Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.stsci.edu/~inr/post/maj1.ps Äàòà èçìåíåíèÿ: Fri Jun 8 23:31:47 2007 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 03:25:41 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: arp 220

Starcounts Redivivus: III. A Possible Detection of the Sagittarius
Dwarf Spheroidal Galaxy at b = \Gamma40 o
S. R. Majewski 1;2;3 , M. H. Siegel 1 ,
W. E. Kunkel 4
I. N. Reid 6 ,
K. V. Johnston 5 ,
I. B. Thompson 7 ,
A. U. Landolt 2;8 ,
C. Palma 1
ABSTRACT
As part of the Selected Areas Starcount Survey (SASS), a CCD survey to
V ? 21, we have obtained V I photometry of two fields at b = \Sigma40 ffi aligned
roughly with an extrapolation of the major axis of the Sagittarius (Sgr) dwarf
spheroidal galaxy. Comparison of the color­magnitude diagram (CMD) for
some of these fields to the CMDs of fields reflected about the Galactic l = 0 ffi
meridian reveals an excess of stars at V o = 17:85 and 0:9 ! (V \Gamma I) o ! 1:1 in
the (l; b) = (11 ffi ; \Gamma40 ffi ) field. The excess stars have colors consistent with the
Sgr red clump, and deeper CMD imaging in these locations shows evidence of
a main­sequence turnoff at V = 21 with the main­sequence extending to the
1 Dept. of Astronomy, University of Virginia, Charlottesville, VA, 22903­0818
(srm4n@didjeridu.astro.virginia.edu, mhs4p@virginia.edu, cp4v@virginia.edu)
2 Visiting Research Associate, The Observatories of the Carnegie Institution of Washington, 813 Santa
Barbara Street, Pasadena, CA 91101
3 David and Lucile Packard Foundation Fellow
4 Las Campanas Observatory, Carnegie Institution of Washington, Casilla 601, La Serena, Chile
(skunk@roses.ctio.noao.edu)
5 Institute for Advanced Study, Olden Lane, Princeton, NJ 08540
6 California Institute of Technology, 105­24, Pasadena, CA 91125 (inr@astro.Caltech.Edu)
7 The Observatories of the Carnegie Institution of Washington, 813 Santa Barbara Street, Pasadena, CA
91101 (ian@ociw.edu)
8 Dept. of Physics & Astronomy, Louisiana State University, Baton Rouge, LA 70803­4001
(landolt@rouge.phys.lsu.edu)

-- 2 --
limit of our data (V = 24). The surface brightnesses we derive from either
the potential excess of red clump stars or the apparent excess of MSTO stars
are consistent with each other and with the results of other surveys at this
latitude. No similar excess appears in our northern Galactic hemisphere fields
near l = 353 ffi ; b = +41 ffi field.
We have obtained spectroscopy of all 30 candidate red clump stars in the
range 0:9 ! (V \Gamma I) o ! 1:1 and 17:75 ! V o ! 17:95. The radial velocity
distribution of the stars, while dissimilar from expectations of Galactic structure
models, does not show a contribution by stars near the Galactocentric radial
velocity seen in other studies near the Sgr core. It is difficult to reconcile a
photometric result that is consistent with other explorations of the Sagittarius
stream with a radial velocity distribution that is apparently inconsistent.
In a companion paper (Johnston et al. 1999), we discuss how some of the
discrepancies are resolved if our potential Sgr detection corresponds to a
different Sgr tidal streamer than that detected by most other surveys.
Subject headings: Galaxy: evolution -- Galaxy: formation -- Galaxy: halo --
Galaxy: structure -- galaxies:individual(Sagittarius) -- galaxies: interactions --
stars: horizontal­branch

-- 3 --
1. Introduction
The discovery in 1994 (Ibata, Gilmore & Irwin 1994, I94) of the Sagittarius (Sgr) dwarf
spheroidal (dSph) galaxy has provided direct observational support for the proposition that
dwarf satellite galaxies can be tidally disrupted by the gravitational potential of the Milky
Way and contribute to both its field star and cluster populations. Stars associated with
Sgr but well beyond its tidal radius of ¸ 4 ffi (Ibata et al. 1997, I97 hereafter) have been
detected by a variety of techniques, some that we utilize in the present study. Deep color
magnitude diagrams (CMDs) of fields oriented along the major axis of Sgr have revealed
the presence of the Sgr main sequence turn off (MSTO) superimposed on the CMD of field
stars (Mateo et al. 1996, hereafter M96; Fahlman et al. 1996, F96; Mateo et al. 1998,
M98). Surveys of RR Lyrae stars have shown concentrations at the distance modulus of Sgr
(M96; Alard 1996, A96; Alcock et al. 1997, A97). The prominent Sgr red clump has also
proven useful for delineating its structure (I94; Sarajedini & Layden 1995, SL95; Ibata et
al. 1995, I95). These stellar Sgr detections are plotted in Aitoff projection in Figure 1 along
with planetary nebulae (Zijlstra & Walsh 1996) and globular clusters associated with Sgr
(Terzan 7, Terzan 8, Arp 2 and M54). Figure 1 demonstrates the growing extent of the Sgr
``tidal stream'' as the joint surveys probe a larger area.
We have begun the Selected Areas Starcounts Survey (SASS) to provide accurate
photometric catalogues for our continuing study of the classical Galactic starcounts problem
(Reid & Majewski 1993). Among our completed SASS fields, Kapteyn Selected Areas
(SA) 107 (l = 6 ffi ; b = 41 ffi ) and SA184 (356 ffi ; \Gamma40 ffi ) were imaged in the BVRI passbands
with an approximately square, 5x5 grid of overlapping CCD images covering 2.1 deg 2 per
SA, of which a subset (1.26 deg 2 ) has been analyzed in the context of Galactic starcounts
already (Siegel et al. 1997, SMRT97 hereafter). To constrain the symmetry of the halo and
Intermediate Population II thick disk near l = 0 ffi , we also observed ``anti­fields'' (ASA107
and ASA184 in Figure 1), reflected about the Galactic l = 0 ffi meridian, in V and I. Because
these ``anti­fields'' lay near the projection of the Sgr major axis, however, we rearranged our
normal square grid of individual CCD images to a linear strip optimized to detect the Sgr
stream, if present at these latitudes. In this way, we might evaluate better the extent to
which our starcount analysis might be sensitive to, and affected by, interception of Sgr­like
tidal structures.
We outline here the identification of possible interloping Sgr stars in our ASA184 field,
at b = \Gamma40 ffi , as first suggested by Siegel et al. (1997). In that report, we presented results
based on excesses in the field star color magnitude diagram that were intriguing, unique
in our data set, and consistent with the presence of Sgr in the one survey area most likely
to intercept the Sgr debris stream, though, admittedly, the photometric evidence was not

-- 4 --
decisive. Since that report, however, additional support for a Sgr interpretation to the
photometric excesses we found has come from the work by M98 who have obtained deep
CMDs in fields near our ASA184 field and who claim unambiguous detection of Sgr stars.
The M98 results (particularly the measured surface brightness of Sgr debris in this part of
the sky) are remarkably consistent with our own results.
Though our findings here may be argued to remain non­decisive, there are several
reasons that motivate making our results available: (1) Our data would seem to corroborate
(with great consistency) the independent work of M98 to trace the dwindling (and more
difficult to follow) density profile of the Sgr debris stream at great angular separations from
the Sgr core. (2) There is great interest in, indeed need for, tracing out the full extent
of the Sgr debris stream; however, because of the huge area covered by Sgr, mapping its
extent will require a huge investment by the community. In some sense, ``every little bit
helps'' (note the number of different groups contributing to the Sgr mapping in Figure 1),
and our findings can help advance this endeavour by providing an additional benchmark
in one region of the sky. (3) We have supplemented our photometry with spectroscopy
of the supposed Sgr red clump stars in the ASA184 field and have produced a radial
velocity distribution. Thus, if the Sgr interpretation is born out, we provide the only
velocity measure of the debris stream at such large distances from the Sgr core. (4) While
the velocity distribution we derive does show a curious trend that is suggestive of the
presence of a ``moving group'', the mean radial velocity of the stars is unlike expectations
extrapolated from previous radial velocity studies of Sgr near its core. This has inspired a
new dynamical model, presented in a companion paper (Johnston et al. 1999, ``Paper II''
hereafter), of the destruction and present state of the Sgr dwarf spheroidal. The new model,
which is unique in that it accomodates all previous Sgr observations (as well as our new and
unexpected results) naturally and self­consistently, explains the low latitude results (those
presented here and by M98) as being derived from a separate, but overlapping Sgr tidal
tail stripped off on a previous perigalactic passage of Sgr. (5) There is value in pointing
out anomalous findings in field star studies, even if only to point out phenomena meriting
further study, to alert colleagues of potential problem areas, and to underscore our present
lack of understanding of the nature of the Galactic stellar populations. We point to the
precedent of Rodgers et al. 1990, whose presentation of an unusual excess population in a
similar direction of the sky bears some resemblance to our report here.

-- 5 --
2. Photometric Observations and Analysis
The fields ASA184 and ASA107 were observed in the V and I passbands as part of the
SASS program in July, 1996 on the Swope 1­meter telescope at Las Campanas using the
thinned 2048 2 TEK5 CCD chip. The large format CCD at the 29.''03 mm \Gamma1 Swope image
scale gives a 23.'5 field of view. In each of these two antifields, a narrow strip of images was
taken approximately perpendicular to the major axis of Sgr, with field centers displaced by
20' to retain some overlap as a check on the photometry. However, some fields are displaced
laterally as much as 20' from this alignment so as to avoid bright foreground stars (see
Figure 2).
Reductions were carried out with the IRAF package CCDRED. Photometry was
performed with the IRAF version of DAOPHOT (Stetson 1987) using a variable PSF.
Photometric calibration was obtained via Landolt (1992) standards, the IRAF package
PHOTCAL, and our own code. Photometry presented here reaches a depth of V = 21 in
nine ASA184 subfields (1.18 deg 2 ), eight ASA107 subfields (1.07 deg 2 ), nineteen SA107
subfields (2.42 deg 2 ) and fifteen SA184 subfields (1.96 deg 2 ). Photometric errors at V = 18:5
are (oe V ; oe V \GammaI ) = (0.02,0.03).
The CMD's of ASA184 and ASA107 are shown in Figure 3. In this paper, we adopt
reddening coefficients, EB \GammaV , of 0.05 for ASA107, 0.07 for for SA107, 0.08 for ASA184
and 0.02 for SA184. Differential reddenings from CCD field to CCD field are estimated
by intercomparing the peak of the blue edge on the (V \Gamma I; V ) CMD in one magnitude
brightness bins centered at V = 16:0; 16:5 and 17:0; the estimated error in the differential
reddening from this technique is 0.01 magnitude. For all four fields, the relative reddening
differences match the Burstein & Heiles maps (1982) well, and, in the end, we set the
absolute reddening from these maps. We have taken E V \GammaI = 1:24E B \GammaV and A V = 3:1E B \GammaV
from Cardelli, Clayton & Mathis (1989).
On first inspection, a distinctive feature in the CMD corresponding to Sgr stars is not
apparent in either ASA field in Figure 3 -- any such signal is lost in the noise of the Galactic
field star population. Any relative contribution by Sgr can be amplified by considering only
stars with colors consistent with those expected for the satellite's populous red clump and
avoiding the prominent, blue, MSTO edge of the field star CMD. We therefore consider
stars with 0:90 ! (V \Gamma I) o ! 1:10. In this restricted color range, the field star magnitude
distribution rises monotonically to fainter magnitude, while the Sgr population should
produce a sharp peak at the apparent magnitude appropriate to the red clump. I97
estimated the FWHM of this peak to be 0.1 magnitudes from a CMD in the center of the
Sgr dSph.

-- 6 --
Figure 4a shows a magnitude histogram of ASA184 stars with 0:90 ! (V \Gamma I) o ! 1:10.
A slight peak is apparent in the ASA184 data at V o = 17:85, the magnitude expected for
the red clump of Sgr at a distance modulus of (m \Gamma M)=16.8 given the proper reddening
correction and using the I97 absolute magnitude, M V =1.04, for the red clump. The
distance modulus would be close to the prediction of I97 for this latitude. This V o = 17:85
overdensity is strongest in a limited range of Galactic longitudes we have sampled. The
histogram for stars in three specific, adjacent ASA184 subfields (numbers 2­4 in Figure
2), centered at (l; b) = (10:9; \Gamma40:2), or (ff, ffi) 2000
= (20:59:58, ­33:21:01), show that the
contrast of this peak is increased from little more than 2oe (Figure 4a) to more than 3oe
(Figure 4b). Figure 3c shows the CMD of these subfields with the SL95 Sgr isochrone
superimposed, shifted to a distance modulus of 16.8. Unfortunately, the dominant field
star population still allows no more than a slight impression of a possible Sgr red clump in
either Figures 3a and 3c.
We compare these results to the magnitude histograms in the same color range for
ASA107, SA184 and SA107 (Figures 4c­4e, respectively). No V o = 17:85 peak is seen in
these other fields, nor in comparison CMD regions -- given by 1:15 Ÿ (V \Gamma I) o ! 1:40 --
in any of the fields including ASA184. Moreover, for either color range no field shows
any other peak of nearly the same significance as the V o = 17:85 peak in ASA184, which
suggests that this feature is unique in the data set. The ASA184 peak is robust to change
in bin size and bin center, generally appearing at the 2­3 oe level of significance.
This statistical peak is not, in itself, strong evidence of the presence of Sgr. In a
large survey parsed into small color­magnitude bins, one might expect to find occasional
deviations from a uniform distribution. However, such a peak happening to fall near the
magnitude and color of the expected Sgr clump at a latitude consistent with the expanding
tidal stream of Sgr (and now consistent with the M98 detection) is suggestive and warrants
further investigation.
If Sgr is present in these subfields of ASA184, it should be possible to identify its more
populous upper main sequence and turnoff stars through deeper imaging. On UT 14­16
July 1997, we imaged two 0.022 deg \Gamma2 ASA184 subfields using the du Pont 2.5­m and the
same CCD used in the Swope observations. One deep CCD image was centered at the
location of the possible red clump excess (ASA184 subfield 3), the other approximately two
degrees west along b = \Gamma40 ffi (in ASA184 subfield 8, which does not show the V o = 17:85
excess in the Swope data). These new photometric data, which have a magnitude limit of
V o = 24:25, were reduced similarly to the shallower Swope data. Figure 5 shows the CMD
of these two subfields. An excess of objects at (V \Gamma I) o = 0:8 and 20:5 ! V o ! 24 appears in
ASA184­3 (Figure 5a) compared to ASA184­8 (Figure 5b). In Figure 5c, the Figure 5a data

-- 7 --
are shown again with the F96 isochrone for Sgr plotted. The excess in Figure 5a is matched
by the Sgr MSTO region if we assume, as before, a Sgr distance modulus of (m \Gamma M) o =
16.8
As further support for the presence of a Sgr MSTO in ASA184­3, we show in Figure
5d the cumulative starcounts over 17:75 ! V o ! 22:25 for ASA184­3 and ASA184­8 in two
color ranges: 0:6 ! (V \Gamma I) o ! 0:8, which would bracket the main contribution of a Sgr
MSTO and upper main sequence in the magnitude range shown, and 1:1 ! (V \Gamma I) o ! 1:3,
which should miss most of a (m \Gamma M) o = 16.8, Sgr isochrone at these magnitudes. We note
that while the starcounts of the redder color range track each other well, the cumulative
starcounts for the MSTO color range do indeed show a break near V o ú 20:75 in the
ASA184­3 subfield which is not seen in the ASA184­8 subfield. A Kolmogorov­Smirnov test
of the null hypothesis that the plotted cumulative distributions for ASA184­3 and ASA184­8
are drawn from the same population gives a signficance level of 0.95 for the redder color
range (which yields a total of 25 stars in ASA184­3 and 22 stars in SA184­8), but a level
of only 0.23 for the MSTO color range (represented by 98 stars in ASA184­3 and 67 stars
in ASA184­8). While this KS significance does not strongly rule out a similar parent
population, 0.23 is at the upper end of KS significances (as low as a few percent) found for
various binnings tested (including wider bin widths with increased sample sizes and better
statistics that include more of the ``SGR MSTO'' region) blueward of (V \Gamma I) o = 0:95. For
the case illustrated in Figure 5d, the significance of the excess from Poissonian statistics is
2:7oe. Further deep imaging over more area in this region of the sky, away from the ``spine''
of fields covered by M98, would be a great help to improving the statistics of these tests
and verifying the width of the Sgr stream at these latitudes.
3. Surface Density Considerations
If we have indeed detected tidal stellar debris of Sgr, we can place some estimate on
the surface brightness of that debris. The implied excess of stars in our red clump CMD bin
is some 10­14 stars in 0.41 deg 2 based on a comparison to our Galactic model (see below),
which matches other parts of the CMD well. Therefore, we adopt 35 red clump stars deg \Gamma2
as an approximate upper limit on the excess.
Note that the lowest isopleth plotted in I95 for red clump stars is 1800 deg \Gamma2 , and this
is expected to be a ¸ 50% completeness contour. Thus, the density of our red clump stars
26 ffi from the Sgr core at M54 appears to be lower by about two orders of magnitude from
the lowest I95 isopleths which extend to about 5 ffi from the Sgr core near M54, where the
red clump densities reach a factor of roughly three higher still (but note that the red clump

-- 8 --
density of SL95 stars 12 arcmin north of M54 is roughly 1800 deg \Gamma2 -- see I97 Figure 2 -- a
factor of two below the I95 and I97 isopleths at this point).
In I97, Sgr maps are given utilizing stars within a magnitude of the MSTO. In this
case, isopleths of MSTO stars at 3600 deg \Gamma2 reach to about 13 ffi from the Sgr core at M54
(where the Sgr MSTO density is a factor of 10 higher). From Figure 5d, we see that our
excess of stars within one magnitude of the MSTO is roughly 14 stars in .022 deg 2 , and this
yields a density of MSTO stars 26 ffi from the Sgr core of ¸ 1200 deg \Gamma2 , a fall off of only a
factor of three across the 13 ffi from the last I97 point 1 . The latter suggests a more gradual
falloff of Sgr light in the last 13 ffi along the stream to the point of our detection than in the
first 13 ffi from the Sgr core south. Such a ``break'' in the surface density of Sgr is precisely
what is observed by M98 in their systematic survey along the length of the Sgr southern
tidal arm.
In M98, the surface density of Sgr MSTO stars is derived to a Galactic latitude
of b = \Gamma48 ffi using a slightly different bin in (I; V \Gamma I) space, which we estimate as
19:6 ! I ! 20:3 and 0:4 ! V \Gamma I ! 0:8. At this latitude, M98 detects a density of
approximately 1200 stars deg \Gamma2 , which includes both background Galactic stars and Sgr
MSTO stars in approximately equal numbers. Our data detect 26 stars in this region of
color­magnitude space in the deep 0.033 deg 2 observation of ASA184­3, which corresponds
to 1200 stars deg \Gamma2 , a result identical to the M98 result. This would place our surface
brightness as identical to M98 at this latitude (29.8 mag arcsec \Gamma2 ).
The surface brightness implied by the red clump stars should produce an identical
result. In I95, the density of Sgr red clump stars in the center of Sgr is set at 1.125 stars
arc­minute \Gamma2 . Our corresponding density of 0.01 stars arc­minute \Gamma2 (assuming only half of
the stars to be Sgr as in M98) would place our surface brightness 5.0­5.7 magnitudes below
the core surface brightness of Sgr, or 30.2­30.9 mag arcsec \Gamma2 -- a value slightly fainter, but
fairly consistent with, the MSTO calculation.
4. Radial Velocities
In August 1997, September 1997, August 1998, and October 1998 we used the du Pont
2.5­m telescope at Las Campanas to obtain spectra of the stars in the 17:75 ! V o ! 17:95
peak of Figure 4b. The celestial coordinates and undereddened colors and magnitudes for
1 Because I97 do not explicitly name the color range used in their isopleth map, it is not clear whether
we have sampled exactly the same region of the Sgr MSTO.

-- 9 --
these stars are listed in Table 1 (the dereddened colors and magnitudes used in our analysis
are obtained by assuming E(V \Gamma I) = 0:1 and A V = 0:25). Typical errors in our V and
V \Gamma I are about 0.015 and 0.020 magnitudes, respectively. Our spectra cover the wavelength
range from 4800­6800 š A at a resolution of 2.5 š A and include the Mgb triplet, MgH band
(5211 š A bandhead), and Na D doublets, which have been shown (Deeming 1960, Friel 1987,
Paltoglou & Bell 1994, Tripicchio et al. 1997) to be useful discriminants between metal
poor, evolved stars and more metal rich dwarfs. The data also encompass several iron lines
useful for evaluating [Fe/H].
The spectra are of sufficient signal­to­noise to allow measurement of line indices and we
anticipated being able to elucidate luminosity classes for our sample of candidate Sgr red
clump stars, segregating foreground Galactic dwarfs. In fact, we found the results of such an
analysis 2 to be ambiguous. The separation of disk dwarfs from red clump stars -- which, for
the Mgb, MgH and NaD features, are expected to have line strengths intermediate between
those of giants and dwarfs of the same metallicity -- is complicated in the case of Sgr red
clump stars by virtue of the apparently high mean metallicity (photometrically determined
to be [Fe/H]=­0.52 by SL95) and large spreads of abundances in this galaxy. As determined
by the high resolution spectroscopic analysis of Smecker­Hane, McWilliam & Ibata (1998),
Sgr has some metal­poor stars of [Fe/H] ú \Gamma1:5; however this same analysis indicates the
existence of Sgr stars as metal rich as [Fe/H]=+0.11. Marconi et al. (1999) report an even
more extreme result in their study of 23 candidate Sgr stars; these authors find no Sgr stars
as metal poor as [Fe/H] = ­1.5, and apparently find a rather high mean abundance, and
stars as metal rich as [Fe/H]=+0.7. The effect of such high abundances on the line strengths
of evolved stars is to counteract the suppression of absorption associated with decreasing
surface gravity. Contributing to this modulation of line strengths by [Fe/H] is an apparent
mean overabundance of magnesium in Sgr. From their table of representative Sgr stars,
Marconi et al. (1999) find a mean [Mg/Fe] overabundance of +0.26. Such ambiguities were
possibly manifest in our analysis 3 , which yielded the bulk of the 30 stars in the sample to
have line strengths between those of metal rich giants and metal poor dwarfs, but, in many
cases, with implied surface gravities dependent on which specific spectroscopic indicator
was selected.
2 Our analysis employed use of (a) Friel's (1987) Mgb+H index, (b) a wider index sensitive to MgH
absorption over 5000­5250 š A (c) the equivalent width of the NaD doublet, fixed to continuum levels at 5875
and 5909 š A , and (d) a variant of the Friel (1987) [Fe/H] measure using five available iron lines, and for
which Friel calibrated stars of luminosity class III and IV.
3 The referee mentions having had similar problems in determining abundances for stars in the main body
of Sgr, which he has found to have quite peculiar element ratios.

-- 10 --
Because of these and other difficulties, in the end we found a line strength analysis of
our spectra to be premature, and beyond the scope of the present contribution. Further
work along these lines will require a larger sample of stars with improved signal­to­noise,
a more thorough calibration incorporating a sizable sample of bona fide Sgr red clump
stars for comparison, a consistent set of photometric colors for both standard and target
stars, and better stellar atmosphere models specifically constructed for red clump stars. As
a clearer picture of the complex Sgr abundance patterns emerges, it may be possible to
take better advantage of the line strength information in our spectra. We can provide our
spectroscopic data to interested parties upon request.
Nevertheless, our spectra are of suitable quality for the measurement of radial
velocities. In Table 1 we present the heliocentric radial velocities derived for our red
clump sample. Since all of our stars are in a small region of the sky, the offset from
heliocentric to Galactocentric frames is nearly constant at +21 km s \Gamma1 , if a rotational
velocity for the local standard of rest of 220 km s \Gamma1 and a solar motion with respect
to that of (U; V; W ) = (\Gamma9; 11; 6) km s \Gamma1 are adopted. Velocities were obtained via a
cross­correlation analysis against well established, bright velocity standards. To minimize
flexure induced systematics, comparison arcs for each target and calibration source were
always taken right after, and at the same telescope position as, the stellar spectra. Strong
spectral lines were left out of the cross­correlation analysis. The location of the peak
(``CCP'') of the cross­correlation curve was taken as the radial velocity difference between
target and calibration spectrum; the CCP value is included in Table 1. The typical error
in the radial velocities is 12 km s \Gamma1 , as determined by repeat measures of a number of
stars (including numerous stars beyond those presented here) on different observing runs
and after accounting for a mean run to run systematic offset velocity determined from all
repeat stars. However, in some cases, due to low signal to noise, a CCP was rather weak
and the radial velocity is rather more suspect. In cases where CCP ! 0:30 (radial velocities
marked as ``::'' in Table 1) or 0:30 ŸCCP! 0:40 (radial velocities marked as ``:'' in Table
1), an observation with higher signal to noise was sought. For the subsequent analysis, we
have adopted radial velocities with CCP – 0:40 over other measured radial velocities for
the same star if the latter velocities have CCP ! 0:40. Otherwise, multiple radial velocity
measures have been averaged.
Because the field surveyed, ASA184, is almost directly below the Galactic center,
the mean radial velocity of stars from both nearby and very distant stars (whose mean
rotational component is predominently perpendicular to the line of sight) are expected to
be around V helio ¸ !0 km s \Gamma1 . For stars near the Galactic center, the line of sight becomes
tangent to their mean rotation around the Galactic center and the radial velocity reflects
more of their mean rotational velocity. In Figure 6 we show the expected mean heliocentric

-- 11 --
radial velocities for stellar populations of different mean rotational rates around the Galactic
center as a function of distance from the Sun in the direction of the field ASA184. We
consider four possible contaminants of our ``Sgr red clump'' sample, which is centered in the
CMD at V o = 17:85 and (V \Gamma I) o = 1:0:
Halo giant stars: At V o = 17:85 and b = \Gamma40 ffi any giant star with (V \Gamma I) o = 1:0 will be
well into the halo. If we adopt stars of the metallicity of M3 and M13 as typical of the halo,
we derive a distance of 37 kpc. The expected mean radial velocity for such contaminants
would be ! \Gamma25 km s \Gamma1 (Figure 6) but with a rather large (¸ 100 km s \Gamma1 ) dispersion.
Horizontal branch stars: With (V \Gamma I) o = 1:0 these would most likely be red clump stars. If
we adopt the I97 absolute magnitude for Sgr red clump stars as typical for ``contaminant''
red clump stars from the halo field, they would be at a distance of about 23 kpc. The mean
heliocentric radial velocity for these stars would be ! \Gamma10 km s \Gamma1 , but again with a large
velocity dispersion if from a random halo field population.
Metal rich dwarfs: If solar abundance, a (V \Gamma I) o = 1:0 dwarf would be 2.2 kpc distant
and 1.4 kpc (some four old disk scaleheights) below the Galactic plane. The expected
mean radial velocity for these stars would be ! 0 km s \Gamma1 , and, while of solar metallicity,
the velocity dispersion would need to be of order 40 km s \Gamma1 or so for them to be at
such large distances from the Galactic mid­plane. One might not expect an abundance
of solar metallicity stars at this z­distance; we present them here as one limit for dwarf
contaminants, to compare to...
Metal poor subdwarfs: For a subdwarf 3 magnitudes below the nominal solar abundance
main sequence, the distance is some 550 pc, and 350 pc below the Galactic plane. Thus,
most dwarf contaminants in the sample are likely to be somewhere between 0.5 and 2.2 kpc
distant, and, from Figure 6, should show mean heliocentric velocities near \Gamma10 km s \Gamma1 for
typically expected, ``thick disk'' asymmetric drifts.
What should we expect as ``contaminants'' of our ``SGR red clump sample'' as predicted
from standard Galactic models? Using star count model B from Reid et al. (1996), which
we have already shown to give reasonable fits to the SA107 and SA184 data (SMRT97; see
also Figure 4), we predict 40­50 stars deg \Gamma2 for 17:75 ! V o ! 17:95 and 0:9 ! (V \Gamma I) 0
! 1:1
at this (l; b) (or 16­20 stars in our spectroscopic sample of 30 stars). This expected number
is in reasonable (though not exact) agreement with the density of stars (a) in this same
color­magnitude bin for the Galactic structure fields ASA107, SA184 and SA107 (Figures
4c­e), and (b) with the density of stars in ASA184 in the magnitude bins adjacent to the
V o = 17:8 bin in Figures 4a and 4b. In our model, about one third of the expected 16­20
stars are predicted to be from the disk and two­thirds from the thick disk; only a few

-- 12 --
stars are expected to be evolved (M V ! 3) stars from a homogeneously distributed halo
population (and these will be predominently giants). Assuming mean rotational velocities
of ¸ 200, ¸ 170, and ¸ 40 km s \Gamma1 for the old thin disk, the thick disk, and the halo,
respectively, then we should predict some ten or so stars centered around V helio ¸ \Gamma10 km
s \Gamma1 (Figure 6) with a dispersion of ¸ ! 50 km s \Gamma1 or so (mostly a combination of the oe U and
oe W parts of the thick disk velocity ellipsoid), some five stars more tightly clumped around
V helio ¸ \Gamma10 km s \Gamma1 (represeting the old thin disk), and a few halo giants broadly spread
around V helio ¸ \Gamma30 km s \Gamma1 . It is not clear what population the ``excess'' dozen or so stars
should represent in the sample of 30 stars.
We find the mean heliocentric radial velocity for all 30 stars in the sample is 4 km s \Gamma1 ,
with an RMS of 58 km s \Gamma1 . At first blush, the distribution of velocities (Figure 7) seems
roughly consistent with expectations for a mix of thin and thick disk stars and a smattering
of halo stars, but from Figure 6 and the starcount model predictions detailed above we
should expect the velocities for any combination of stellar populations to skew towards
negative V helio , rather than positive V helio as is found (Figure 7). We note, for example, that
16 of the 30 stars have V helio ? 0 (the median is 9.5 km s \Gamma1 ), when the majority of stars
should have V helio ! 0.
Despite the skew to V helio ? 0, the lack in our data of a distinct radial velocity signal
separate from that of the Galactic disk stars is unfortunate in that it does not allow us
immediately to isolate specific members of any excess population above the predicted
Galactic stellar populations, and, indeed, may even call into question the reality of such
an excess population in the spectroscopic sample. However, we call attention to a color
dependence of the radial velocity distribution which does seem to point to the presence of
a distinct radial velocity signal. As shown in Figure 8, there is a dichotomy in the radial
velocity distributions on either side of (V \Gamma I) o ¸ 1:0, with a difference in the mean velocity
of stars of 47 km s \Gamma1 . The bluer stars show a mean V helio = \Gamma17 km s \Gamma1 and velocity
dispersion of 67 km s \Gamma1 while the redder stars show a mean velocity of V helio = 30 km s \Gamma1
and much smaller velocity dispersion of only 27 km s \Gamma1 . The velocity distribution and
number of stars in the bluer population are very similar to the expectations for Galactic
stellar populations as outlined above.
On the other hand, the velocity dispersion for the redder stars is much less than
expected for even a random sample of only Galactic disk dwarf stars in this direction
of the sky; however, the red star velocity dispersion is more typical of expectations
for halo ``moving groups'', and we note that if this redder sample should also contain
some contribution by Galactic stars, they would tend to increase the velocity dispersion.
Moreover, the mean radial velocity of the redder sample is more than 40 km s \Gamma1 higher than

-- 13 --
expectations for Galactic disk stars. A Kolomogorov­Smirnov test that the radial velocity
distributions on either side of (V \Gamma I) o = 1:0 are drawn from the same population gives
only 1% confidence.
The relevance of this velocity­color disparity may lie in a closer inspection of the
Sgr CMD in SL95 (see their Figures 4 and 6b) which seems to show (1) the Sgr red
clump concentrated to 1:0 ! (V \Gamma I) o ! 1:15 and (2) a relative dearth of stars for
0:9 ! (V \Gamma I) o ! 1:0. Perhaps a more appropriate color bin for our selection of Sgr
red clump stars should have been 1:0 ! (V \Gamma I) o ! 1:15, and we should expect Sgr red
clump stars primarily in the (V \Gamma I) o ? 1:0 part of the spectroscopic sample, but not the
(V \Gamma I) o ! 1:0 part of the spectroscopic sample. While this would seemingly dovetail nicely
with the color dependence of the radial velocities described above, we must also point out
that this parsing of the spectroscopic sample by color yields no consistent excess in counts
by our photometric data (both the red and blue sample show a similar low­significance
excess). That is, to identify the 1:0 ! (V \Gamma I) o ! 1:1 stars that show the much tighter
velocity dispersion as a ``moving group'' we have to believe simultaneously that the
major contribution of the Galaxy to our spectroscopic sample just happened to be in the
0:9 ! (V \Gamma I) o ! 1:0 sample. This is an unlikely assertion and stresses the need for a larger
statistical sample to understand these apparent contradictions.
We conclude from this analysis that if the Sgr red clump population is indeed present
in our spectroscopic sample, its mean velocity is likely to be rather low -- a mean velocity
of V helio = 30 km s \Gamma1 or V GSR = 51 km s \Gamma1 if we adopt the 1:0 ! (V \Gamma I) o ! 1:1 stars as
representative, and lower if we do not. We note that most of the previous velocity surveys
of Sgr near its core determined mean velocities greater than V GSR = 150 km s \Gamma1 , whereas
we have only two stars with velocities that large. As we discuss in the companion paper,
this discrepancy with previous radial velocity work does not, in itself, discount our having
detected the Sgr red clump in our spectroscopic sample.
5. Discussion
5.1. Sagittarius or Not Sagittarius?
We are presented with the following set of apparently conflicting circumstances: (1)
the implication of our model and starcount data presented in Figure 4b is that there is a
50­100% excess of stars in the ASA184 CMD in the magnitude and color range of interest;
(2) this particular part of the CMD is a reasonable match to the expected location of the
prominent Sgr red clump, and a Sgr­like isochrone and the appearance of a prominent

-- 14 --
Sgr­like MSTO is implied by our deep CMD in this part of the sky, as presented in Figure
5; (3) the surface brightness of stars in these two regions are consistent with each other,
consistent with an extrapolation of the Sgr extent into this field, and consistent with the
results of M98; but (4) the radial velocity distribution of stars in this field, while dissimilar
from expectations for that of Galactic stars, is not grossly so when the sample is considered
in total; (5) there is a color trend in the velocity distribution, which is not supported by the
location of the count excess, but which does have an intriguing correspondence to the color
range of Sgr red clump stars in SL95; (6) we find the spectral line strength indices to yield
ambiguous gravity/abundance information.
At this point, we are uncertain of how to account for (5) and both (5) and (6) clearly
require further observations to resolve. We propose three ways to reconcile (1)­(4): (a) Our
photometric results are merely a statistical fluke in the large SASS project and these stars
are all Galactic; (b) we have found a stellar system with an extended spatial distribution
that has a CMD extremely similar to Sgr, is at about the same distance as Sgr, has about
the same surface brightness as Sgr in this part of the sky, but that is not Sgr; or (c), we
have indeed identified Sgr, but in our field Sgr has a very different radial velocity than seen
near the center of the dwarf galaxy.
Given the consistency of our photometric results with each other and with M98,
we find proposition (a) unlikely, although not impossible. A larger survey with deep
imaging would address this issue conclusively. Proposition (b) would require the unique
circumstance of two distinct stellar systems overlapping in surface density and position,
which is also unlikely, but not impossible. In the companion paper (Johnston et al. 1999),
we demonstrate how proposition (c) is viable if our observations are revealing a different
tidal tail of Sgr than observed at higher Galactic latitudes. With semi­analytical modeling
of the destruction of the Sgr dwarf and by following the fate of debris shed by perigalacticon
passages of the dwarf on a variety of orbits, Johnston et al. conclude that lines of sight
progressively farther from the Sgr core should intercept additional debris streamers of Sgr
with differing radial velocities and distances. The possible detection of Sgr here, with the
distance and velocity characteristics we derive, match remarkably well to the predictions for
a tidal streamer torn from Sgr three perigalactic passages ago. Spectroscopy of Sgr stars in
the nearby M98 fields would be an important test of the various propositions put forth here
and in Johnston et al. (1999).

-- 15 --
5.2. Sgr in the Northern Hemisphere
From the lack of a Sgr detection in our ASA107 field at b = +40 ffi , it is tempting to
impose an upper limit to Sgr's possible extent in this direction. However, we hedge on such
an interpretation. In the first place, the orbital path of the dSph is not precisely known
above the Galactic equator. It is not only possible, but likely, that we missed the orbit
entirely and that the ASA107 data do not constrain the length or age of the tidal stream.
It is worth noting that in a more careful treatment of the Sgr disruption, Johnston (1998)
shows that it would take a mere 3 Gyr for Sgr tidal debris to wrap entirely around 360 ffi , if
it has a mass of 10 8 M fi .
Most studies of the Sgr stream have concentrated on its southern extent since studying
the northern tail means working into the Galactic plane, a difficult prospect. Nevertheless,
observations at b ? 10 ffi are possible and a Sgr component might be expected at b – 10 ffi
given the 25 ffi southern extent we may have detected in this paper and that is seen by M98.
5.3. Implications for Galactic Structure Studies
Finally, it is important to place our likely detection of Sgr within the context of
the original aims of our SASS survey, which were to constrain global density models of
Galactic stellar populations. From the mere existence of the Sgr system, it is evident that
simple triaxial models and symmetric analytic expressions are insufficient to describe the
outer populations of the Milky Way. The recent work by Johnston (1998) is particularly
relevant to this discussion. She finds that after 10 Gyr, the fraction of the sky that would
be covered by the debris of a disrupted, 10 8 M fi Sgr would be on the order 5000 deg 2 ,
while the LMC should cover some 10,000 deg 2 . However, to the extent that there do exist
more dynamically relaxed disributions of stars, the expansive diaspora of tidal debris may
lurk as a mere tittle against an overwhelming stellar background (e.g., see Johnston 1998,
Figure 8). With a wider canvas of SASS fields, we hope to evaluate the veracity of simple
analytical approaches to modeling the structure of the Milky Way, and determine the level
of need for superposition of irregularities in model Galactic density distributions.
This research was supported by NSF grants AST­9412265 to IT and SRM, AST­9412463
to INR, and AST­9528177 to AUL. We thank Rodrigo Ibata for useful comments in the
refereeing stage. We dedicate this work to the memory of our friend and colleague, Jerry
Kristian.

-- 16 --
REFERENCES
Alard, C. 1996, ApJ, 458, L17
Alcock, C., Allsman, R. A., Alves, D. R., Axelrod, T. S., Becker, A. C., Bennett, D. P.,
Cook, K. H., Guern, J. A., et al. 1997, ApJ, 474, 217
Burstein, D. & Heiles, C. 1982, AJ, 87, 1165
Cardelli, J. A., Clayton, G. C. & Mathis, J. S. 1989, ApJ, 345, 245
Deeming, T. J. 1960, MNRAS, 121, 52
Fahlman, G. G., Mandushev, G., Richer, H. B., Thompson, I. B. & Sivaramakrishnan, A.
1996, ApJ, 459, L65 (F96)
Friel, E. D. 1987, AJ, 93, 1388
Ibata, R. A., Gilmore, G. & Irwin, M. J. 1994, Nature, 370, 194 (I94)
Ibata, R. A., Gilmore, G. & Irwin, M. J. 1995, MNRAS, 277, 781 (I95)
Ibata, R. A., Wyse, R. F. G., Gilmore, G., Irwin, M. J. & Suntzeff, N. B. 1997, AJ, 113,
634 (I97)
Johnson, J. A. & Bolte, M. 1998, AJ, 115, 693
Johnston, K. V. 1998, ApJ, 495, 297
Johnston, K. V., Majewski, S. R., Siegel, M. H., Kunkel, W. E., Reid, I. N. 1999 submitted
Landolt, A. U. 1992, AJ, 104, 340
Majewski, S. R. 1994, in IAU Symposium 161: Astronomy from Wide Field Imaging, eds.
H. T. MacGillivray et al., (Dordrecht: Kluwer), p. 425
Marconi, G., Bonifacio, P., Pasquini, L. & Molaro, P. 1998, astro­ph 9812268
Mateo, M., Mirabal, N., Udalski, A., Szyman'ski, M., Ka/lu`zny, J., Kubiak, M., Krzemi'nski,
W. & Stanek, K. Z. 1996, ApJ, 458, L13 (M96)
Mateo, M., Olszewski, E. W. & Morrison, H. L. 1998, ApJ, 508, L55 (M98)
Monet, D. G., Dahn, C. C., Vrba, F. J., Harris, H. C., Pier, J. R., Luginbuhl, C. B. &
Ables, H. D. 1992, AJ, 103, 638
Olling, R. P. & Merrifield, M. R. 1998, MNRAS, 297, 943
Paltoglou, G. & Bell, R. A. 1994, MNRAS, 268, 793
Reid, I.N. & Gizis, J.E. 1997, AJ, 113, 2246
Reid, N. & Majewski, S. R. 1993, ApJ, 409, 635

-- 17 --
Reid, I. N., Yan, L., Majewski, S., Thompson, I. & Smail, I. 1996, AJ, 112, 1472
Rodgers, A. W., Harding, P. & Walker, I. 1990, AJ, 100, 1841
Sarajedini, A. & Layden, A. C. 1995, AJ, 109, 1086 (SL95)
Siegel, M. H., Majewski, S. R., Reid, I. N. & Thompson, I. 1997, in Proper Motions
and Galactic Astronomy, ed. R. M. Humphreys, PASP Conf. Ser. Vol. 127, 175
(SMRT97)
Siegel, M. H., Majewski, S. R., Reid, I. N., Thompson, I., Landolt, A. U., & Kunkel, W. E.
1997, BAAS, 29, 1341
Smecker­Hane, T., McWilliam, A. & Ibata, R. 1998, BAAS, 30, 916.
Stetson, P. B. 1987, PASP, 99, 191
Tripicchio, A., Severino, G., Covino, E., Terranegra, L. & Garc ' ia L'opez, R. J. 1997, A&A,
327, 681
Zijlstra, A. A. & Walsh, J. R. 1996, A&A, 312, L21
This preprint was prepared with the AAS L A T E X macros v4.0.

-- 18 --
TABLE 1.
Star ff ffi V V \Gamma I Obs Date Exptime RV helio CCP
J2000.0 J2000.0 mmddyy (sec) (km=sec)
01 21:00:05.7 ­33:30:15 18.028 1.049 081197 2700 ­137.5 :: 0.25
02 20:59:35.2 ­33:31:39 18.146 1.161 081197 1800 ­10.8 0.45
03 20:59:41.3 ­33:32:11 18.024 1.043 081198 900 99.2 : 0.33
102898 1800 114.1 0.61
04 21:01:04.8 ­33:34:30 18.199 1.075 102898 1440 ­30.4 0.70
05 21:01:33.7 ­33:39:55 18.115 1.073 081497 1350 10.6 :: 0.22
081298 1800 19.6 0.86
06 20:50:48.2 ­33:50:57 18.099 1.073 081298 1200 ­113.6 0.75
07 21:00:22.4 ­33:29:59 18.141 1.170 102898 1080 17.9 0.73
08 20:59:26.2 ­33:33:17 18.022 1.007 082397 900 110.6 :: 0.28
081298 1350 113.2 0.90
09 20:59:58.6 ­33:35:31 18.027 1.034 081497 1350 ­37.4 : 0.37
081497 1350 ­36.9 : 0.37
081298 900 ­51.0 0.70
10 21:00:42.2 ­33:49:59 18.159 1.140 082497 900 ­14.0 0.44
11 20:59:26.2 ­33:49:47 18.023 1.140 081298 900 45.9 0.77
12 21:00:07.5 ­33:09:56 18.066 1.057 081298 900 37.5 0.73
13 21:00:12.7 ­33:20:34 18.130 1.188 081298 900 31.5 0.64
14 21:01:01.5 ­33:23:30 18.065 1.135 082497 900 61.2 : 0.30
15 21:00:22.6 ­33:10:05 18.016 1.156 081298 900 67.1 0.59
16 20:59:28.9 ­33:13:27 18.050 1.084 081397 1800 ­3.5 0.51
081397 1800 ­4.0 0.50
17 20:59:57.6 ­33:14:41 18.177 1.004 081298 900 ­145.9 : 0.32
102898 1800 ­73.7 0.63
18 21:01:11.0 ­33:17:23 18.035 1.064 081298 3600 ­15.5 0.72
102898 1800 ­23.1 0.83
19 21:01:17.1 ­33:21:56 18.018 1.164 081398 1440 22.7 0.77
102898 1800 ­2.7 0.75

-- 19 --
TABLE 1 (continued).
Star ff ffi V V \Gamma I Obs Date Exptime RV helio CCP
J2000.0 J2000.0 mmddyy (sec) (km=sec)
20 20:59:37.2 ­33:21:36 18.058 1.001 081297 1800 ­34.9 :: 0.20
102898 1800 9.1 0.58
21 21:00:43.9 ­33:27:08 18.077 1.110 082397 900 65.3 : 0.32
081398 1800 61.3 0.85
24 21:01:15.3 ­33:33:56 18.146 1.189 102898 1800 46.9 0.60
25 21:00:04.8 ­33:25:06 18.072 1.143 082397 900 10.2 0.45
081398 1080 7.4 0.73
27 21:00:48.0 ­33:03:07 18.154 1.167 081398 2720 36.0 0.80
28 20:59:38.8 ­33:06:46 18.081 1.027 081398 2720 ­32.9 : 0.35
30 21:00:08.2 ­32:51:50 18.021 1.066 082497 900 ­10.0 :: 0.24
081398 1260 ­51.5 0.69
31 21:00:32.4 ­33:03:22 18.075 1.032 102898 1350 ­37.4 0.47
32 21:00:16.3 ­33:05:57 18.090 1.167 081498 1440 33.2 0.68
33 21:00:06.3 ­33:06:54 18.058 1.092 081397 1350 ­65.2 :: 0.22
102898 1800 ­40.4 0.89
34 21:00:07.5 ­33:09:56 18.040 1.032 081498 1440 14.5 0.58

-- 20 --
Figure Captions
Fig. 1.--- Aitoff projection showing the location of the suspected Sgr globular clusters (circled
crosses), previous detections of stars in the Sgr stellar stream (solid dot for M96, dotted lines
delineate the survey areas of A96 and A97, and solid contour for the approximate outer
isophote of I97), Sgr planetary nebulae (crosses -- the two crosses outside the I97 isophote
are only possibly associated), our SA fields (squares) and our anti­SA fields (solid strips).
The long solid line shows the survey of M98.
Fig. 2.--- Placement of the CCD fields in the ASA184 strip. Subfield 6 was displaced to
avoid a bright star and subfield 9 is intended to provide a check on the photometric tie­in
between subfields 6 and 7.
Fig. 3.--- Dereddened color­magnitude diagrams (CMDs) of the (a) ASA184 and (b) ASA107
fields. Panel (c) shows the CMD of subfields 2, 3 and 4 of ASA184. Although the Sgr red
clump may be present in the field, the overwhelming contribution of the field star population
makes it difficult to discern. A Sgr isochrone from SL95 is superimposed on top of the data
in panel (c).
Fig. 4.--- Magnitude histograms of the (a) ASA184 , (c) ASA107 , (d) SA184 and (e) SA107
fields in the dereddened color interval 0:90 ! (V \Gamma I) o ! 1:10. In each panel, the dark solid
lines are our data normalized to 1 deg \Gamma2 for the color range 0:9 Ÿ (V \Gamma I) o ! 1:1, and,
for comparison, our computer model predictions for this color range (dashed lines) and the
data in a redder color bin (1:15 Ÿ (V \Gamma I) o ! 1:4, dotted lines). A few representative error
bars are given in each panel. We have determined EB \GammaV reddenings for each field as 0.08
(ASA184), 0.05 (ASA107), 0.02 (SA184), and 0.07 (SA107). Note the weak V = 17:8 peak
in ASA184. Panel (b) narrows ASA184 to the 2, 3 and 4 subfields. The peak at V o = 17:8
is much clearer in this smaller ASA184 sample.
Fig. 5.--- Dereddened color­magnitude diagrams of (a) the ASA184­3 subfield ((l; b) =
(10:9 ffi ; \Gamma40:2 ffi )), which shows the strongest Sgr detection and (b) the ASA184­8 subfield
((l; b) = (13:0 ffi ; \Gamma40:1 ffi )), which is nearly two degrees west (approximately along b = \Gamma40 ffi ),
away from the detection. Panel (c) shows the ASA184­3 CMD with the F96 isochrone
superimposed. Panel (d) shows the cumulative starcount distributions from the deeper du
Pont CCD data for ASA184­3 (solid lines) and ASA184­8 (dashed curves) for two color
ranges: 0:6 ! (V \Gamma I) o ! 0:8 (upper curves), which should include the bulk of the Sgr main
sequence turn off, and 1:1 ! (V \Gamma I) o ! 1:3 (lower curves), which should not contain many
Sgr stars in the magnitude range shown (17:75 ! V o ! 22:25). The discrepancy between the
shapes of the two upper curves suggests the presence of the Sgr MSTO in ASA184­3 (see
text).

-- 21 --
Fig. 6.--- Expected mean heliocentric radial velocities for stars of various mean rotational
velocities about the Galactic center as a function of distance (kpc) in the direction of ASA184.
The curves are drawn assuming a mean rotation for the local standard of rest (\Theta o ) of 220 km
s \Gamma1 , a solar motion of (U; V; W ) = (\Gamma9; 11; 6) km s \Gamma1 , and a solar Galactocentric distance
of 8.0 kpc. For stars of V o = 17:85 and (V \Gamma I) o = 1:0 (the magnitude and color of the
``Sgr red clump'' sample), we mark the approximate distances of stars of different possible
luminosities: solar abundance dwarf (absolute magnitude M V = 6:16 adopted from Reid &
Gizas 1997), a 3.0 magnitude subluminous subdwarf (M V as suggested by the subdwarfs in
Monet et al. 1992), a red clump star of similar luminosity to those in Sgr (adopting the
I97 M V = 1:04), and a red giant star of the abundance ([Fe/H] ¸ \Gamma1:55) of M3 and M13
(absolute magnitude of M V ¸ 0 derived from the data in Johnson & Bolte 1998). The effect
of lowering \Theta o is to offset the curves systematically in the positive direction; e.g., the Olling
& Merrifield (1998) recommended \Theta o = 184 km s \Gamma1 increases the expected radial velocities
shown by 5 km s \Gamma1 .
Fig. 7.--- Histogram of radial velocities for the Sgr red clump sample.
Fig. 8.--- The distribution of stars in radial velocity and color. Note the skew of the redder
stars in this sample. While it is tempting to assign these stars to Sgr, this assignment does
not necessarily square with the distribution of the count excess (see text).

-- 22 --

-- 23 --

-- 24 --

-- 25 --
3
magnitude
subdwarf
solar
metallicity
dwarf
Sgr­like
red
clump
star
M3/M13
red
giant
­10 0 10 20 30 40 50
Distance (kpc)
­50
0
50
100
150
200
Heliocentric
Radial
Velocity
(km/sec)
v rot =50
v rot =100
v rot =150
v rot =200

-- 26 --