Skip to main content

Full text of "Stellar over-densities in the halo: the extent of the Virgo over-density"

See other formats


Mon. Not. R. Astron. Soc. 000, [Th?? (2008) Printed 21 January 2009 



(MN I^TeX style file v2.2) 



Stellar over-densities in the halo: the extent of the Virgo 
over-density* 



o 
o 

c 



< 

O 

Oh 
6 

o3 



Stefan C. Keller 1 !, Gary S. Da Costa 1 and Sayuri L. Prior 1 

1 Research School of Astronomy and Astrophysics, Australian National University, Cotter Rd, Weston, ACT 2611, Australia 



ABSTRACT 

We map the three dimensional extent of the Virgo Over-density by combining distance 
information from R R Lyrae variables and projected spatial information from SEKBO 
(IKeller et al.ll2008l) and Sloan Digital Sky Survey (SDSS) DR6 photometry. The Virgo 
Over-density is seen to comprise two filaments 14.5° x 3° and 10° x 3° and a circular 
structure 3° in diameter. Together the three features span 38° of right ascension and 
declinations of +2 to -15°. RR Lyrae variables place the two filamentary features at 
heliocentric distances of 20 and 17 kpc respectively, with projected dimensions of 5 x 1 
kpc and 3x1 kpc. 

Key words: Galaxy: halo — Galaxy: structure 



> 

in 

(N 



O 

o 



1 INTRODUCTION 

One of the most important processes yet to be understood 

■ by modern astrophysics is that of galaxy formation. The 
halo of the Galaxy offers arguably the best opportunity to 
study galaxy formation due to its relative quiessence and our 
ability to obtain accurate kinematics and detailed chemical 
ab undances for a larg e number of stars. The seminal work 
of lEggen et al . (1962) proposed galaxy formation occurred 

■ in the wake of a monolith ic colla pse of a proto-galactic 
cloud. Later, ISearle fc Zinnl (|l978l ) challanged this notion 
by proposing that the halo formed via the accretion of nu- 

' merous small entities over an extended period. 

Today, galaxy formation is discussed within the con- 

■ text of LambdaCDM cosmology. This predicts hierarchical 
structure formation in a way that qualitatively parallels the 
Searle & Zinn scenario. Observational evidence of the dual- 
ity in the Galactic halo llFreeman fc Bland-Hawthornl 120021 : 
ICarollo et~ail 120071 ; iBell et al l 120071 ) suggests at least two 
formation mechanisms. The kinematics and star formation 
history of the inner halo (R <20 kpc) suggest formation 
from rapid collapse at an early ep och. The outer hal o, on 
the other hand, possess kinematics (|Carollo et alj|2007l ) and 
a level of substructure (|Bell et al.ll2007l ) consistent with as- 
sembly from disrupted dwarf galaxies. 



* Based, in part, on observations obtained from Mount Stromlo's 
1.3 meter Telescope (Great Melbourne Telescope), 
f E-mail: stefan@mso.anu.edu.au 



1.1 The Sagittarius dwarf and the Virgo 
over-density 

The most prominent on-going accretion event in the Milky 
Way halo is the dissolution of the Sagittarius dwarf (Sgr, 
llbata et al.ll 1994 ) . The leading and trailing arms of Sgr tidal 
debri s are seen to w r ap ar o und the sky ([ Maiewsk i et al.l 
| 2003l: iNewberg et al l |2002| , 120071 ; iBelokurov et all 120061 : 
iKeller et al.ll2008l ). The debris of Sgr are the largest con- 
tribution to substructure in the outer halo yet to be found. 

Arguably the next most significant outer halo substruc- 
ture is the Virgo Over-density (VOD). T he VOD was dis - 
covered in the Sloan Digitial Sky Survey (lYork et alj|2000h 
as a diffuse over-densit y of F-type main-sequence s tars span- 
ning some 1000 de g 2 iNewberg et alJliool . l2007t) and dis- 
tances of 6-20 kpc l|juric et al. 20081 ). The region has a co- 
incident series of over-densities in RR Lyrae variable stars 
I Vivas fc Zinnl 120061 : IKeller et al1l2008l) at dist ances of 16- 
19 kp c. The studies of iDuffau et al.1 |2006l) and lPrior et~ai] 
(2008) found subsets of RR Lyraes show a common radial 
velocity with very low velocity dispersion (consistent to mea- 
surement uncertainties). This is as expected from kinemat- 
ically cold tidal debris. The moving group was termed the 
Virgo Stellar Stream to distinguish it from the general stel- 
lar over-density. 

The nature of the VOD remains uncertain. Does it rep- 
resent a separate accretion event, or is it related to the 
spatia lly overlapping Sgr debris? iMartmez-Delgado et al.l 
(2007) proposed that the VOD results from the confluence 
of the lea ding and trailing a rms of Sgr as seen in the n-body 
models of lLaw et al.l (120051 ). The Law et al. models predict 
highly negative radial velocities for the Sgr m aterial in this 
region . Thi s is contrary to the ob servations of IDuffau et al.l 
(2006|) and lNewberg et all (|2007l ) who find most stars lie at 



2 Keller et al. 



a radial velocity of +100-130 kms INewberg et all (|2007i ) 
also notes that the models fail to predict the high stellar den- 
sity seen in the VOD, and that the VOD should be spatially 
offset from the Sgr le ading arm as we observe. This leads 
iNewberg et al.l (|2007h to conclude t hat the VOD is a halo 
substructure of separate origin to Sgr. I Vivas et al.l (|2008l ) ex- 
amines the radial velocities of a sample of RRLs in the direc- 
tion of the VOD at distances of less than 13 kpc. lVivas et al.l 
finds that the Virgo Stellar Stream extends to ~ 12 kpc, but 
none of the additional RRLs can be ascr ibed to the l eading 
arm of Sgr. Nevertheless, the study of iPrior et al.l ([2008) 
find RRLs at negati ve radial velocities which they speculate 
could be Sgr debris (IPrior et al.ll2008bl) . 



It is evident from 



Newberg et all (I2002TI; [Duffau et all 



(I2006T): iBelokurov et al.l (|2006l ); INewberg et al.1 (|2007h 



I Juric et al.1 |2008t) that the centre of the VOD lies 
where to the south of the region surveyed by SPS S. The 
VOD was traced southwards by iKeller et al.1 (120081 ). They 
identified two RR Lyrae over-densities in the VOD region: 
one at a heliocentric distance of 16 kpc, 8° south-e ast of the 
centre of the VOD defined bv lDuffau et al.1 (120061), and an- 
other at 19 kpc and 24° south-east of this centre. Prior et"ahl 
(2008) explore the extent of the VOD by looking at a sam- 
ple of SEKBO fields that sparsely sample the vicinity of the 
VOD. This study reveals that the VOD is a large diffuse 
feature, covering at least 760 deg 2 of sky. 

In the present study, we draw together the SEKBO and 
SDSS data sets to constrain the 3-d extent of the VOD with 
the aim of clarifying its nature. We use the distance infor- 
mation available from the SEKBO RR Lyrae candidates and 
the spatial extent determined from the combined SEKBO 
and SDSS DR6 data sets. In section [4] we conclude with 
some remarks regarding the nature of the VOD. 



1.2 Luminosity function excess 

We utilise the Sloan Digital Sk y Survey Dat a Release 6 
jAdelman-McCarthv et al]|200Sl ) and SEKBO l|Keller et all 
2008) photometric databases to construct luminosity func- 
tions over the combined survey area. W e have transformed 
the Johnson V,R of the SEKBO survey l|Keller et al.ll2002T l 
to SDS S g,r f or consistency using the transformations of 
iLuptonl ((2005). In order to uniformly combine luminosity 
functions from the two datasets we needed to consider mag- 
nitudes that do not suffer from substantial observational 
incompleteness. As can be seen from the example colour- 
magnitude diagrams in Figure [1] observational incomplete- 
ness becomes appreciable in the SEKBO data set at go ~ 20 
(there is some variation in depth amongst the SEKBO 
fields). This is almost 2 magnitudes brighter than the limit 
of SDSS photometry. 

To construct the luminosity functions seen in Figure [1] 
we utilise the colour range 0.1 < (g — r)o < 1.0 and mag- 
nitude r ange 14 < go < 19.7 5. Reddening is as estimated 
from the Schlegel et al. I (|1998| ) reddening maps. This colour 
range excludes the near-by red disk dwarf star sequence that 
is not of interest to the present study. The size of individual 
SEKBO fields ranged from 0.81 to 0.60deg 2 . The areal cov- 
erage of each field was dependant on the dither pattern and 
operational status of CCDs. The SDSS data are compared 
over a ldeg 2 field. To correct for the variation in sample size, 
we normalised the luminosity functions over the magnitude 



range 14.5 < go < 15.5. As can be seen from an exam- 
ple overlap region in Figure [TJ the SDSS and corresponding 
SEKBO luminosity functions agree within uncertainties. 

We then compared the observational data with syn- 
thetic lumino sity functions de rived from the Besancon 
galaxy model (|Robin et al.ll2003T ) . The simulations were con- 
structed for ldeg 2 fields with a distance range of 0-120 kpc 
with magnitude and colour cuts applied as per observations. 
The Besancon model is a dynamica lly self-cons i stent, para- 
metric description of the Galaxy. As I Juric et al. (2008)) note, 
a parametric model is susceptible to the choice of input ob- 
servations since these are typically based upon small areal 
coverage and many include unrecognised stella r substruc- 
ture. T he solution of lBelokurov et al (I2006T ) and lJuric et al.l 
(2008) is to directly map the stellar number density using 
distances derived for main-sequence stars from photometric 
parallax. The technique requires highly accurat e colours in 
order to precisely derive an absolute magnitude. I Juric et al.l 
(2008) show, magnitudes determined to a precision of 0.01- 
0.02 mag. result in an uncertainty in distance of ±10%. This 
is a result of the extremely steep absolute magnitude - colour 
relation inherent in the photometric parallax method. In ad- 
dition, considerable systematics enter with the choice of a 
photometric parallax relation (up to ±70% in distance for 
the bluest halo stars). 

Our approach to discerning ha l o sub struc ture is un- 
like th at utilised bv lBelokurov et"al] (|2006T ) and I Juric et al.l 
(2008) that have focused on the density of f-type main- 
sequence stars. The SEKBO data lacks the global photomet- 
ric accuracy required for such a technique. Due to propaga- 
tion of uncertainties in standardisation and colour transfor- 
mation from the non-standard MACHO B & R passbands, 
the global accuracy of SE KBO phtometry i s ±0.03 mag- 
nitudes in V & R bands j|Keller et al.l 120081 ) for g < 16. 
This rises to ±0.15 magnitudes at g ~ 19.75 magnitudes ( 
with some variation from field to field due to variable survey 
depth). 

For this reason, we have chose to construct our search 
for halo substructure via the luminosity function of the ob- 
sevred data. This approach has also bee n applied in the 
study of lDuffau et all (|2006h . iDuffau et al.l has examined the 
region of sky between 176°s;RA^210° and -4^Dec^±4° in 
2° x 2° fiel ds. The outlin e of si gnificant over-density as de- 
scribed by IDuffau et ail (|2006l ) is shown in Figu r e [H O ur 
approach has also been applied by INewberg et all l|2007l ) to 
two fields, one centred on RA=186° and Dec=0° and one 
further south at RA=191°, Dec=-7.8°. Both fields reveal 
significant luminosity function excess relative to fields at 
the same galactic latitude but 180° displaced in longitude. 
Nonetheless, the disadvantage of the our technique has the 
drawback of removing distance information from the view 
of substructure. However, in Section [2] we merge distance 
determinations from the distribution of RR Lyraes to con- 
strain the distance to the substructures. 

Localised excess over the Besancon model shows regions 
that depart from an axially symmetric model. While features 
covering several thousand square degrees (quadrupole mo- 
ments, for instance) may suffer from uncertainty due to the 
validity or otherwise of the model, we can consider features 
on scales of several hundred square degrees as representing 
departures from the local smooth background density. 

The bottom panel of Figure [2] shows a field well sepa- 



The Virgo over-density 3 




Figure 1. Left: A comparison of SEKBO and SDSS colour-magnitude diagrams for a 1° X 1° region centred on (RA, Dec) 192°, -4°. 
Right: a comparison of the luminosity functions derived from the SEKBO (solid line) and SDSS (dashed line) fields with Poisson 
uncertainties shown. Note the agreement between the luminosity functions derived from the two data sets is within uncertainties until 
the completeness limit of the SEKBO data at go ~ 19.75. 




Figure 2. Three 1° X 1° fields progressively further away from regions of marked excess. The top row shows a field centred on the VOD 
(RA, Dec) 192°, -4° and the middle and bottom rows are for fields consecutively 8° north l|Keller et al. 2002). T he right panels com pare 
the luminosity function observed (solid line) with that predicted from the Besancon Galaxy model (dashed line. iRobin et al. 2003). 



4 Keller et al. 



rated from substructure in the southern Galactic cap. Here 
we see the agreement of the observed LF with the Besancon 
model LF to within the Poisson uncertainties. As we proceed 
through to the area of the VOD there is a marked excess of 
stars rising towards the faintest bin consid ered (at po=19-5 
in Figure [2] top, see also lPrior et al.1 l|2008h ). 

The cumulative difference between the observed and 
predicted luminosity functions is then expressed in terms 
of significance relative to the Poisson uncertainties inherent 
in the observed sample. That is, 

1/2 



N„ 



+ N 2 ) 

' "model) 



(1) 



where S is the significance of a field, N a i, s is the observed 
number of objects to magnitude bin i, N mo dei is the pre- 
dicted number of objects to bin i from Besancon galaxy 
model normalised to that observed in the magnitude range 
14.5< go < 15.5. 

The resulting map of significance of departure from the 
luminosity function predicted by the Besancon model is seen 
in Figure T he SPSS data cover th e region of the 'Field 
of Streams' of IBelokurov et al.l (|2006h . namely 120° <RA< 
220° and from the southern SDSS limit (varies from Dec=- 
1° to -4°) to Dec=+34°. In the region of overlap between 
SDSS and SEKBO we utilise the SDSS dataset in light of 
the superior photometric precision of the SDSS data. 

Figure|3]expands the northern region of overlap between 
SDSS and SEKBO. A series of 'point' sources are seen - these 
mark globular clusters in the field (distances < 100 kpc, after 
this the horizontal branch falls out of the magnitude range 
considered for the luminosity function). The main source of 
the over-densi ty in the north is the l eading arm of Sgr (se e 
for example, IBelokurov et al.ll2006l : iFellhauer et al.1 [2006). 
Enmeshed with the linear feature of the Sgr leading arm is 
a large, diffuse region of excess between RA=170-195° and 
Dec<+7°. This diffuse region is also coin cident with the 
broad VOD detection of lJuric et a l. (2008) that is seen to 
cover ~1000 deg 2 (centred at 192°, +2°). 

Three significant new features are also seen in the ex- 
tension to the south-east afforded by the SEKBO data. A 
linear feature A (198°, -10°) forms a band approximately 3° 
wide and 14.5° long. A second linear feature B (192°,-2°), is 
3° wide and 10° long. The third significant feature C, seen 
at (178°, 0°) is approximately 3° in diameter. The three fea- 
tures are br oadly consi s tent w ith the area of over-density 
discussed in lPrior et all (|2008h . 



2 DISTANCES TO SUBSTRUCTURES FROM 
RR LYRAE VARIABLES 

The above detections of halo substructure are based upon 
the excess or deficit of main-sequence and red giant branch 
populations over the Besancon galaxy model. The range in 
luminosity exhibited by both stellar populations does not 
afford any distance information at the low surface densities 

seen here. 

The iKeller et all l|200Sl) study of the RR Lyrae over- 
density from SEKBO presents the radial distribution of 
RR Lyrae candid ates (76±7% of which are RR Lyraes, 
Prior et al. I l2008f> across the VOD region. Using the 
Keller et al.l ( 20081 ) data set, we can construct a plot of the 



o 

Q. 

"c7> 
c 
CD 
T3 

CD 
_Q 



0.1 



rr 

DC 



0.01 



- Vs ' 


• 










RA: 'l70-2'l 6°' : 




• 
• 












\^ • 














_ • 

: 




• 

•• 


s s 






- 




• 






• 














•v 














.\67% 

*\ 


Clump 


1 




2 




3 


• \ 


Feature 


B 




A 






• \ 


-, , i 















10 



Heliocentric distance r (kpc) 



100 



Figure 5. A plot of the number density of RR Lyrae candidates 
in the RA range 170-210° as a function of heliocentric distance. 
RR Lyrae candidates are from lKeller et al.l (2008). The solid line 
shows the average power law relation of halo density. From Monte- 
Carlo simulations we are able to derive 1 and 2 a intervals (dashed 
lines) above this average relation based on random samples of 
appropriate sample size. The grey shaded regions indicate the 
locations of three significant over-dense regions. 



number density of RR Lyrae candidates as a function of 
distance for the range in right ascension 170-210°. 

The calculation of the RRL number den s ity fo llows the 
technique dev eloped bv | Wetterer fc McGrawl l| 19961 ) and im- 
plemented in IKeller et al. (|2008f ) ftp which we refer the 
reader for details). The following equation describes the lo- 
cal space density as a function of galactocentric distance: 



p(R) = 



dN 



A-KR 2 f(R) dR ' 



(2) 



where f(R) is the fraction of the total halo volume at R that 
is sampled by the survey (,f(R) is analogous to a solid an- 
gle) and N is the number of RRL as a function of distance. 
Whereas the solid angle is constant as one looks through the 
halo, f{R) varies as a function of Galactocentric radius and 
hence must be calculated numerically for each field. To ac- 
count for the effects of completeness we multiply each f(R) 
by the Monte-Carlo derived RRLafr completeness profile for 
the field as a function of galactocentric radius. This gives the 
effective volume of the halo that is sampled by each field. 
We use Equation [2] to calculate the local space density for 
each candidate. A transformation to heliocentric distance is 
then made. 

The result is seen in Figure [5] From Monte-Carlo sim- 
ulations of appr opriate samp l e size using the halo density 
profile defined in IKeller et"atl (|2008l ) (the solid line in Fig. [5] 
for R <45 kpc) we have determined the 1 and 2<r significance 
contours. These are shown in Figure [5] as the dashed lines. 
Figure [5] shows three peaks of greater than 2a significance; 
at 17, 20 and 34 kpc (heliocen tric). 

From IKeller et all (|2008l ). figure 20, we see that the 
VOD Clump 1 spatially extends over the linear feature B 
and Clump 2 is seen to coincide with the southern portion of 



The Virgo over-density 5 




Figur e 3. The significan ce of cumulated excess of the observed luminosity function compared to that predicted by the Besancon galaxy 
model llRobin et ajj|2003h to a limiting magnitude of </o=19.75 in 1° X 1° fields. Darker shading represents more significant excess over 
the model, as described by the greyscale colour bar to the right. The significance is expressed as the number of sigma over the model 
(see text for details). 



NGC5466 M3 




220 200 180 160 140 120 

FW 



Figure 4. Left: A close-up of Figure[3]to show the central "Field of Streams" region. Numerous point-like features are associated with 
known globular clusters as shown. The leading arm of Sagittarius debris is seen as the broad feature from 120°, +20° to 190°, +8° . Thre e 
new features A, B, and C are indicated. The dashed ellipses mark the regions of excess RR Lyrae density from iKeller et al. I <200St). 
Right: A further enl argement of the regi on around the Virgo Over-den sity (Featu re B) to show previou s detections by Ip^iffauet al] 
i2006l ) (solid outline - ). Ivivas fc Zinr] ]2006T) (triangle), [ivezic et al] l l2005l) (star) and lNewberg et alj <2007ft (circles). 



feature A, as shown in Figure [4] The more distant, and less 
numerous, Clump 3 does not present a well defined spatial 
centroid. Feature C does not exhibit a corresponding excess 
in RRLs. At their determined distances feature A (20 kpc) 
has projected dimensions of 5 x 1 kpc and feature B (17 kpc) 
dimensions of 3 x 1 kpc. 



3 DISCUSSION OF SUBSTRUCTURES 

An extensive literature has developed that examines the stel- 
lar over-densities in this region. Our study is the first to 
resolve the VOD region into a series of distinct substruc- 
tures. The filling factor of the three f eatures is significan tly 
less than the 76 deg 2 dete r mined by iPrior et al.l (|2008h or 
~ lOOOdeg 2 by ljuric et al] <|2008h . We believe th is can be 
understood by the spatial sampling of lPrior et al] and their 
extrapolation to an elliptical morphology. The three features 
span some 38° of sky simil ar to the major axis size of 45° 
determined by IPrior et all It is by increasing our spatial 
resolution that we are now able to see the details of sub- 
structure in this region. 



In the following section we discuss our finds in the light 
of the existing literature for the region. 



3.1 Feature A 

iNewberg et~aH l|2007h examines a 1.5° radius field centred 
at RA=191°, Dec=-7.8°. this corresponds to the peak over- 
density seen in the southerly 'outrigger' scan of SEGUE. 
This field overlaps with the north-west corner of feature A. 
SDSS spectroscopy reveals an marginally significant excess 
at radial velocities betw een VgS-R=+1Q0— l-130 kms~ 1 . Such 
an excess is also seen bv lNewberg et all (|2007l ) towards the 
VOD (feature B, see below). This suggests that feature A 
and the VOD share a common spatial velocity and poten- 
tially a common origin. 



3.2 Feature B - the VOD 

Feature B is identified as the VOD. The VOD corre- 
spond s to the "12.4hr" clump described by IVivas fc Zinnl 
(2006) and first detected as an overdensity in RRLs by 



6 Keller et al. 



IVivas et all i|200ll ). It was independently s een as an over- 
densit y of F-type main-sequence stars by iNewberg et al.l 
(2002) at RA=190°, Dec=0 and heliocentric distance of 
18 kpc. Using a 2.3° wide band centred on declination - 
1:10:48, I Vivas fc Zirml (120061 ) find a peak density located at 
RA=186° and a heliocentric distance of 17 kpc. 

Feature B is also r ecovered in the stu dy of RRL can- 
didates from SPS S by llvezic et al] (|2005h . The study of 
llvezic et all (|2005h finds the VOD cent red at RA~ 1 90° in 
a 2.5°-wide band centred on Dec=0°. Ivczic et all (|2005T ) 
derive a mean heliocentric distance of 18 kpc from the lu- 
min osity of the clump c andidate RRLs. 

IDuffau" et alJ (|2006h presents spectroscopic follow-up of 
the "12.4hr" clump and finds 9 of 18 RRLs studied share 
a common radial velocity of + 100kms~ 1 with velocity dis- 
persion less than measurement uncertainties. This moving 
group was termed the Virgo Stellar Stream to differentiate 
the population from the region of over-density. In an ex- 
ami nation of SPSS pho tometry for the adjoining region of 
sky, IDuffau et al.l (J2006) saw evidence for an excess in the 
luminosity function (the same technique as applied in the 
present study) over an area of at least 106 square degrees of 
sky centred on RA=186° and Dec=-1°. 

Since the over-density is located at the southern edge 
of the SPSS fi e ld the spatial extent remains unbound. In 
INewberg et alJ (|2007h one of the three SEGUE 'outrig- 
ger' scans apparently shows that the VOE extends beyond 
the Dec=-4° l i mit o f SDSS to Dec< -15°. In addition, 
INewberg et""aH (|2007h examines the radial velocities of stars 
in a 1.5° radius centred in RA=186°, Dec=0°. At magni- 
tudes corresponding to main-sequence members of the VOD 
with a distance of 18 kpc there is a suggestion of a population 
of stars with radial velocities between + 1 00— 1-130 kms— 1 a s 
seen in the study of IDuffau et aD (|2006l ) . I Vivas et all (|2008T l 
presents evidence that the VSS population extends towards 
us to around 12 kpc. 

The present study places the VOD in a clearer spa- 
tial context by providing broader sky coverage. Our derived 
mean distance of 17 kpc is in good agreement with previous 
studies. As can be seen in Figure [4] there is also agreement 
between previous positions of VOD presented in the litera- 
ture and our map of o ver-density in this r egion. 

The discovery by IWalsh et af] (120081 ) of VirZ, a dwarf 
galaxy candidate at the north-western tip of feature B is 
intriguing. Walsh et al. find an optimal distance of ~ 40 kpc 
but considerable uncertainty is accommodated by the small 
number of stars populating the giant branch. The possibility 
of a connection between VirZ and feature B is one that must 
await verification from deeper photometric and kinematic 
studies. 



4 A FIELD OF COMPLEXITY — THE 

NATURE OF THE VIRGO OVER-DENSITY 

There are a number of possible interpretations for the mor- 
phology we see in Figure [3] Structures A, B & C could rep- 
resent individual accretion events or they may be 'hotspots' 
of increased density along a single stream. If describing a 
coherent structure, the distances to features A & B do not 
indicate a significant line of sight inclination for the stream. 
The proximity of the features to Sgr and their align- 



ment parallel to the Sgr leading arm stream could point 
to an origin from Sgr. In this scenario, A, B, & C may 
trace the trailing arm of Sgr some ~ 270° downstream from 
Sgr. All n-body models in the literature dLaw et al.l 120051 : 
iHelmi fc Wrutell200lt iFellhauer et ail kooe), while they dif- 
fer on the distance to which the trailing arm extends, place 
the returning trailing arm at a distance of ~ 20 kpc over 
the RA range of features A & B. However, these models do 
not predict the substantial stellar density seen in the two 
features. There is also no evidence for a return of the trail- 
ing arm from the distribution of M-giants. However, this 
material is expected to be derived fr om disruption fo ur or- 
bital periods, or (3-3.5) Gyrs ago l|Law et alJ I2005T 1 . and 
given that the the bulk of M-giants are 2-3 Gyr or younger 
(|Lavden fc Saraiedini|[200ol : |KeUeJll999h we would expect a 
marked reduction in the contrast of M-giant tracers along 
this section of stream. 

The possibility of a common origin for features A, B, & 
C in Sgr is one that requires clarification as this is critical 
for studies of the orbit of Sgr and, flowing from this, our 
ability to constrain the shape of the dark matter halo. Once 
we have clarified what material is associated with Sgr, we 
can then realise the potential of Sgr as a test particle of the 
Galaxy's dark matter halo shape, q(= c/a). q is fundamen- 
tal to Galactic dynamics. Recently, the RAVE consortium 
■ 1| — — | 

ijSiebert et alj |2008) was unable to estimate q due to uncer- 
tainties in halo+disk mass and disk scale length. Consider- 
ing the distance and extension of the Sgr debris, they repre- 
sent the best opportunity to constrain q. Tracing Sgr debris 
through the VOD region will provide substantial new kine- 
matic information that will allow much tighter constraints 
on the models. In turn, this may resolve the discrepancy 
in determination of q (current literature ranges from oblate 
q = 0.8, to prolate q = 5/3). 

Should subsequent observations and modelling reveal 
that the substructure revealed by the present study origi- 
nates from Sgr, this will represent an important constraint 
on the number and mass of accretion events in the outer 
halo over the last 3-4 Gyrs. The accretion history may then 
be compared to expectations from cosmological simulations. 
The small scale height of the disk suggests an anom alously 
low late-time merger rate (see e.g. iParrv et al.ll2008h — our 
observations will form additional evidence for, or against, 
this departure from the expectations of LambdaCDM. 

In light of the apparent complexity of the VOD region, 
additional radial velocities are critically needed. They will 
enable us to clarify if features A, B & C form a coherent 
common stream. With revised n-body modelling, it will then 
be possible to test if these features can be accommodated 
by an origin in the Sgr debris stream. We aim to obtain 
further data with the AAT's AAOmega spectrograph. In 
the near future it will be possible to apply the techniques of 
the present study t o the hemispheric coverage of SkyMapper 
jKeller et al.ll2007l ). 



ACKNOWLEDGMENTS 

This research has been supported in part by the Aus- 
tralian Research Council through Discovery Project Grants 
DP0343962 and DP0878137. 



REFERENCES 

Adelman-McCarthy, J. K. et al. 2008, ApJS, 175, 297 

Bell, E. F., et al. 2008, ApJ, 680, 295 

Belokurov, V., et al. 2006, ApJL, 642, L137 

Carollo, D., et al. 2007, Nature, 450, 1026 

Duffau, S., Zinn, R., Vivas, A. K., Carraro, G., Mendez, 

R. A., Winnick, R., & Gallart, C. 2006, ApJL, 636, L97 
Eggen, O. J., Lynden-Bell, D., & Sandage, A. R. 1962, ApJ, 

136, 748 

Fellhauer, M., Belokurov, V., Evans, N. W., Wilkinson, 
M. I., Zucker, D. B., Gilmore, G., Irwin, M. J., Bramich, 

D. M., Vidrih, S., Wyse, R. F. G., Beers, T. C, & 
Brinkmann, J. 2006, ApJ, 651, 167 

Freeman, K. & Bland-Hawthorn, J. 2002, ARAA, 40, 487 
Helmi, A. & White, S. D. M. 2001, MNRAS, 323, 529 
Ibata, R. A., Gilmore, G., & Irwin, M. J. 1994, Nature, 
370, 194 

ivezic, Z., Vivas, A. K., Lupton, R. H., & Zinn, R. 2005, 

AJ, 129, 1096 
Juric, M., et al. 2008, ApJ, 673, 864 
Keller, S. C. 1999, AJ, 118, 889 

Keller, S. C, Bessell, M. S, Cook, K. H., Geha, M., & 

Syphers, D. 2002, AJ, 124, 2039 
Keller, S. C, Da Costa, G. S., & Bessell, M. S. 2001, AJ, 

121, 905 

Keller, S. C, Murphy, S., Prior, S., DaCosta, G., & 

Schmidt, B. 2008, ApJ, 678, 851 
Keller, S. C, et al. 2007, Publications of the Astronomical 

Society of Australia, 24, 1 
Law, D. R., Johnston, K. V., & Majewski, S. R. 2005, ApJ, 

619, 807 

Layden, A. C. & Sarajedini, A. 2000, AJ, 119, 1760 
Lupton, R. H. 2005, Transformations between SDSS 

magnitudes and UBVRcIc, http://www.sdss.org/dr4/- 

algorithms/sdssUBVRITransf orm . html 
Majewski, S. R., Skrutskie, M. F., Weinberg, M. D., & 

Ostheimer, J. C. 2003, ApJ, 599, 1082 
Martmez-Delgado, D., Penarrubia, J., Juric, M., Alfaro, 

E. J., & Ivezic, Z. 2007, ApJ, 660, 1264 

Newberg, H. J., Yanny, B., Cole, N., Beers, T. C, Re 
Fiorcntin, P., Schneider, D. P., & Wilhelm, R. 2007, ApJ, 
668, 221 

Newberg, H. J., et al. 2002, ApJ, 569, 245 

Parry, O. H., Eke, V. R., & Frenk, C. S. 2008, preprint 

(astro-ph/0806.4189) 
Prior, S. L., Da Costa, G. S., & Keller, S. C. 2008b, in prep 
Prior, S. L., Da Costa, G. S., Keller, S. C, & Murphy, S. 

2008, ApJ Accepted, preprint (astro-ph/0810.0057) 
Robin, A. C, Reyle, C, Derriere, S., & Picaud, S. 2003, 

A&A, 409, 523 
Sales, L. et al. 2008, MNRAS, 389, 1391 
Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 

500, 525 

Searle, L. & Zinn, R. 1978, ApJ, 225, 357 

Siebert, A., et al. 2008, MNRAS, 391, 793 

Vivas, A. K., et al. 2008, AJ, 136, 1645 

Vivas, A. K. & Zinn, R. 2006, AJ, 132, 714 

Vivas, A. K., et al. 2001, ApJL, 554, L33 

Walsh, S., Willman, B., & Jerjen, H. 2008, preprint (astro- 

ph/0807.3345) 
Wetterer, C. J. & McGraw, J. T. 1996, AJ, 112, 1046 



The Virgo over- density 7 
York, D. G., et al. 2000, AJ, 120, 1579