1: \documentclass[12pt,preprint]{aastex}
2: %% You can insert a short comment on the title page using the command below.
3: \slugcomment{Submitted to ApJ Letters. May 9th 2006}
4:
5: \shorttitle{Stellar spectra in M33}
6: \shortauthors{McConnachie et al.}
7:
8: \begin{document}
9:
10: \title{The stellar halo and outer disk of M33}
11:
12: \author{Alan W. McConnachie$^1$\email{alan@uvic.ca}}
13: \author{Scott C. Chapman$^{1,2}$\email{schapman@submm.caltech.edu}}
14: \author{Rodrigo A. Ibata$^{3}$\email{ibata@astro.u-strasbg.fr}}
15: \author{Annette M. N. Ferguson$^{4}$\email{ferguson@roe.ac.uk}}
16: \author{Mike J. Irwin$^{5}$\email{mike@ast.cam.ac.uk}}
17: \author{Geraint F. Lewis$^{6}$\email{gfl@Physics.usyd.edu.au}}
18: \author{Nial R. Tanvir$^{7}$\email{nrt@star.herts.ac.uk}}
19:
20: \affil{$^1$Department of Physics and Astronomy, University of Victoria, Victoria, B.C., V8P 1A1, Canada}
21: \affil{$^2$California Institute of Technology, Pasadena, CA 91125, U.S.A.}
22: \affil{$^3$Observatoire de Strasbourg, 11, rue de l'Universite, F-67000, Strasbourg, France}
23: \affil{$^4$Institute for Astronomy, University of Edinburgh, Royal Observatory, Blackford Hill, Edinburgh EH9 3HJ, U.K.}
24: \affil{$^5$Institute of Astronomy, Madingley Road, Cambridge, CB3 0HA, U.K.}
25: \affil{$^6$Institute of Astronomy, School of Physics, A29, University of Sydney, NSW 2006, Australia}
26: \affil{$^7$Physical Sciences, University of Hertfordshire, Hatfield, AL10 9AB, U.K.\\}
27:
28: \begin{abstract}
29: We present first results from a Keck/DEIMOS spectroscopic survey of
30: red giant branch (RGB) stars in M33. The radial velocity distributions
31: of the stars in our fields are well described by three Gaussian
32: components, corresponding to a candidate halo component with an
33: uncorrected radial velocity dispersion of $\sigma \simeq
34: 50$\,km\,s$^{-1}$, a candidate disk component with a dispersion
35: $\sigma \simeq 16$\,km\,s$^{-1}$, and a third component offset from
36: the disk by $\sim 50$\,km\,s$^{-1}$, but for which the dispersion is
37: not well constrained. By comparing our data to a model of M33 based
38: upon its HI rotation curve, we find that the stellar disk is offset in
39: velocity by $\sim 25$\,km\,s$^{-1}$ from the HI disk, consistent with
40: the warping which exists between these components. The spectroscopic
41: metallicity of the halo component is [Fe/H] $\simeq -1.5$,
42: significantly more metal-poor than the implied metallicity of the disk
43: population ([Fe/H] $\simeq -0.9$), which also has a broader colour
44: dispersion than the halo population. These data represent the first
45: detections of individual stars in the halo of M33 and, despite being
46: $\sim 10$ times less massive than M31 or the Milky Way, all three of
47: these disk galaxies have stellar halo components with a similar
48: metallicity. The color distribution of the third component is
49: different to the disk and the halo, but is similar to that expected for
50: a single, coeval, stellar population, and could represent a stellar
51: stream. More observations are required to determine the true nature of
52: this intriguing third kinematic component in M33.
53: \end{abstract}
54:
55: \keywords{ galaxies: individual (M33/Triangulum) --- Local Group --- galaxies: stellar content --- galaxies: structure}
56:
57:
58: \section{Introduction}
59:
60: M33, the Triangulum Galaxy ($1^h 33^m 51^s, 30^\circ 39^\prime
61: 36^{\prime\prime}$), is the third brightest Local Group galaxy, with
62: an integrated luminosity of $M_V = -18.9$. It is a late-type
63: spiral, Sc II-III, with two open spiral arms and no clear evidence of
64: any bulge component (\citealt{bothun1992,minniti1993,mclean1996}). The
65: overwhelming majority of light in M33 is distributed in an exponential
66: disk component (\citealt{devaucouleurs1959}), and the optical disk is
67: tilted at nearly $30^\circ$ to the strongly warped HI envelope
68: (\citealt{rogstad1976}). The mass-to-light ratio of its nucleus is
69: small ($M/L < 0.4$), ruling out the presence of a supermassive black
70: hole (\citealt{kormendy1993,lauer1998}) like that implied to exist in
71: the Milky Way (MW; \citealt{genzel1997}) and M31
72: (\citealt{dressler1988,kormendy1988}).
73:
74: \cite{chandar2002} analyse the kinematics of stars clusters in M33,
75: and find two distinct kinematic populations, which they argue
76: represent a disk and halo population, although individual stars
77: belonging to a halo component in M33 had not been directly
78: observed. In a companion paper by A. Ferguson et al. {\it in
79: preparation} (see also \citealt{ferguson2006}), we present our wide
80: field photometry of this galaxy taken with the Isaac Newton Telescope
81: Wide Field Camera (INT~WFC), and show that a low level stellar
82: component dominates the radial profile of M33 at large radius,
83: implying the presence of an extended stellar halo in M33. Unlike its
84: massive neighbor, M31, there is no evidence of substructure in M33 to
85: a surface brightness threshold of $\mu_V \simeq 30$\,mags\,sq.arcsec.
86:
87: In this {\it Letter}, we present first results from a spectroscopic survey
88: of red giant branch (RGB) stars in M33 using the DEep Imaging
89: Multi-Object Spectrograph (DEIMOS) on Keck~II
90: (\citealt{faber2003}). Section~2 describes our observations. Section 3
91: analyzes the radial velocity distributions in our fields. Section
92: 4 analyzes the metallicities of our target stars, and Section 5
93: discusses our results and concludes. We assume a distance to M33 of
94: $809 \pm 24$\,kpc (\citealt{mcconnachie2004a,mcconnachie2005a}).
95:
96: \section{Observations}
97:
98: Since 2002, our group has been conducting a spectroscopic survey of
99: M31 using DEIMOS on Keck~II
100: (\citealt{ibata2004,ibata2005,mcconnachie2004b,chapman2005,chapman2006}). In
101: September -- October 2005, we extended the scope of this survey to
102: M33. The two fields analyzed here were taken at the end of the night
103: at high airmass, during imperfect weather conditions. Their location
104: relative to the disk of M33 can be seen in Figure~\ref{rotcurve}, and
105: are positioned $\sim 38^\prime$ (9\,kpc) along the southern major axis
106: to study the kinematics of stars at the edge of the M33 disk. In the
107: same way as our M31 study, we used the `mini-slitlet' DEIMOS setup for
108: these fields, which significantly increases the multipexing
109: capabilites of this instrument
110: (\citealt{ibata2005,chapman2006}). Targets were selected from the INT
111: survey using a color -- magnitude box defined by $20.5 < i < 22.0$ and
112: $1.0 < \left(V - i\right)_o < 4.0$. Other targets brighter than $i =
113: 22$ were automatically chosen to fill in the available space of the
114: spectrograph detector.
115:
116: The contours in Figure~\ref{rotcurve} show the expected mean stellar
117: heliocentric radial velocities for a kinematic model of M33, based
118: upon the HI rotation curve of M33 derived by \cite{corbelli2003} and
119: shown in her Figure~5. We assume a stellar distribution in M33
120: described by $\rho\left(R, z\right) \propto
121: \exp\left(-R/R_e\right)\exp\left(-z/z_e\right)$, with $R_e =
122: 5.8^\prime$ (1.4\,kpc; \citealt{regan1994}) and $z_e =
123: 0.2$\,kpc. Changing $z_e$ within the range $\left[0,1\right]$\,kpc
124: makes no significant difference to any of our results. Stars occupy
125: circular orbits with $v_z = 0$ and $v_R$ defined by interpolation of
126: the HI rotation curve, linearly extrapolated beyond $R = 77^\prime$
127: (18\,kpc) where necessary. The model is then corrected for the
128: position angle ($\theta \simeq 23^\circ$) and inclination ($i \simeq
129: 54^\circ$) of the stellar disk of M33. The mean velocity along any
130: sight-line through the disk is calculated as $\bar{v_l} = \int
131: \mathbf{v} \bullet \hat{\mathbf{L}}~\rho(R, z)~dl / \int \rho(R,
132: z)~dl$, and corrected to the heliocentric frame. $\hat{\mathbf {L}}$
133: is a unit vector pointing along the direction of our line of sight.
134:
135: The data were processed and reduced in the same way as the main M31
136: survey (\citealt{ibata2005,chapman2006}). Due to the generally poor
137: observing conditions, the data analyzed here consist of $280$\,stars
138: with a cross correlation coefficient $> 0.02$, a signal-to-noise
139: ($S/N$) in the continuum $> 2$, a Tonry-Davis coefficient
140: (\citealt{tonry1979}) $> 2$, and a velocity error $v_{err} <
141: 25$\,km\,s$^{-1}$ ($\overline{v}_{err} \simeq 10$\,km\,s$^{-1}$).
142:
143: \section{Kinematics}
144:
145: The top two panels of Figure~\ref{3fields} shows the heliocentric
146: radial velocity ($v_h$) distributions in our two fields. The dashed
147: lines indicate $v_h = 0$, and the dot-dashed lines indicate the
148: systemic velocity of M33, $v_h = -179$\,km\,s$^{-1}$. The dot-dashed
149: histograms show the expected heliocentric velocity distributions of
150: the MW stars that lie in each field using the MW model of
151: \cite{robin2003}, and which satisfy our selection criteria for M33 RGB
152: stars. The histograms have been scaled to match the area of the DEIMOS
153: fields. They show that any contribution from MW foreground stars at
154: $v_h < -80$\,km\,s$^{-1}$ is small for the purposes of this {\it
155: Letter}, and so this velocity cut is adopted in the following
156: kinematic analysis. Stars with $v_h > -80$\,km\,s$^{-1}$ are
157: predominantly MW foreground, and the small numbers we observe further
158: suggest that the contamination at more negative velocity will be
159: negligible.
160:
161: The third panel of Figure~\ref{3fields} shows the velocity
162: distributions of Fields 157 and 158 added together to boost
163: statistics. We expect the main stellar components in M33 will be well
164: fit by individual Gaussians, and any stellar halo component in M33
165: will have a velocity distribution centered on $v_h =
166: -179$\,km\,s$^{-1}$. Since M33 is $\sim 10$\,times less massive than
167: M31, we expect that the M33 halo velocity dispersion will be $\sim
168: \sqrt{10}$ that of M31 ($\sigma_{M31} \sim 145$\,km\,s$^{-1}$;
169: \citealt{chapman2006}). The broad Gaussian overlaid in the third panel
170: of Figure~\ref{3fields} represents a M33 halo component with a
171: dispersion of $\sigma = 50$\,km\,s$^{-1}$, scaled to match the
172: approximate number of stars observed. We also expect that the
173: prominent peak of stars in these fields is due to the dominant disk of
174: M33. Therefore, we have subtracted the halo Gaussian from the $v_h$
175: distribution in the third panel and fitted a Gaussian to this peak,
176: centered at $v_1 = -102 \pm 2$\,km\,s$^{-1}$ with an uncorrected
177: dispersion $\sigma_1 = 16 \pm 3$\,km\,s$^{-1}$ (corrected dispersion
178: of $\sim 12.5$\,km\,s$^{-1}$). The residuals of this fit show a
179: prominent grouping of stars at $v \sim -150$\,km\,s$^{-1}$. We fit a
180: third Gaussian to this feature, with a peak $v_2 = -150 \pm
181: 4$\,km\,s$^{-1}$ and a dispersion $\sigma_2 = 10 \pm 4$\,km\,s$^{-1}$.
182: The latter quantity is more poorly constrained than given by the
183: formal error, however, as its dispersion is covariant with the
184: dispersion of the disk and halo, which have been fitted
185: independently. All 3 Gaussians and their summed total are overlaid in
186: the third panel of Figure~\ref{3fields}.
187:
188: The fourth panel of Figure~\ref{3fields} shows the expected
189: distribution of stars at the positions of Fields 157 and 158, for the
190: model of M33 shown in Figure~\ref{rotcurve}, combined with a halo
191: model with a radial velocity dispersion of $50$\,km\,s$^{-1}$. The
192: radial velocity dispersion of the disk is set to $16$\,km\,s$^{-1}$,
193: to mimic our observations. The normalization of both components is
194: chosen to approximately match the corresponding peaks in the data . A
195: comparison between model and data shows (1) the peak at $v_2 =
196: -150$\,km\,s$^{-1}$ does not naturally arise as a result of our
197: line-of-sight through the main M33 disk sampling stars at different
198: circular velocities, (2) the stellar disk is offset in $v_h$ from the
199: HI disk by $\sim 20 - 25$\,km\,s$^{-1}$, implying a deprojected offset
200: of $\sim 25 - 30$\,km\,s$^{-1}$. For comparison, the asymmetric drift
201: between the stars and HI in the MW and M31 is of order $15 -
202: 20$\,km\,s$^{-1}$. Scaling for mass, the expected asymmetric drift in
203: M33 is not likely to be more than $\sim 10$\,km\,s$^{-1}$. We believe
204: that the velocity offset between the two disks is due to differential
205: warping, specifically that the HI disk is warping into our
206: line-of-sight at large radius from M33. This has been implied, quite
207: spectacularly, by \cite{rogstad1976} (see their Figure~14), and is
208: consistent with the magnitude of the velocity offset.
209:
210: \section{Metallicities}
211:
212: Figure~\ref{cmd} shows the color -- magnitude diagrams (CMDs) for the
213: spectroscopically observed M33 RGB stars in Fields 157 and 158,
214: divided by the radial velocity cuts indicated so as to kinematically
215: separate the three candidate components as best as possible. Victoria
216: -- Regina isochrones from \cite{vandenberg2005}, with $BVRI$ color --
217: $T_{eff}$ relations as described by \cite{vandenberg2003},
218: corresponding to a 13 Gyr stellar population with metallicities of
219: [Fe/H] $= -2.0, -1.3, -0.8$ and $-0.5$, are also overlaid.
220:
221: The differences between the CMDs in Figure~\ref{cmd} are striking, and
222: appear to validate our kinematic deconvolution. The candidate M33 halo
223: stars are predominantly bluer and have a smaller color dispersion than
224: the candidate disk stars, indicating a difference in the age and/or
225: metallicity properties of the populations. Assuming both populations
226: are $13$\,Gyrs old implies a median photometric metallicity for the
227: halo of [Fe/H] $= -1.3 \pm 0.1$, with a dispersion of $\sigma_{{\rm
228: [Fe/H]}} = 0.21$. Given the small number of stars which contribute to
229: this measurement, this is consistent with the Ferguson et al. (2006)
230: metallicity measurement for the halo, of [Fe/H] $= -1.5 \pm 0.1$. The
231: disk has an implied photometric metallicity of [Fe/H] $= -0.9 \pm 0.1$
232: and a broader dispersion of $\sigma_{{\rm [Fe/H]}} = 0.35$. The median
233: color (metallicity) of the `third component' is similar to the disk,
234: although its color dispersion is remarkably narrow ($\sigma_{{\rm
235: [Fe/H]}} = 0.15$). This population is difficult to isolate
236: kinematically, and our sample could be contaminated by both of the
237: other two components. However, the locus of its stars on the CMD does
238: not clearly associate it with either the candidate disk or halo
239: population, and is reminiscent of a CMD for a single stellar
240: population. If this population was due to MW foreground contamination,
241: we would not expect its stars to occupy so well defined a locus in
242: this diagram. The most probable explanation is that these stars belong
243: to a stellar stream. These stars do not, however, define a spatially
244: obvious structure, but given our limited field size and that our
245: choice of slitlet position will affect the spatial distribution, this
246: test is inconclusive at best.
247:
248: We calculate the spectroscopic metallicity of the halo stars by
249: measuring the equivalent width of the CaT from stacked spectra, using
250: an identical method to \cite{ibata2005} and \cite{chapman2006}. We
251: find that [Fe/H] $= -1.5 \pm 0.3$, in good agreement with the
252: photometric measurement. The large uncertainty reflects the
253: relatively low signal-to-noise of the stacked spectrum. We are unable
254: to measure the metallicities of the other two components in this way,
255: since their velocities cause the CaT to lie on top of strong sky lines
256: which prevents accurate equivalent width measurements.
257:
258: \section{Discussion}
259: \label{final}
260:
261: M33 is considered an example of the quintessential disk galaxy, with
262: no evidence for any bulge or spheroid component. However, the radial
263: profile for M33 presented in Ferguson et al. (2006), based upon our
264: INT-WFC observations, reveals for the first time the presence of a
265: spatially extended, low density component. Using Keck/DEIMOS, we have
266: identified a population of RGB stars which have velocities centered on
267: the systemic velocity of M33 and which appear to have a large
268: dispersion ($\sigma \sim 50$\,km\,s$^{-1}$). These stars are bluer
269: than other RGB stars in M33, with a small color dispersion. Their
270: average metallicity is [Fe/H] $\sim -1.5$, and we attribute them to
271: the newly discovered halo component of M33. The haloes of the MW and
272: M31 have similar metallicities to this
273: (\citealt{eggen1962,chapman2006,kalirai2006b}), even although both
274: these galaxies are $\sim 10$\,times more massive than M33. It is
275: important to emphasise that there is no evidence of a trend linking
276: halo metallicity to galaxy mass or luminosity for the Local group
277: galaxies, as has been suggested from recent Hubble Space Telescope
278: studies of galaxies within $\sim 10$\,Mpc
279: (\citealt{mouhcine2005a,mouhcine2005b}). Instead, all the large Local
280: Group spiral galaxies appear to have very similar halo metallicities.
281:
282: We also observe stars in the disk of M33. Their $v_h$ distribution is
283: well described by a Gaussian with an uncorrected dispersion of $\sigma
284: \simeq 16$\,km\,s$^{-1}$, corresponding to a corrected dispersion of
285: $\simeq 12.5$\,km\,s$^{-1}$. These stars are redder and have a broader
286: color dispersion than the halo stars. The stellar disk is offset in
287: $v_h$ from the HI disk in our simple model by $\sim 20 -
288: 25$\,km\,s$^{-1}$, which is fully consistent with the extreme warping
289: of the HI disk implied by \cite{rogstad1976}.
290:
291: Finally, we present evidence for a third kinematic feature in M33,
292: revealed by the $v_h$ distribution of the southern fields. The
293: velocity dispersion of this component is not well constrained by our
294: data. A comparison to our simple model of the disk of M33 shows that
295: this feature does not arise as a result of our line-of-sight through
296: M33 sampling stars at different circular velocities in the main
297: disk. The locus of these stars in the CMD has a remarkably small color
298: dispersion, reminiscent of a single stellar population and
299: inconsistent with belonging to a foreground population. We believe the
300: most likely explanation for this feature is that it is a stellar
301: stream. However, more spectroscopic observations of fields at
302: different locations around the disk of M33 are required before the
303: true nature of this intriguing feature is revealed.
304:
305: \acknowledgments The data presented herein were obtained at the
306: W.M. Keck Observatory, which is operated as a scientific partnership
307: among the California Institute of Technology, the University of
308: California and the National Aeronautics and Space Administration. The
309: Observatory was made possible by the generous financial support of the
310: W.M. Keck Foundation. We thank E.~Corbelli for supplying us with her
311: data for the HI rotation curve of M33, and the anonymous referee for
312: constructuve comments . AWM would like to thank J.~Navarro and
313: S.~Ellison for financial support. AMNF is supported by a Marie Curie
314: Excellence Grant from the European Commission under contract
315: MCEXT-CT-2005-025869. GFL acknowledges support through ARC DP0343508.
316:
317: \clearpage
318:
319: \begin{thebibliography}{29}
320: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
321:
322: \bibitem[{{Bothun}(1992)}]{bothun1992}
323: {Bothun}, G.~D. 1992, \aj, 103, 104
324:
325: \bibitem[{{Chandar} {et~al.}(2002){Chandar}, {Bianchi}, {Ford}, \&
326: {Sarajedini}}]{chandar2002}
327: {Chandar}, R., {Bianchi}, L., {Ford}, H.~C., \& {Sarajedini}, A. 2002, \apj,
328: 564, 712
329:
330: \bibitem[{{Chapman} {et~al.}(2005){Chapman}, {Ibata}, {Lewis}, {Ferguson},
331: {Irwin}, {McConnachie}, \& {Tanvir}}]{chapman2005}
332: {Chapman}, S.~C., {Ibata}, R., {Lewis}, G.~F., {Ferguson}, A.~M.~N., {Irwin},
333: M., {McConnachie}, A., \& {Tanvir}, N. 2005, \apjl, 632, L87
334:
335: \bibitem[{{Chapman} {et~al.}(2006){Chapman}, {Ibata}, {Lewis}, {Ferguson},
336: {Irwin}, {McConnachie}, \& {Tanvir}}]{chapman2006}
337: ---. 2006, astro-ph/0602604
338:
339: \bibitem[{{Corbelli}(2003)}]{corbelli2003}
340: {Corbelli}, E. 2003, \mnras, 342, 199
341:
342: \bibitem[{{de Vaucouleurs}(1959)}]{devaucouleurs1959}
343: {de Vaucouleurs}, G. 1959, \apj, 130, 728
344:
345: \bibitem[{{Dressler} \& {Richstone}(1988)}]{dressler1988}
346: {Dressler}, A. \& {Richstone}, D.~O. 1988, \apj, 324, 701
347:
348: \bibitem[{{Eggen} {et~al.}(1962){Eggen}, {Lynden-Bell}, \&
349: {Sandage}}]{eggen1962}
350: {Eggen}, O.~J., {Lynden-Bell}, D., \& {Sandage}, A.~R. 1962, \apj, 136, 748
351:
352: \bibitem[{{Faber} {et~al.}(2003){Faber}, {Phillips}, {Kibrick}, {Alcott},
353: {Allen}, {Burrous}, {Cantrall}, {Clarke}, {Coil}, {Cowley}, {Davis}, {Deich},
354: {Dietsch}, {Gilmore}, {Harper}, {Hilyard}, {Lewis}, {McVeigh}, {Newman},
355: {Osborne}, {Schiavon}, {Stover}, {Tucker}, {Wallace}, {Wei}, {Wirth}, \&
356: {Wright}}]{faber2003}
357: {Faber}, S.~M., {Phillips}, A.~C., {Kibrick}, R.~I., {Alcott}, B., {Allen},
358: S.~L., {Burrous}, J., {Cantrall}, T., {Clarke}, D., {Coil}, A.~L., {Cowley},
359: D.~J., {Davis}, M., {Deich}, W.~T.~S., {Dietsch}, K., {Gilmore}, D.~K.,
360: {Harper}, C.~A., {Hilyard}, D.~F., {Lewis}, J.~P., {McVeigh}, M., {Newman},
361: J., {Osborne}, J., {Schiavon}, R., {Stover}, R.~J., {Tucker}, D., {Wallace},
362: V., {Wei}, M., {Wirth}, G., \& {Wright}, C.~A. 2003, in Instrument Design and
363: Performance for Optical/Infrared Ground-based Telescopes. Edited by Iye,
364: Masanori; Moorwood, Alan F. M. Proceedings of the SPIE, Volume 4841, pp.
365: 1657-1669 (2003)., ed. M.~{Iye} \& A.~F.~M. {Moorwood}, 1657--1669
366:
367: \bibitem[{{Ferguson} {et~al.}(2006){Ferguson}, {Irwin}, {Chapman}, {Ibata},
368: {Lewis}, \& {Tanvir}}]{ferguson2006}
369: {Ferguson}, A., {Irwin}, M., {Chapman}, S., {Ibata}, R., {Lewis}, G., \&
370: {Tanvir}, N. 2006, astro-ph/0601121
371:
372: \bibitem[{{Genzel} {et~al.}(1997){Genzel}, {Eckart}, {Ott}, \&
373: {Eisenhauer}}]{genzel1997}
374: {Genzel}, R., {Eckart}, A., {Ott}, T., \& {Eisenhauer}, F. 1997, \mnras, 291,
375: 219
376:
377: \bibitem[{{Ibata} {et~al.}(2004){Ibata}, {Chapman}, {Ferguson}, {Irwin},
378: {Lewis}, \& {McConnachie}}]{ibata2004}
379: {Ibata}, R., {Chapman}, S., {Ferguson}, A.~M.~N., {Irwin}, M., {Lewis}, G., \&
380: {McConnachie}, A. 2004, \mnras, 351, 117
381:
382: \bibitem[{{Ibata} {et~al.}(2005){Ibata}, {Chapman}, {Ferguson}, {Lewis},
383: {Irwin}, \& {Tanvir}}]{ibata2005}
384: {Ibata}, R., {Chapman}, S., {Ferguson}, A.~M.~N., {Lewis}, G., {Irwin}, M., \&
385: {Tanvir}, N. 2005, \apj, 634, 287
386:
387: \bibitem[{{Kalirai} {et~al.}(2006){Kalirai}, {Gilbert}, {Guhathakurta},
388: {Majewski}, {Ostheimer}, {Rich}, {Cooper}, {Reitzel}, \&
389: {Patterson}}]{kalirai2006b}
390: {Kalirai}, J.~S., {Gilbert}, K.~M., {Guhathakurta}, P., {Majewski}, S.~R.,
391: {Ostheimer}, J.~C., {Rich}, R.~M., {Cooper}, M.~C., {Reitzel}, D.~B., \&
392: {Patterson}, R.~J. 2006, astro-ph/0605170
393:
394: \bibitem[{{Kormendy}(1988)}]{kormendy1988}
395: {Kormendy}, J. 1988, \apj, 325, 128
396:
397: \bibitem[{{Kormendy} \& {McClure}(1993)}]{kormendy1993}
398: {Kormendy}, J. \& {McClure}, R.~D. 1993, \aj, 105, 1793
399:
400: \bibitem[{{Lauer} {et~al.}(1998){Lauer}, {Faber}, {Ajhar}, {Grillmair}, \&
401: {Scowen}}]{lauer1998}
402: {Lauer}, T.~R., {Faber}, S.~M., {Ajhar}, E.~A., {Grillmair}, C.~J., \&
403: {Scowen}, P.~A. 1998, \aj, 116, 2263
404:
405: \bibitem[{{McConnachie} {et~al.}(2004{\natexlab{a}}){McConnachie}, {Irwin},
406: {Ferguson}, {Ibata}, {Lewis}, \& {Tanvir}}]{mcconnachie2004a}
407: {McConnachie}, A.~W., {Irwin}, M.~J., {Ferguson}, A.~M.~N., {Ibata}, R.~A.,
408: {Lewis}, G.~F., \& {Tanvir}, N. 2004{\natexlab{a}}, \mnras, 350, 243
409:
410: \bibitem[{{McConnachie} {et~al.}(2005){McConnachie}, {Irwin}, {Ferguson},
411: {Ibata}, {Lewis}, \& {Tanvir}}]{mcconnachie2005a}
412: ---. 2005, \mnras, 356, 979
413:
414: \bibitem[{{McConnachie} {et~al.}(2004{\natexlab{b}}){McConnachie}, {Irwin},
415: {Lewis}, {Ibata}, {Chapman}, {Ferguson}, \& {Tanvir}}]{mcconnachie2004b}
416: {McConnachie}, A.~W., {Irwin}, M.~J., {Lewis}, G.~F., {Ibata}, R.~A.,
417: {Chapman}, S.~C., {Ferguson}, A.~M.~N., \& {Tanvir}, N.~R.
418: 2004{\natexlab{b}}, \mnras, 351, L94
419:
420: \bibitem[{{McLean} \& {Liu}(1996)}]{mclean1996}
421: {McLean}, I.~S. \& {Liu}, T. 1996, \apj, 456, 499
422:
423: \bibitem[{{Minniti} {et~al.}(1993){Minniti}, {Olszewski}, \&
424: {Rieke}}]{minniti1993}
425: {Minniti}, D., {Olszewski}, E.~W., \& {Rieke}, M. 1993, \apjl, 410, L79
426:
427: \bibitem[{{Mouhcine} {et~al.}(2005{\natexlab{a}}){Mouhcine}, {Ferguson},
428: {Rich}, {Brown}, \& {Smith}}]{mouhcine2005a}
429: {Mouhcine}, M., {Ferguson}, H.~C., {Rich}, R.~M., {Brown}, T.~M., \& {Smith},
430: T.~E. 2005{\natexlab{a}}, \apj, 633, 821
431:
432: \bibitem[{{Mouhcine} {et~al.}(2005{\natexlab{b}}){Mouhcine}, {Rich},
433: {Ferguson}, {Brown}, \& {Smith}}]{mouhcine2005b}
434: {Mouhcine}, M., {Rich}, R.~M., {Ferguson}, H.~C., {Brown}, T.~M., \& {Smith},
435: T.~E. 2005{\natexlab{b}}, \apj, 633, 828
436:
437: \bibitem[{{Regan} \& {Vogel}(1994)}]{regan1994}
438: {Regan}, M.~W. \& {Vogel}, S.~N. 1994, \apj, 434, 536
439:
440: \bibitem[{{Robin} {et~al.}(2003){Robin}, {Reyl{\'e}}, {Derri{\`e}re}, \&
441: {Picaud}}]{robin2003}
442: {Robin}, A.~C., {Reyl{\'e}}, C., {Derri{\`e}re}, S., \& {Picaud}, S. 2003,
443: \aap, 409, 523
444:
445: \bibitem[{{Rogstad} {et~al.}(1976){Rogstad}, {Wright}, \&
446: {Lockhart}}]{rogstad1976}
447: {Rogstad}, D.~H., {Wright}, M.~C.~H., \& {Lockhart}, I.~A. 1976, \apj, 204, 703
448:
449: \bibitem[{{Tonry} \& {Davis}(1979)}]{tonry1979}
450: {Tonry}, J. \& {Davis}, M. 1979, \aj, 84, 1511
451:
452: \bibitem[{{VandenBerg} {et~al.}(2005){VandenBerg}, {Bergbusch}, \&
453: {Dowler}}]{vandenberg2005}
454: {VandenBerg}, D.~A., {Bergbusch}, P.~A., \& {Dowler}, P.~D. 2005,
455: astro-ph/0510784
456:
457: \bibitem[{{VandenBerg} \& {Clem}(2003)}]{vandenberg2003}
458: {VandenBerg}, D.~A. \& {Clem}, J.~L. 2003, \aj, 126, 778
459:
460: \end{thebibliography}
461:
462: \clearpage
463:
464: \begin{figure}
465: \begin{center}
466: \includegraphics[angle=270, width=8cm]{fig1}
467: \caption{A tangent plane projection of RGB stars in M33 with $I <
468: 22$, from our INT WFC survey (Ferguson et al. 2006). The
469: locations of the Keck/DEIMOS fields are indicated. Contours show
470: the mean heliocentric radial velocity of a model M33 disk, based
471: upon the Corbelli (2003) HI rotation curve.}
472: \label{rotcurve}
473: \end{center}
474: \end{figure}
475:
476: \clearpage
477:
478: \begin{figure}
479: \begin{center}
480: \includegraphics[angle=0, width=8cm]{fig2}
481: \caption{First and second panels: heliocentric radial velocity
482: distributions for our M33 fields. The dashed lines represent zero
483: velocity, and the dot-dashed line represents the systemic velocity
484: of M33, $v_h = -179$\,km\,s$^{-1}$. The dot-dashed histograms show
485: the heliocentric velocity distribution of the expected MW
486: foreground stars in each field, from the MW model by
487: \citealt{robin2003}, which satisfy our color selection
488: criteria. These histograms have been scaled to match the area of
489: the DEIMOS fields, and the small number of stars at $v_h >
490: -80$\,km\,s$^{-1}$, which are most likely foreground, suggest that
491: contamination at more negative velocity will be negligible. Third
492: panel: Same as first two panels, except now the two fields have
493: been added together to boost statistics. Three Gaussian velocity
494: distributions and their summed total are overlaid. See text for
495: details. Fourth panel: the expected heliocentric radial velocity
496: distribution of stars in our fields, using our M33 disk model and
497: an additional Gaussian component, representing a smooth stellar
498: halo.}
499: \label{3fields}
500: \end{center}
501: \end{figure}
502:
503: \clearpage
504:
505: \begin{figure*}
506: \begin{center}
507: \includegraphics[angle=270, width=15cm]{fig3}
508: \caption{CMDs of spectroscopically observed RGB stars from Fields
509: 157 and 158, selected to satisfy the velocity cuts indicated
510: ($\sigma_1 = 16$\,km\,s$^{-1}$, $\sigma_2 =
511: 10$\,km\,s$^{-1}$). These velocity cuts predominantly sample stars
512: which belong to the candidate halo (left panel), disk (middle
513: panel) and the `third component' (right panel). Victoria -- Regina
514: isochrones (VandenBerg et al. 2005), corresponding to 13\,Gyr old
515: stars and with metallicities of [Fe/H] $= -2.0, -1.3, -0.8$ and
516: $-0.5$, are overlaid.}
517: \label{cmd}
518: \end{center}
519: \end{figure*}
520:
521: \end{document}
522: