astro-ph0411029/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: \newcommand\Msun {M_{\odot}\ }
3: \newcommand{\vdag}{(v)^\dagger}
4: \newcommand{\myemail}{etolstoy@astro.rug.nl}
5: 
6: \def\etal{\rm et al.~}
7: \def\eg{\rm e.g., }
8: \def\ie{\rm i.e., }
9: 
10: \begin{document}
11: 
12: \title{Two distinct ancient components in the Sculptor Dwarf
13: Spheroidal Galaxy: First Results from DART\altaffilmark{1}}
14: 
15: \author{Eline Tolstoy\altaffilmark{2}, M.J. Irwin\altaffilmark{3}, 
16: A. Helmi\altaffilmark{2}, G. Battaglia\altaffilmark{2}, 
17: P. Jablonka\altaffilmark{4}, 
18: V. Hill\altaffilmark{4}, K.A. Venn\altaffilmark{5}, 
19: M.D. Shetrone\altaffilmark{6}, B. Letarte\altaffilmark{2}, 
20: A.A. Cole\altaffilmark{2},
21: F. Primas\altaffilmark{7}, P. Francois\altaffilmark{4}, 
22: N. Arimoto\altaffilmark{8}, K. Sadakane\altaffilmark{8},
23: A. Kaufer\altaffilmark{9}, T. Szeifert\altaffilmark{9}, 
24: T. Abel\altaffilmark{10}, 
25: }
26: 
27: \altaffiltext{1}{Based on FLAMES and UVES observations 
28: collected at the European Southern Observatory, 
29: proposals 71.B-0641 and 171.B-0588}
30: 
31: \altaffiltext{2}{Kapteyn Institute, University of Groningen, 
32: Postbus 800, 9700AV Groningen, the Netherlands}
33: \altaffiltext{3}{Institute of Astronomy, University of Cambridge, 
34: Madingley Road, Cambridge CB3 0HA, UK} 
35: \altaffiltext{4}{Observatoire de Paris, section de Meudon, 
36: 5 Place Jules Janssen, F-92195 Meudon cedex, France}
37: \altaffiltext{5}{Macalester College, Saint Paul, MN 55105, USA}
38: \altaffiltext{6}{University of Texas, McDonald Observatory, USA}
39: \altaffiltext{7}{European Southern Observatory, 
40: Karl-Schwarzschild str 2, D-85748 Garching bei M\"{u}nchen, Germany}
41: \altaffiltext{8}{National Astronomical Observatory, 2-21-1 Osawa, 
42: Mitaka, Tokyo 181-8588, Japan}
43: \altaffiltext{9}{European Southern Observatory, Alonso de Cordova 3107, 
44: Santiago, Chile}
45: \altaffiltext{10}{Penn State University, State College, PA, USA}
46: 
47: \begin{abstract}
48: 
49: We have found evidence for the presence of two distinct ancient
50: stellar components (both $\geq$ 10~Gyr old) in the Sculptor dwarf
51: spheroidal galaxy.  We used the ESO Wide Field Imager (WFI) in
52: conjunction with the VLT/FLAMES spectrograph to study the properties
53: of the resolved stellar population of Sculptor out to and beyond the tidal
54: radius. We find that two components are discernible in the spatial
55: distribution of Horizontal Branch stars in our imaging, and in the
56: [Fe/H] and v$_{hel}$ distributions for our large sample of
57: spectroscopic measurements. They can be generally described as a
58: ``metal-poor'' component ([Fe/H]$< -1.7$) and a ``metal-rich''
59: component ([Fe/H]$> -1.7$).  The metal-poor stars are more spatially
60: extended than the metal-rich stars, and they also appear to be 
61: kinematically distinct.  These results provide an important insight
62: into the formation processes of small systems in the early universe
63: and the conditions found there.  Even this simplest of galaxies
64: appears to have had a surprisingly complex early evolution.
65: 
66: \end{abstract}
67: 
68: \keywords{dwarf galaxies: general ---
69: dwarf galaxies: individual(\objectname{Sculptor Dwarf Spheroidal})}
70: 
71: \section{Introduction}
72: 
73: The Sculptor (Scl) Dwarf Spheroidal galaxy (dSph) is a close companion
74: of the Milky Way at high galactic latitude ({\it b}~=~$-$83$^o$).
75: Located at a distance of 72 $\pm 5$~kpc (Kunkel \& Demers 1977), with
76: v$_{hel} =$ 109.9 $\pm 1.4$ km/s (Queloz, Dubath \& Pasquini 1995) it
77: has a low total (dynamical) mass, 1.4 $ \pm 0.6 \times 10^7 \Msun$
78: (Queloz \etal 1995).  Irwin \& Hatzidimitriou (1995) made the first
79: modern comprehensive determination of the physical properties of Scl.
80: They determined a core radius, r$_c = 5.8 \pm 1.6$ arcmin, and a tidal
81: radius, r$_t = 76 \pm 5$ arcmin with a position angle of 99 $\pm$ 1
82: degrees, and confirmed it to be a moderately luminous galaxy in the
83: dSph class (M$_V = -10.7 \pm 0.5$), with a modest central surface
84: brightness ($\Sigma_{o,V} = 23.5 \pm 0.5$ mag/arcsec$^2$).  Like most
85: dSph galaxies Scl does not appear to have much, if any, HI gas
86: (\eg Bouchard,
87: Carignan \& Mashchenko 2003).
88: 
89: Observations of Scl have revealed a sizeable population of RR
90: Lyrae variable stars (\eg Kaluzny \etal 1995), clearly indicating that
91: its stellar population contains a globular cluster age component.  Scl
92: also shows tantalizing evidence in its extended horizontal branch (HB)
93: for an unusual enrichment history (\eg Hurley-Keller, Mateo \& Grebel
94: 1999; Majewski \etal 1999).  
95: Monkiewicz \etal (1999) have detected stars several magnitudes
96: below the oldest possible main-sequence turn-offs. Their accurate
97: photometry allowed them to conclude that the mean age of Scl is
98: similar to that of a globular cluster, but that there is an age spread
99: within this epoch indicative of an extended star formation history.
100: Thus the entire star formation history of Scl apparently lasted only a
101: few Gyr, and after this initial period of activity more than 10~Gyr
102: ago Scl appears to have been dormant.
103: 
104: There have been several previous spectroscopic studies of individual
105: stars in Scl, firstly to determine the kinematic 
106: properties (\eg Armandroff \& Da Costa 1986; Aaronson \& Olszewski
107: 1987; Queloz et al. 1995) which have indicated a moderate
108: mass-to-light ratio M/L$= 9 \pm 6 $(M/L$_V$)$_\odot$.  This value is
109: significantly higher than that typical for a globular cluster, but
110: much lower than found for many dSphs (\eg Draco and Ursa Minor).
111: There have also been studies of the abundances of small samples of
112: individual Red Giant Branch (RGB) stars at low and high resolution
113: (Shetrone \etal 2003; Tolstoy \etal 2001, 2003).  These stars were
114: selected from the central region of Scl, and although they showed a
115: wider range in [Fe/H] than would be naively expected based on
116: colour-magnitude diagram (CMD) analysis, no conclusions could be drawn
117: on the large scale spatial variations of these properties.
118: 
119: 
120: Here we present the first results from the DART (Dwarf Abundances and Radial
121: velocities Team) large programme at ESO.  We present the results for
122: v$_{hel}$ and [Fe/H] measurements from our FLAMES spectroscopy of
123: 401 RGB stars for the first galaxy in our sample, Scl dSph.
124: The relatively high signal/noise, S/N ($\approx$ 10-20 per pixel) of our
125: data has enabled us to derive both accurate metallicites ($\approx$
126: 0.1 dex from internal errors) and radial velocities (to $\approx \pm$2
127: km/s).  This is the first time that a large sample of both types of
128: measurement has been made in a single galaxy. A more detailed
129: description of our data and data reduction techniques will be provided
130: in subsequent papers (Irwin \etal in prep; Hill \etal in prep;
131: Battaglia \etal in prep). Here we provide an overview of our first
132: intriguing results.
133: 
134: \section{Observations}
135: 
136: \subsection{Imaging}
137: 
138: The ESO Wide Field Imager (WFI) observations were obtained from the
139: ESO archive for the central region of Scl, and more extended
140: observations of outer fields were made by us.  We used the pipeline
141: reduction software developed by the Cambridge Astronomical Survey Unit
142: for processing mosaic camera imaging (Irwin \& Lewis 2001; Irwin \etal
143: 2004).  These observations were used to study the photometric
144: properties of Scl stars out to the tidal radius and as target
145: selection for VLT/FLAMES observations of stars along the RGB.
146: 
147: \subsection{Spectroscopy}
148: 
149: All VLT/FLAMES observations were made in Medusa mode, giving $\sim$120
150: fibres to be placed over a 25 arcmin diameter field of view.  Most of
151: the observations were taken with the Low Resolution grating LR8
152: (which includes the Ca~II triplet (CaT) lines), apart from the central
153: pointing where High Resolution gratings were used (see Hill \etal 
154: in prep).  The CaT method is well calibrated to provide an accurate
155: estimate of the [Fe/H] abundance of an RGB star (Rutledge, Hesser \&
156: Stetson 1997; Cole \etal 2004) and also to provide an accurate radial
157: velocity measurement.  The FLAMES fields were placed at varying
158: distances from the centre of the galaxy to beyond the tidal radius.
159: 
160: The FLAMES data were all reduced, extracted and wavelength calibrated
161: using the GIRBLDRS pipeline\footnote{available at SouceForge,
162: http://girbldrs.sourceforge.net/} provided by the FLAMES consortium
163: (Geneva Observatory, Blecha \etal 2003).  For sky subtraction,
164: velocity and equivalent width (EW) estimation, we developed our own
165: software which was thoroughly checked on
166: multiple observations of the same fields taken at different times.
167: Wavelength shifts were generally found to be negligible ($\approx \pm$1
168: km/s).  Each target spectrum was automatically continuum-corrected and
169: cross-correlated with a zero-continuum Gaussian model CaT template.
170: With a velocity determined it is then straightforward to estimate the
171: EW of the individual CaT lines. We found that for reliable [Fe/H]
172: determinations from CaT calibrations a S/N $\geq$ 10 is required, and
173: for reliable v$_{hel}$ a S/N $\geq$ 7.5 is sufficient. All the plots
174: in this paper use the S/N $> 10$ data set, as we always consider both
175: the [Fe/H] and the v$_{hel}$ measurements.
176: 
177: We were able to verify the zero point accuracy for both velocity and
178: abundance determinations by comparison with independent
179: previous measurements of stars in Scl, which we had deliberately
180: reobserved.  We also have overlap between our own HR
181: and LR samples to verify consistency of velocity and [Fe/H]
182: measurements using different gratings and based on different abundance
183: determination methods.
184: 
185: \section{Analysis}
186: 
187: \subsection{Spatial Distribution}
188: 
189: Previous studies have already suggested that the spatial distribution
190: of the Horizontal Branch stars in Scl shows signs of a population
191: gradient, with the red horizontal branch stars (RHB) being more
192: tightly concentrated than the blue horizontal branch (BHB) stars (\eg
193: Hurley-Keller \etal 1999; Majewski \etal 1999).  Unlike previous wide
194: field studies, our WFI imaging data extends beyond the nominal tidal
195: radius and with an average 5-$\sigma$ limiting magnitude of V$=$23.5
196: and I$=$22.5 also probes well below the horizontal branch.  This has
197: enabled us to unequivocally demonstrate that the BHB and RHB stars
198: have markedly different spatial distributions (see Figure 1).
199: 
200: These RHB and BHB populations were selected from the WFI CMD shown in
201: Figure~2 using the selection boxes marked.  It is clear that the BHB
202: region is almost completely uncontaminated by foreground stars,
203: whereas the RHB zone includes a significant underlying foreground
204: component. To make a quantitative estimate of the contamination of the
205: RHB sample by foreground stars we selected a region of similar CMD
206: foreground occupancy, well away from the Scl stellar locus (also
207: outlined in Figure~2).  The radial density distributions of the RHB
208: component were then corrected by the expected foreground contamination
209: and the results for both the RHB and BHB distributions together with
210: their density ratio as a function of radial distance, are shown in the
211: lower panels of Figure 1.  The different spatial occupancy of the two
212: populations is striking and provides strong evidence that we are
213: seeing two distinct components.  The characteristics of the BHB and
214: RHB are also consistent with different ages (\eg $\leq$2~Gyr), or
215: different metallicities ($\Delta$[Fe/H]$\sim$0.7 dex), from
216: theoretical modelling of globular cluster horizontal branches by Lee
217: et al. (2001).
218: 
219: \subsection{Metallicity Distribution}
220: 
221: Our VLT/FLAMES spectroscopic sample provides measurements of
222: [Fe/H] for 401 individual candidate RGB-selected stars with final
223: S/N$>$10, of which 308 have a high membership probability
224: (these are plotted in Figure~3), defined to be 80 km/s $<$
225: v$_{hel}$ $<$ 150 km/s \ie approximately within 3$\sigma$ of
226: v$_{sys}$ (see Figure 4).  At the high Galactic latitude of Scl
227: v$_{hel}$ is a good estimator of membership probability assuming
228: v$_{sys} = 110$ km/s, and $\sigma \approx$ 10 km/s.
229: 
230: For those RGB stars which were determined to have a high probability
231: of membership, we show the distribution of [Fe/H] measurements as a
232: function of elliptical radius (the equivalent distance along the
233: semi-major axis) from the centre of Scl in the upper panel of
234: Figure~3.  A well-defined metallicity gradient is apparent with a
235: similar scale size to the RHB versus BHB spatial distributions.  The
236: lower panel of this figure highlights the abrupt change in the [Fe/H]
237: distribution that occurs at about r$= 0.2$~deg (2r$_c$) and also the
238: broad range in [Fe/H] over the whole galaxy.  Although both inner and
239: outer distributions cover a broad range of values, the inner is
240: centred at [Fe/H]$= - 1.4$ with a range of $\pm 0.6$dex and the outer
241: distribution is centred at [Fe/H] $= -2$ with a similar range of $\pm
242: 0.6$dex.  The drop in $<$[Fe/H]$>$ at 2r$_c$ from the centre also
243: corresponds to the transition between the dominance of the RHB and BHB
244: populations seen in Figure 1. This is not seen if the colour of the
245: RGB alone is taken as an indication of metallicity.
246: 
247: \subsection{Kinematics}
248: 
249: In the upper panel of Figure~4 we plot the sample of v$_{hel}$
250: measurements versus elliptical radius for which the continuum S/N $>
251: 10$.  The distinction between foreground stars and members of Scl is
252: generally unambiguous.  The clustering around the systemic velocity of
253: v$_{sys} = 110$ km/s is clear and is highlighted by the dashed line
254: and the dotted 3$\sigma$ deviation limits.  Although present,
255: foreground contamination can be seen to be almost
256: negligible, even in the outer regions, which simplifies subsequent
257: analysis.
258: 
259: We have divided our ``member'' sample into metal-poor ([Fe/H$< -1.7$)
260: and metal-rich ([Fe/H]$>-1.7$) distributions. This difference is
261: quantified in the lower panels of Figure~4 which shows the v$_{hel}$
262: distributions over three radial zones for the different metallicity
263: components separately.  The variation in v$_{sys}$ and $\sigma$ in
264: these three regions is computed using the maximum likelihood approach
265: described in Hargreaves \etal (1994), but modified to directly compute
266: the likelihood function over a 2D search grid.  We found that
267: v$_{sys}$ does not vary significantly between different 
268: populations or regions, but $\sigma$ does (see
269: Figure 4).
270: 
271: One obvious feature of the velocity distributions is the difference in
272: the kinematic properties of the two metallicity components.  In
273: general the more metal poor component has a higher velocity dispersion
274: than the more metal rich component. This difference is most striking
275: in the central panel ($0.2^\circ <$ r $< 0.5^\circ$) of the triptych
276: where $\sigma$ is 11.0$\pm$1.0 km/s and 6.8$\pm$0.8 km/s 
277: respectively.  The swap
278: in the dominance of the components is stark in the outer zone (r $>$
279: 0.5), where there are hardly any stars at all in the metal-rich
280: category and provides compelling evidence that we are indeed
281: observing two dynamically different components in Scl.
282: 
283: 
284: \section{Conclusions}
285: 
286: Our FLAMES results show that Scl contains two distinct stellar
287: components, one metal-rich, $-0.9 >$ [Fe/H] $> -1.7$, and one
288: metal-poor, $-1.7 >$ [Fe/H] $> -2.8$. The metal-rich population is
289: more centrally concentrated than the metal poor, and on average
290: appears to have a lower velocity dispersion, $\sigma_{metal-rich} = 7
291: \pm 1$ km/s, whereas $\sigma_{metal-poor} = 11 \pm 1$ km/s 
292: (see Figure~4).  Our
293: WFI results show that there is a comparable difference over the same
294: scale in the spatial distribution of BHB stars and 
295: RHB stars.  Plausibly combining these results enables us 
296: to distinguish these
297: two distinct components on the basis of kinematics (Figure 4),
298: metallicity (Figure 3) and spatial distribution (Figure 1).
299: These large samples allow us for
300: the first time to quantify these large scale trends in a statistically
301: meaningful fashion.
302: 
303: There are indications that this is a common feature of dSph galaxies.
304: Our preliminary analysis of v$_{hel}$ and [Fe/H]
305: measurements in the other galaxies in our sample (Fornax and Sextans
306: dSph, Battaglia et al., in prep) also shows very similar characteristics to
307: Scl, especially in the most metal poor component.  The mix of
308: populations between the inner and outer regions is clearly changing.
309: The radial velocity studies of Wilkinson \etal (2004) and Kleyna \etal 
310: (2004) have also considered the possibility that kinematically distinct
311: components exist in Ursa Minor, Draco and Sextans.
312: 
313: What mechanism could create two ancient stellar components in a small
314: dwarf spheroidal galaxy?  A simple possibility is that
315: the formation of these dSph galaxies began with an initial burst
316: of star formation, resulting in a stellar population with a mean
317: [Fe/H] $\leq - 2$.  Subsequent supernovae explosions from this initial
318: episode could have been sufficient to cause gas (and
319: metal) loss such that star formation was inhibited until the remaining
320: gas could sink deeper into the centre and begin star formation again
321: (\eg Mori, Ferrara \& Madau 2002).  Thus the subsequent generation(s)
322: of stars would 
323: inhabit a region closer to the centre of the galaxy, and have a
324: higher average metallicity ($\geq -1.4$) and different kinematics. 
325: Another possible cause is external influences, such as
326: minor mergers, or accretion of additional gas.  It might also be
327: that events surrounding the Epoch of Reionisation, strongly influenced the
328: evolution of these small galaxies (\eg Bullock, Kravtsov \& Weinberg
329: 2001) and resulted in the stripping or photoevaporation of the outer
330: layers of gas in the dSph, meaning that subsequent more metal enhanced
331: star formation occured only in the central regions.
332: 
333: The full abundance analysis of the FLAMES HR data (Hill et al. in
334: prep) will provide more details of the chemical enrichment history of
335: Scl. This will hopefully enable us to distinguish between two
336: episodes of star formation, each of which created the separate
337: kinematic components seen here, or more continuous star formation,
338: manifested as a gradient in velocity dispersion and
339: metallicity from the centre of the galaxy.
340: 
341: Here we have shown that even dSphs, the smallest galactic systems, did
342: not form simply, as might be assumed.  The wealth of data presented
343: here points to the presence of dynamically distinct components
344: in these galactic building blocks, or templates, and
345: provides a stimulus to theories of the formation and evolution of
346: dSphs.
347: 
348: \acknowledgments
349: 
350: ET gratefully acknowledges support from a fellowship of the Royal
351: Netherlands Academy of Arts and Sciences, and thanks Renzo Sancisi
352: for a careful reading of the draft.
353: KAV thanks the NSF for support through a CAREER award, AST-9984073.
354: MDS would like to thank the NSF for partial support under AST-0306884.
355: 
356: \begin{thebibliography}{}
357: 
358: \bibitem[]{} Aaronson M. \& Olsezewski E.W. 1987 AJ, 94, 657
359: 
360: \bibitem[]{} Armandroff T.E. \& Da Costa G.S. 1986 AJ, 92, 777
361: 
362: \bibitem[]{} Blecha A., North P., Royer F. \& Simond, G. 2003 
363: {\it BLDR Software Reference Manual 1.09}, Doc. No. 
364: VLT-SPE-OGL-13730-0040 
365: 
366: \bibitem[]{} Bouchard A., Carignan C. \& Mashchenko S. 2003 AJ, 126, 1295
367: 
368: \bibitem[]{} Bullock J.S., Kravtsov A.V. \& Weinberg D.H. 2001 ApJ, 548, 33
369: 
370: \bibitem[]{} Cole A.A., Smecker-Hane T.A., Tolstoy E., Bosler T.L. 
371: \& Gallagher J.S. 2004 MNRAS, 347, 367
372: 
373: \bibitem[]{} Hargreaves J.C., Gilmore G., Irwin M.J. \& 
374: Carter D. 1994 MNRAS, 269, 957
375: 
376: \bibitem[]{} Hurley-Keller D., Mateo M., \& Grebel E.K. 1999 ApJL, 523, 25
377: 
378: \bibitem[]{} Irwin M. \& Lewis J. 2001 NewAR, 45, 105
379: 
380: \bibitem[]{} Irwin M. \etal 2004, 
381: %in {\em Optimizing Scientific return from astronomy through 
382: %Information technologies}, 
383: {\em Proc. SPIE} {\bf 5493}
384: 
385: \bibitem[]{} Irwin M. \& Hatzidimitriou D. 1995 MNRAS, 277, 1354
386: 
387: \bibitem[]{} Kaluzny J., Kubiak M., Szymanski M., Udalski A.,
388: Krzeminski W. \& Mateo M. 1995 A\&AS, 112, 407
389: 
390: \bibitem[]{} Kleyna \etal 2004 MNRAS, in press (astro-ph/0409066)
391: 
392: \bibitem[]{} Kunkel W.E. \& Demers S. 1977 ApJ, 214, 21
393: 
394: \bibitem[]{} Lee Y-W., Yoon S-J., Rey S-C. \& Chaboyer B. 2001 
395: {\it Astrophysical Ages \& Time Scales}, eds.
396: T. von Hippel \etal
397: 
398: \bibitem[]{} Majewski, S.R., Siegel, M.H., Patterson, R.J. \&
399: Rood, R. 1999 ApJL, 520, 33
400: 
401: \bibitem[]{} Monkiewicz J., Mould J.R., Gallagher J.S., Clarke J.T.,
402: Trauger J.T., Grillmair C. \& WFPC2 IDT 1999 PASP, 111, 1392
403: 
404: \bibitem[]{} Queloz D., Dubath P. \& Pasquini L. 1995 A\&A, 300, 31
405: 
406: \bibitem[]{} Rutledge G.A., Hesser J.E. \& Stetson P.B. 1997 PASP, 109, 907 
407: 
408: \bibitem[]{} Shetrone M.D., Venn K., Tolstoy E., Primas F., Hill V. \&
409: Kaufer A. 2003 AJ, 125, 684
410: 
411: \bibitem[]{} Tolstoy E., Irwin M.J., Cole A.A., Pasquini L., Gilmozzi R.
412: \& Gallagher J.S. 2001 MNRAS, 327, 918
413: 
414: \bibitem[]{} Tolstoy E., Venn K., Shetrone M.D., Primas F., Hill V.,
415: Kaufer A. \& Szeifert T. 2003 AJ, 125, 684
416: 
417: \bibitem[]{} Wilkinson M., Kleyna J.T., Evans N.W., Gilmore G.,
418: Irwin M.J. \& Grebel E.K. 2004 ApJL, 611, 21
419: 
420: \end{thebibliography}
421: 
422: %Figure 1
423: \begin{figure}[!ht]
424: {\centering \scalebox{0.6}{
425: \includegraphics[angle=270,scale=.75]{fig1a.eps}} \par}
426: \vskip 0.5cm
427: {\centering \scalebox{0.6}{
428: \includegraphics[angle=270,scale=0.67]{fig1b.eps}} \par}
429: 
430: \caption{
431: The distribution of Horizontal Branch stars from WFI
432: imaging of the Scl dSph.  The upper panels show the different spatial
433: distributions of blue horizontal branch stars (BHB), and red
434: horizontal branch stars (RHB) as selected from the M$_{v}$, V$-$I
435: Colour-Magnitude Diagram (CMD) shown in Figure 2.  Also shown, to
436: illustrate the foreground contamination in the RHB distribution, are a
437: CMD-selected sample of foreground stars to match the RHB contamination
438: density (see Figure 2).  The ellipse is the tidal radius of Scl as
439: defined by Irwin \& Hatzidimitriou (1995).  The lower panels
440: quantitatively show the spatial variation of number density per unit
441: area of these three population components (BHB in blue, RHB in red and
442: foreground in green) and the ratio of BHB to RHB number densities
443: after correcting the RHB component for foreground contamination. The
444: physical scale at the distance of Sculptor is 1 degree to 1.4~kpc.
445: }
446: \end{figure}
447: 
448: 
449: %Figure 2
450: \begin{figure}[!ht]
451: {\centering \scalebox{0.6}{
452: \includegraphics[scale=.8]{fig2.eps}} \par}
453: \caption{
454: The Colour-Magnitude Diagram 
455: for the WFI coverage of Scl shown in Figure~1.  The green and
456: magenta symbols are the Red Giant Branch stars which were observed
457: with VLT/FLAMES and for which we have accurate v$_{hel}$ and [Fe/H]
458: measurements (S/N$>$10).  The potential members of Scl are shown as
459: magenta dots and non-members shown as green stars.  Also shown are the
460: regions used to define the BHB, RHB and foreground comparison 
461: sample (FG).
462: }
463: \end{figure}
464: 
465: 
466: %Figure 3
467: \begin{figure}[!ht]
468: {\centering \scalebox{0.6}{
469: %\hskip -0.5cm
470: \includegraphics[scale=0.87]{fig3a.eps}} \par}
471: \vskip -5.5cm
472: %\hskip 0.5cm
473: {\centering \scalebox{0.6}{
474: \includegraphics[angle=270,scale=.65]{fig3b.eps}} \par}
475: %\vskip 0.3cm
476: \caption{ 
477: VLT/FLAMES spectroscopic measurements of [Fe/H] for the 308 potential RGB 
478: velocity members of Scl versus radius 
479: (top panel).  The lower panel shows the distribution of [Fe/H] for: the 
480: entire sample of Scl member RGB stars (dotted line); the 97 stars 
481: within the central, r$=$0.2 degree region (red solid line); and the 211
482: stars beyond r$>$0.2 degrees (blue solid line).
483: }
484: \end{figure}
485: 
486: 
487: %Figure 4
488: \begin{figure}[!ht]
489: \vskip -2.5cm
490: %\hskip -0.9cm
491: {\centering \scalebox{0.6}{
492: \includegraphics[scale=0.87]{fig4a.eps}} \par}
493: \vskip -0.3cm
494: {\centering \scalebox{0.6}{
495: \includegraphics[angle=270,scale=.7]{fig4b.eps}} \par}
496: \caption{
497: The top panel shows v$_{hel}$ as a function of elliptical radius for
498: all stars satisfying S/N $>$ 10.  The
499: 308 stars which are potential members are plotted as red stars
500: ([Fe/H] $>$ -1.7) and blue circles ([Fe/H] $<$ -1.7), while the 93 
501: green crosses are assumed to be non-members.  The
502: lower panels quantify how the velocity distribution changes as a
503: function of metallicity and radius. The 
504: low metallicity distribution is plotted 
505: in blue, and the higher metallicity in red, as in the upper panel.
506: }
507: \end{figure}
508: 
509: \end{document}
510: