskip2pt$\mathchar"201$}}} % lower case
\def\del{\nabla}
\def\delv{\bb\nabla}
\def\r{${\rm r^{1/4}}$}
\def\jla{J_{\lambda}}
\def\jmu{J_{\mu}}
\def\jnu{J_{\nu}}
\def\pomega{\varpi}
\def\sigla{\sigma_{\lambda}}
\def\sigmu{\sigma_{\mu}}
\def\signu{\sigma_{\nu}}
\def\dotsfill{\leaders\hbox to 1em{\hss.\hss}\hfill}
\def\sun{\odot}
\def\earth{\oplus}
\def\Gyr{{\rm\,Gyr}}
\def\kmsd{{\rm\,km/s/degree}}
\def\Rod#1{\noindent{\bf[$\diamondsuit$ #1]}}
\def\Mic#1{\noindent{\bf[$\heartsuit$ #1]}}
\def\Fran#1{\noindent{\bf[$\clubsuit$ #1]}}
\def\other#1{\noindent{\bf[$\spadesuit$ #1]}}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% END OF LATEX DEFINITIONS %%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{document}
\title{Building up the globular cluster system of the Milky Way. \\
The contribution of the Sagittarius galaxy}
\author{Michele Bellazzini, Francesco R. Ferraro}
\affil{INAF - Osservatorio Astronomico di Bologna, Via Ranzani 1, 40127,
Bologna, ITALY}
\email{bellazzini@bo.astro.it, ferraro@apache.bo.astro.it}
\and
\author{Rodrigo Ibata}
\affil{Observatoire de Strasbourg, 67000 Strasbourg, France}
\email{ibata@newb6.u-strasbg.fr}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{abstract}
We demonstrate that there is a clear statistical correlation between
the (X,Y,Z,$V_r$) phase-space distribution of the outer halo Galactic
globular clusters (having $10\kpc \le R_{GC}\le 40\kpc$) and the
orbital path of the Sagittarius dwarf spheroidal galaxy (Sgr dSph), as
derived by Ibata \& Lewis. At least 4 of the sample of 35 globular
clusters in this distance range were formerly members of the Sgr
galaxy (at the 95\% confidence level), and are now distributed along
the Sgr Stream, a giant tidal structure that surrounds the Milky Way.
This is the first instance that a statistically significant structure
associated with the Sgr dSph
has been detected in the globular cluster population of the Galactic
halo. Together with the four well-known globular clusters that are
located near the center of this tidally-disrupting dwarf galaxy, these
clusters constitute $\simgt 20$\% of the population of outer halo
($R_{GC} \ge 10\kpc$) clusters. The Sgr dSph was therefore not only
an important contributor to the halo field star population, but it
also had a significant role in the building-up of the globular cluster
system of the Milky Way.
\end{abstract}
%% Keywords
\keywords{(Galaxy:) globular clusters: general - Galaxy: halo -
Galaxy: formation - Galaxy: structure - galaxies: dwarf}
{}{}
\section{Introduction}
The quest to find ordered structures in the system of satellites of
the Milky Way galaxy dates back to 30 years ago when possible
alignments of globular clusters and/or dwarf galaxies along wide
streams were first noted \citep{hm69,ly76,k77,k79,ly82}. The mounting
consensus for scenarios in which the accretion of satellites has a
major role in the formation of the outer halo of the Galaxy
\citep{sz78,z93} prompted a new burst of such kind of studies since the
mid '90s up to the present day \citep{maj94,lynlyn,f95,dana,palma}.
Despite the many interesting
suggestions, none of the quoted studies was able to provide a
conclusive proof of the reality of the alignments, mainly because of
the overwhelming difficulty to assess the statistical significance of
structures formed by inherently small numbers of objects.
However, recent theoretical and observational achievements may help us
to look into the problem from a different and more fruitful
perspective:
\begin{enumerate}
\item{} N-body simulations of the process of Galaxy assembly, starting from
standard cosmological conditions (Cold Dark Matter - CDM),
strongly suggest that the hierarchical
merging of satellites is the main driver of galaxy formation
\cite[see][and references therein]{DMsub,klyp}.
\item{} Independent of the assumed cosmology, it has been demonstrated that
the accretion/disruption of satellites into the halo of a
larger galaxy may leave
long-lived relics in the form of streams of stellar (and/or dark
matter) remnants that remain aligned to the orbital path of the
parent satellite \citep{katy1,katy2,carb,orb,lucio}.
\item{} Convincing observational evidence of the clumpy and ``filamentary''
nature of the Galactic halo have been provided by many different groups
\citep[e.g][]{sdss,vivas,dohm,yanny,ivez,carb,ami1,ami2,maj99}.
\end{enumerate}
The {\em in vivo} example of a satellite accretion/disruption is
provided by the Sagittarius dwarf Spheroidal galaxy \cite[Sgr
dSph;][]{s1,s2}, which is currently merging with the Milky Way, and is
carrying its own globular cluster system (i.e., M~54, Ter~8, Arp~2 and
Ter~7, previously believed to be normal Galactic globulars). There is
now clear observational evidence that the Sgr dSph is loosing stars
under the strain of the Milky Way tidal field. These tidally-removed
stars are found along a huge (and quite coherent) stream extending all
over the sky \citep[Sgr Stream, see][and references
therein]{carb,dohm,david,sdss}, tracing the orbit of the parent
galaxy.
\citet[][hereafter IL98]{orb}, and \citet{carb} have simulated the
evolution of the Sgr dSph over several orbital periods (P $\sim 1$
Gyr), computing the orbit of the galaxy as well as the phase-space
distribution of the debris under different assumptions about the
flattening of the CDM halo. The initial conditions of the simulations
were based on the known position and radial velocity of Sgr dSph and
on its proper motion as estimated by \citet{s2,carb}. The orbit has a
planar rosette structure, with the pole of the orbit located at
[$\ell=90^\circ$, $b=-13^\circ$] (i.e. a nearly polar orbit), and peri- and
apo-Galactic distances of $15\kpc$ and $60\kpc$ respectively. The
derived orbit has been successfully compared with the observed position
of the Sgr Stream \citep{carb,orb-sds}, providing also remarkable
indications that the dark halo of the Milky Way is nearly spherical.
In this framework it is a tantalizing application to look for other
halo globulars that may be correlated with the orbital path of the Sgr
dwarf, and which could be lying in the Sgr Stream. In particular, we
look for the phase-space coincidence of outer halo globulars with the
computed orbit of the Sgr dSph from 1~Gyr ago up to the present day,
searching for the most recent episodes of globular cluster loss,
i.e. the ones whose traces are most likely to be still detectable.
\section{Looking for structures}
For our comparison we selected from the catalogue by \citet{harris}
the 35 globular clusters in the range of galactocentric distance
$10\kpc \le R_{GC}\le 40\kpc$. Among these, 33 have also measured
radial velocity $V_r$. For sake of brevity and clarity we will call
this sample the Outer Halo Sample (OHS), in the following. With this
selection we avoid the central part of the Galactic halo where it is
less likely that ordered structures can survive for a long time, and
we leave out of the sample the handful of clusters lying outside of
$R_{GC}\ge 60\kpc$, a region that lies beyond the Sgr Stream according
to the IL98 orbit\footnote{There is a 28 kpc wide gap in the
radial distribution of Galactic globular clusters
\cite[see, e.g.][and references therein]{z85}.
There are only five
clusters beyond $R_{GC}=40$ kpc, namely Pal~14, Eridanus, Pal~3, NGC~2419,
Pal~4, and AM1. Their respective galactocentric distances are $R_{GC}=$
65, 83, 90, 98, 99, and 117 kpc, much beyond the apogalacticon of the
Sgr orbit. The adopted outer radial threshold \cite[$R\le 40 \kpc$, quite
similar to the one adopted by][i.e., $R_{GC}\le 36 \kpc$]{palma}
provides the selection of a homogeneous sample without significant gaps.}.
The adopted OHS {\em does not include the known
Sgr globulars}, to avoid the detection of the obvious signal of their
clustering around the center of the Sgr galaxy.
In Figure~1 we show the OHS clusters (small solid circles) and the Sgr
orbit in the planes formed by the rectangular Galactocentric
coordinates\footnote{This Galactocentric coordinate system is defined
such that the origin lies at the Galactic Center; at the Solar
position, $(-8,0,0)$, the $Y$-axis points in the direction of Galactic
rotation; while the $Z$-axis towards the North Galactic Pole.}
($X,Y,Z$, in kpc) and in the $R_{GC}$ [kpc] vs. $V_r$ [km/s] plane.
The large full circles are the known Sgr globulars, which we also show
in the plots for completeness. Note that these clusters lie around
the end of the orbit corresponding to the present time ($t=0$). We
highlight (with encircled solid circles) six more clusters that lie
remarkably close to the orbit in all the considered planes. These
clusters are: Pal~12 \citep[whose association to the Sgr Stream has
been already established by][]{pal12,dav12},
NGC~4147, NGC~5634, NGC~5053,
Pal~5 and Ter~3. Is this association {\em real} or could it be the
mere occurrence of a chance alignment? Though chance alignments in
the four-dimensional phase space (X,Y,Z,$V_r$) are not expected to be
very likely, the key point is to quantify the probability that the
observed structure could have originated from a statistical
fluctuation. To do this we will compare the observed distribution -
and its phase space distance to the Sgr orbit - with synthetic samples
(having the same dimension as the OHS) extracted from a model
representing an unstructured parent halo.
\begin{figure*}
\figurenum{1}
\centerline{\psfig{figure=fig1.ps}}
\caption{The distribution of Galactic globular clusters with $10\kpc
\le R_{GC}\le 40\kpc$ (OHS; small full circles) in the Y vs. Z, X
vs. Z, X vs. Y, and $V_r$ vs. $R_{GC}$ planes. The large full circles
are the known Sgr globulars M~54, Arp~2, Ter~8 and Ter~7 while the
encircled points mark the clusters that are near the Sgr orbital path
in all of the planes shown here (NGC~4147, Pal~12, NGC~5634, Pal~5,
NGC~5053, and Ter~3). The continuous line is the orbit of the Sgr
dSph since the present time (at the position of the Sgr clusters) to 1
Gyr ago, according to \citet{orb}.}
\end{figure*}
The most conservative comparison that can be made is with a model that
closely resembles the observed radial and velocity distribution of the
OHS. Figure~2 (upper panel) shows that the cumulative radial
distribution of the OHS is well reproduced by a spherical halo model
with a density distribution $\propto R^{-1.6}$. A Kolmogorov-Smirnov
(KS) test shows that the probability that the OHS is drawn from the
$\Phi \propto R^{-1.6}$ model is $\simeq 90$\%. On the other hand,
the probability that the same sample is drawn from the other two
models shown for comparison ($\Phi \propto R^{-1.0}$, and $\Phi
\propto R^{-2.5}$) is $\le 15$\%. Doubts may be cast on the
appropriateness of a spherical model. It may be conceived that if the
{\em true} parent halo is flattened, some excess of clustering of the
observed points along an orbit with low inclination may artificially
emerge in the comparison with a spherical model. This is clearly not
the case, however, since the IL98 orbit is nearly polar, i.e. it is
almost perpendicular to the Galactic Plane (see Figure~1). In the
lower panel of Figure~2 it is shown that the observed distribution of
radial velocity of the OHS is well reproduced by a Gaussian
distribution with $<V_r> = -38\kms$ and $\sigma_V =
175\kms$. According to a KS test the probability that the observed
sample is drawn from the model distribution is $\simeq 90$\%.
In the following simulations we extract all the synthetic samples from
a spherical and isotropic model with $\Phi(R_{GC}) \propto
R_{GC}^{-1.6}$ and with the Gaussian distribution of radial velocities
shown in Figure~2. For each simulated cluster (as well as for all the
OHS ones) we computed the spatial distance from the nearest point in
the Sgr orbit ($D_{orb}$, in kpc) and the difference between their
radial velocity and the one predicted from the computed orbit at that
point ($\Delta V_r = V_r(obs) - V_r(orb)$).
\begin{figure*}
\figurenum{2}
\centerline{\psfig{figure=fig2.ps}}
\caption{Upper panel: the cumulative radial distribution of the OHS
globulars (thick continuous line) is compared to three different model
distributions in the considered radial range ($10\kpc \le R_{GC}\le
40\kpc$). According to a Kolmogorv-Smirnov test, the probability that
the observed distribution is drawn from the model $ \Phi \propto
R^{-1.6} $ is $P \simeq 90$\%, while the other two models shown score
$P \le 15$\%. Lower panel: the observed cumulative distribution of
radial velocities for the same clusters (thick continuous line) is
compared with a Gauss distribution with $<V_r> = -38\kms$ and
$\sigma_V = 175\kms$. Also in this case the probability that the
observed distribution is drawn from the considered model is $P \simeq
90$\%. }
\end{figure*}
In Figure~3, the $D_{orb}$ values of the selected clusters (large
filled circles) are plotted against their $\Delta V_r$. The encircled
points are the six clusters highlighted in Figure~1. A sample of 10000
synthetic clusters (dots) extracted from the adopted model is also
shown, for comparison, in the upper panel of Figure~3. The OHS
clusters show a remarkable over-density toward the Sgr orbit, that
lies in the origin of the axis in the considered plane. The dashed
dotted lines enclose the points whose observed radial velocity is
within $\pm 60\kms$ of the velocity predicted by the IL98 orbit. Note
that the expected velocity dispersion of the Sgr debris along the Sgr
Stream is $\sigma \sim 60\kms$, according to \citet{orb-sds}. The
continuous vertical segments are placed at $D_{orb}= $6, 12, and 18
kpc.
The lower panel of Figure~3 is arranged in the same way, but in this
case the dots represent the distribution of the points in the
best-fitting \citet{carb} simulation that retains a bound core to the
present day. The boldface dots are the particles that remain bound to
Sgr {\em or were bound less than 3 Gyr ago}, while the ordinary dots
are particles that were already unbound at that time. Note the
remarkable similarity with the distribution of the OHS clusters. In
particular, the correlation between the particles that have flown in
the Sgr Stream in recent times (less than 3 Gyr ago) and the OHS
clusters with $|\Delta V_r| < 60\kms$ at any $D_{orb}$ is quite
striking.
We define $N_{D<x}$ as the number of synthetic or real clusters having
$-60\kms < \Delta V_r < +60\kms$ and $D_{orb}<x \kpc$. For example,
it can be seen from Figure~3 that there are five OHS clusters with
$|\Delta V_r| < 60\kms$ and $D_{orb} < 6\kpc$, thus $N_{D<6}=5$. In
the same way we find $N_{D<12}=10$ and $N_{D<18}=14$ from the observed
sample. Now the question is {\em what is the probability that a
sample having $N_{D<x}$ greater or equal to the observed one
($N_{D<x}^{obs}$) is drawn from the assumed unstructured model?} To
answer this question we randomly extracted 10000 samples of 32
synthetic clusters from the assumed model, and for each of them we
measured $N_{D<6}$, $N_{D<12}$, and $N_{D<18}$.
\begin{figure*}
\figurenum{3}
\centerline{\psfig{figure=fig3.ps}}
\caption{The distribution of the OHS globular clusters about the
predicted orbit of Sgr. The parameter $D_{orb}$ is the distance of the
cluster from the nearest point of the orbit, while $V_r(obs)-V_r(orb)$
is the difference between the observed radial velocity and the
prediction of the computed orbit (at the nearest point in the orbit).
The observed distribution (large full circles) is compared with the
distribution of 10000 points randomly drawn from our assumed halo
model (upper panel) and with the distribution of the Sgr debris as
computed by \citet{carb} (lower panel). This N-body simulation
displayed in the lower panel is the model by \citet{orb} that gives
the best fit to the positions and velocities of the carbon stars
observed in the Sgr stream, while retaining a bound center (their
dwarf galaxy model ``D1'', and halo model ``H2'', integrated in a
Galactic potential with circular velocity $v_c=200\kms$ at $50\kpc$).
The boldface dots show particles that remain bound to Sgr or were
bound less than $3\Gyr$ ago, whereas the ordinary dots were already unbound
at that time. Clearly, several clusters in the OHS follow the
expected trend of the recently disrupted Sgr debris. The encircled
points represent the clusters similarly put in evidence in Figure~1.
The horizontal dashed-dotted lines enclose the range $-60\kms \le
\Delta V_r \le 60\kms$, while the continuous vertical segments are
located at $D_{orb} = 6$, 12 and 18~kpc.}
\end{figure*}
Figure~4 reports the distributions of $N_{D<6}$ (upper panel),
$N_{D<12}$ (middle panel), and $N_{D<18}$ (lower panel) for the 10000
simulated samples. The respective observed values are indicated by a
dashed line. In all of the considered cases the mean and modal
$N_{D<x}$ values of the distribution are much smaller than the
observed ones. $N_{D<6}\ge 5$ occurs for only 200 simulated samples in
10000 (2\% of the cases), $N_{D<12}\ge 10$ occurs only for 42 samples
in 10000 ($0.4$\%), and $N_{D<18}\ge 14$ occurs only in 2 cases in
10000 ($0.02$\%). It can be concluded that it is highly improbable
that the observed clustering of the OHS globulars around the Sgr orbit
occur by chance alignment. It is important to remark again that the
model from which the simulated samples have been extracted has the
same $R_{GC}$ and $V_r$ distribution of the observed sample. Thus {\em
the Sgr orbit appears to be a strongly preferred subset of the
phase-space for the Galactic globulars} in the considered range of
$R_{GC}$.
Finally, we note that the simulated samples that have a significant
probability of realization (e.g., $P\ge 20$\%) have $N_{D<6}\le 3$,
$N_{D<12}\le 6$, and $N_{D<18}\le 8$, significantly lower than the
observed values. Hence, the significance of the detected phase-space
structure cannot be due to the actual correlation of just a couple of
clusters with the Sgr orbit. According to the distributions shown in
Fig.~4, the probability that $N_{D<18}\ge 10$ is $P\simeq 5$\%.
Therefore it can be stated, with 95\% confidence, that at least 4 real
associations are needed to produce the observed signal of
$N_{D<18}=14$.
The above test demonstrates that the observed phase-space clumping
around the orbit of the Sgr dwarf galaxy is highly unlikely to have
occurred by a chance coincidence, if the halo globular cluster
population is indeed distributed according to the simple spherical
model described above. To address this concern, we performed a second
test, using ``bootstrapped'' artificial data sets constructed from the
real OHS. The artificial samples were constructed by rotating the
position of each real globular cluster by a random azimuthal angle
about the Galactic center, and introducing a random flip perpendicular
to the Galactic plane. This implicitly assumes, of course, that the
globular cluster distribution is symmetric about the Galactic center,
as well as above and below the Galactic plane. A slight complication
arises from the fact that we do not know the proper motion of many of the
clusters, so we cannot deduce the heliocentric radial velocity that
the cluster should have at its new position. We therefore have to
assume some model for the halo velocity distribution; we take a simple
Gaussian model, and draw random realizations of the total space
velocity $\vec{V}$ consistent with the observed heliocentric velocity
$V_r$ (that is, we calculate the conditional probability of $\vec{V}$
given $V_r$). Having defined thereby the 3-dimensional velocity
vector, we rotate the vector to the new position, and project it along
the line of sight to the Sun (corrections for the peculiar motion of
the Sun and for Galactic rotation are made). With a Gaussian model
that has $\sigma=175\kms$ and $V_r=-38\kms$, we find that out of 10000
artificial data sets, $F(N_{D<6} \ge 5) = 399$, $F(N_{d<12} \ge 10) =
102$, and $F(N_{D<18} \ge 14) = 18$. For comparison, for a model with
$\sigma=110\kms$ (or $\sigma=150\kms$) and $V_r=0\kms$, the statistics
are as follows: $F(N_{D<6} \ge 5) = 463 (425)$, $F(N_{D<12} \ge 10) =
140 (121)$, and $F(N_{D<18} \ge 14) = 35 (28)$, where the result in
brackets refers to the higher velocity dispersion model. According to
these tests using ``bootstrapped'' random data sets, the observed
phase-space clumping is highly unlikely to have occurred by chance, in
agreement with our previous test. The fact that the observed
structure has a slightly lower statistical significance with the
``bootstrapped'' samples than was deduced from the previous test with
its perfectly isotropic and isothermal random samples is not
surprising. This is presumably a consequence of occasional chance
reappearances of (part of) the structure of the real OHS in the
``bootstrapped'' samples.
\begin{figure*}
\figurenum{4}
\centerline{\psfig{figure=fig4.ps}}
\caption{Results from the random drawing of 10000 synthetic GC samples
(equivalent to the observed one) from the assumed halo model.}
\end{figure*}
As a final remark, we note that the region of the $\Delta V_r -
D_{orb}$ plane with $|\Delta V_r| < 60\kms$ and $D_{orb}<18$ kpc is
also populated by particles that become unbound more than 3 Gyr
ago. Thus it has to be considered the possibility that some of the OHS
clusters found in that region drifted into the Sgr Stream in more
ancient times. For instance, this can be the case for NGC~5634, as we
argue elsewhere \citep{n5634}. On the other hand, the discussion of
the individual OHS clusters is beyond the scope of the present
analysis.
\subsection{Proper Motions}
Thanks to the painstaking effort of a few teams of astronomers,
estimates of the proper motions (hereafter PMs) of 41 Galactic
globular clusters are now available \cite[see][for complete lists and
references]{dana,palma}. Most of these measures have quite sizeable
uncertainties, ranging from $\sim 0.2 $ mas/yr to $\sim 2$ mas year.
Moreover, subtle - and unaccounted - systematics may still affect
them, as suggested by the large differences between independent
estimates of the motion of the same cluster. In the list by
\citet{palma} there are 14 cluster with more than one PM estimate. The
average absolute difference among independent estimates for the same
cluster are $1.72\pm 1.29$ mas/year in $\mu_{\alpha}cos \delta$ and
$2.90\pm 2.50$ mas/year in the $\mu_{\delta}$ component. Differences
much larger than the quoted errors (up to $\sim 10$ times) are common
and a reasonable assumption of the {\em minimum} uncertainty of the
proper motion estimates is probably $\sim 1$ mas/year in each
component. Hence, while available PMs may be very valuable, for
instance, to characterize the {\em kind} of orbit followed by a given
globular, they are expected to have a modest constraining power in the
application presented in this paper. Such constraining power is also
lowered by the fact that PMs predicted by the model we are comparing
with carry their own, non negligible, uncertainty (typically $\sim
0.85$ mas/yr in each component, which we estimate from the intrinsic
dispersions in distance, position and velocity in the N-body stream
models, assuming a 20\% uncertainty in the Galactic circular velocity
at $50\kpc$). Radial velocities, being unaffected by uncertainties in
the distance, are more reliably predicted by the model and are easily,
and much more accurately, measured. For these reasons we preferred to
limit the bulk of our analysis to the the (X,Y,Z,$V_r$) phase-space.
\begin{figure*}
\figurenum{5}
\centerline{\psfig{figure=fig5.ps}}
\caption{The predicted direction of proper motions are compared with
the observed ones for the subset of candidate Sgr Stream cluster for
which estimates of proper motion are available \cite[from the list
by][]{palma}. When two independent estimates of proper motion are
available for the same cluster we report both: one as a large filled
circle and the other as a small filled circle, and the two symbols are
joined by a long dashed line. The large dot of Pal~5 is the
estimate quoted by \citet{palma} as Cudworth et al. (2001), the small
dot is from \citet{shweitz}.
The solid line is the
locus of perfect agreement between predictions and observation, the
dotted lines are displaced by $\pm 45$ deg. The error-bars are derived
from the propagation of the uncertainties in the PM estimates of
clusters as reported by \citet{dana,palma} and by the uncertainties in
the model predictions, as described in \S2.1. Panel (a): comparison
between PMs including the S.r.m. Panel (b): comparison between PMs
corrected for the S.r.m.}
\end{figure*}
Yet it may be interesting to compare the PMs predicted by the IL98
model for the clusters that are candidate members of the Sgr Stream
(listed in Table 1, see below), with the observational estimates. The
orientations (directions) of the PM vectors are expected to be less
sensitive to systematics with respect to the actual moduli. In Fig.~5
we compare the predicted and observed {\em directions} of the PM
vectors of the seven clusters listed in Tab.~1 for which PM estimates
are available. Four of these seven clusters have two independent PM
estimates: in these cases we report the comparison with both values
(see caption). We stress that this kind of comparison {\em cannot have
a serious impact on our previous conlusions} except if both the
following conditions were simultaneously fulfilled: (a) if accurate PM
estimates were available for all the considered clusters (or the large
majority of them), and (b) if less than 3-4 clusters were found to
have observed PMs in agreement with the predictions. This case would
imply that the correlation among OHS clusters and the Sgr orbit we
detect in the (X,Y,Z,$V_r$) phase-space is largely due to a very
unlikely (but not impossible) chance alignment in four dimensions. On
the other hand some (up to $\sim 10$, actually) of the clusters listed
in Table 1 {\em are expected not to belong to the detected structure}.
In panel (a) the predicted proper motions include a correction for the
Solar reflex motion (hereafter S.r.m.), assuming a Galactocentric
distance of $8\kpc$, a circular velocity of $220\kms$ and the peculiar
motion of the Sun as derived by \citet{dehnen}. The predicted PM
directions of all the seven clusters agree with the observed values to
within $\pm 45\deg$. The adoption of alternative PM estimates does not
change significantly the observed correlation. However, the result of
this comparison may be affected by the inclusion of the Solar reflex
motion. Hence, in panel (b) we present the comparison between the
plain (uncorrected for the S.r.m) model predictions and the observed
PM directions after correction for the S.r.m. The result is very
similar to that presented in panel (a) despite of the larger errors
involved. The only cluster whose predicted and observed PM directions
differ by more than $45 \deg$ is Pal~12, e.g. the only one for which
independent evidences of membership to the Sgr Stream are already
available \citep{dav12}.
In conclusion, it is found that the available proper motion estimates
are fully compatible with the results presented in \S2. Given the high
degree of correlation among predicted and observed PM directions shown in
both panels of Fig.~5 one may also be tempted to draw firm conclusion
on the actual membership of individual clusters. Nevertheless, in our
view, this may still be an hazard.
If one looks at
the amplitude of the individual corrections it is found that (a) all
the considered clusters have at least one component of the PM vector
that is corrected by more than 50 \% of the observed value, (b) 4 out
of seven clusters have at least one component whose S.r.m correction
is larger that 100 \% of the observed values and, (c) corrections as
large as 200 - 400 \% are indeed applied in some cases. Given the
remarkable uncertainties in the observed PMs and the uncertainties
involved in the S.r.m correction it is easy to conclude that currently
measured PMs are not good tracers of the actual motion of the
considered clusters and are instead predominantly measures of the
Solar reflex motion. These considerations fully support the approach
followed in \S2, e.g. using only positions and radial velocity measurements
in the search for statistical structures in the globular cluster system of
the Milky Way.
\subsection{Comparison with previous analysis}
The possible association of Galactic globulars with the Sgr dSph
galaxy has been considered before \citep{irwin, dana, pal12, pal13,
palma}. All these studies strongly rely on PM estimates. Furthermore
they are centered on testing the actual association of individual
clusters to the Sgr galaxy and its remnants. On the other hand our
approach and our results are statistical in nature. We find a
statistically significant clustering of OHS clusters along the Sgr
orbit but we have no possibility to check individual membership,
except for the limited test shown in \S2.1. We can only provide a
list of possible members, ranked according to their distance to the
Sgr orbit in the (X,Y,Z,$V_r$) phase-space (e.g., Tab.~1). Hence, the
comparison with the above quoted studies can be made only in a broadly
general sense.
\citet{dana} performed a thorough study of the orbits of the Galactic
globulars for which PM estimates are available. In doing this, they
identify three metal-poor clusters having orbits typical of the thick
disc (instead that of the halo, as expected) and they check the
hypothesis that the peculiar orbits of these clusters (namely
NGC~6254, NGC~6626 and NGC~6752) could be due to their former
membership to the Sgr dSph, concluding that this possibility is
unlikely. All these clusters have $R_{GC}<10$, and are thus are not
included in the OHS considered here and we cannot comment further on
the \citet{dana} results.
Following the suggestion of \citet{irwin}, \citet{pal12} modeled the
(PM based) orbits of the Sgr dSph and of Pal~12 and concluded that the
available observations are compatible with the possibility that the
cluster was torn from the galaxy $\sim 1.4 - 1.7$ Gyr ago. Our
analysis confirms this result.
\citet{pal13} obtained an estimate of the PM of Pal~13 and compared
the general characteristics of the integrated orbit of the cluster
(total energy, angular momentum, eccentricity, perigalactic and
apogalactic radii) to those of five satellite galaxies of the Milky
Way, including Sgr. They conclude that a common origin for Pal~13 and
Sgr is unlikely. However we note that (a) Sgr provides (by far) the
best match to the orbital parameters of Pal~13 with respect to any
other considered galaxy (see Siegel et al.'s Tab.~8), the maximum
discrepancy being $\sim 40$\% in angular momentum, and (b) the
orbital parameters are provided without any uncertainty, as if the
adopted PMs were perfectely known. In our view this shortcoming
seriously weakens the conclusion by these authors. Here we adopt the
same PM estimate obtained by \citet{pal13}, reaching the conclusion
that Pal~13 is a possible member of the Sgr Stream by direct
comparison with a model that includes the Sgr Stream itself.
\citet{palma} (hereafter PMJ02) performed a global search for
structures made of Galactic satellites, adopting the technique of the
``orbital poles'', originally introduced by \citet{lynlyn}. In this
kind of analysis one searches for intersections among the great
circles described by the familiy of possible poles of the orbit of any
given satellite. As stated by PMJ02 ``any set of objects having
common origin and mantaining a common orbit, no matter how spread out
in the sky, will have great circle pole families (GCPFs) that
intersects at the same pair of antipodal points on the celestial
sphere''. Adding the information from PMs, PMJ02 limit the valid pole
family of each satellite to an arc of its great circle of orbital
poles (arc segment pole family, ASPF). With this technique they select
a set of candidate-former-members of the Sgr galaxy (Ter~7, Ter~8,
Arp~2, M~54, M~53, NGC~5053, Pal~5, M~5, Pal~12 and NGC~6356) by
requiring that their GCPF come within 5 deg from the Sgr ACPF and
$6\le R_{GC} \le 36$ Kpc. The first four clusters in their list are
the well known present members of the galaxy. They rise doubts of the
clear association of each of the other candidate members. For
instance, they exclude NGC~5053 and M~53 because their metal content
is lower than any of the known Sgr globulars\footnote{This is not
completely true since M~53 has the same metallicity as Ter~8, i.e.
$[Fe/H]=-1.99$ \cite[see][]{mont,taft}. See also below, for further
discussion about non-kynematic criteria of selection of candidate Sgr
members.}. With the same arguments adopted by \citet{pal13} for
Pal~13 they argue that Pal~5 and Pal~12 are unlikely members of the
family, but they cannot rule out the possibility. Finally, they
conclude that despite the counter-arguments presented they mantain
their original list of candidates, with the exception of NGC~5053.
Elsewhere in the paper, they also note that NGC~5466 has energy and
angular momentum within 1-$\sigma$ of those of the Sgr galaxy.
NGC~6356, M~5 and M~53 have $R_{GC} < 10$ kpc, hence are not included
in our sample. The possible association of NGC~5466 with the Sgr dSph is
confirmed by our analysis. The possible association of NGC~5053, Pal~5
and Pal~12 is also confirmed by the present work. PMJ02 also suggested
a possible association of NGC~4147 with a group of clusters whose
orbits may be related to those of the Magellanic Clouds, but they
state that the result is very uncertain. Furthermore we note that on
the same basis the possible association of M~53 (elsewhere classified
as a candidate for association with Sgr) with the Small Magellanic
Cloud and Ursa Minor is also suggested by PMJ02. We conclude that
there is no serious disagreement between the present analysis and the
one by PMJ02.
The present contribution, however, {\em provides the first proof that
the orbit of Sgr is a preferential subset of phase-space for globular
clusters inhabiting the outer halo of the Milky Way}, and we place
this result on a sound statistical basis, a result not accomplished by
any of the previous studies. The success of the present analysis is
due to the direct comparison of the positional and kinematical
properties of the OHS clusters with a realistic model of the orbit of
the Sgr dSph and its relics, that has been previously tested against
independent observations \citep{carb,orb-sds}.
Finally, we shall shortly comment on ``non-kinematic'' criteria that
have been used to assess the association of a given cluster to the Sgr
galaxy. Many authors \cite[see, e.g.][and references
therein]{dana,palma} have discussed the likeliness of the membership
of their candidates on the basis of the similarity of their
metallicity, Horizontal Branch (HB) morphology and/or structural
parameters with those of the four known Sgr clusters. In our opinion
this approach may not be very useful. It is quite clear that Sgr was
a much larger and complex system in the past and its present status
(as well as its present GC system) may be not fully representative of
its original range of properties. If, by chance, the cluster Ter~7
had been lost by the Sgr galaxy in the previous perigalactic passage,
this kind of criteria would have {\em erroneously} classified it as a
bad candidate for Sgr membership, since Ter~7 is much more metal rich
($[Fe/H]\simeq -0.5$) than any of the other clusters that are still
present in the main body of Sgr dSph ($[Fe/H]<-1.5$). Moreover, the
existence of population and/or metallicity gradients may have driven
the preferential loss of metal poor populations \cite[see][and
references therein]{alard,sdss,n5634}, making the actual low
metallicity limit of the Sgr stellar population quite uncertain. The
judgement of the likelihood of membership on the basis of the HB
morphology is even less justified since it is well known that clusters
of different morphologies do actually co-exist in real galaxies
\citep[see, e.g.,][and references therein, for the case for
Fornax]{fornax}.
A much more reliable discriminant may be provided, in the future, by
the detailed abundance patterns of the clusters. For instance, the
behaviour of $\alpha$ elements as a function of $[Fe/H]$ is determined
by the star formation history of the parent galaxy \citep[see][and
references therein]{mcw} that was (probably) not the same in the Sgr
dSph and in the Galactic Halo. In this context it is interesting to
note that Pal~12, Ru~106 \citep{brown} and Pal~5 \citep{smith} have
been found to be less $\alpha$-enhanced than Galactic Halo globulars
of similar metallicity.
For the above reasons, we rely only on phase-space parameters in our
analisys, recalling that by searching the lost relics of the Sgr
system we are trying the reconstruct its original properties, not the
present day ones.
\section{Conclusions}
We have demonstrated that there is a coherent structure in the
phase-space distribution of the outer halo globular cluster of the
Milky Way, strongly correlated with the orbital path of the Sgr dSph
galaxy. This correlation cannot have originated by chance, as a
random realization of an unstructured halo, if such halo has the same
distribution of galactocentric distance and radial velocity as the
considered OHS clusters.
Several of the OHS globular clusters have spatial positions and radial
velocities compatible with the hypothesis that
they are following the same orbit as the Sgr dSph, i.e. they probably
belong to the Sgr Stream. Furthermore the spread in phase space
around the Sgr orbit of such clusters is similar to that predicted for
the Sgr Stream population, according to the numerical simulations by
\citet{carb}.
We conclude, with 95\% confidence, that {\em at least 4 (out of 32) of
the OHS clusters are physically associated with the Sgr Stream and
were former members of the Sgr dSph}. It should be noted however,
that the analysis presented in this contribution is best suited to
identify clusters that became unbound relatively recently. In
principle, more OHS globular clusters could be associated with the Sgr
stream.
In Table~1 we report the list of the selected globulars having
$|\Delta V_r|<60$ Km/s and $D_{orb}<18\kpc$, in order of increasing
$D_{orb}$. The proper motions predicted by the IL98 model are also
reported in Table~1. Since the adopted analysis is statistical in
nature it is clear that some of the clusters listed in Tab.~1 may be just
occasional interlopers of the Sgr orbit (up to 10 over 14 in the worst
case, according to our estimate). The comparison with the available
proper motion estimates shown in \S2.1 does not provide any strong
indication in this regard. While the
actual membership to the Sgr Stream of each individual cluster can be
firmly established only with very accurate proper motion measures (as
will be achieved by dedicated space missions like GAIA or SIM), it is
reasonable to assume that the clusters with the smaller $D_{orb}$ are
the ones for which the membership is more likely (see \S2.1). We make
the hypothesis that the six clusters more closely associated with the
Sgr orbital path (i.e. those put in evidence in Figures~1 and 3) are
former members of the Sgr dSph. In this case the number of globular
clusters originally in the Sgr galaxy is $N_{cl}=10$. If we adopt for
the Sgr dSph the same {\em globular cluster specific frequency}
\citep[$S_N=N_{cl}10^{0.4(M_V+15)}$][]{hvdb} of the only other
Galactic dSph that has a globular cluster system, i.e. Fornax
$S_N\simeq 26$, we obtain an estimate of the total absolute magnitude
of Sgr before the occurrence of any tidal stripping, $M_V\simeq -14$.
\cite{sdss} estimated that in the Sgr Stream there are as many stars
as in the present undisrupted body of the galaxy, thus the {\em ab
initio} total luminosity of the Sgr dSph was roughly two times the
present value. Hence the total absolute magnitude at that time was
$M_V-\log(2)\simeq -14.1$ \citep[where $M_V=-13.4$, from][]{M98}. The
excellent agreement between the two independent estimates of the
initial $M_V$ fully supports the plausibility of the proposed
scenario.
According to the results presented, it emerges that the Sgr dSph was
not only an important contributor of halo field stars
\citep{ivez,yanny,sdss} but it also had a significant role in the
building-up of the globular clusters system of the Milky Way.
\acknowledgments
M.B. and F.R.F. acknowledge the financial support to this research by
the italian {Ministero dell'Universit\'a e della Ricerca Scientifica}
(MURST) through the grant p. 2001028879, assigned to the project {\em
Origin and Evolution of Stellar Populations in the Galactic Spheroid}.
This research has made use of NASA's Astrophysics Data System Abstract
Service.
%%THEBIBLIOGRAPHY
\begin{thebibliography}{300}
%
\bibitem[Alard(2001)]{alard} Alard, C., \aap, 377, 389
\bibitem[Bellazzini, Ferraro \& Ibata(2002)]{n5634} Bellazzini, M., Ferraro,
F.R., \& Ibata, R., 2002, \aj, in press, (astro-ph/0205148)
\bibitem[Brown, Wallerstein \& Zucker]{brown} Brown, J.A., Wallerstein, G., \&
Zucker, D., 1997, \aj, 114, 180
\bibitem[Buonanno et al.(1998)]{fornax} Buonanno, R., Corsi, C.E., Zinn, R.,
Fusi Pecci, F., Hardy, E., \& Suntzeff, N.B., 1998, \apj, 501, L33
\bibitem[Da Costa \& Armandroff(1995)]{taft} Da Costa, G.S., \& Armandroff,
T.E., 1995, \aj, 109, 2533
\bibitem[Dehnen \& Binney(1998)]{dehnen} Dehnen, W., Binney J., 1998, \mnras,
298, 387
\bibitem[Dinescu, Girard \& van Altena(1999)]{dana} Dinescu, D.I., Girard, T.M.,
\& van Altena, W.F., 1999, \aj, 117, 1792
\bibitem[Dinescu et al.(2000)]{pal12} Dinescu, D.I., Majewski, S.R.,
Girard, T.M., \& Cudworth, K.M., 2000, \aj, 120, 1892
\bibitem[Dohm-Palmer et al.(2001)]{dohm} Dohm-Palmer, R.C., Helmi, A., Morrison,
H., Mateo, M., Olszewski, E.W., Harding, P., Freeman, K.C., Norris, J.,
\& Shectman, S.A., 2001, \apj, 555, L37
\bibitem[Fusi Pecci et al.(1995)]{f95} Fusi Pecci, F., Bellazzini, M., Cacciari,
C., Ferraro, F.R., 1995, \aj, 110, 1664
\bibitem[Harris(1996)]{harris} Harris, W.E. 1996, \aj, 112, 1487
\bibitem[Harris \& Van den Bergh(1981)]{hvdb} Harris, W.E., \& Van den Bergh,
S., 1981, \aj, 86, 1627
\bibitem[Helmi(2001)]{ami2} Helmi, A., 2001, \nat, 412, 25
\bibitem[Helmi et al.(1999)]{ami1} Helmi, A., White, S., de Zeeuw, P.T., Zhao,
H., 1999, \nat, 402, 53
\bibitem[Hodge \& Michie(1969)]{hm69} Hodge, P.W., \& Michie, R.W., 1969, \aj,
74, 581
\bibitem[Ibata et al.(1994)]{s1} Ibata, R.A., Irwin, M.J., \& Gilmore, G., 1994,
Nature, 370, 194
\bibitem[Ibata et al.(1997)]{s2} Ibata, R.A., Wyse, R.F.G., Gilmore, G.,
Irwin, M.J., \& Suntzeff, N.B., 1997, \aj, 113, 634
\bibitem[Ibata \& Lewis(1998)]{orb} Ibata, R.A., \& Lewis, G.F., 1998, \apj,
500, 575 (IL98)
\bibitem[Ibata et al.(2001a)]{carb} Ibata, R.A., Lewis, G.F., Irwin, M.,
Totten, E., \& Quinn, T., 2001a, \apj, 551, 294
\bibitem[Ibata et al.(2001b)]{orb-sds} Ibata, R.A., Irwin, M., Lewis, G.F., \&
Stolte, A., 2001b, \apj, 547, L136
\bibitem[Irwin(1999)]{irwin} Irwin, M. 1999, in IAU Symp. 192,
The Stellar Content of Local Group Galaxies,
eds. P. A. Whitelock \& R. D. Cannon (San Francisco: ASP), 409
\bibitem[Ivezic et al. (2000)]{ivez} Ivezic, Z., et al., 2000, \aj, 120, 963
\bibitem[Johnston, Hernquist \& Bolte(1996)]{katy1} Johnston, K.V.,
Hernquist, L., \& Bolte, M., 1996, \apj, 465, 278
\bibitem[Johnston et al.(1999)]{katy2} Johnston, K.V., Zhao, H., Spergel, D.N.,
\& Hernquist, L.,1999, \apj, 512, L109
\bibitem[Klypin et al.(1999)]{klyp} Klypin, A., Kravtsov, A.V., Valenzuela, O.,
\&, Prada, F., 1999, \apj, 522, 8
\bibitem[Kunkel(1979)]{k79} Kunkel, W.E., 1979, \apj, 228, 711
\bibitem[Kunkel \& Demers (1977)]{k77} Kunkel, W.E., Demers, S., 1977, \apj,
214, 21
\bibitem[Lynden-Bell(1976)]{ly76} Lynden-Bell, D., 1976, \mnras, 174, 691
\bibitem[Lynden-Bell(1982)]{ly82} Lynden-Bell, D., 1982, The Observatory,
102, 201
\bibitem[Lynden-Bell \& Lynden-Bell(1995)]{lynlyn} Lynden-Bell, D., \&
Lynden-Bell, R.M., 1995, \mnras, 275, 429
\bibitem[Majewski(1994)]{maj94} Majewski, S.R., 1994, \apj, 431, L17
\bibitem[Majewski et al.(1999)]{maj99} Majewski, S.R., Siegel, M.H., Kunkel, W.E.,
Reid, I.N., Johnston, K.V., Thompson, I.B., Landolt, A.U., \& Palma,
C., 1999, \aj, 118, 1709
\bibitem[Mart\'inez-Delgado et al.(2001)]{david} Mart\'inez-Delgado, D.,
Aparicio, A., G\'omez-Flechoso, M.A., \& Carrera, R., 2001, \apj, 549,
L199
\bibitem[Mart\'inez-Delgado et al.(2002)]{dav12} Mart\'inez-Delgado, D., Zinn,
R., Carrera, R., \& Gallart, C., 2002, \apj, 537, L19
\bibitem[Mateo(1998)]{M98} Mateo, M., 1998, \araa, 36, 435
\bibitem[Mayer et al.(2001)]{lucio} Mayer, L., Governato, F., Colpi, M., Moore,
B., Quinn, T., Wadsley, J., Stadel, J., Lake, G., 2001, \apj, 547, L127
\bibitem[Moore et al.(1999)]{DMsub} Moore, B., Ghigna, S., Governato, F., Lake,
G., Quinn, T., Stadel, J., \& Tozzi, P., 1999, \apj, 524, L19
\bibitem[McWilliam(1997)]{mcw} McWilliam, A., 1997, \araa, 35, 503
\bibitem[Montegriffo et al.(1998)]{mont} Montegriffo, P., Bellazzini, M.,
Ferraro, F.R., Martins, D., Sarajedini, A., \& Fusi Pecci, F., 1998,
\mnras, 294, 315
\bibitem[Newberg et al.(2002)]{sdss} Newberg, H.J., et al., 2002, \apj, 569, 245
\bibitem[Odenkirchen et al.(1997)]{oden} Odenkirchen, M., Brosche, P., Geffert,
M., Tucholke, H.-J., 1997, NewA, 2, 477
\bibitem[Odenkirchen et al.(2002)]{pal5} Odenkirchen, M., et al., 2002,
(astro-ph/0012311)
\bibitem[Palma, Majewski \& Johnston(2002)]{palma} Palma, C., Majewski, S.R., \&
Johnston, K.V., 2002, \apj, 564, 736
\bibitem[Scholz et al.(1998)]{scholz} ScholzR.-D., Irwin, M., Odenkirchen, M.,
\& Meusinger, H., 1998, \aap, 333, 531
\bibitem[Schweitzer et al.(1993)]{shweitz} Schweitzer, A.E., Cudworth, K.M., \&
Majewski, S.R., 1993, in The globular cluster - galaxy connection,
G. Smith and J. Brodie Eds., S. Francisco: ASP, ASP Conf. Series,
48, 113
\bibitem[Searle \& Zinn(1978)]{sz78} Searle, L., \& Zinn, R., 1978, \apj, 225,
351
\bibitem[Siegel at al.(2001)]{pal13} Siegel, M.H., Majewski, S.R., Cudworth,
K.M., \& Takamiya, M., 2001, \aj, 121, 935
\bibitem[Smith, Sneden \& Kraft(2002)]{smith} Smith, G.H., Sneden, C., \& Kraft,
R.P., 2002, \aj, 123, 1502
\bibitem[Vivas et al.(2001)]{vivas} Vivas, A.K., et al., 2001, \apj, 554, L33
\bibitem[Yanny et al.(2000)]{yanny} Yanny, B., et al., 2000, \apj, 540, 825
\bibitem[Zinn(1985)]{z85} Zinn, R., 1985, \apj, 293, 424
\bibitem[Zinn(1993)]{z93} Zinn, R., 1993, in The Globular Cluster - Galaxy
Connection, G.H. Smith \& J.P. Brodie Eds., (ASP, San Francisco), ASP
Conf. Ser., 48, 31
\end{thebibliography}
\clearpage
%%%%%%%%%%%%%%%% TABLE 1
\begin{deluxetable}{lcccccc}
\tablecolumns{7}
\tablewidth{0pc}
\tablecaption{OHS clusters in the Sgr Stream}
\tablehead{
%\multicolumn{5}{lcccc}{}\\
\colhead{Name}& \colhead{D$_{orb}$ \footnotesize{[kpc]}}
&\colhead{$\Delta V_r$ \footnotesize{[$km/s$]}}
& \colhead{$\mu_{\alpha}cos ~\delta$\footnotesize{[mas/year]}} &
\colhead{$\mu_{\delta}$\footnotesize{[mas/year]}}
&\colhead{ $\mu_{l}cos ~b$\footnotesize{[mas/year]}} &
\colhead{$\mu_b$\footnotesize{[mas/year]}}\\ }
\startdata
Pal 12 & 3.00 & -11.2 & 0.67& -2.41&-2.324 & -0.934 \\
NGC 4147 & 3.72 & 49.1 & -2.99& -0.73&-2.284 & -2.061 \\
Pal 5 & 4.01 & -36.0 & -4.68& 0.25&-3.468 & 3.146 \\
NGC 5634\tablenotemark{a}& 4.72 & -48.6 & -4.64 & 0.25&-3.579 & 2.968 \\
NGC 5053 & 5.30 & -55.7 & -3.78& -0.28&-3.754 & 0.494 \\
Terzan 3 & 6.18 & -24.8 & -4.09& -0.29&-2.858 & 2.941 \\
NGC 5466 & 8.55 & 21.9 & -3.97& -0.16&-3.818 & 1.104 \\
NGC 288 & 9.88 & 17.2 & 3.37& -1.22& 1.235 & -3.368 \\
NGC 7089 & 10.14 & -18.9 & 1.47& -2.23&-2.037 & -1.727 \\
NGC 6426 & 11.59 & -54.4 & -4.15& -0.25&-2.891 & 2.981 \\
NGC 5824 & 13.90 & 52.5 & -4.34& -0.09&-3.023 & 3.119 \\
Pal 13 & 14.21 & 40.4 & 3.47& -1.14& 1.695 & -3.237 \\
Pal 2 & 14.95 & 16.9 & 3.99& 0.20& 2.933 & 2.707 \\
Rup 106 & 17.37 & -19.4 & -4.67& 0.24&-3.439 & 3.176 \\
&&&&\\
% &&Present members of the Sgr dSph&&\\
&&&&\\
NGC 6715 & 0.45 & -4.2 & \nodata & \nodata& \nodata & \nodata \\
Terzan 7 & 1.89 & 34.0 &\nodata & \nodata& \nodata & \nodata \\
Terzan 8 & 2.04 & -2.0 & \nodata & \nodata& \nodata & \nodata \\
Arp 2 & 5.13 & -26.6 & \nodata & \nodata& \nodata & \nodata \\
\enddata
\tablecomments{Distance from the nearest point in the orbital path
followed by the Sgr dSph during the last Gyr ($D_{orb}$) and
difference in radial velocity between the considered cluster and
prediction of the computed orbit at that point ($\Delta V_r$) for the
OHS clusters having $|\Delta V_r| < 60 $ km/s and $D_{orb}< 18$ kpc.
The clusters are ordered with growing $D_{orb}$. The known members of
the Sgr system are also reported at the end of the table, for
comparison. The first six clusters of the list are the ones put in
evidence in Fig.~1 and Fig.~3. The four known members of the Sgr
globular cluster system are also reported for comparison in the last four rows
of the table. The proper motions predicted by the
IL98 model [under the hypothesis that (a) the clusters belong to the
Sgr stream, and (b) they have been lost during the last orbit of Sgr,
see \S2] are also reported, in both the equatorial and galactic
reference system. The intrinsic dispersion of the modelled
stream, combined with estimated 20\% uncertainties in the Galactic
mass model, translate to uncertainties of $\sim 0.85$ mas/yr on each
PM component.}
\tablenotetext{a}{\citet{n5634} argue that NGC 5634 was likely lost
before the last peri-centric passage of Sgr, so the proper motion
predictions listed here may be substantially in error.}
\end{deluxetable}
\end{document}